article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
reduction in the spatial dimensions in nanoparticles results in a number of interesting properties such as , for example , the widening of the band gap , reduction in melting temperature , increased magnetic moments etc .
the fundamental understanding of the size dependence of properties of nanoparticles , i.e. at what size the quantum size effects set in , is important for tailoring these systems for possible nanotechnological applications .
studies along these lines continue to be frontline research in nanoscience@xcite .
fullerenes are hollow cages formed by sp@xmath8 bonded carbon atoms .
they are finite analogues of graphene with 12 pentagonal defects that introduce curvature and make them closed structures .
unlike the most nanoparticles however , fullerenes are distinctly different in that they have all the atoms on the surface whereas solid spherical nanoparticles have roughly n@xmath9 atoms on the surface .
this feature results in rapid increase in their volume as the number of atoms ( i.e. its size ) increases . as dipole polarizability is proportional to volume
, fullerenes are most suitable for investigating quantum size effects manifested in polarizability .
the large dipole polarizability of carbon fullerenes compared to that of metallic or dielectric spheres of same size have been previously interpreted as quantum size effect @xcite .
a number of studies on the polarizability of fullerenes have been devoted to understanding its size dependence@xcite .
these studies have been mainly based on classical models or have employed drastic approximations to simplify the polarizability calculations . unfortunately , the predictions of these models are not consistent with each other and show rather large variance . for example , the various predictions of the polarizability of the c@xmath7 ( see table [ table : pol ] ) are : 7100 ( pariser - parr - pople model@xcite ) , 2726 ( bond - order - bond polarizability@xcite ) , 9548 ( penn model - linear response theory@xcite ) , 17817 ( tight binding@xcite ) . the predictions from different models of fullerene polarizabilities and associated quantum size effects differ by almost an order of magnitude .
the most reliable methods to determine electronic structure are the first - principles - based quantum - chemical methods . in the preset work ,
we use the first - principles density - functional methods to accurately determine the first order electric response of carbon fullerenes of sizes from c@xmath0 to c@xmath7 by computing their static dipole polarizabilities .
these calculations provide the most accurate prediction of the dielectric response of fullerenes in the size range c@xmath0-c@xmath7 and show substantial quenching of quantum size effects by c@xmath5 in contrast to current understanding .
moreover , the fullerene polarizabilities computed herein also provide an accurate input to the models used in determination of the ultraviolet ( uv ) spectrum for possible detection of of hyper - fullerene particles in space@xcite .
the first essential step in obtaining an accurate estimate of polarizability of fullerene involves accurate determination of the fullerene structure .
we optimized the geometry of carbon fullerenes using recent fully analytic and variational formulation of density functional ( adft ) theory@xcite .
being free of numerical grids , the adft implementation is computationally very efficient and allows calculations of matrix elements and energies accurate to machine precision .
its functional form is restricted but its space of atomic parameters is rich .
it has been successfully used to study electronic properties of boron nitrides , aluminum nitride tubes@xcite , and provides dipole moments , total and atomization energies that are comparable to those obtained using some of the most sophisticated density functional models . in this work , we optimized the geometries of c@xmath0 , c@xmath2 , c@xmath3 , c@xmath5 , c@xmath6 , and c@xmath7 fullerenes from the 60@xmath10 family of icosahedral fullerenes .
additionally , two fullerenes ( c@xmath1 and c@xmath4 ) from the 180@xmath10 family are also optimized .
the structures ( cf . [ fig : str ] ) are optimized at the _ all _ electron level using the triple - zeta - quality ( 6 - 311 g * * ) orbital basis@xcite and an exchange - correlation basis set that includes @xmath11 functions , and the exchange - correlation parametrization that gives an exact c@xmath0 geometry@xcite .
these approximations give accurate geometries in a very efficient manner , which we verified by reoptimizing the larger c@xmath2 fullerene using even larger nrlmol basis ( 35 basis functions per atom ) and using the more complex and popular perdew - burke - ernzerhof ( pbe)@xcite generalized gradient approximation ( gga ) .
both models give identical c@xmath2 geometries within a root - mean - square value of 10@xmath12 a.u .
for the forces . for the polarizability calculations
the fullerenes structures are held rigid i.e. , the ionic ( vibronic ) contribution to dipole polarizabilities is ignored .
this is a tenable assumption as our earlier studies on c@xmath0 and c@xmath13 show that in fullerenes electronic part determines the total polarizability with negligible contributions from the ionic vibrations @xcite . | the static dielectric response of c@xmath0 , c@xmath1 , c@xmath2 , c@xmath3 , c@xmath4 , c@xmath5 , c@xmath6 , and c@xmath7 fullerenes is characterized by an all - electron density - functional method .
first , the screened polarizabilities of c@xmath0 , c@xmath1 , c@xmath2 , and c@xmath3 , are determined by the finite - field method using gaussian basis set containing 35 basis functions per atom . in the second set of calculations ,
the unscreened polarizabilities are calculated for fullerenes c@xmath0 through c@xmath7 from the self - consistent kohn - sham orbitals and eigen - values using the sum - over - states method .
the approximate screened polarizabilities , obtained by applying a correction determined within linear response theory show excellent agreement with the finite - field polarizabilities .
the static dipole polarizability per atom in c@xmath7 is ( 4 ) three times larger than that in c@xmath0 ( 1.344 ) .
our results reduce the uncertainty in various theoretical models used previously to describe the dielectric response of fullerenes and show that quantum size effects in polarizability are significantly smaller than previously thought . |
with the start of the large hadron collider ( lhc ) a new energy regime is accessible , either to confirm the standard model ( sm ) or to verify new physics at the tev scale .
famous possible sm extensions are heavy gauge bosons ( e.g. @xmath0 ) , supersymmetry or kaluza - klein resonances in models with extra dimensions . beside the outstanding discovery of the higgs boson , processes involving top - quarks , gauge bosons of the weak interaction and jets
are of particular interest .
the experimental identification will rely on their characteristic decay products with leptons or ( bottom- ) jets as characteristic examples .
in addition to this sm processes , ( bottom- ) jets are also involved in many decays originating from signals for physics beyond the sm .
in particular new resonances , e.g. heavy gauge bosons , decaying into @xmath1-jets or light quark jets require a detailed sm - based prediction to prove a possible deviation from the sm .
this needs a detailed theoretical understanding of the corresponding processes like bottom - quark pair , single bottom - quark and di - jet production .
these processes were studied in the past . the differential cross section for bottom - quark pair production
is known to next - to - leading order ( nlo ) accuracy in quantum chromodynamics ( qcd ) @xcite .
for massless single bottom production and di - jet production the nlo qcd can be found in . + it is well known that weak corrections can also be significant because of the presence of possible large sudakov logarithms .
these effects were studied intensively for several processes like weak boson and top - quark pair production @xcite . earlier work on sudakov logarithms in four - fermion processes can be found in . a study for @xmath1-jet production
can be found in .
in the following processes with initial state photons are neglected and we distinguish between quark - induced processes ( with two quarks in the initial state , one of these being @xmath2 ) , gluon - induced contributions ( with one or two gluons in the initial state ) and in the case of @xmath1-jet production also pure bottom - quark induced processes .
the lo gluon induced processes proceed through qcd amplitudes only while the four quark processes can be seperated into qcd @xmath3 , electroweak qcd @xmath4 and mixed qcd - electroweak @xmath5 contributions .
sample diagrams are shown in fig .
[ fig : quarkborns ] and fig .
[ fig : gluonborns ] . here
the light quarks ( denoted generically by @xmath6 and @xmath7 ) and the bottom - quark ( @xmath1 and @xmath8 ) are taken as massless . since jets close to the beam pipe escape detection
, we require a minimal transverse momentum @xmath9 of 50 gev .
the leading order cross section is then obtained from @xmath10 where the factorization scale dependence is suppressed and @xmath11 and @xmath12 are the partonic momentum fractions . the parton distribution functions ( pdf s ) for parton @xmath13 in hadron @xmath14
are denoted by @xmath15 .
the sum runs over all possible parton configurations @xmath16 in the initial state . for @xmath1-jet production three types of partonic processes
will be distinguished at leading order : @xmath17 furthermore we distinguish between single @xmath1-tag and double @xmath1-tag events .
the lo @xmath18-distribution for single @xmath1-tag events at the lhc is presented in fig .
[ fig : pt - bjet - lhc](a ) .
the gluon - induced processes dominate in the low energy regime ( @xmath19 gev ) . for @xmath18 larger than 500 gev the quark - induced processes take over and finally dominate the distribution .
this `` cross over '' of gluon- and quark - induced contributions is a consequence of the different behaviour of lhc quark and gluon luminosities .
the relative contributions from the exchange of electroweak bosons and bottom - induced processes are always below @xmath20 and therefore negligible .
a similiar picture is observed for the differential cross section for double @xmath1-tag events at the lhc ( fig .
[ fig : pt - bjet - lhc](b ) ) . here
the `` cross over '' of quark- and gluon - induced contributions is around @xmath21 tev . for low @xmath18
the pure bottom - induced processes are as important as quark - induced contributions .
this seems surprising , because the parton luminosities of bottom - quarks in a proton should be highly suppressed compared to the light flavours .
there are two reasons for the relatively large cross section of the purely bottom - induced processes .
first , the partonic cross sections of @xmath22 and @xmath23 scattering are strongly enhanced for large @xmath24 values . in this region
the parton processes with bottom - quarks in the initial state can be several orders of magnitude larger than the quark antiquark - induced process .
second , the bottom - quark pdf is essentially obtained by multiplying the gluon distribution in the proton with the splitting function of a gluon into a bottom - quark pair . because of the high gluon luminosities at low energies , the bottom - quark pdf becomes of the order of a few percent relative to the pdf s of the light flavours .
this , together with the large partonic contributions is responsible for the relatively large bottom - induced differential cross section .
the argumentation implies that the main contribution from @xmath22 , @xmath25 and @xmath23 scattering comes from the low @xmath18 region , while for high @xmath18 values these effects are small
. it might be interesting to study whether the @xmath1-pdf could be further constrained using @xmath23 production at low @xmath18 . + as shown above , the leading order contributions from electroweak gauge boson exchange are always negligible for the study of @xmath1-jet production .
this is also true in the context of nlo corrections with an expected size of serveral percent relative to the leading order distributions .
moreover , we have shown that the qcd contributions from processes with two bottom - quarks in the initial state are unimportant for the study of @xmath18-distributions at large transverse momentum .
in particular with regard to the sudakov logarithms becoming important at high energies this approximation is justified .
consequently the weak @xmath26 corrections to @xmath27 , @xmath28 and @xmath29 will not be taken into account . for further details
we refer to .
( -8.7,2.8)(a ) ( -1.25,2.8)(b ) for di - jet production the partonic final state consists of light quarks and/or gluons .
similiar to @xmath1-jet production we distinguish between quark- and gluon induced processes .
@xmath30 the lo @xmath18-distribution and its composition at the lhc is shown in fig .
[ fig : di - jetlo ] . for di - jet events(upper figure ) at the lhc ( @xmath31 tev ) and
the relative compsosition normalized to the born cross section ( lower figure).,width=302 ] the differential cross section is dominated by contributions coming from the qcd processes .
the contributions of @xmath5 and @xmath4 are at the permille level for transverse momenta up to 1 tev . for higher @xmath18 values
the electroweak contributions remain insignificant while the mixed qcd - electroweak contributions yield up to 5% to the differential cross section at @xmath32 2 tev . in context of
the sudakov logarithms the four gluon process will not receive weak corrections within the sm at nlo accuracy .
therefore in the lower figure of fig .
[ fig : di - jetlo ] the relative impact of this process on the @xmath18-distribution is shown in magenta .
for @xmath33 gev @xmath34 provides the major contribution to the differential cross section . with increasing @xmath18 values the relative contribution drops but remains non - negligible for the depicted @xmath18 region . for @xmath35 tev 90% of the cross section
is delivered by the remaining qcd processes , which receive nlo weak corrections .
+ the discussion shows that the contributions of @xmath4 are always negligible for di - jet events at the lhc .
investigating the nlo @xmath36 corrections to di - jet production the @xmath5 and @xmath3 contributions can not be seperated and therefore qcd and weak corrections must be calculated .
moreover it was shown , that the relative contribution of the partonic channel @xmath34 is of the order of 5 - 20% in the high energy regime , which justifies the calculation of the weak corrections .
as before we subdivide the partonic channels in contributions from quark - induced and gluon - induced processes .
for the calculation of the next - to - leading order weak corrections the t hooft - feynman gauge with gauge parameters set to 1 is used .
the longitudinal degrees of freedom of the massive gauge bosons @xmath37 and @xmath38 are thus represented by the goldstone fields @xmath39 and @xmath40 . since all incoming and outgoing partons are massless there are no contributions from the goldstone boson @xmath39 and the higgs boson ( also not @xmath40 for di - jet production ) .
ghost fields do not contribute at the order under consideration .
the virtual contributions to quark - induced processes contain infrared and ultraviolet singularities , while the gluon - induced ones are infrared finite .
concerning @xmath26 corrections to qcd processes only wave function renormalisation is needed and no mass or coupling renormalisation has to performed
. for the qcd corrections to @xmath5 processes the renormalisation is performed in the @xmath41 scheme .
the quark - induced contributions consist of virtual and real corrections and to handle the ir singularities we use the dipole subtraction formalism by .
in addition the results for @xmath1-jet production were cross - checked by implementing a phase space slicing method @xcite .
the comparison between slicing and subtraction method is shown in fig .
[ fig : cut - study ] + for the subsequent discsusion @xmath42 tev will be adopted .
the relative corrections for @xmath1-jet production to the leading order @xmath18-distributions are shown in fig .
[ fig : pt - nlo - bjet - relative - lhc ] . for single @xmath1-tag events ( upper figure )
the relative corrections are always negative and of the order of a few permille up to @xmath43 for 50 gev @xmath44 250 gev . for 250 gev @xmath45
1 tev the weak nlo contributions vary between @xmath43 and @xmath46 , compared to the leading order distribution , a consequence of the sudakov logarithms . in the high energy regime ( @xmath47 tev ) the relative corrections amount to @xmath48 and will even reach @xmath49 for @xmath50 tev . + the lower figure in fig .
[ fig : pt - nlo - bjet - relative - lhc ] shows the relative corrections for double @xmath1-tag events at the lhc . despite the strong suppression of @xmath51 at leading order ( see fig .
[ fig : pt - bjet - lhc](b ) ) a small remnant of the enhancement from virtual top - quarks is visible in the double @xmath1-tag case .
with increasing @xmath18 the sudakov logarithms dominate the shape of the weak nlo contributions and yield relative corrections between @xmath43 and @xmath52 ( 250 gev @xmath45 1 tev ) . at the highest @xmath18-values considered relative corrections up to @xmath49
are observed .
[ fig : pt - cut - nlo - bjet - relative - lhc ] ( upper figure ) shows the integrated cross section for single @xmath1-tag events at the lhc together with an estimate of the statistical error based on an integrated luminosity of 200 @xmath53 . the same composition in shape and magnitude
is observed as for the differential distribution .
the statistical error estimate matches the size of the weak corrections up to @xmath54 tev . for higher @xmath18-values
the rate drops quickly and it will be difficult to observe the effect of the weak corrections . for double @xmath1-tag events ( lower figure )
we find again the smoothing of the @xmath55-threshold for the relative weak corrections , while the composition of the curve for @xmath56 gev is very similiar to the already discussed differential distribution . considering the statistical error , the slight increase from virtual top - quarks will not be observable at the lhc . for @xmath18-values between @xmath57-@xmath58 gev the weak corrections are larger than the statistical error , above @xmath59 tev they are comparable or smaller . + in fig .
[ fig : pt - lhc10 ] we show results for the lhc operating at @xmath60 tev .
the absolute cross sections are by more than a factor two lower , due to the lower parton luminosities .
however the impact of the electroweak corrections is nearly the same and qualitatively similiar results are obtained for the relative nlo corrections .
-tag ( upper figure ) and double @xmath1-tag ( lower figure ) events at the lhc ( @xmath42 tev).,title="fig:",width=302 ] -tag ( upper figure ) and double @xmath1-tag ( lower figure ) events at the lhc ( @xmath42 tev).,title="fig:",width=302 ] at the lhc for single @xmath1-tag ( upper figure ) and double @xmath1-tag events ( lower figure ) .
the red lines give an estimate on the statistical uncertainty as described in the text.,title="fig:",width=302 ] at the lhc for single @xmath1-tag ( upper figure ) and double @xmath1-tag events ( lower figure ) .
the red lines give an estimate on the statistical uncertainty as described in the text.,title="fig:",width=302 ] ( -1.3,10.45)(a ) ( 7.2,10.45)(b ) + for di - jet production we present a preliminary result for the realtive weak corrections to the @xmath18-distribution in fig .
[ fig : pt - dijet - nlo ] .
we find similiar results like for the @xmath1-jet case .
however the magnitude of the weak corrections is smaller . for @xmath61
gev the relative corrections are negative and of the order of a few percent .
for @xmath62 1 tev the nlo weak corrections become of the order of @xmath48 up to @xmath63 at @xmath64 tev .
tev).,scaledwidth=49.0% ] ( 0.8,7)preliminary
we have presented numerical predictions for the weak nlo corrections to @xmath1-jet and di - jet production at the lhc . for @xmath1-jet production
we find the nlo weak corrections decrease the lo @xmath18-distribution by -10% for transverse momenta around 1 tev for single @xmath1-tag as well as double @xmath1-tag events .
these effects are larger than the anticipated statistical uncertainty .
the preliminary result for di - jet production indicates that the weak corrections lower the lo @xmath18-distribution by several percent for high transverse momenta at the lhc .
[ sec : results ] p. nason , s. dawson and r. k. ellis , nucl .
b * 327 * , 49 ( 1989 ) [ erratum - ibid .
b * 335 * , 260 ( 1990 ) ] .
w. beenakker , h. kuijf , w. l. van neerven and j. smith , phys .
d * 40 * ( 1989 ) 54 .
r. k. ellis and j. c. sexton , nucl .
b * 269 * ( 1986 ) 445 .
w. t. giele , e. w. n. glover and d. a. kosower , nucl .
b * 403 * , 633 ( 1993 ) [ arxiv : hep - ph/9302225 ] .
j. h. khn , a. kulesza , s. pozzorini and m. schulze , nucl .
b * 727 * , 368 ( 2005 ) [ arxiv : hep - ph/0507178 ] .
j. h. khn , a. kulesza , s. pozzorini and m. schulze , jhep * 0603 * , 059 ( 2006 ) [ arxiv : hep - ph/0508253 ] . j. h. khn , a. kulesza , s. pozzorini and m. schulze , nucl .
b * 797 * , 27 ( 2008 ) [ arxiv:0708.0476 [ hep - ph ] ] .
s. moretti , m. r. nolten and d. a. ross , phys .
b * 639 * , 513 ( 2006 ) [ erratum - ibid .
b * 660 * , 607 ( 2008 ) ] [ arxiv : hep - ph/0603083 ] .
s. moretti , m. r. nolten and d. a. ross , nucl .
b * 759 * , 50 ( 2006 ) [ arxiv : hep - ph/0606201 ] .
w. beenakker , a. denner , w. hollik , r. mertig , t. sack and d. wackeroth , nucl .
b * 411 * ( 1994 ) 343 .
j. h. khn , a. scharf and p. uwer , eur .
j. c * 45 * ( 2006 ) 139 [ arxiv : hep - ph/0508092 ] .
j. h. khn , a. scharf and p. uwer , eur .
j. c * 51 * ( 2007 ) 37 [ arxiv : hep - ph/0610335 ] . w. bernreuther , m. fuecker and z. g. si , phys .
d * 74 * , 113005 ( 2006 ) [ arxiv : hep - ph/0610334 ] .
j. h. kuhn and a. a. penin , arxiv : hep - ph/9906545 . j. h. kuhn , a. a. penin and v. a. smirnov , eur . phys .
j. c * 17 * , 97 ( 2000 ) [ arxiv : hep - ph/9912503 ] .
j. h. kuhn , s. moch , a. a. penin and v. a. smirnov , nucl .
b * 616 * , 286 ( 2001 ) [ erratum - ibid .
b * 648 * , 455 ( 2003 ) ] [ arxiv : hep - ph/0106298 ] .
b. feucht , j. h. kuhn , a. a. penin and v. a. smirnov , phys .
lett . * 93 * , 101802 ( 2004 ) [ arxiv : hep - ph/0404082 ] .
b. jantzen , j. h. kuhn , a. a. penin and v. a. smirnov , nucl .
b * 731 * , 188 ( 2005 ) [ erratum - ibid .
b * 752 * , 327 ( 2006 ) ] [ arxiv : hep - ph/0509157 ] .
m. beccaria , g. montagna , f. piccinini , f. m. renard and c. verzegnassi , phys .
d * 58 * ( 1998 ) 093014 [ arxiv : hep - ph/9805250 ] .
p. ciafaloni and d. comelli , phys .
b * 446 * ( 1999 ) 278 [ arxiv : hep - ph/9809321 ] .
v. s. fadin , l. n. lipatov , a. d. martin and m. melles , processes , phys .
d * 61 * ( 2000 ) 094002 [ arxiv : hep - ph/9910338 ] .
m. beccaria _
_ , phys . rev .
d 61 ( 2000 ) 011301 ; d 61 ( 2000 ) 073005 .
a. denner and s. pozzorini , eur .
j. c * 18 * ( 2001 ) 461 [ arxiv : hep - ph/0010201 ] .
a. denner and s. pozzorini , eur .
j. c * 21 * ( 2001 ) 63 [ arxiv : hep - ph/0104127 ] .
j. h. kuhn , a. scharf and p. uwer , arxiv:0909.0059 [ hep - ph ] . | simultaneously with the turning - on of the large hadron collider ( lhc ) the data taking for jet production rates will start .
the study of these production rates is an important test of the standard model in the new energy regime accessible at the lhc . in order to discriminate possible extensions of the standard model
accurate theoretical predictions are needed .
we investigate the next - to - leading order weak corrections for bottom - quark jet and di - jet production at the lhc and find detectable effects in transverse momentum distributions . |
the nonlinear schrdinger equation ( nls ) is a fundamental dispersive partial differential equation .
nls can be used in many physical nonlinear systems such as quantum many body systems , optics , hydrodynamics , acoustics , quantum condensates , and heat pulses in solids . in this article
, we consider the following nonlinear schrdinger equation with nonlinear boundary condition on the ( right ) half - line .
@xmath9 where @xmath10 is a complex valued function , the real variables @xmath11 and @xmath12 are space and time coordinates , and @xmath13 denote partial derivatives with respect to time and space .
the constant parameters satisfy @xmath14 , and @xmath15 .
when @xmath16 , the boundary condition reduces to the classical homogeneous neumann boundary condition . when @xmath17 , the boundary condition is the classical homogeneous robin boundary condition .
when @xmath18 and @xmath19 are both non - zero as in the present case , the boundary condition can be considered as a nonlinear variation of the robin boundary condition .
our main goal is to solve the classical local well - posedness problem for .
more precisely , we will prove the local existence and uniqueness for together with the continuous dependence of solutions on the initial data @xmath20 , which is taken from an @xmath21based inhomogeneous sobolev space @xmath5 with @xmath22 we will also deduce a blow - up alternative for the solutions of in the @xmath23-sense . the well - posedness problem will be considered in the function space @xmath24 , which is the set of those elements in @xmath25;h^s(\mathbb{r}_+))\cap c(\mathbb{r}_+^x;h^{\frac{2s+1}{4}}(0,t))\ ] ] that are bounded with respect to the norm @xmath26 .
this norm is defined by @xmath27}\|f(\cdot , t)\|_{h^s(\mathbb{r_+})}+\sup_{x\in\mathbb{r}_+}\|f(x,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}.\ ] ] it is well - known that the trace operators @xmath28 and @xmath29 are well - defined on @xmath5 when @xmath30 and @xmath31 , respectively . therefore , both @xmath32 and @xmath33 make sense if @xmath31 .
hence , we will assume the compatibility condition @xmath34 when @xmath31 on the initial data to comply with the desire that the solution be continuous at @xmath35 .
now we can state our main result .
[ mainresult1 ] let @xmath36 be arbitrary , @xmath37 , @xmath38 , @xmath14 , @xmath39 together with @xmath40 whenever @xmath31 .
we in addition assume the following restrictions on @xmath41 and @xmath19 : * if @xmath42 is integer , then @xmath43 if @xmath41 is an odd integer and @xmath44\ge s-1 $ ] if @xmath41 is non - integer .
* if @xmath42 is non - integer , then @xmath45 if @xmath41 is an odd integer and @xmath44\ge [ s]$ ] if @xmath41 is non - integer .
* @xmath46 if @xmath19 is an odd integer and @xmath47\ge \left[\frac{2s-1}{4}\right]$ ] if @xmath19 is non - integer . then , the following hold true . * local existence and uniqueness : there exists a unique local solution @xmath48 of for some @xmath49 $ ] . * continuous dependence : if @xmath50 is a bounded subset of @xmath5 , then there is @xmath51 ( depends on the diameter of @xmath50 ) such that the flow @xmath52 is lipschitz continuous from @xmath50 into @xmath53 . *
blow - up alternative : if @xmath54 is the set of all @xmath55 $ ] such that there exists a unique local solution in @xmath53 , then whenever @xmath56 , it must be true that @xmath57 .
if @xmath58 or @xmath41 is even , then the assumptions on @xmath41 given in ( a1)-(a2 ) in theorem [ mainresult1 ] are redundant .
the same remark applies to @xmath19 when @xmath59 or @xmath19 is even . in the above theorem , when @xmath60 , the equation is understood in the @xmath21sense .
however , if @xmath61 , the equation should be understood in the distributional sense , namely in the sense of @xmath62 . for low values of @xmath42 , the boundary and
the initial condition can be understood in the sense of definition 2.2 in @xcite . to the best of our knowledge ,
the model given in has only been studied in the case that @xmath63 , and @xmath64 by ackleh - deng @xcite . in @xcite ,
the main equation was only linear .
more precisely , the authors studied the following .
@xmath65 ackleh - deng @xcite proved that if @xmath66 , then there is @xmath51 such that possesses a unique local solution @xmath67 .
moreover , it was shown in @xcite that ( large ) solutions with negative initial energy blow - up if @xmath68 and are global otherwise .
therefore , @xmath69 was considered the critical exponent for .
obtaining local existence and uniqueness consisted of two steps .
first , the authors studied the linear schrdinger equation with an inhomogeneous neumann boundary condition on the half - line .
secondly , they used a contraction argument once the representation formula for solutions was restricted to the boundary point @xmath70 . in other words , the contraction argument was used on a function space which included only time dependent elements .
unfortunately , the same technique can not be applied in the presence of the nonlinear source term @xmath71 in the main equation .
the reason is that even if the representation formula can still be restricted to the point @xmath70 , the sought after fixed point in the representation formula would also depend on the space variable .
therefore , one can no longer use a simple contraction argument on a function space which includes only time dependent elements .
we are thus motivated to use a contraction argument on a function space which includes elements that depend on both time and space variables .
of course , this requires nice linear and nonlinear space - time estimates .
the drawback of the technique used in @xcite is that the initial data has been assumed to be too smooth compared to the regularity of the solutions obtained .
it is well - known from the theory of the linear schrdinger equation that solutions are of the same class as the initial state . from this point of view
, the generation of @xmath72 solutions with @xmath73 data seems far from optimal .
we are thus inclined to obtain a regularity theory which shows that @xmath23 initial data generates @xmath23 solutions . regarding nonlinear boundary conditions , we are aware of very few other results for schrdinger equations , see for example @xcite and @xcite . in @xcite ,
the authors study the schrdinger equation with nonlinear , attractive , and dissipative boundary conditions of type @xmath74 where @xmath75 is a monotone function with the property that the corresponding evolution operator generates a strongly continuous contraction semigroup on the @xmath4-level .
the more recent paper @xcite studies schrdinger equation with wentzell boundary conditions .
this work also uses the fact that the wentzell boundary condition provide a semigroup in an appropriate topology . in the present case , due to the fact that @xmath18 is not a purely imaginary number ,
the problem does not have a monotone structure , and the method of @xcite , @xcite can not be applied here .
a common strategy for proving well - posedness of solutions to pdes with nonlinear terms relies on two classical steps : ( 1 ) obtain a good linear theory with non - homogeneous terms ; ( 2 ) establish local well - posedness for the nonlinear model by a fixed point argument . obtaining a good linear theory with non - homogeneous terms is a subtle point for boundary value problems , especially those with low - regularity boundary data .
one might attempt to extend the boundary data into the domain and homogenize the boundary condition .
however , this approach in general requires a high regularity boundary data ( @xcite , @xcite ) , as opposed to the rough boundary situation as in the present paper for low values of @xmath42 .
there are different approaches one can follow to study a linear pde with an inhomogeneous boundary data on the half - line without employing an extension - homogenization approach , though .
for example , colliander - kenig @xcite used a technique on the kdv equation by replacing the given initial - boundary value problem with a forced initial value problem where the forcing is chosen in such a way that the boundary condition is satisfied by inverting a riemann - liouville fractional integral .
holmer @xcite applied this technique on nonlinear schrdinger equations with inhomogeneous dirichlet boundary conditions on the half line .
a second approach is to obtain norm estimates on solutions by using a representation formula , which can be easily obtained through a laplace / fourier transform .
this technique has been used for example by kaikina in @xcite for nonlinear schrdinger equations with inhomogeneous neumann boundary conditions and by bona - sun - zhang in @xcite for inhomogeneous dirichlet boundary conditions . in @xcite ,
the well - posedness result assumes the smallness of the given initial - boundary data while the results of @xcite have global character in this sense .
although nonlinear schrdinger equations with inhomogeneous boundary conditions have been studied to some extent , most of these papers were devoted to inhomogeneous dirichlet boundary conditions ; see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
there are relatively less results on inhomogeneous neumann boundary conditions ; see @xcite , @xcite , @xcite , @xcite , @xcite . in @xcite and @xcite ,
well - posedness is obtained under smooth boundary data .
relatively less smooth boundary data was treated in @xcite using strichartz estimates , but the regularity results were not optimal . in @xcite ,
the smallness of initial and boundary data was crucial . in @xcite ,
the focus was on the existence of weak solutions , and questions concerning continuity in time , uniqueness , and continuous dependence on data were not studied . in the present paper , we draw a more complete and optimal well - posedness picture where the spatial domain is half - line . in this paper
, we will follow a step - by - step approach to prove theorem [ mainprob ] : _ step 1 _ : we will first study the linear schrdinger equation with inhomogeneous terms both in the main equation and in the boundary condition . this problem is written in . our aim in this step is to derive optimal norm estimates with respect to regularities of the initial state @xmath20 , boundary data @xmath76 , and nonhomogeneous source term @xmath77 .
this linear theory is constructed in section [ sectionlin ] by adapting the method of @xcite to nonhomogeneous neumann boundary conditions .
_ step 2 _ : in the second step , we will replace the nonhomogeneous source term @xmath78 in with @xmath79 as in .
we will use a contraction mapping argument to prove the existence and uniqueness of local solutions together with continuous dependence on data .
the blow - up alternative will be obtained via a classical extension - contradiction argument .
this step is treated in sections [ locext ] - [ blowsec ] .
_ step 3 _ : in this step , we will replace the boundary data @xmath80 in with @xmath81 , and @xmath77 with @xmath82 .
arguments similar to those in step 2 will eventually give the well - posedness in the presence of nonlinear boundary conditions .
the only difference is that the contraction argument must be adapted to deal with the nonlinear effects due to the nonlinear boundary source .
this is given in section [ sectionnonlin2 ] .
step 2 is indeed optional .
one can directly run the contraction and blow - up arguments with nonlinear boundary conditions .
however , it is useful to include the general theory of nonlinear schrdinger equations with inhomogeneous neumann boundary conditions to study other related problems in the future .
in this section , we study the nonhomogeneous linear schrdinger equation with nonhomogeneous neumann boundary condition
. we will later apply this linear theory to obtain the local well - posedness for nonlinear schrdinger equations first with inhomogeneous neumann boundary conditions and then with nonlinear boundary conditions . in order to obtain a sufficiently nice linear theory ,
we adapt the method presented for nonhomogeneous dirichlet boundary conditions in @xcite to the case with nonhomogeneous neumann boundary conditions .
we consider the following linear model @xmath83 where @xmath77 and @xmath76 lie in appropriate function spaces .
suppose @xmath84 , @xmath8 in .
it is well - known from the trace theory that both @xmath33 and @xmath85 make sense when @xmath31 .
therefore , one needs to assume the zeroth order compatibility condition @xmath86 when @xmath87 in order to get continuous solutions at @xmath35 .
as the value of @xmath42 gets higher , one needs to consider more compatibility conditions .
for example , if @xmath88 ( @xmath89 ) , then the @xmath90-th order compatibility condition is defined inductively : @xmath91 @xmath92 provided that @xmath77 is also smooth enough for traces to make sense .
if one wants to add the end point cases @xmath93 to the analysis , then _ global compatibility conditions _ must be assumed ( see for example @xcite for a discussion of local and global compatibility conditions in the case of dirichlet boundary conditions ) .
we will first deduce a representation formula for solutions of the following linear model with an inhomogeneity on the boundary .
@xmath94 we will study the above model by constructing an evolution operator which acts on the boundary data
. we will start by taking a laplace ( in time ) - fourier ( in space ) transform of the given model . in order to do that
, we will first extend the boundary data to the whole line utilizing the following lemma .
[ extension]let @xmath6 , @xmath95 with @xmath96 if @xmath31 .
then , there exists @xmath97 with compact support in @xmath98 which extends @xmath76 so that @xmath99 also has compact support in @xmath98 and @xmath100 for some @xmath101 which is independent of @xmath102 . if @xmath103 , we have @xmath104 .
now we take the zero extension of @xmath76 onto @xmath105 , say we get @xmath106 .
then we set @xmath107 if @xmath108 , then @xmath109 . in this case , we first take an extension @xmath110 of @xmath76 onto @xmath105 so that @xmath111 by using the fact that @xmath112 secondly , the restriction @xmath113 will satisfy @xmath114 .
now we can take the zero extension , say @xmath115 , of @xmath116 onto @xmath105 so that @xmath117 with @xmath118 independent of @xmath102 . by the previous inequalities , we get @xmath119 with @xmath118 independent of @xmath102 .
then we pick a function @xmath120 so that @xmath121 on @xmath122 and @xmath123 on @xmath124 .
now we consider @xmath125 , which is of course in @xmath126 , since @xmath126 is a banach algebra when @xmath31 .
finally , we set @xmath127 note that @xmath128 where the positive constant @xmath118 does not depend on @xmath102 , since all the extensions in the above paragraph and the multiplication by @xmath129 are continuous operators between corresponding sobolev spaces whose norms do not depend on the initial domain @xmath122 .
moreover , we set up @xmath130 in such a way that its average is zero .
hence , its antiderivative @xmath99 is compactly supported and therefore belongs to the space @xmath131 since @xmath132 is compactly supported with support in @xmath133 by the poincar inequality we have @xmath134 hence @xmath135 for some @xmath101 .
now we consider the following model , which is an extended - in - time version of .
@xmath136 where @xmath130 is the extension of @xmath76 , as in lemma [ extension ] .
we first take the laplace transform of in @xmath12 to get @xmath137 with @xmath138 , where @xmath139 denotes the laplace transform of @xmath140 .
the solution of is @xmath141 where @xmath142 solves @xmath143 together with @xmath144 .
then , @xmath145 where @xmath146 ( fixed ) , solves . by passing to the limit in @xmath147 as @xmath148 and applying change of variables ,
we can rewrite @xmath10 as follows : @xmath149 note , that @xmath150 is a solution of .
we define @xmath151 for @xmath152 and zero otherwise .
let @xmath153 be the inverse fourier transform of @xmath154 , that is @xmath155 for @xmath156 .
similarly , we define @xmath157 for @xmath152 and zero otherwise .
let @xmath158 be the inverse fourier transform of @xmath159 , that is @xmath160 for @xmath156 .
now , for @xmath161 , we can write @xmath162(x):=[w_{b,1}(t)h_{e}](x)+[w_{b,2}(t)h_{e}](x)\ ] ] where @xmath163(x):=\int_{-\infty}^\infty\exp(-i\beta^2 t+i\beta x)\hat{\phi}_{h_{e}}(\beta)d\beta\ ] ] and @xmath164(x):=\int_{-\infty}^\infty\exp(i\beta^2 t-\beta x)\hat{\psi}_{h_{e}}(\beta)d\beta.\ ] ] note that we can extend @xmath165 to @xmath105 without changing its definition .
for such an extension we have the following lemma : [ wb1lem ] @xmath166(x)$ ] solves the initial value problem @xmath167 by direct calculation , we have @xmath168[w_{b,1}(t)h_{e}](x)=0,\ ] ] and @xmath169 we deduce from the above lemma that we can get space time estimates on @xmath165 by using the well - known linear theory of schrdinger equations on @xmath170 these estimates are given in section [ rtheory ] .
we extend @xmath171(x)$ ] to @xmath105 by setting @xmath164(x):=\int_{-\infty}^\infty\exp(i\beta^2 t-\beta |x|)\hat{\psi}_{h_{e}}(\beta)d\beta.\ ] ] however , if @xmath31 , then this extension would not be differentiable at @xmath70 . therefore ,
if @xmath31 , we can not directly use the linear theory of schrdinger equations on @xmath105 to estimate various norms of the term @xmath172 .
this makes it necessary to obtain space - time estimates for @xmath172 directly by using its definition .
the relation between regularities of @xmath173 and the regularity of the boundary data @xmath76 is given by the following lemma .
[ phivsh ] let @xmath174 , @xmath8 such that @xmath96 if @xmath31 .
then , @xmath175 .
@xmath176 upon change of variables , the last term in can be rewritten and estimated as follows .
@xmath177 where we use the relationships , @xmath178 and @xmath179 in the first inequality .
the last estimate combined with lemma [ extension ] implies that @xmath180 .
we can repeat the same argument for @xmath181 , too .
a given pair @xmath182 is said to be _ admissible _ if @xmath183 for @xmath184 .
now , we will present several space - time estimates for the second part of the evolution operator @xmath185 .
let @xmath174 and @xmath36 .
then , there exists @xmath101 ( independent of @xmath102 ) such that @xmath186}\|w_{b,2}(\cdot)h_{e}\|_{h^s}\le c(1+t ) \|h\|_{h^{\frac{2s-1}{4}}(0,t)}\ ] ] for any @xmath8 with @xmath96 if @xmath31 .
we can rewrite @xmath187(x)$ ] as @xmath188(x):=\int_{-\infty}^\infty k_t(x , y){\psi}_{h_e}(y)dy=:\mathcal{k}(t)\psi_{h_e}\ ] ] where @xmath189 it is proven in ( * ? ? ?
* proposition 3.8 ) that @xmath190 for an admissible @xmath182 .
similarly , taking one derivative in @xmath11 variable , one gets @xmath191\|_{l^q(0,t;l^r)}\lesssim \|\partial_x[\psi_{h_e}]\|_{l^2}.\ ] ] now , one can interpolate and use the proof of lemma [ phivsh ] to obtain @xmath192 for @xmath193 $ ] . for larger @xmath42
, one can differentiate and interpolate again .
finally , follows by taking @xmath194 in .
now , follows from and lemma [ extension ] .
let @xmath174 and @xmath36 .
then , there exists @xmath101 ( independent of @xmath102 ) such that @xmath195 for any @xmath8 with @xmath96 if @xmath31 .
this result is an application of theorem 4.1 in @xcite .
for @xmath196 ( integer ) @xmath197 upon interpolation , the result follows in the case that @xmath76 , @xmath130 , and @xmath198 are smooth , then a density argument finishes the proof .
now , follows from and lemma [ extension ] .
we take an extension of @xmath20 to @xmath105 , say @xmath199 such that @xmath200 .
therefore , @xmath201 solves the problem @xmath202 where @xmath203 is the evolution operator for the linear schrdinger equation .
similarly , if @xmath204 is an extension of @xmath77 , then the solution of the non - homogeneous cauchy problem @xmath205 can be written as @xmath206 therefore , if we define @xmath207_e(t))\ ] ] with @xmath208 and @xmath209 then @xmath210 will solve @xmath211 in the formula we have given , @xmath212 and @xmath213 make sense only if @xmath214 . in other cases , we take those boundary traces equal to zero in the representation formula
. we will utilize the following space and time estimates on @xmath105 for the evolution operator of the linear schrdinger equation @xcite . note that these estimates can be directly applied to the first part @xmath215 of the boundary evolution operator .
[ rtheory1 ] let @xmath216 , @xmath36 , @xmath217 , and @xmath218 .
then , there exists @xmath219 such that @xmath220}\|u(\cdot , t)\|_{h^s}+\sup_{x\in\mathbb{r}}\|u(x,\cdot ) \|_{h^{\frac{2s+1}{4}}(0,t)}\le c \|\phi\|_{h^s}.\ ] ] [ varpar ] let @xmath36 , @xmath221 , and @xmath222 then , for any @xmath223 , there exists a constant @xmath224 such that @xmath225}\|u(\cdot , t)\|_{h^s}+\sup_{x\in\mathbb{r}}\|u(x,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}\le c \|f\|_{l^{1}(0,t;h^{s})}.\ ] ] combining lemmas [ wb1lem]-[varpar ] , we have the following regularity theorems for the linear model .
[ regthm1]let @xmath36 , and @xmath226 .
then , there exists @xmath101 ( independent of t ) such that for any @xmath8 with @xmath96 if @xmath31 , @xmath227 satisfies @xmath228}\|u(\cdot , t)\|_{h^s(\mathbb{r}_+)}+\sup_{x\in\mathbb{r}_+}\|u(x,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}\\ \le c(1+t)\|h\|_{h^{\frac{2s-1}{4}}(0,t)}.\end{gathered}\ ] ] let @xmath36 , @xmath229 , @xmath8 , @xmath230 , @xmath231 , and if @xmath87 , we assume the zeroth order compatibility condition @xmath232
. then there exists @xmath101 ( independent of @xmath102 ) such that the solution @xmath233 of satisfies @xmath228}\|u(x,\cdot)\|_{h^s(\mathbb{r}_+)}+\sup_{x\in\mathbb{r}_+}\|u(x,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}\\ \le c\left(\|u_0\|_{h^s(\mathbb{r}_+)}+(1+t)\|h\|_{h^{\frac{2s-1}{4}}(0,t)}+\|f\|_{l^{1}(0,t;h^s(\mathbb{r}_+))}\right).\end{gathered}\ ] ] the optimal local smoothing estimate for the schrdinger evolution operator is @xmath234 ; see for instance @xcite .
this is why we consider the space @xmath24 defined in section [ intro ] as our solution space .
it is shown in @xcite and @xcite that the natural space for the boundary data @xmath76 is @xmath235 , when one considers dirichlet boundary conditions . since one can formally think that one derivative in the space variable is equivalent to @xmath236 derivatives in the time variable , we are inclined to consider @xmath237 as the natural space for the boundary data @xmath76 when we consider neumann boundary conditions .
in this section , we study nonlinear schrdinger equations with nonhomogeneous neumann type boundary data . more precisely , we consider the following model : @xmath238 where @xmath71 , @xmath239 , @xmath240 , @xmath231 , and @xmath241 here , we consider two problems .
the first one is the open - loop well - posedness problem when @xmath76 is taken as a time dependent function in the sobolev space @xmath242 .
the second one is the closed - loop well - posedness problem when @xmath76 is taken as a function of @xmath243 in the form @xmath244 with @xmath245 . in order to prove the local existence of solutions we will use the contraction mapping argument . for the contraction mapping argument
, we will use the following operator on a closed ball @xmath246 in the function space @xmath53 for appropriately chosen @xmath247 and @xmath55 $ ] .
@xmath248(t):=w_\mathbb{r}(t)u^*_0-i\int_0^tw_\mathbb{r}(t-\tau)f(u^*(\tau))d\tau\\+w_b(t)([h - g - p(u^*)]_e)\end{gathered}\ ] ] with @xmath249 and @xmath250(t)=-i\partial_x\int_0^tw_\mathbb{r}(t-\tau)f(u^*(\tau))d\tau|_{x=0}.$ ] here , @xmath212 and @xmath213 make sense only if @xmath214 . for @xmath251 ,
we take these boundary traces equal to zero in . in order to use the banach fixed point theorem
, we have to show that @xmath252 maps @xmath246 onto itself , and moreover that it is a contraction on the same set .
therefore , we will estimate each term in with respect to the norm defined in . by lemma [ rtheory1 ] , @xmath253 *
if @xmath42 is integer , then assume that @xmath43 if @xmath41 is an odd integer and @xmath44\ge s-1 $ ] if @xmath41 is non - integer .
* if @xmath42 is non - integer , then assume that @xmath45 if @xmath41 is an odd integer and @xmath44\ge [ s]$ ] if @xmath41 is non - integer .
see lemma 4.10.2 @xcite for @xmath42 being an integer and lemma 3.10(2 ) @xcite for @xmath41 being an even number .
therefore , we will only consider the cases with @xmath42 being a non - integer , and @xmath41 being an odd integer or non - integer .
let us first consider the case @xmath259 . by the chain rule ( theorem a.7 @xcite ) , @xmath260 since @xmath261 , we have @xmath262 where the last inequality follows by the sobolev embedding @xmath263 for @xmath264 . also , @xmath265 it follows that @xmath266 . on the other hand , @xmath267 , where the inequality follows by the sobolev s embedding @xmath268 for @xmath30 .
hence , we have just shown that @xmath269 .
now , consider the case @xmath270 for some positive integer @xmath271 and @xmath272 .
then , @xmath273 where @xmath274 is a sum of the terms of type @xmath275 where @xmath90 ranges from @xmath276 up to @xmath277 and @xmath278 . by the fractional version of the leibniz rule @xcite , we can write @xmath279 together with @xmath280 , @xmath281 . by using the chain rule ,
the first term is estimated as @xmath282 together with @xmath283 , @xmath284 . here , we choose @xmath285 sufficiently large so that @xmath286 .
therefore , @xmath287 and @xmath288 .
if @xmath276 ( therefore @xmath289 ) , then the second term can be estimated as @xmath290 . in the last estimate ,
if @xmath291 , then we choose @xmath292 as @xmath293 , otherwise we can use any @xmath294 . if @xmath295 , then using hlder s inequality @xmath296 where @xmath297 and @xmath298 .
hence , it follows that we always have @xmath299 .
the third term can be easily estimated as @xmath300 regarding the fourth term , the case @xmath276 is trivial .
so let us consider the case @xmath295 . in this case , applying the leibniz formula , we have @xmath301 for some @xmath302 such that @xmath303 . but the right hand side of the last inequality is dominated by @xmath304
hence , it follows that @xmath305 . by the above estimates , we deduce .
now , we consider the case @xmath270 for some positive integer @xmath271 and @xmath272
. then , @xmath308 where @xmath309 is a sum of the terms of type @xmath310 where @xmath90 ranges from @xmath276 up to @xmath277 , @xmath278 , and @xmath311 s are equal to @xmath233 or @xmath312 , except one , which is equal to @xmath313 .
now the @xmath4-norm of the term @xmath314 can be estimated in a manner similar to using the fractional leibniz rule , except we also use several applications of young s inequality to separate the products involving @xmath233 and @xmath312 .
what remains is to estimate the term @xmath315 $ ] , which can also be done as in using the fractional leibniz rule . in order to do this
, we also use the observation @xmath316 which easily follows from the fact that @xmath317 and the sobolev embedding @xmath306 for @xmath264 .
the assumption @xmath318 and @xmath319 are needed to guarantee that @xmath77 is sufficiently smooth .
the assumption @xmath318 guarantees that @xmath77 is at least @xmath320 , which is what one needs in the case @xmath42 is an integer ( see remark 4.10.3 @xcite ) . since @xmath77 is @xmath321 when @xmath41 is even , no assumption was necessary in this case .
if @xmath42 is fractional , the proof uses the @xmath322-th derivative , which forces us to make the second assumption @xmath319 .
similarly , @xmath324d\tau\right\|_{x_t^s}\\
\le \int_0^t\|f(u^*(\tau))-f(v^*(\tau))\|_{h^s}d\tau\\ \lesssim \int_0^t(\|u^*(\tau)\|_{h^s}^p+\|v^*(\tau)\|_{h^s}^p)\|u^*(\tau)-v^*(\tau)\|_{h^s}d\tau\\ \lesssim \int_0^t(\|u(\tau)\|_{h^s(\mathbb{r}_+)}^p+\|v(\tau)\|_{h^s(\mathbb{r}_+)}^p)\|u(\tau)-v(\tau)\|_{h^s(\mathbb{r}_+)}d\tau\\ \lesssim t(\|u\|_{x_t^s}^p+\|v\|_{x_t^s}^p)\|u - v\|_{x_t^s}.\end{gathered}\ ] ] for @xmath108 , we have the assumption @xmath327 , and therefore @xmath328 vanishes at @xmath70 .
moreover , the following estimate holds true .
@xmath329_e)\|_{x_t^s}\le c(1+t)\|h - g - p(u^*)\|_{h^{\frac{2s-1}{4}}(0,t)}\\ \le c(1+t)\left(\|h\|_{h^{\frac{2s-1}{4}}(0,t)}+\|g\|_{h^{\frac{2s-1}{4}}(0,t)}+\|p(u^*)\|_{h^{\frac{2s-1}{4}}(0,t)}\right).\end{gathered}\ ] ] now , let @xmath337 , @xmath338 and @xmath102 be small enough that @xmath339 .
now , if necessary we can choose @xmath102 even smaller so that @xmath252 becomes a contraction on @xmath340 , which is a complete space .
hence , @xmath252 must have a unique fixed point in @xmath246 when we look for a solution whose lifespan is sufficiently small .
[ propext]let @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath8 , and @xmath232 whenever @xmath31 .
we in addition assume ( a1)-(a2 ) given in lemma [ fest ] .
then , has a local solution @xmath48 for some @xmath341}$ ] . in the previous section ,
we have proved uniqueness in a fixed ball in the space @xmath24 .
this does not immediately tell us that the solution must also be unique in the entire space .
fortunately , this latter statement is also true . in order to show this ,
let @xmath342 be two solutions of .
then , @xmath343ds\\ + w_b(t)([p(u_2^*)-p(u_1^*)]_e)\end{gathered}\ ] ] for a.a .
@xmath344 $ ] .
regarding continuous dependence on data , let @xmath50 be a bounded subset of @xmath349 .
let @xmath350 and @xmath351 .
let @xmath352 be two solutions on a common time interval @xmath353 corresponding to @xmath354 and @xmath355 , respectively .
then @xmath356 satisfies @xmath357 now , using the linear theory together with the nonlinear @xmath23 estimates on the differences , we have @xmath358 where @xmath359 choosing @xmath360 , which depends on @xmath20 and @xmath76 ( i.e. , on the bounded set @xmath50 ) , as in the proof of the local existence , and @xmath361 accordingly small enough , we obtain @xmath362 hence , we have the following result . in this section
, we want to obtain a condition which guarantees that a given local solution on @xmath364 $ ] can be extended globally .
let s consider the set @xmath54 of all @xmath55 $ ] such that there exists a unique local solution in @xmath53 .
we claim that if @xmath56 , then @xmath365 in order to prove the claim , assume to the contrary that @xmath366 then @xmath367 and @xmath368 such that @xmath369 and @xmath370 for a fixed @xmath371 , we know that there is a unique local solution @xmath372 on @xmath373 $ ] .
now , we consider the following model .
we know from the local existence theory that the above model has a unique local solution @xmath375 on some interval @xmath376 $ ] for some @xmath377.$ ] now , choose @xmath371 sufficiently large that @xmath378 . if we set @xmath379$},\end{array }
\right.\ ] ] then @xmath233 is a solution of on @xmath380 $ ] where @xmath378 , which is a contradiction . in this section , we study the most general nonlinear model given in .
we define the operator @xmath252 as in , except that we take @xmath381 therefore , the solution operator we have to use for the contraction argument takes the following form .
the proofs of local well - posedness and blow - up alternative now follows similar to the proofs in sections [ locext ] - [ blowsec ] .
the only additional work in this part would be to get nonlinear @xmath23 estimates on the boundary trace @xmath383 , which is of course possible with assumptions on @xmath19 , which are almost equivalent to the assumptions we made on @xmath41 .
indeed , we will assume that @xmath46 if @xmath19 is an odd integer and @xmath47\ge \left[\frac{2s-1}{4}\right]$ ] if @xmath19 is non - integer .
let @xmath387 . applying the cauchy - schwartz inequality we get @xmath388 . on the other hand @xmath389 , which implies @xmath390 , hence @xmath391 . by interpolation theorem (
* theorem 12.4 , proposition 2.3 ) , @xmath392 , in which @xmath393 .
hence we obtain @xmath394 let us first consider the case @xmath19 being an odd integer . in this case , we assume @xmath46 . now ,
if @xmath395 , then we can choose @xmath396 so that @xmath397 . if @xmath398 , then we can choose @xmath399 sufficiently small so that we again have @xmath400 .
secondly , let us consider the situation for @xmath64 being a non - integer . in this case , we assume @xmath47\ge \left[\frac{2s-1}{4}\right]$ ] . if @xmath395 then we choose @xmath396 so that @xmath401 . if @xmath398 , then we choose @xmath399 sufficiently small so that @xmath402=\left[\frac{2s-1}{4}\right].$ ] now , given @xmath403 , we know that @xmath404 in particular belongs to the space @xmath405 for any @xmath406 $ ] .
so , let us take an extension of @xmath407 , say @xmath408 , so that @xmath409 see .
now , @xmath410 is an extension of @xmath411 , and therefore @xmath412 by lemma [ fest ] . by using the inequality
, we have @xmath413 combining the above estimates , we arrive at @xmath414 finally , we deduce that @xmath415_e\|_{x_t^s}\le c(1+t)\||u(0,\cdot)|^ru(0,\cdot)\|_{h^{\frac{2s-1}{4}}(0,t)}\\ \le c(1+t)t^{\frac{4\epsilon}{2s+3 + 4\epsilon}}\||u(0,\cdot)|^ru(0,\cdot)\|_{h^{\frac{2s-1}{4}+\epsilon}(0,t)}\\ \le c(1+t)t^{\frac{4\epsilon}{2s+3 + 4\epsilon}}\|u(0,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}^{r+1}\\ \le c(1+t)t^{\frac{4\epsilon}{2s+3 + 4\epsilon}}\sup_{x\in\mathbb{r}_+}\|u(x,\cdot)\|_{h^{\frac{2s+1}{4}}(0,t)}^{r+1}\\ \le c(1+t)t^{\frac{4\epsilon}{2s+3 + 4\epsilon}}\|u\|_{x_t^s}^{r+1}.\end{gathered}\ ] ] we can estimate the differences similarly .
namely , for any given @xmath416 , we have @xmath417_e\|_{x_t^s}\\ \le c(1+t)t^{\frac{4\epsilon}{2s+3 + 4\epsilon}}\left(\|u\|_{x_t^s}^{r}+\|v\|_{x_t^s}^{r}\right)\|u - v\|_{x_t^s}.\end{gathered}\ ] ] now , let @xmath420 , @xmath338 and @xmath102 small enough that @xmath421 now , if necessary we can choose @xmath102 even smaller so that @xmath252 becomes a contraction on @xmath340 , which is a complete space .
hence , @xmath252 must have a unique fixed point in @xmath246 when we look for a solution whose lifespan is sufficiently small . in order to prove uniqueness , we proceed as in section [ uniq ] , taking into account that the boundary forcing now depends on the solution itself .
so , let @xmath342 be two solutions of .
then , @xmath343ds\\ + w_b(t)\left([h(u_1(0,\cdot))-h(u_2(0,\cdot))+p(u_2^*)-p(u_1^*)]_e\right)\end{gathered}\ ] ] for a.a .
@xmath344 $ ] . then , @xmath422 now , choosing @xmath361 sufficiently small , we can subtract the last term above from the left hand side , estimate the rest of terms at the right hand side as in section [ uniq ] , and then use the gronwall s inequality to obtain @xmath346 .
the proof of continuous dependence can be done as in section [ contdep ] by taking into account that @xmath76 is now a function of @xmath243 .
for this closed loop problem , the estimate takes the following form .
@xmath423 for sufficiently small @xmath361 .
of course , for the closed loop problem @xmath50 is taken as a subset of @xmath5 with finite diameter .
the proof of the blow - up alternative is almost identical to the proof given in section [ blowsec ] , and is therefore omitted here .
the only modification is that the parameter @xmath424 in the proof given in section [ blowsec ] now depends only on @xmath425 .
t. cazenave , _ semilinear schrdinger equations _ , courant lecture notes in mathematics , 10 .
new york university , courant institute of mathematical sciences , new york ; american mathematical society , providence , ri , 2003 .
kenig , g. ponce , and l. vega , _ well - posedness and scattering results for the generalized korteweg - de vries equation via the contraction principle _
pure appl .
46 ( 1993 ) , no .
4 , 527620 .
i. lasiecka , r. triggiani , _ well - posedness and sharp uniform decay rates at the _ @xmath426__-level of the schrdinger equation with nonlinear boundary dissipation _
_ , j. evol .
( 2006 ) , no .
3 , 485537 . j.l .
lions , e. magenes , _ non - homogeneous boundary value problems and applications .
i. translated from the french by p. kenneth .
die grundlehren der mathematischen wissenschaften , band 181 .
springer - verlag , new york - heidelberg , 1972 .
t. zsar , v.k .
kalantarov , i. lasiecka , _ uniform decay rates for the energy of weakly damped defocusing semilinear schrdinger equations with inhomogeneous dirichlet boundary control _
, journal of differential equations , volume 251 , issue 7 , 1 october 2011 , pages 1841 - 1863 t. zsar , _ global existence and open loop exponential stabilization of weak solutions for nonlinear schrdinger equations with localized external neumann manipulation _ , nonlinear analysis : theory , methods & applications , volume 80 , march 2013 , pages 179 - 193 t. zsar , _ well - posedness for nonlinear schrdinger equations with boundary forces in low dimensions by strichartz estimates _ , journal of mathematical analysis and applications , volume 424 , issue 1 , 1 april 2015 , pages 487 - 508 | in this paper , we study the initial boundary value problem for nonlinear schrdinger equations on the half - line with nonlinear boundary conditions of type @xmath0 @xmath1 , @xmath2 .
we discuss the local well - posedness when the initial data @xmath3 belongs to an @xmath4-based inhomogeneous sobolev space @xmath5 with @xmath6 .
we deal with the nonlinear boundary condition by first studying the linear schrdinger equation with a time - dependent inhomogeneous neumann boundary condition @xmath7 where @xmath8 .
this latter problem is studied by adapting the method of bona - sun - zhang @xcite to the case of inhomogeneous neumann boundary conditions . |
in newmann - penrose formalism @xcite , the six components of the maxwell field tensors are represented by three complex scalars @xmath2 , @xmath3and @xmath4 . in the tetrad frame
we have chosen , @xcite they are @xmath5\ ] ] , @xmath6\ ] ] and @xmath7\ ] ] . in projections of quantities in the tetrad frame ,
the number of @xmath8 minus the number of @xmath9 is called the spin weight , while the number of @xmath10 minus the number of @xmath11 is the boost weight . as a consequence
, we see that @xmath3 has spin and boost weight equal to zero and thus behaves as a spin zero ( scalar ) field . similarly , @xmath2 and @xmath4 are spin + 1 and -1 fields . next , the maxwell equations , with source by the dirac current ( @xcite ) take on the form @xmath12 @xmath13 @xmath14 and @xmath15 where we have substituted @xmath16 , @xmath17 and @xmath18 .
the radial - temporal operator @xmath19and the angular operator @xmath20 , and @xmath21 is the conformal time @xmath22 with @xmath23 equal to the scale factor .
we can easily decouple the maxwell equations to read @xcite
@xmath24\phi_{0 } & = & \sin r\mathsf{\mathsf{\mathcal{d}}}_{-1}^{\pm}j_{0}^{\pm}+\mathcal{l}_{\textrm{1}}^{\pm}j_{\pm1}=s_{0}\label{eq : maxwell eq1}\end{aligned}\ ] ] and @xmath25\phi_{1 } & = & \pm\sin r\mathsf{\mathsf{\mathcal{d}}}_{0}^{\mp}j_{\pm1}\mp\mathcal{l}_{\textrm{0}}^{\mp}j_{0}^{\pm}=s_{\pm1}\label{eq : maxwell eq2}\end{aligned}\ ] ] in eq .
( [ eq : maxwell eq2 ] ) , it is seen that @xmath26 is the source for the radial electric field and @xmath27 for the radial magnetic field .
the eigenfunctions of the angular operator are the well known spin - weighted spherical harmonics satisfying @xmath28 ; we have identified these functions as the spherical harmonics formed with the jacobi polynomials _ viz .
_ @xmath29 while @xmath30 are usual spherical harmonics .
these are normalized .
so , @xmath31 , are complete in @xmath32 , @xmath33 , and satisfy the spin lowering operation @xmath34 .
the green s function of the scalar equation can be determined by working out the solution of the eq .
( [ eq : maxwell eq1 ] ) with point source : @xmath35g_{0}\left(\eta,\,\ , r,\,\,\theta,\,\,\phi\right)=\delta\left(r - r'\right)\delta\left(\eta-\eta'\right)\delta\left(\theta-\theta'\right)\delta\left(\phi-\phi'\right)\ ] ] by standard technique , we can make an eigenfunction expansion of the green s function@xcite as @xmath36 where the radial eighenfunctions are the normalized , appropriately weighted gegenbauer polynomials@xcite @xmath37 the addition theorem of spherical harmonics gives @xmath38 with @xmath39 the @xmath40 integral can be done by the method of residues to give @xmath41 , with these results we can rewrite the summation as @xmath42}{\left(k+1\right)}\left(\sin r\sin r'\right)^{l}\nonumber \\ & & c_{k - l}^{l+1}\left(\cos r'\right)c_{k - l}^{l+1}\left(\cos r\right)c_{l}^{\frac{1}{2}}\left(\cos\rho\right)\\ & = & -\frac{1}{4\pi}\underset{k=0}{\sum^{\infty}}\sin\left[\left(k+1\right)\left(\eta-\eta'\right)\right]c_{k}^{(1)}\left(\cos\rho\right)\end{aligned}\ ] ] where we have used the addition theorem of gegenbauer polynomial from reference @xcite and @xmath43 .
now @xmath44}{\sin\rho},\,\,\ , so\,\ , we\,\ , find\nonumber \\
g_{0 } & = -\frac{1}{4\pi\sin\rho } & \left[\delta\left(\eta-\eta'-\rho\right)-\delta\left(\eta-\eta'+\rho\right)\right]\label{eq : g0}\end{aligned}\ ] ] representing the retarded and advanced green s function .
thus , the solution for the retarded scalar field is @xmath45 we can even solve for lienard - wiechert like field in closed frw space time for a point source moving along the trajectory @xmath46given by @xmath47$]to find @xmath48}\frac{1}{\left[1+\frac{d\rho\left(\eta'\right)}{d\eta}\right]}\right|_{\eta'=\eta-\rho\left(\eta'\right)}\ ] ] in the limit to flat frw space time given by @xmath49 and @xmath50gives @xmath51 which is the familiar lienard - wiechert solution .
in this regard , let us solve for the field in the very early universe when the dirac particles are copiously produced so that all the available states are fully occupied .
when we sum the source on the left of eq .
( [ eq : maxwell eq1 ] ) , the second term containing @xmath52 will give zero .
the consequence is that there is no source for the radial magnetic field .
this is general in that whenever the @xmath53 states are paired , their contributions cancel out . in this case
, there is an isotropic source of electric field only as @xmath54 .
thus the maxwell eq .
( [ eq : maxwell eq2 ] ) reduces to @xmath55 & = & \underset{kl}{\sum}\left|\phi_{\pm\frac{1}{2}}\right|^{2}\frac{2\left(2l+1\right)}{4\pi}\\ or,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\frac{\partial}{\partial r}\sin^{2}r\,\ , f_{r\eta } & = & \underset{kl}{\sum}\frac{\left(2l+1\right)}{4\pi}\left(\left|\phi_{\frac{1}{2}}\right|^{2}+\left|\phi_{-\frac{1}{2}}\right|^{2}\right)\\ and\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\frac{\partial}{\partial r}\sin^{2}r\,\ , f_{r\eta } & = & \underset{kl}{\sum}\frac{\left(2l+1\right)}{4\pi}\left(\left|\phi_{\frac{1}{2}}\right|^{2}-\left|\phi_{-\frac{1}{2}}\right|^{2}\right)\end{aligned}\ ] ] the solution is @xmath56,\nonumber \\\end{aligned}\ ] ] where @xmath57are appropriately weighted and normalized jacobi polynomials , is just a statement of gauss law , as the right handside is the co - moving number of enclosed particles . here to first order wkb , @xmath58 . the field intensity @xmath0are shown in fig ( [ fig : f vs r for different masses ] ) for differnt values of m and in fig .
( [ fig : f vs r for different k ] ) for some values of the comoving momentum @xmath59 of the dirac particles.we take those values to be representative of the proton and electron . at any finite time after the big bang , we see that the protons produce a stronger field than the electron .
the two fields do not have the same magnitude even for the lowest state @xmath60 , and the heavier mass consistently produces stronger field of higher momentum also .
these are exact solutions of the field equations , and we feel that this effect will have strong consequence on the formation of hydrogen and other atoms .
next , we discuss the solutions in the case after particle - antiparticle anhilation when the electron number becomes drastically reduced
. then there will not always be pairing of the @xmath53 states .
for this , we can use the well known generalized clebsch gordon expansion @xmath61 where @xmath62 are the wigner 3-j symbol .
hence , we find @xmath63 contributes only the even angular momentum states . similarly , @xmath64 contributes the odd states .
so , we can conclude that the radial electric field is composed of even angular momentum states and the odd ones compose radial magnetic field .
these solutions and the solutions for @xmath65 will be considered in future work .
we have solved some of the maxwell equations in frw spacetime with the source given by dirac field .
the variables are all separable . for the scalar field representing the components @xmath0 and @xmath1 , we are able to solve the green s function which reduces to the familiar ones of electrodynamics in the appropriate limit to flat case . the retarded part is used to derive the lienard wiechert type solution for moving point source . in section [ sec : radial - maxwell - fields ] , we consider the source given by dirac potentials . in the case that all the available states of the dirac particle are filled ,
the magnetic field @xmath1 is found to vanish .
when we compare the electric filed strength @xmath0 due to dirac particles of mass ratio 10000 , they are somewhat different with the lighter one contributing more . at this lowest level
the fields due to proton and electron are not exactly opposite of each other .
this should have observable effects on the formation of atoms and particularly on the time of decoupling .
when we consider individual azimuthal states of dirac particles , the source term vanishes .
when we consider dirac particles after the particle - antiparticle annihilation , the number density is greatly decreased so that all the available states can not be occupied . in particular ,
if all the @xmath53 azimuthal states are not paired , then the source for @xmath0 and @xmath1 do not vanish . in that case
@xmath0 is composed of even angular momentum ( @xmath66 ) states and @xmath1 is composed of odd states .
so , there is the possibility of the existence of a primordial magnetic field generated just after @xmath67 annihilation . | some solutions of the maxwell equations with dirac particles for the source in frw spacetime are discussed .
the green s function of the equation for the radial component of the maxwell fields , @xmath0 and @xmath1is solved .
green s function is found to reduce to that of minkowskian spacetime in the appropriate limit .
also , the lienard - wiechert type solution is derived .
also , the solutions with the dirac particle current is also presented .
it is found that the @xmath0is composed of even angular momentum states while the odd states constitue @xmath1 .
pacs : 03.65.pm , 04.20.cv , 04.98.80.jk keywords : dirac equation , np formalism , frw space - time , maxwell field . |
it is widely believed that accretion flow onto black holes drives major activities of astrophysical black holes , such as active galactic nuclei ( agns ) , galactic black hole candidates ( bhcs ) , and possibly gamma - ray bursts ( grbs ) .
it is also a common belief that the basic accretion processes and radiation properties can well be described by the standard - disk model by shakura & sunyaev ( 1973 ) . on the other hand ,
the observational facts which can not fit the standard - disk picture are also being accumulated recently .
good examples are intense high - energy emission from black holes , which indicates the presence of very hot plasmas around black holes with temperature of @xmath0k .
this leads to the ideas of accretion disk corona and/or radiatively inefficient flow ( riaf ) .
the standard - disk picture breaks down not only in the low - luminosity regimes but also in the high - luminosity regimes , in which mass - accretion rates , @xmath1 , becomes comparable to or exceeds the critical mass - accretion rate , @xmath2 , with @xmath3 being the eddington luminosity .
apparently , they look similar to those of the standard - type disks , since the disk remains to be optically thick and thus emits blackbody - like emission , just as the standard - type disks do .
flow dynamics is , however , distinct .
what makes the super - critical accretion flows distinct from the standard - disk - type flow is the presence of photon trapping ( begelman 1978 ) .
photon trapping occurs , when photon diffusion time , time for photons to travel from the equatorial plane to the surface , exceeds the accretion timescale . under such circumstances photons generated via the viscous process
are advected inward with gas flow without being able to go out from the surface immediately .
we can define the trapping radius , inside which photon trapping is significant ; @xmath4 where @xmath5 is the mass - accretion rate normalized by the critical mass - accretion rate , @xmath6 is the half thickness of the disk , @xmath7 is the radius , and @xmath8 is the schwarzschild radius with @xmath9 being the black hole mass ( ohsuga et al .
2002 , hereafter referred to as paper i , see also begelman 1978 for the spherical case ) .
it is claimed that photon trapping effects were incorporated in the slim - disk model by abramowicz et al .
( 1988 ; for a review , see kato , fukue , & mineshige 1998 , section 10.1 ) . we have shown previously , however , that the slim - disk model does not fully consider the photon trapping effects , since it is a radially one - dimensional model , whereas the photon trapping is basically a multi - dimensional effect ( see paper i ) .
we thus need at least two - dimensional treatment .
indeed , we have demonstrated in paper i that the photon trapping is grossly underestimated in the slim - disk model .
let us recall what lacks in the slim - disk model more explicitly .
the equation of energy balance , @xmath10 , is solved in the slim - disk model , where @xmath11 , @xmath12 , and @xmath13 are the viscous heating , radiative cooling , and advective cooling rates per unit surface , respectively .
the problem resides in that the radiative cooling is evaluated under the usual diffusion approximation ( in the vertical direction )
. this approximation may be justified , if radial inflow of gas is totally negligible so that photons can mainly diffuse in the vertical direction . however
, this is not the case , when the diffusion timescale is longer than the accretion timescale .
this leads to over - estimation of @xmath12 and , hence , under - estimations of @xmath13 , compared with the correct value .
more importantly , photon trapping modifies spectral energy distribution ( sed ) , as well .
we have found that large photon trapping yields spectral softening , because hard photons which are created deep inside the disk are more effectively trapped than soft photons ( ohsuga , mineshige , & watarai 2003 ) . since our previous formulations are only partly multi - dimensional , we , as a next step , need to perform fully two - dimensional analysis of the super - critical accretion flows .
this is a major motivation of the present study .
when considering photon trapping effects , we should also pay attention to the fact that the super - critical accretion flow becomes geometrically thick .
the multi - dimensional gas motion , such as convective or large - scale circulation , might occur .
further , strong outflow might also be generated at the disk surface via radiation pressure force .
such complex flow motion will influence radiation energy distribution through the advective energy transport , which , in turn , affects the flow motion via radiation pressure force .
we need to carefully solve such strong coupling between radiation and matter .
two - dimensional radiation - hydrodynamical ( 2d - rhd ) simulations of accretion disks were initiated by eggum , coroniti , & katz ( 1987 ) , who assumed the equilibrium between gas and radiation .
the improved simulations , in which the energy of gas and radiation are separately treated , were performed by kley ( 1989 ) , okuda , fujita , & sakashita ( 1997 ) , fujita & okuda ( 1998 ) , kley & lin ( 1999 ) , and okuda & fujita ( 2000 ) .
the simulations of super - critical flows around black holes were pioneered by eggum , coroniti , & katz ( 1988 ) , which again assumed the equilibrium between gas and radiation , and was improved by okuda ( 2002 ) .
eggum et al . ( 1988 ) showed that mass accretion onto the black hole occurs at a super - critical rate , although the mass - accretion as well as mass - outflow rate were still variable in their simulations ; that is , quasi - steady state had not been achieved by their simulations .
this is the same also in the simulations by okuda ( 2002 ) , in which the resultant luminosity slowly decreases with time .
he also found that the mass - accretion rate is sub - critical around the black hole , although the mass is injected from the outer boundary at a super - critical rate .
note that the calculation times of these simulations were only 0.6 s and 1.6 s ( in physical unit ) , respectively , for the black hole mass of @xmath14 .
notably , these are shorter than the viscous timescale , @xmath15 with @xmath16 being the viscosity parameter .
more recently , okuda et al . ( 2005 ) performed long - term 2d - rhd calculations of the super - critical accretion .
the luminosity as well as the mass - accretion rate seems to be quasi - steady in their simulations , although the flow structure is not steady yet .
since the sum of the mass - accretion rate at the inner boundary and mass - outflow rate at the outer boundary is much smaller than the mass - input rate at the disk boundary , the mass within the computational domain continues to increase .
in addition , the photon - trapping did not appear in their simulations , whereas the mass - accretion rate exceeds the critical value .
to summarize , despite the interesting simulations being made so far by two groups the quasi - steady structure of the super - critical accretion flows still remains to be an open issue .
further , interesting issues related to super - critical flow , including the photon - trapping effects , the dependence of the observed luminosity on various viewing angles , and the effects of metalicity on the flow structure , have not been investigated previously .
this is a motivation of the present study . here
, we report for the first time the quasi - steady structure of the super - critical disk accretion flows around black holes , which were revealed by 2d - rhd simulations . through the present simulations we mainly aim at understanding the dynamics of the viscous flow in the vicinity of the black hole of @xmath17 .
basic equations and our model are explained in 2 , and the numerical methods are described in 3 .
we will then display the quasi - steady flow structure and study the photon trapping effects in the simulated flow in 4 .
finally , 5 and 6 are devoted to discussion and conclusions .
we solve the full set of rhd equations including the viscosity term .
we use spherical polar coordinates @xmath18 , where @xmath7 is the radial distance , @xmath19 is the polar angle , and @xmath20 is the azimuthal angle . in the present study , we assume the flow to be non self - gravitating , reflection symmetric relative to the equatorial plane ( with @xmath21 ) , and axisymmetric with respect to the rotation axis ( i.e. , @xmath22 ) .
we describe the gravitational field of the black hole in terms of pseudo - newtonian hydrodynamics , in which the gravitational potential is given by @xmath23 as was introduced by paczynski & wiita ( 1980 ) .
the basic equations are the continuity equation , @xmath24 the equations of motion , @xmath25 @xmath26 @xmath27 the energy equation of the gas , @xmath28 and the energy equation of the radiation , @xmath29 here , @xmath30 is the gas mass density , @xmath31 is the velocity , @xmath32 is the gas pressure , @xmath33 is the internal energy density of the gas , @xmath34 is the blackbody intensity , @xmath35 is the radiation energy density , @xmath36 is the radiation flux , @xmath37 is the radiation pressure tensor , @xmath38 is the absorption opacity , @xmath39 is the viscous force , and @xmath40 is the viscous dissipative function , respectively .
the radiation force , @xmath41 , is given by @xmath42 where @xmath43 is the total opacity with @xmath44 being the thomson scattering cross - section and @xmath45 being the proton mass .
as the equation of state , we use @xmath46 where @xmath47 is the specific heat ratio .
the temperature of the gas , @xmath48 , can then be calculated from @xmath49 where @xmath50 is the boltzmann constant and @xmath51 is the mean molecular weight , respectively . to complete the set of equations we apply flux limited diffusion ( fld ) approximation developed by levermore and pomraning ( 1981 ) . in this framework ,
the radiation flux is written as @xmath52 with the flux limiter , @xmath53 , and the radiation pressure tensor is expressed in terms of the radiation energy density via @xmath54 where @xmath55 is the eddington tensor . here
, the flux limiter is given by @xmath56 with using the dimensionless quantity , @xmath57 .
the components of the eddington tensor are @xmath58 where @xmath59 is the eddington factor , @xmath60 and @xmath61 is the unit vector in the direction of the radiation energy density gradient , @xmath62 this approximation holds both in the optically thick and thin regimes . in the optically thick limit ,
we find @xmath63 and @xmath64 because of @xmath65 . in the optically thin limit of @xmath66 ,
on the other hand , we have @xmath67 .
these give correct relations in the optically thick diffusion limit and optically thin streaming limit , respectively .
for the absorption opacity , we consider the free - free absorption , @xmath68 , and bound - free absorption , @xmath69 , @xmath70 where @xmath68 and @xmath69 are given by @xmath71 ( rybicki & lightman 1979 ) , and @xmath72 with @xmath73 being the metalicity ( hayashi , hoshi , & sugimito 1962 ) , respectively . here , we assume that only the @xmath74-component of the viscous stress tensor , which plays important roles for the transport of the angular momentum and heating of the disk plasma , is non zero , @xmath75 in the present study , where @xmath76 is the dynamical viscosity coefficient .
then , the radial and polar components of the viscous force are null ( @xmath77 ) , and the right hand side of equation ( [ mom_varphi ] ) is described as @xmath78 the viscous dissipative function is given by @xmath79 ^ 2.\ ] ] finally , we prescribe the dynamical viscosity coefficient as a function of the pressure @xmath80 where @xmath81 is the keplerian angular speed .
it is modified @xmath16 prescription of the viscosity , which is proposed by shakura & sunyaev ( 1973 ) . in this form ,
the dynamical viscosity coefficient is proportional to the total pressure because of @xmath63 in the optically thick regime . in the optically thin regime , by contrast
, we find @xmath82 ( or kinematic viscosity is @xmath83 with @xmath84 being isothermal sound velocity ) , since @xmath53 vanishes in this limit . here
, we need to remark that the realistic formalism about the viscosity should be investigated from the magneto - hydrodynamical point of view , since the dominant sources of viscosity would be chaotic magnetic fields and turbulence in the gas flow ( e.g. , machida , hayashi , & matsumoto 2000 ; stone & pringle 2001 ) .
we numerically solve the set of rhd equations shown in the previous section by using an explicit - implicit finite difference scheme on the eulerian grids .
our methods and boundary conditions are similar to those of okuda ( 2002 ) , but we adopt a different initial condition and have carried out long - term calculations in order to examine a quasi - steady structure of the supercritical accretion flows . since we assume the axisymmetry as well as the reflection symmetry , the computational domain can be restricted to one quadrant of the meridional plane .
the domain consists of spherical shells of @xmath85 and @xmath86 , where @xmath87 and @xmath88 are the radial coordinates of the inner and outer boundaries of the computational domain , respectively , and is divided into @xmath89 grid cells , amounting to @xmath90 meshes .
the @xmath91 grid points in radial direction are equally spaced logarithmically , while the @xmath92 grids are equally distributed in such a way to achieve @xmath93 .
all the physical quantities are defined at each cell center .
we divide the numerical procedure for the finite - difference equations into the following steps : [ 1 ] hydrodynamical terms for ideal fluid , [ 2 ] advection term in the energy equation of the radiation , [ 3 ] radiation flux term in the radiation energy equation , [ 4 ] gas - radiation interaction terms in the energy equations of the radiation and gas , and [ 5 ] viscous terms in the momentum equation and energy equations of the gas .
steps [ 1 ] and [ 2 ] are solved with the explicit method while the steps [ 3 ] , [ 4 ] , and [ 5 ] are treated based on the implicit method . in the first step ,
we use the computational hydrodynamical code , the virginia hydrodynamics one .
it is based on the piecewise parabolic method ( colella & woodward 1984 ) . the equations ( [ mass_con])-([gase ] ) except the viscous terms and the gas - radiation interaction terms of equation ( [ gase ] ) are solved in this step . in the second step , an integral formulation is used to generate a conservative differencing scheme for the advection term of the equation ( [ rade ] ) .
the energy transport by the radiation flux term is solved in the third step .
the radiation energy density is updated again with using the bicgstab method for a matrix inversion . in the fourth step ,
we solve the gas - radiation interaction terms in equation ( [ gase ] ) .
all the terms on the right hand side of equation ( [ rade ] ) except for the radiative flux term is also treated in this step .
the radiation energy and gas energy is advanced simultaneously .
the method used in this step is basically the same as that described by turner & stone ( 2001 ) . in the final step
, we updates the azimuthal component of the velocity by solving the viscosity term in the equation ( [ mom_varphi ] ) with the gauss - jordan elimination for a matrix inversion .
the viscous dissipative function is also calculated in this step . throughout the present study
, we assume @xmath94 , @xmath95 , @xmath96 , and @xmath97 .
the size of the computational domain is set to be @xmath98 .
[ here , it is noted that our conclusion does not change so much , even if we employ @xmath99 . ]
we employ the absorbing inner boundary condition so that the density , gas pressure , and velocity are smoothly dumped ( i.e. kato , mineshige , & shibata 2004 ) . here
, it is noted that our simulation results are not sensitive to the inner boundary condition nor to the location of the inner boundary .
the results does not change even if we use the free boundary conditions , where all the matter and waves can transmit freely .
the radiation flux is set to be @xmath100 at the inner boundary ; that is , we apply the condition of the optically thin limit at the inner boundary .
the outer boundary at @xmath101 is divided into two parts : the disk part ( @xmath102 ) and the part above the disk ( @xmath103 ) . through the outer - disk boundary
we assume that matter is continuously injected into the computational domain .
we set the injected mass - accretion rate ( mass - input rate ) so as to be constant at the boundary . in the present study
, we explore the cases of @xmath104 , @xmath105 , and @xmath106 , where @xmath107 is the mass - input rate normalized by the critical mass - accretion rate .
the injected matter is supposed to rotate with sub - keplerian speed , and it has a specific angular momentum corresponding to the keplerian angular momentum at @xmath108 , since our main purpose of the present study is to investigate the viscous accretion flows within @xmath17 . by setting the keplerian radius ( @xmath109 ) to be much smaller than the radius of the outer boundary , we can prevent that the outer boundary conditions directly affect the accretion flow within @xmath110 . with such a large outer boundary we can reproduce complex motions like circulation around @xmath109 in the disk ; that is , the matter can transiently go out across the radius of @xmath110 and return . at the outer boundary region above the accretion disk
we use free boundary conditions and allow for matter to go out but not to come in .
if the radial velocity is negative at the outermost grid , it is automatically set to be zero .
we also employ the radiation flux in the optically thin limit , @xmath111 , at the outer boundary . with respect to the the rotation axis we assume @xmath30 , @xmath32 , @xmath112 , and @xmath35 to be symmetric , while @xmath113 and @xmath114 are antisymmetric .
on the equatorial plane , on the other hand , @xmath30 , @xmath32 , @xmath112 , @xmath114 , and @xmath35 are symmetric and @xmath113 is antisymmetric .
we start the calculations with a hot , rarefied , and optically - thin atmosphere .
there is no cold dense disk in the computational domain , initially .
the initial atmosphere is constructed to approximately achieve hydrostatic equilibrium in the radial direction ; namely , its density profile is given by @xmath115,\ ] ] where @xmath116 is the density at the outer boundary .
we employ @xmath117 and @xmath118 in the present study .
since this atmospheric gas is finally ejected out of the computational domain , it does not affect the resultant quasi - steady structure .
the time step is restricted by the courant - friedrichs - levi condition .
we set the time step as @xmath119,\ ] ] where @xmath120 is a parameter , and @xmath121 and @xmath122 are the cell sizes in the radial and polar directions , respectively .
although we only show the results for the cases with @xmath123 , we also simulated the cases with @xmath124 or @xmath125 , confirming that our conclusions do not alter by this change .
we first represent time evolution of 2d - rhd simulations .
overall evolution is divided into two distinct phases : the accumulation phase and the quasi - steady phase .
the mass is continuously injected through the outer - disk boundary , and creates continuous gas inflow because of the gravity force by the central black hole . since angular momentum of the injected mass is set to be equal to the keplerian angular momentum at @xmath108 , it is natural that the gas tends to accumulate around the regions with the radius of @xmath109 by degrees ( see figure [ phase1 ] ) .
this is the accumulation phase . since eggum et al .
( 1988 ) and okuda ( 2002 ) started calculations with a cold dense disk , this phase did not appear in their simulations .
eventually , the viscosity starts to work so that the angular momentum of the gas can be transported outward , which drives inflow gas motion in a quasi - steady fashion .
this is the quasi - steady phase .
such a two - step evolution is clear in figure [ evolve ] . in this figure
, we show the time evolution of normalized mass - accretion rate onto the central black hole , @xmath126 ( thick solid curve ) , the luminosity , @xmath127 ( thin solid curve ) , the viscous heating rate , @xmath128 ( dotted curve ) , and the total mass contained within the computational domain , @xmath129 ( dashed curve ) . here , we set the mass - input rate of @xmath130 and the metalicity of @xmath131 .
both of the accretion rate and luminosity steadily increases until @xmath132 s. we also notice that the mass - accretion rate is much smaller than the mass - input rate ( @xmath130 ) , and that the total mass within the computational domain rapidly increases with time , indicating mass accumulation really occurring in this phase .
then starts the quasi - steady accretion phase ( at @xmath133s ) , when the mass - accretion rate exceeds critical value , @xmath134 , and all the physical quantities stay nearly constant .
[ the constant @xmath135 implies that the sum of the mass - accretion rate at the inner boundary and mass - outflow rate at the outer boundary is equal to the mass - input rate .
] we can thus conclude that we could for the first time succeed in reproducing the quasi - steady state of the super - critical accretion flows with 2d - rhd simulations . the critical time
( @xmath136 7s ) separating these two phases roughly coincides with the viscous timescale [ equation ( [ tvis ] ) ] .
it is natural that the quasi - steady structure is obtained within 10 sec , which is the viscous timescale at the keplerian radius ( @xmath137 ) , although the size of the computational domain ( @xmath138 ) is much larger .
this is because that the injected matter accretes from the outer boundary to the keplerian radius with the free - fall velocity so that the evolutionary timescale of the outer regions is given by the free - fall timescale at @xmath139 , @xmath140 .
this is certainly shorter than the viscous timescale at @xmath108 .
let us , next , examine the quasi - steady structure in some details .
figure [ rho ] displays the cross - sectional view of the density distributions ( with colors ) , overlaid with the velocity vectors ( with arrows ) at @xmath141 . here ,
@xmath142- and @xmath143-axis are @xmath144 and @xmath145 , respectively .
we understand with this figure that the flow structure is roughly divided into two regions : the disk region around the equatorial plane ( characterized by orange color ) and the outflow region above the inflow region ( characterized by blue color ) .
roughly , the boundary is at @xmath146 ; that is the disk is geometrically and optically thick , as was predicted by the slim - disk model .
however , the density distribution definitely deviates from that of the slim - disk model , since it is neither smooth nor plane parallel in the vertical direction .
we can even see a number of cavities in this figure .
the flow pattern is also complex , though the slim - disk model predicts the simple convergence flow .
we found the prominent circular motion within the disk and the strong outflow which is generated at the disk surface .
the patchy structure around the boundary between the dense inflow region and the rarefied outflow region seems to be caused by the kelvin - helmholtz ( k - h ) instability ( discussed later ) .
the less dense gas in the outflow region penetrates into the disk body because of the k - h instability , thus forming the cavities .
to understand the dynamics of our entire calculation domain , we show the 2d density distribution in the whole region in figure [ all_rho ] .
we see from this figure that the density is larger around the equatorial plane ( @xmath147 ) than at the large altitudes ( @xmath148 ) in the outer region , @xmath149 .
this reflects that the injected mass accretes along the equatorial plane . on the other hand ,
the viscous accretion disk forms inside , @xmath17 .
we focus the viscous accretion flows in the present study .
the situation is apparently similar to that studied by chakrabarti ( 1996 ) , although they were concerned with subcritical flow and their study was not multi - dimensional nor radiation - hydrodynamical simulations .
figure [ center ] show the 2d distributions of radiation energy density ( upper left ) , the ratio of the radiation energy to the internal energy of gas ( upper right ) , gas temperature ( lower left ) , and the radial velocity normalized by the escape velocity ( lower right ) , respectively , on the @xmath150 plane .
as shown in the upper left panel , the radiation energy distribution roughly coincides with the gas density distribution .
that is , the radiation energy tends to be larger around the equatorial plane than that around the rotation axis . since the radiation energy distribution is smoothed due to the radiative diffusion within the disk , there is no cavity found in the radiation energy distribution , which makes a marked difference from the density distribution ( see figure [ rho ] ) .
radiation energy greatly exceeds the gas energy in the entire region , including the outflow region in our simulations ( see the upper right panel ) .
this confirms that most of the regions is the radiation - pressure dominated and strong radiation pressure supports the geometrically thick disk and drives the outflow .
the gas temperature distribution shown in the lower left panel shows relatively low temperatures in the disk region , compared with those in the outflow region , although the viscous heating rate is much larger in the former than in the latter .
this can be understood , because radiative cooling is more effective in dense region as a consequence of its strong density dependence .
the gas in the disk region is heated up by the viscous heating , and the processed energy is effectively converted to the radiation energy .
therefore , gas temperature does not rise so much within the disk .
conversely , the gas can not emit effectively in the outflow region , since both free - free emissivity and bound - free emissivity are more sensitive to the density than the gas temperature , @xmath151 .
it can be understood by the comparison between the lower left and the upper right panels .
the radiation temperature ( @xmath152 ) in the outflow region is lower than that in the disk region , on the contrary to the gas temperature profile .
the lower right panel indicates the radial velocity normalized by the escape velocity .
the gas moves toward the black hole or flows outward slowly in the disk regions ( see the blue area ) .
the white color indicates that the velocity exceeds the escape velocity in this area .
it is found that the gas is accelerated through the radiation pressure and is blown away to a large distance ( see also the upper right panel ) .
such a flow component will be identified as a strong disk wind .
note that the outflow presented here is distinct in nature from that known as a bounce jet ( chen et al .
1997 ) , which arises because of a bounce of free - falling low angular momentum material , when it goes through the centrifugal barrier at small @xmath7
. the outflow will also produce large absorption in the emergent spectra .
we also notice strong velocity shear at the boundary between the disk region and the outflow region . this complex density profile around the disk surface as shown in figure [ rho ] is explained as a consequence of the k - h instability .
the growth timescale of the k - h instability is roughly given by @xmath153 where @xmath154 is the wavenumber , while @xmath155 is the density ratio of the disk region to the outflow region at the disk boundary .
here , we assume the incompressible fluid as well as @xmath156 and neglect the gravity .
also , the viscosity is not considered , since @xmath157-component of the viscous stress tensor is set to be zero in the present simulations . by setting @xmath158 , @xmath159 , and @xmath160
, we find @xmath161 s. note that this is shorter than the escape time , @xmath162 s , for @xmath108 and @xmath163 . that is , there is ample time for the k - h instability to grow before the material flows outward .
the mass - accretion rates as a function of the radius are displayed in figure [ mdot ] for the case of @xmath130 and @xmath131 .
the solid , dotted , and dashed curves indicate the @xmath164 profiles at elapsed times of @xmath165 s , 30 s , and 50 s , respectively . here , the mass - accretion rate at each radius is evaluated by @xmath166 \sin\theta d\theta.\ ] ] it is found that the mass - accretion rates are not constant in the radial direction but decrease inward , as the flow approaches the black hole .
roughly , we find @xmath167 .
in addition , we find that the radial @xmath168 profile remains nearly the same after 7s , from which we can conclude that the flow is in a quasi - steady state .
this @xmath168 change is caused by a cooperation between a wind mass loss around the disk surface and the circular motion deep inside the disk .
note that accretion rates are assumed to be constant in space in the slim - disk model formulation .
the multi - dimensional numerical simulations of riafs have shown that the accretion disks are convectively unstable , thereby the circular motion being driven within the flow ( e.g. , igumenshchev & abramowicz 1999 ; stone , pringle , & begelman 1999 ; mckinney & gammie 2002 ) . the convection might cooperate with the k - h instability in generation of the circular motion as well as the cavities found in our simulations
. the simulations of riafs have also revealed that @xmath168 increases with radius as @xmath169 , which is similar to our value , and was attributed to the circular motion as well as the bipolar outflows ( stone , pringle , & begelman 1999 , see also narayan , igumenshchev , & abramowicz 2000 for self - similar solution ) .
here , we need to emphasize that although the flow structures look similar at first glance , the basic physical processes are distinct from those of riaf simulations .
that is , entropy is carried by fluid in the riaf , whereas it is mostly by photons in the present case .
in addition , since the riaf simulations do not basically take into account of the radiative cooling , they would overestimate the driving force of the outflows .
the photon - trapping characterizes the super - critical accretion flows . it works to reduce the energy conversion efficiency , @xmath170 ( see , e.g. paper i ; ohsuga et al .
2003 ) . as a result
, the flow luminosity becomes insensitive to the mass - accretion rate when @xmath171 .
however , simple stream lines were assumed in the previous studies on super - critical flows , including the slim - disk model . even if we do not consider possible multi - dimensional gas motions , photon trapping works to some degree , leading to a reduction in the energy conversion efficiency .
such multi - dimensional gas motion and associated reaction of the radiation can be calculated in the present rhd simulations , since we have calculated the radiation energy transport , being coupled with the gas dynamics .
figure [ trap ] plots the luminosity as a function of the mass - accretion rate onto the black hole .
the filled squares , circles , and triangles indicate the results for @xmath172 , @xmath173 , and 0 , respectively , with different mass - input rates , @xmath104 , @xmath105 , and @xmath106 from the left to the right .
we also indicated in the same figure the luminosity calculated based on the slim - disk model ( watarai et al .
2000 ) and the one based on model a of paper i with dashed and dotted curves , respectively . in paper
i , a simple model for the accretion flow is employed , and the luminosity with carefully taking account of the photon - trapping is calculated by solving energy transport inside the accretion flows .
[ more precisely , the luminosity plotted in figure [ trap ] is the corrected one , which fixes initial small errors in their model a ( see figure 1 in paper i ) . ] here , it should be stressed that the mass - accretion rate onto the black hole is not input parameter but is calculated dynamically in the present simulations , although it was a parameter in both of model a of paper i and the slim - disk model .
the resultant @xmath164 profile was shown in figure [ mdot ] .
it is evident in figure [ trap ] that the calculated luminosity agrees more with model a of paper i , rather than the slim - disk model , in all the cases .
this proves that the two - dimensional effects of photon - trapping is really significant in the super - critical accretion flows .
this result is , in a sense , surprising . in model
a in paper 1 , we assumed that the viscous heating occurs only in the vicinity of the equatorial plane , although the gas might be heated up at the high altitude .
since the photons emitted at deep inside the disk tends to be more effectively trapped in the flow , we anticipated that the photon - trapping effects would be reduced in the realistic situation , compared with model a. we also argued in paper i that photon trapping effects may be attenuated by the presence of large - scale circulation motion , which could help photon diffusion motion and thus considerably reduce photon traveling time to the surface of the flow .
however , our current results show significant photon trapping effects even when we explicitly include complex flow motions .
figure [ trap ] also shows that the mass - accretion rate onto the black hole increases with an increase of the metalicity ( for fixed mass - input rates ) . in our simulations
the gas with higher metalicity has larger absorption opacities so that the gas energy can be more effectively converted into the radiation energy , yielding smaller gas pressure than in the metal - poor case .
the gas could be effectively blown away by the strong gas pressure .
however , there is a counter effect .
the gas with large absorption opacity enhances the radiation energy .
enhanced radiation flux and large opacity can drive strong radiation - pressure driven outflows . the physical reason will be investigated in future .
let us see more explicitly how significant photon trapping is .
figure [ flux ] compares the distributions of the radial component of the radiative flux in the comoving frame ( upper panel ) and that in the inertial frame ( lower panel ) at @xmath141 .
other parameters are the same ; the mass - accretion rate is @xmath130 and the metalicity is @xmath131 .
the former radiative flux ( @xmath174 ) is roughly proportional to the radial gradient of the radiation energy distribution . on
the other hand , the latter flux ( @xmath175 ) includes the advective transport of radiation energy ( @xmath176 ) in addition to the former flux ; namely , we approximately have @xmath177 .
in other words , differences between two panels represent how significant photon trapping is .
in fact , we see a significant difference between the two panels ; whereas the blue area ( indicating large radiation flux ) is restricted to the vicinity of the black hole in the upper panel , it is more widely spread over the entire disk region in the lower panel .
this is a direct manifestation of the photon - trapping effects .
since the super - critical accretion flows are geometrically and optically thick , the observed images and luminosity should strongly depends on the viewing ( inclination ) angle .
we calculate the intensity map with using the monochromatic radiation transfer equation , @xmath178,\end{aligned}\ ] ] where @xmath179 is the specific intensity , @xmath180 is the lorentz factor , @xmath181 @xmath182 and @xmath183 are the directional cosine in the inertial and comoving frame , respectively , with @xmath184 and @xmath185 being the azimuthal and polar angles of radiation propagation in the inertial frame . here
, we assumed isotropic scattering .
the calculated effective temperature ( @xmath186 ) maps at @xmath187 are shown in figure [ image ] for various viewing angles : @xmath188 , @xmath189 , @xmath190 , and @xmath191 .
the other parameters are @xmath130 and @xmath131 . on the observer s screen ,
the black hole lies at the center of the cartesian coordinate @xmath192 .
since we use the spherical computational domain with the size of @xmath193 in the present study , the blue four corners of the maps indicate outside of the domain . in the face - on view ( see the upper left panel ; @xmath194 ) , the effective temperature exceeds @xmath195k within @xmath110 and amounts @xmath196k in the central region .
such a high temperature region disappears in the upper right panel ( the case with @xmath197 ) . in this panel , the most luminous region
is found on the upper side ( not at the center ) and is elongated in the vertical direction .
these are caused by an occultation of the innermost part of the flow by the outer parts ( see fukue 2000 , watarai et al .
2005 for self - occultation effects based on the slim - disk model ) .
since the accretion flow is both geometrically and optically thick , the innermost region can not be seen for large inclination angles , @xmath198 . the emission from the super - critical accretion flows is mildly collimated due to the same reason .
figure [ lth ] represents the viewing angle - dependence of the isotropic luminosity ( normalized by bolometric luminosity ) for @xmath131 ( circles ) and @xmath172 ( squares ) , respectively .
the other parameters are @xmath199 s and @xmath130 . here , the isotropic luminosity is calculated by assuming isotropic radiation field , @xmath200 , with @xmath201 being the distance from an observer and @xmath202 being a revised factor .
the luminosity calculated by solving radiation transfer equation does not always coincide with that evaluated by the fld method . here
, we assume @xmath203 to fit these luminosities .
as expected , the isotropic luminosity is quite sensitive to the viewing angle , indicating that the emission from the flow is mildly collimated . if the flow shape would be flat and ( geometrically ) thin , and if no collimation would occur , the observed luminosity should vary along the cosine curve , as is indicated by the dotted curve .
we also find that the angle dependence of the isotropic luminosity is more enhanced in large - metalicity case of @xmath172 than that of @xmath131 .
this is because of different density distribution . as have already mentioned
, the mass - outflow rate is smaller in the high metalicity case than in the low metalicity case . as a result , the density contrast between the disk region and the outflow region above the disk gets larger with an increase of the metalicity .
accordingly , the hot innermost region of the high metal flows are easier to be observed owing to smaller optical depth in the outflow region for small viewing angles .
the super - critical accretion flows would be identified as very high @xmath127 objects in the face - on case , since the emission is mildly collimated in the polar direction and the bolometric luminosity itself exceeds the eddington luminosity .
the former effect is enhanced in the case of high metalicity ( large absorption opacity ) .
it is also true that super - critical flow may be identified as fairly sub - critical source , if the viewing angle is large ( watarai et al . 2005 ) .
here , we directly compare the structure of the super - critical accretion flows based on our 2d simulations with that of the slim - disk model . figure [ slim ] represents the time - average density profile ( top panel ) , radial and rotation velocity profiles ( middle panel ) , and radiative temperature profile ( bottom panel ) on the equatorial plane in the quasi - steady accretion phase of @xmath204 s. the adopted parameters are @xmath205 and @xmath131 .
the resultant profiles of the one - dimensional numerical study for the slim - disk model are represented by the dashed curves ( watarai et al .
as have already mentioned in section 4.1 , the injected matter falls with the free - fall velocity outside the keplerian radius ( shaded region ) . a viscous accretion disk forms in the inner part ( white region ) .
the free - falling matter penetrates slightly within the keplerian radius , thus a separation radius appears around 70 @xmath206 .
we focus on the viscous accretion regime in this subsection .
the density profile is shown by the solid curve in the top panel .
the density peak around @xmath207 is made by accumulation of the injected matter , where occurs because of the centrifugal barrier .
we compare the density profile in the viscous accretion regime with that of numerical and self - similar solutions of the slim - disk model . in this panel
, it is found that the slope of the density profile of our simulations is steeper than that of the numerical solution of the slim - disk model .
we also find that the slope is close to that of the self - similar solution of the slim disk , @xmath208 ( watarai et al .
1999 , see also spruit et al .
1987 ) , whereas the profile becomes flatter in the vicinity of the black hole ( @xmath209 . here , we note that the slope in our simulations is steeper than that of the riafs .
the riaf simulations showed the density profile to be @xmath210 ( mckinney & gammie 2002 ) . in the middle panel , the radial and rotation velocities
are plotted by the filled circles and the solid curve .
the dotted curve indicates the keplerian velocity , @xmath211 .
it is found that the rotation velocity nicely agrees with the keplerian velocity as well as the result of the numerical solution of the slim disk .
however , the radial velocity profile is not smooth and largely deviates from that of the slim - disk value , whose slope agrees with that of the keplerian velocity ( dotted curve ) in the free - fall regime .
the radial velocity remarkably varies at @xmath212 , and the radial velocity sometimes becomes negative , around @xmath213 , @xmath214 , and @xmath215 .
such a complex @xmath112 profile is formed by the prominent circular motions around the equatorial plane ( see figure [ rho ] ) . the bottom panel shows the radiation temperature profile , where the radiation temperature is defined as @xmath216 with @xmath217 being the radiation constant .
as shown in this panel , the profile nicely agrees with that of the one - dimensional numerical solution and is flatter than that of the self - similar solution of the slim - disk model , @xmath218 . to sum up
, we can thus conclude that the quasi - steady structure of the super - critical accretion flows are apparently similar to but , precisely speaking , deviates from the prediction of the slim - disk model .
next , we represent again the effective temperature profile in order to directly compare with the prediction of the slim - disk model , although the 2d images of the effective temperature have already shown in figure [ image ] .
figure [ slim_teff ] represents the effective temperatures at @xmath219 as a function of @xmath220 for @xmath221 and @xmath222 , where @xmath220 and @xmath223 are the horizontal and vertical coordinates on the observer s screen ( see figure [ image ] ) .
the other parameters are @xmath224 , @xmath225 , and @xmath199 s. it is found that the slope of the effective temperature profile for @xmath221 becomes flatter than that of the slim - disk model , @xmath226 , within @xmath227 .
in contrast , we found that the profile is consistent with the slim - disk model in the outer region .
as have already mentioned , the effective temperature as well as the luminosity is quite sensitive to the viewing angle , since the accretion flow is both geometrically and optically thick .
we found that the effective temperature profile for @xmath228 is almost flat .
the slim - disk model succeeds in reproducing the observed behavior of ultra - luminous x - ray sources ( ulxs ) and narrow - line seyfert 1 galaxies ( nls1s ) , which are thought to be candidates of near- or super - critical flows ( watarai , mizuno , & mineshige 2001 ; mineshige et al .
2000 ; kawaguchi 2003 ) .
the seds are produced based on the blackbody emission with the effective temperature coupled with the some modification .
we , here , note that the temperature profile obtained by our current simulations show some deviations from that of the slim - disk model .
one characteristic feature of the super - critical accretion flow is its generation of the radiation - pressure driven wind .
our simulations show that a large quantity of gas is blown away by the strong radiation pressure above and below the accretion disk
. such outflow material , which is thought to be highly ionized by the radiation from the accretion flow , would absorb the continuum emission .
therefore , the super - critical accretion objects have the blueshifted absorption lines by the highly ionized ions .
such blueshifted absorption in high - ionization lines was observed in the uv spectra of broad absorption line quasars ( weymann , carswell , & smith 1981 ; becker et al . 2000 ) .
moreover , pounds et al .
( 2003a , 2003b ) and reeves , obrien , & ward ( 2003 ) reported by the blueshifted x - ray absorption lines that the highly ionized matter with outflow velocity of order @xmath229 in quasars , and the flow is likely to be optically thick to the electron scattering within @xmath137 . in our simulations , the typical outflow velocity is @xmath230 and the optical depth of the outflow regions is around @xmath231 ( this estimated value does not depend on the black - hole mass because of @xmath232 and @xmath233 . ) these are remarkably close to the observed results , and therefore , our simulations basically account for the origins of high - velocity outflows .
detailed comparison with the observations is left as future work .
we reveal by 2d - rhd simulations that the black hole can swallow the gas at the super - critical accretion rate , although the luminosity exceeds the eddington luminosity .
this result implies that the black hole can rapidly grow , and the growth timescale is given by @xmath234 .
thus , if the supermassive black holes in the galactic nuclei are built up by the super - critical accretion , such a growing phase would be quite short because of a short growth timescale .
umemura ( 2001 ) suggested the radiation - hydrodynamical model for the formation of the supermassive black holes , in which the supermassive black holes grow in the ultraluminous infrared galaxies ( ulirgs ) due to the mass accretion caused by the radiation drag .
based on this model , kawakatu , umemura , & mori ( 2003 ) also claim that proto - quasars , which have relatively small black hole and show only narrow emission - lines , appear just after the ulirg phase , @xmath235 yr . however , such proto - quasar phase might not be observed if the black hole grows at super - critical accretion rate in ulirg phase .
a seed black hole in the ulirg can evolve to a supermassive black hole via the super - critical accretion by the end of the ulirg phase , since the mass - accretion rate from the circumnuclear regions can exceed the critical rate due to the effective radiation drag .
thus , we presumably observe the grown - up quasars after the ulirg phase , and proto - quasars are obscured by huge amount of the dust in the ulirg .
as shown in figure [ mdot ] , our simulations show that the black hole can effectively grow in the case that the accreting gas is metal rich .
this implies that super - critical accretion flows tend to emerge in the metal rich objects like starburst galaxies or star - forming regions .
this tendency is also consistent with the observed results that the nls1s , which is though to be near- or super - critical accretion objects , are metal rich objects ( nagao et al .
2002 ; shemmer & netzer 2002 ; shemmer et al .
it is proposed that nls1s accrete at super - critical rates based on the study of the optical band ( kawaguchi 2003 ; collin & kawaguchi 2004 ) .
the growth of the black hole may gradually slow down , since the density as well as the absorption opacity of the super - critical accretion flows are small around the massive black hole , if the normalized mass - accretion rate does not change so much .
observational constraints are proposed by yu & tremaine ( 2002 ) , by which most of luminous quasars must be sub - critical accretion phase .
they showed that the luminous quasars have the energy conversion efficiency of @xmath236 , based on the study of the local black hole density and the luminosity function of the quasars .
since this efficiency is much smaller than 0.1 in the super - critical accretion flows , most of luminous quasars must be sub - critical accretion phase ( see also soltan 1982 ) . throughout the present study , we use the frequency - integrated energy equation of the radiation
[ see equation ( [ rade ] ) ] . by solving the monochromatic radiation transfer equation , we can calculate the effective temperature profiles and show them in figure [ image ] .
however , the emergent seds of the super - critical accretion flows are not a simple superposition of blackbody spectra with various effective temperatures ( ohsuga et al . 2003 ) .
the photons generated deep inside the disk are difficult to reach the disk surface and thus moves downward with gas .
furthermore , most of photons generated in the vicinity of the black hole will be immediately swallowed by the black hole and can not contribute to the emergent sed .
frequency dependent radiation - hydrodynamical simulations are necessary to resolve this issue . according to the current simulations the hot outflow with temperature of @xmath237 k appears above the disk .
the compton @xmath143-parameter of this outflow region is comparable to or larger than unity , meaning that the inverse compton scattering can not be ignored .
seed soft photons emitted from the disk region should be compton up - scattered , producing high - energy , non - thermal emission which contributes to the emergent sed .
the corona above the super - critical accretion disk has been claimed to fit the observed seds of nls1s and very - high state of bhcs ( wang & netzer 2003 ; kubota & done 2004 ) , although the formation mechanism of the corona is poorly known .
the hot outflow shown in our simulations might resolve this issue . here
, we should note that the comptonization process is not included in our simulations .
if the compton cooling were effective , the hot outflow might be cooled to some extent .
the bulk comptonization may also contributes the production of the high - energy photons because of large radial velocities , @xmath238 , near the black hole .
the detailed study of the sed with comptonization is , however , beyond the scope of this paper and should be explored in future work .
we assumed that only the @xmath74-component of the viscous stress tensor is non - zero while all the other components vanish , because the @xmath74-component plays an essential role of in the angular momentum transport within the disk .
if the @xmath157- as well as @xmath239-component were non zero , the structure near the disk surface might change .
figure [ rho ] clearly shows abrupt velocity changes across the boundary between the disk and the outflow regions , which is thought to promote the k - h instability .
the growth of this instability might be suppressed by the @xmath157- and @xmath239-components of viscosity .
these components might also partially suppress the circular motion in the disk .
stone et al ( 1999 ) discovered by the simulations of riafs that the flow structure differs from that given by igumenshchev & abramowicz ( 1999 ) . it implies that the inclusion of the components of the viscous stress tensor suppresses the convection , since the @xmath157-component is excluded in stone et al .
( 1999 ) . this will be examined in future study .
more importantly , we need coupled mhd and rhd simulations , since the source of disk viscosity is likely to be of magnetic origins ( hawley , balbus , & stone 2001 ; machida , matsumoto , & mineshige 2001 ; balbus 2003 for a review ) . recently , local radiation mhd simulations of the accretion flow have been performed by turner et al .
( 2003 ) and turner ( 2004 ) .
we definitely need global simulations to include global field effects .
finally , we need to comment on the fld approximation .
it is known that the fld approximation does not always give good results for the regions with moderate optical thickness .
the fld approximation is thought to be valid in the disk region , which has a large optical depth , but the spherical shape of the radiation energy density distribution above the disk might be artificial .
in addition , the radiation drag force can not be treated in the fld approximation , whereas it might play an important role for the transport of the angular momentum within the outflow region .
since the timescale of the angular momentum transport via the radiation drag , which is given by @xmath240 ( with @xmath241 being the mass within the outflow region ) , is about ten times longer than the escape time , @xmath242 , roughly ten percent of the angular momentum could be extracted in the outflow region . [
the exact expressions for the radiation drag are found in the literature ( umemura , fukue , & mineshige 1997 ; fukue , umemura , mineshige 1997 ; ohsuga et al .
1999 ) . ]
radiation drag arises where there exists a large velocity difference between radiation sources and the irradiated matter . in the fld approximation , however , the radiation flux is determined solely by the local gradient of the radiation energy , and photon trajectories can not be properly considered .
thus , the fld approximation can not treat the radiation drag force , in principle
. it would be better to solve the radiation transfer equations without using this approximation .
begelman ( 2002 ) suggested that strong density inhomogeneities on scales much smaller than the disk scale height is formed due to the photon bubble instability in the radiation pressure - dominated accretion disks , and the disk can remain geometrically thin even as the the maximum luminosity exceeds eddington luminosity by a factor of one hundred .
however , the fld approximation is not suitable method for investigating the radiation fields in such inhomogeneous structure . the detailed study of the photon bubble instability should be explored in future work .
by performing the 2d - rhd simulations , we , for the first time , investigate the quasi - steady structure of the super - critical accretion flows around the black hole with particular attention being paid on the photon - trapping effects .
we have obtained several new findings : \(1 ) the quasi - steady structure of the super - critical flow is divided into two parts : the disk region ( with mostly inflow ) and the outflow region above the disk .
the two regions are separated by a sharp density jump . the gas outflow driven by the strong radiation pressure
is produced around the rotation axis .
further , there exists velocity shears at the boundary , which causes k - h instability around the disk surface , producing the patchy structure as well as the circular motion within the disk region .
convection may also be responsible for such inhomogeneous structure .
\(2 ) the photon - trapping plays an important role in the super - critical accretion regime .
the advective energy transport is substantial , and the large amount of photons generated inside the disk is swallowed by the black hole without radiated away .
\(3 ) our 2d - rhd model shows some differences from the slim - disk model .
the slim - disk model assumes a simple convergence flow , while our simulations revealed rather complex gas motion and structure .
we also found that the mass - accretion rate is not constant in space but decreases as matter accretes , roughly as @xmath243 , as a result of wind mass loss and circular motion .
the calculated luminosity of the flows agrees more with the prediction of paper i rather than that of the slim - disk model .
\(4 ) the emission of the super - critical accretion flows is moderately collimated .
the apparent luminosity could become more than ten times larger than the eddington luminosity .
the super - critical accretion flows would be identified as high @xmath127 objects in the face - on view , but not , if the viewing angle is large , for which self - occultation tends to reduce the total luminosity and the maximum flow temperature .
( 5 ) the mass - accretion rate increases with increase of the absorption opacity ( metalicity ) of the accreting matter .
it implies that the black hole tends to grow up faster in the metal rich regions as in starburst galaxies or star - forming regions .
in addition , the growth of the black hole may gradually slow down , since the density as well as the absorption opacity of the super - critical accretion flows are small around the massive black hole , if the normalized mass - accretion rate is kept constant .
the authors would like to thank the anonymous referee for important comments and suggestions .
the calculations were carried out at department of physics in rikkyo university .
this work is supported in part by research fellowships of the japan society for the promotion of science for young scientists , 02796 ( ko ) , by the grant - in - aid of the ministry of education , culture , science , and sports , 14740132 , by the promotion and mutual aid corporation for private schools of japan ( mm ) , by the grants - in - aid of the ministry of education , science , culture , and sport , ( 14079205 , 16340057 ) , and by a grant - in - aid for the 21st century coe center for diversity and universality in physics " ( sm ) .
abramowicz , m. a. , czerny , b. , lasota , j. p. , & szuszkiewicz , e. 1988 , apj , 332 , 646 balbus , s. a. 2003 , ara&a , 41 , 555 becker , r. h. , white , r. l. , gregg , m. d. , brotherton , m. s. , laurent - muehleisen , s. a. , & arav , n. 2000 , apj , 538 , 72 begelman , m. c. 1978 , mnras , 184 , 53 begelman , m. c. 2002 , apj , 568 , l97 chakrabarti , s. k. 1996 , apj , 464 , 664 chen , x. , taam , r. e. , abramowicz , m. a. , & igumenshchev , i. v. 1997 , mnras , 285 , 439 colella , p. & woodward , p. 1984 , j. comput .
phys . , 54 , 174 collin , s. & kawaguchi , t. 2004 , a&a , 426 , 797 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 fujita , m. & okuda , t. 1998 , pasj , 50 639 fukue , j. 2000 , pasj , 52 , 829 fukue , j. , umemura , m. , & mineshige , s. 1997 , pasj , 49 , 673 hawley , j. f. , balbus , s. a. & stone , j. m. 2001 , apj , 554 , l49 hayashi , c. , hoshi , r. , & sugimoto , d. 1962 , prog .
suppl . , 22 , 1 igumenshchev , i. v. & abramowicz , m. a. 1999 , mnras , 303 , 309 kato , s. , fukue , j. , & mineshige , s. 1998 , black - hole accretion disks ( kyoto : kyoto univ . press ) kato , y. , mineshige , s. , & shibata , k. 2004 , apj , 605 , 307 kawaguchi , t. 2003 , apj , 593 , 69 kawakatu , n. , umemura , m. , & mori , m. 2003 , apj , 583 , 85 kley , w. 1989 , a&a , 222 , 141 kley , w. & lin , d. n. c. 1999 , apj , 518 , 833 kubota , a. & done , c. 2004 , mnras , 353 , 980 levermore , c. d. & pomraning , g. c. 1981 , apj , 248 , l321 machida , m. , hayashi , m. r. , & matsumoto , r. 2000 , apj , 532 , l67 machida , m. , matsumoto , r. , & mineshige , s. 2001 , pasj , 53 , l1 mckinney , j. c. & gammie , c. f. 2002 , apj , 573 , 728 mineshige , s. , kawaguchi , t. , takeuchi , m. , & hayashida , k. 2000 , pasj , 52 , 499 nagao , t. , murayama , t. , shioya , y. , & taniguchi , y. 2002 , apj , 575 , 721 narayan , r. , igumenshchev , i. v. , abramowicz , m. a. 2000 , apj , 539 , 798 ohsuga , k. , mineshige , s. , mori , m. , & umemura , m. 2002 , apj , 574 , 315 ohsuga , k. , mineshige , s. , & watarai , k. 2003 , apj , 596 , 429 ohsuga , k. , umemura , m. , fukue , j. , & mineshige , s. 1999 , pasj , 51 , 345 okuda , t. 2002 , pasj , 54 , 253 okuda , t. & fujita , m. 2000 , pasj , 52 , l5 okuda , t. , fujita , m. , & sakashita , s. 1997 , pasj , 49 , 679 okuda , t. , teresi , v. , toscano , e. , & molteni , d. 2005 , mnras , 357 , 295 paczynsky , b. & wiita , p. j. 1980 , a&a , 88 , 23 pounds , k. a. , king , a. r. , page , k. l. , & obrien , p. t. 2003b , mnras , 346 , 1025 pounds , k. a. , reeves , j. n. , king , a. r. , page , k. l. , obrien , p. t. , & turner , m. j. l. 2003a , mnras , 345 , 705 reeves , j. n. , obrien , p. t. , & ward , m. j. 2003 , apj , 593 , 65 rybicki , g. b. & lightman , a. p. 1979
, radiative processes in astrophysics ( new york : john wiley & sons , inc . )
shakura , n. i. & sunyaev , r. a. 1973 , a&a , 24 , 337 shemmer , o. & netzer , h. 2002 , apj , 567 , l19 shemmer o. , netzer , h. , maiolino , r. , oliva , e. , croom , s. , corbett , e. , & di fabrizio , l. 2004 , apj , 614 , 547 soltan , a. 1982 , mnras , 200 , 115 spruit , h. c. , matsuda , t. , inoue , m. , & sawada , k. 1987 , mnras , 229 , 517 stone , j. m. & pringle , j. e. 2001 , mnras , 322 , 461 stone , j. m. , pringle , j. e. , & begelman , m. c. 1999 , mnras , 310 , 1002 turner , n. j. 2004 , apj , 605 , l45 turner , n. j. & stone , j. m. 2001 , apjs , 135 , 95 turner , n. j. , stone , j. m. , krolik , j. h. , & sano , t. 2003 , apj , 593 , 992 umemura , m. 2001 , apj , 560 , l29 umemura , m. , fukue , j. , & mineshige , s. 1997 , apj , 479 , l97 yu , q. & tremaine , s. 2002 , mnras , 335 , 965 wang , j .- m . & netzer , h. 2003 , a&a , 398 , 927 watarai , k. & fukue , j. 1999 , pasj , 51 , 725 watarai , k. , fukue , j. , takeuchi , m. , & mineshige , s. 2000 , pasj , 52 , 133 watarai , k. , mizuno , t. , & mineshige , s. 2001 , apj , 549 , l77 watarai , k. , ohsuga , k. , takahashi , r. , & fukue , j. 2005 , pasj , in press weymann , r. j. , carswell , r. f. , & smith , m. g. 1981 , ara&a , 19 , 41 | the quasi - steady structure of super - critical accretion flows around a black hole is studied based on the two - dimensional radiation - hydrodynamical ( 2d - rhd ) simulations .
the super - critical flow is composed of two parts : the disk region and the outflow regions above and below the disk . within the disk region the circular motion as well as the patchy density structure
are observed , which is caused by kelvin - helmholtz instability and probably by convection .
the mass - accretion rate decreases inward , roughly in proportion to the radius , and the remaining part of the disk material leaves the disk to form outflow because of strong radiation pressure force .
we confirm that photon trapping plays an important role within the disk .
thus , matter can fall onto the black hole at a rate exceeding the eddington rate .
the emission is highly anisotropic and moderately collimated so that the apparent luminosity can exceed the eddington luminosity by a factor of a few in the face - on view . the mass - accretion rate onto the black hole increases with increase of the absorption opacity ( metalicity ) of the accreting matter .
this implies that the black hole tends to grow up faster in the metal rich regions as in starburst galaxies or star - forming regions . |
supersymmetry is an attractive possibility beyond the standard model . because of the relations supersymmetry imposes among the dimensionless couplings , the quadratic divergences in the higgs sector are cut - off by the superpartner mass scale .
the cancellations stabilize the hierarchy between the planck scale and the weak scale .
the minimal supersymmetric standard model ( mssm ) is consistent with gauge coupling unification suggested by grand unified theories .
also , it is interesting that one of the hallmarks of supersymmetry , a light higgs boson ( @xmath7 gev ) , is favored by global fits to precision electroweak data @xcite . if low energy supersymmetry exists , superpartners will be copiously produced at the next generation of supercolliders .
for example , the large hadron collider ( lhc ) could produce thousands of gluino pairs in a single day .
each gluino would then , in a typical scenario , decay to the lightest supersymmetric particle ( lsp ) through a cascade of charginos , neutralinos , jets and leptons .
detailed information about the superpartner masses can be determined from such processes @xcite . a future linear collider ( lc )
would likewise prove to be an invaluable tool in measuring the parameters of the soft - supersymmetry breaking lagrangian , as precision measurements of the superpartner spectrum , cross - sections , and branching ratios are possible @xcite .
the constrained kinematics , tunable center - of - mass energy , and beam polarization allow for direct and model independent interpretations of the measurements .
also , in many processes the backgrounds are quite small . in this paper
we examine the prospect of testing supersymmetry via a precise measurement of the lepton - slepton - gaugino vertex .
the linear collider provides a suitably clean experimental environment . among the relations which account for the cancellations of quadratic divergences
, supersymmetry relates the lepton - slepton - gaugino coupling to the usual gauge coupling .
although bare ( or ) couplings enjoy the relations imposed by supersymmetry , the effective gauge and gaugino couplings are not equal because supersymmetry is broken .
in particular , all non - singlet nondegenerate supermultiplets such as the quark - squark supermultiplets contribute to the splitting .
even arbitrarily heavy split multiplets will contribute , although the contribution will be suppressed by the heavy mass scale .
hence , measurements of the type we consider here not only provide for detailed tests of supersymmetry , but can also elucidate important features of the scale and pattern of supersymmetry breaking @xcite .
for example , the ( s)quark contribution to the splitting of the u(1 ) and su(2 ) gaugino / gauge ( s)lepton couplings grows logarithmically with the squark mass , as @xmath8 this correction is obtained by evolving the couplings according to the renormalization group equations ( rge s ) of the effective theory @xcite below the squark mass threshold . when @xmath9 the correction to the su(2 ) ( u(1 ) ) coupling is about 2% ( 0.7% ) .
this gives rise to an enhancement of the @xmath0-channel slepton or gaugino production cross - section of about 8% ( 2.8% ) .
if large statistics are available and systematic errors can be controlled , we can ( assuming the mssm ) constrain the squark mass scale by this measurement .
notice that a modest hierarchy between the squark and slepton masses is typical of the simplest gauge mediated models , which predict @xmath10 . as another example
, it has been proposed that the scalar masses of the first two generations might be very heavy ( e.g. 20 tev ) so that the fcnc constraints on the scalar masses are alleviated @xcite .
this large hierarchy in the superpartner spectrum will be reflected in the splitting of the gaugino / gauge couplings of the third generation ( s)fermions .
we restrict our attention to the measurement of the first generation lepton - slepton - gaugino coupling at an @xmath11 linear collider .
much study has been undertaken to determine how accurately we can expect to measure these couplings . two relevant processes to consider are chargino production and slepton production .
these processes involve the lepton - slepton - gaugino coupling by way of the @xmath0-channel exchange of a sfermion or gaugino .
the measurement of the gaugino coupling via the chargino production process has been considered in ref .
@xcite . in that study
the authors worked under the assumption that only chargino production was available .
it was shown that the @xmath12 coupling could be determined to about 20% . incorporating knowledge about the sneutrino mass allows for an improved @xmath13-@xmath14 determination of the @xmath15 coupling @xcite .
a more precise determination is possible by studying slepton production . in ref .
@xcite it was demonstrated that by measuring @xmath16 production the u(1 ) @xmath17 coupling could be determined to about @xmath18 .
note that this is the same order of magnitude as the splitting due to heavy squarks ( see eq .
( [ ln msq ] ) ) . in ref .
@xcite it was estimated that a @xmath19 measurement of the u(1 ) gaugino coupling could be achieved by measuring @xmath20 production at an @xmath21 collider .
we perform a full one - loop calculation of the slepton production cross - section within the mssm .
since we are interested in the sensitivity to the squark mass scale , we assume the squark mass spectrum is well approximated by a single mass scale , @xmath22 , and allow for a large hierarchy between the squark and slepton masses .
we include only ( s)quark loops in the calculation , because the correction is enhanced by a color factor and the number of generations .
the remaining corrections are small , and if we did include them we expect our conclusions would not change . in the next section
we discuss the radiative correction calculation .
we point out that , to a good approximation , the one - loop @xmath0-channel amplitudes can be rewritten in the same form as the tree - level amplitudes , with the replacement of the tree - level parameters with renormalized effective parameters .
hence we introduce the effective coupling , the effective masses , and the effective mixing matrix . in section 3
we show our numerical results .
we show the renormalization scale and squark mass dependence of the various production modes , for the different electron polarizations .
we compare the tree - level , leading - log , effective , and full cross - section calculations .
the results show that the effective theory approximation describes the full result very well .
we also discuss the @xmath23-term dependence . in section 4
we discuss how well we can measure the squark loop correction to the coupling , and thereby constrain the squark mass , assuming both slepton and chargino production are possible .
we show the statistical significance of the results by combining our knowledge of the superpartner masses and cross - sections .
the slepton production cross - section is a sensitive function of the slepton mass .
the uncertainty in the slepton mass measurement is quite important in this analysis .
we emphasize that the chargino production cross - section does not have this problem of the final - state mass uncertainty . in the last section ,
section 5 , we give our conclusions .
in this section we discuss the calculation of the cross - section of @xmath24 , where @xmath25 , including one - loop ( s)quark corrections . the full result is explicitly given in appendix a ; here we restrict ourselves to outline the general features of the calculation . as a specific example , we consider the process @xmath26 .
the tree - level amplitude is given by @xmath27u\ , \ ] ] where @xmath28 is the center - of - mass energy , @xmath29 is the @xmath0-channel momentum , and @xmath30 .
the sine of the weak mixing angle is denoted @xmath31 , @xmath32 is the neutralino mixing matrix , and the @xmath33 wave functions are denoted @xmath34 and @xmath35 . to evaluate the one - loop amplitude , we treat all the parameters appearing in this tree - level expression as running ( dimensional reduction with modified minimal subtraction @xcite ) quantities , and add the contributions from the one - loop diagrams ( see fig . 1 ) .
note that the ( s)quark loop corrections do not give rise to external wave - function renormalization .
we take the gauge boson pole masses and the standard model electromagnetic coupling @xmath36 as inputs .
the parameters are related to the corresponding input parameters by familiar expressions .
( hatted symbols are renormalized quantities .
we do not hat the soft supersymmetry breaking parameters . ) for example , the running gauge boson masses are given by @xmath37 where the @xmath38 are the transverse parts of the renormalized gauge - boson self - energies .
the gauge coupling in the full theory @xmath39 is related to that of the effective theory ( i.e. the standard model in the scheme ) as for to conversion is omitted because we systematically ignore the gauge boson loops . ]
@xmath40 the su(2)@xmath41 and u(1)@xmath42 gauge couplings are then obtained in terms of the gauge boson running masses , @xmath43 note that the argument of the gauge - boson self - energies is the external momentum squared .
they implicitly depend on the renormalization scale @xmath44 .
explicit expressions for @xmath45 are given in @xcite .
since we only include the ( s)quark loop corrections , for consistency the parameters only run by virtue of the ( s)quark contributions to their rge s . when @xmath46 , @xmath47 , @xmath48 , @xmath49 are needed , we use @xmath50 , @xmath51 , @xmath52 , @xmath53 . the scale choice of each parameters has been made because the relations @xmath54 , @xmath55 , @xmath56 hold when @xmath57 , and @xmath58 . with these counter - terms and the one - loop diagram contributions presented in appendix a , we obtain the full one - loop amplitude .
the differential cross - section is then computed , @xmath59 to obtain the total cross - section , we integrate over the @xmath0-channel momentum numerically .
the cross - sections for the other selectron and sneutrino production processes are obtained similarly . in appendix b ,
we show how the one - loop corrected @xmath0-channel amplitude can be well approximated by a tree - level form , @xmath60u\ , \label{eff amp}\ ] ] where @xmath61 is the effective bino coupling defined as @xmath62 and @xmath63 is the bino - bino component of the neutralino two - point function defined in eq .
( [ eqb4 ] ) .
the @xmath64 and @xmath65 are the effective neutralino mixing matrix and neutralino masses obtained by diagonalizing the effective neutralino mass matrix @xmath66 , defined in eq .
( [ eqb6 ] ) .
we note that @xmath67 are the chargino and neutralino pole masses . also , @xmath68 , @xmath64 , and @xmath65 are scale independent quantities to @xmath69 .
for example , the scale dependence of @xmath70 is cancelled by the implicit scale dependence of @xmath63 .
the physical correction appears as a violation of the tree - level relation @xmath71 , where @xmath72 is the effective coupling in the standard model .
it is obtained by eqs .
( 3 - 6 ) , including only quark contributions in the self - energies .
the @xmath73 terms in @xmath74 and @xmath75 correspond to the change in the beta functions above and below the squark threshold .
also , the leading logarithms of the corrections @xmath76 at @xmath77 are exactly those of eq . 1 , showing that the rge approach of ref .
@xcite gives the proper results .
the other effective couplings @xmath78 , @xmath79 and @xmath80 are given by analogous formulae ( see eq .
( [ eqb10 ] ) ) .
the expressions for @xmath79 and @xmath80 are different .
however , they approach each other in the large @xmath81 limit .
the difference between these effective couplings arises from finite electroweak symmetry breaking contributions proportional to the vacuum expectation value of the higgs bosons . by gauge symmetry ,
the leading @xmath82 correction is the same for the different effective couplings . in section 3
, we will see that the various finite corrections are not large compared to the leading @xmath82 correction .
in contrast to the @xmath0-channel amplitudes , the squark corrections to the @xmath28-channel amplitudes are proportional to @xmath83 . since we take as inputs the standard model coupling , and the @xmath84- and @xmath85-boson pole masses , the decoupling theorem applies , and the effect of the heavy squarks is negligible in the large squark mass limit .
part of the radiative corrections we are considering are closely related to the corrections to the neutralino and chargino masses @xcite .
since the chargino and neutralino masses will be measured in future collider experiments , we will in some of the discussion take the pole masses @xmath86 , @xmath87 , @xmath88 as inputs .
we define @xmath89 , @xmath90 , @xmath91 as the ( 1,1 ) , ( 2,2 ) , and ( 3,4 ) components of the effective neutralino mass matrix @xmath92 . @xmath90 and @xmath93 could also be defined as the ( 1,1 ) , ( 2,2 ) components of the effective chargino mass matrix @xmath94 .
these definitions differ by @xmath95 , as do the different definitions of the wino effective couplings .
the grand unification condition leads to the one - loop relations @xmath96 .
we comment that there is a @xmath97 correction to the @xmath98 ratios .
we set @xmath99 , but we do not impose any such relation in our analysis .
we next study the numerical dependence of the one - loop corrected cross - sections of @xmath24 ( @xmath25 ) on the squark mass .
we consider the case where the initial electron is completely longitudinally polarized .
we therefore treat the following eight modes , @xmath100 in the squark sector , we set @xmath101 where @xmath102 , and @xmath103 are the soft - breaking masses of the left - handed , right - handed up , and right - handed down - type squarks , respectively , and @xmath104 , @xmath105 are the soft - breaking higgs - squark - squark trilinear couplings , in the convention of ref .
we assume generation independence for these parameters .
the renormalization scale independence of the one - loop cross - section serves as an important check of the calculation . in fig .
2 we show the @xmath44 dependence of the tree - level cross - sections ( defined as the first terms of eqs .
( [ eqa2][eqa6 ] ) ) and the corrected ones , for @xmath106 gev , @xmath107 gev , @xmath108 gev , @xmath109 , @xmath110 gev , @xmath111 gev , and @xmath112 gev .
the tree - level cross - sections vary linearly with @xmath113 .
this dependence is due to the running of @xmath114 and @xmath115 .
one can see that this @xmath44 dependence almost vanishes after the inclusion of the one - loop contributions .
the remaining @xmath44 dependence of the corrected cross - sections , which comes from residual higher - order corrections in the numerical calculation , is very small for @xmath116 tev .
the renormalization scale dependence suggests the uncertainty in the calculation due to higher order corrections .
we next consider the squark mass dependence of the corrected cross - sections . as stated , we take the standard model electromagnetic coupling and the @xmath84 and @xmath85 boson pole masses as inputs . here
we also take the three pole masses @xmath117 , @xmath87 , @xmath88 ) , and @xmath53 as inputs .
we assume @xmath118 , in which case @xmath119 , @xmath120 , and @xmath121 hold , where @xmath122 , @xmath123 , and @xmath91 are the ( 1,1 ) , ( 2,2 ) , and ( 3,4 ) elements of the effective neutralino mass matrix @xmath124 , defined in eq .
( [ eqb6 ] ) .
we show in fig . 3 the @xmath22 dependence of the cross - sections for @xmath125 gev , @xmath126 gev , @xmath127 gev , @xmath109 , @xmath110 gev , @xmath128 , @xmath129 , and @xmath112 gev . here
we normalize the cross - sections to the tree - level values .
the tree - level cross - sections are functions of the standard model gauge couplings @xmath130 ( at @xmath131 gev ) and the tree - level @xmath132 and @xmath49 .
the latter parameters are found by inverting the tree - level relations between @xmath133 and the input masses @xmath134 and @xmath135 .
hence , the tree - level cross - sections are independent of the squark mass scale .
the numerical values of the tree - level cross - sections are listed in table 1 .
.total tree - level cross - sections , in fb , of @xmath24 for the various modes .
input parameters are given in the text . [
cols="<,^,^,^,^,^",options="header " , ] the one - loop corrected cross - sections of the modes which have a @xmath0-channel contribution are similar to tree - level ones at @xmath136 gev , but increase linearly with @xmath137 .
because the effective masses are equal to the input pole masses to @xmath138 , the squark mass dependence of the one - loop corrected cross - sections is primarily due to the difference between the effective theory gauge and gaugino couplings . the two channels which have @xmath139 contributions ( @xmath140 ) show the largest @xmath81 dependence . the destructive interference between the @xmath28- and @xmath0-channel amplitudes accounts for the enhancement of the @xmath141 correction relative to the @xmath142 case . other channels @xmath143
, @xmath144 show smaller @xmath81 dependence from @xmath145 contributions .
nevertheless , these @xmath81 dependences are significantly larger than the renormalization scale dependence of the corrected cross - sections ( see fig . 2 ) . in contrast , the remaining channels , which have only @xmath28-channel contributions , show very little @xmath81 dependence , as explained in section 2 . we now consider two approximations to the full calculation
first , we consider the effective theory approximation ( eta ) , where the effective masses , couplings , and mixing matrices are used in the @xmath0-channel amplitudes ( see eq .
( [ eff amp ] ) ) .
we make a further approximation in the eta by fixing the momenta of the effective parameters to an average @xmath0-channel momentum , @xmath146 .
the second approximation is the leading log approximation ( lla ) , which is the eta in leading logarithm approximation .
the full calculation differs from the eta by @xmath147 terms which arise when factorizing the amplitude ( and by @xmath148 corrections due to fixing @xmath149 ) .
the lla differs additionally by constant and @xmath150 corrections .
4(a ) and ( b ) shows the full correction , the eta , and the lla , for @xmath151 and @xmath152 @xmath153 , respectively .
4(b ) we use the effective coupling @xmath154 , rather than @xmath155 , which introduces additional @xmath156 differences between the full and eta calculations . figs . 4 show that the full calculation is well approximated by the eta , even for small squark masses .
the lla does correctly describe the logarithmic increase in the full correction ( as the name implies ) , but there is a constant difference of 0.6% [ 1.0% ] seen between the full and lla curves in fig .
4(a ) [ ( b ) ] .
this constant difference is largely accounted for by the correct matching of the effective coupling , where , at zero momentum , and for zero quark mass , the logarithms in eq .
( [ ln msq ] ) are replaced as @xcite @xmath157 in fig . 4
there is a slight discrepancy between the eta and the full results at very large @xmath22 .
this is due to the resummation of the effective mass corrections inherent in the derivation of the effective amplitude . before leaving this section , we comment on the dependence of the cross - sections on the parameter @xmath158 , which can give sub - leading corrections through the left - right mixing of third generation squarks .
one might expect that the @xmath23 dependence could become significant for small ( @xmath159 ) values of @xmath49 .
5 shows that for @xmath160 the dependence is much smaller than the @xmath81 dependence . for the case
@xmath129 , the cross - section is almost independent of @xmath23 if @xmath161 tev .
for @xmath162 , there is more dependence .
for example , for @xmath163 tev , the cross - section increases by about 3 per cent as @xmath23 varies from @xmath164 to @xmath165 .
in the previous section , we have shown the squark mass dependence of the slepton production cross - section @xmath167 .
we found the cross - sections are well described by that of an effective theory , where the tree - level relations between the fermion - sfermion - gaugino couplings and the gauge couplings are renormalized . in this section
, we consider the determination of the squark mass @xmath81 through the measurement of the slepton production cross - section .
as described in section 2 , the sfermion and chargino / neutralino production cross - sections depend on the gaugino mass parameters @xmath46 and @xmath47 , the higgsino mass parameter @xmath48 , the ratio of the vacuum expectation values @xmath49 , and the sfermion mass @xmath168 .
the cross - sections also depend on the effective fermion - sfermion - gaugino coupling @xmath169 , which is nicely parameterized by @xmath73 . by constraining @xmath170 we can determine @xmath22 , if the rest of the parameters are known accurately enough . it has been demonstrated that an accurate determination of @xmath171 and @xmath168 is indeed possible if sfermions @xmath172 are produced and dominantly decay into a charged lepton @xmath173 and a chargino or neutralino @xmath174 .
the simple kinematics of the two body decays result in a flat energy distribution of the final state leptons between the endpoint energies @xmath175 and @xmath176 , which are simple functions of @xmath177 , @xmath168 , and @xmath171 .
the measured endpoint energies therefore determine @xmath168 and @xmath178 . in ref .
@xcite , the results of detailed mc simulations were presented for @xmath179 and @xmath180 .
it was shown that the @xmath181 and @xmath182 masses could be measured to better than 1% .
recently , baer et al .
@xcite performed a mc study for the case that left - handed sfermions are produced and decay into a gaugino - like chargino or neutralinos . in their example
called point 3 , @xmath142 production is followed by @xmath183 . the decay mode @xmath184 @xmath185 @xmath186 ( @xmath187 @xmath188 is background free and the measured electron endpoint energies allow for a 1% measurement of @xmath189 and @xmath190 process @xcite . the mass measurement from the jet energy distributions in chargino decay @xmath191 gives only a few percent measurement of the masses due to the uncertainty in the jet momentum . ] . the results of ref .
@xcite encourage us to consider their example point 3 .
the chosen parameter set corresponds to @xmath192 gev , @xmath193 gev , @xmath194 gev and @xmath195 gev , and the lightest chargino and neutralinos are gaugino - like .
their study suggests that we can take @xmath196 @xmath86 , and @xmath197 as well constrained input parameters . for 20 fb@xmath198 of luminosity ,
their mc simulations show that at 68% cl , @xmath199 ) = ( 1.5 gev , 2.5 gev ) . in the following
we estimate the statistical significance of the radiative correction to the production cross - section . to most effectively constrain the squark mass
we would take into account all the possible channels .
however , for the purposes of this paper , it suffices for us to focus solely on the @xmath200 production cross - section , because it is larger than 1 pb for a left - handed electron beam , and larger than the other sparticle production cross - sections at @xmath112 gev .
the sensitivity of the sneutrino production cross - section on the chargino mixing is determined by the sign of @xmath48 . for @xmath129
, the gaugino / higgsino mixing is suppressed , and hence the cross - section depends mildly on @xmath49 .
also , there is very little @xmath135 dependence in this case . for @xmath162 ,
the mixing gives rise to significant @xmath49 and @xmath135 dependence .
the heavy chargino and neutralinos are produced once @xmath201 , and we assume that these masses can be constrained to a reasonable range or @xmath85 and a wino - like neutralino or chargino ( through its suppressed higgsino component ) .
the dominant decay process gives maximally 8 jets in the final state , and the background to the process is small .
if the energy distribution of the gauge bosons in these decays could be measured , the endpoint technique could also be applied here . ] . also , if @xmath48 is not too large , the splitting of the heavy chargino and neutralino masses is quite different for the different signs of @xmath48 .
if the splitting can be measured , the sign of @xmath48 might be determined .
for example , at point 3 with @xmath202 gev and @xmath129 ( @xmath162 ) , @xmath203 ( 27 ) gev .
the reasonably large production cross - sections of the heavier chargino and neutralinos ( @xmath204 fb at @xmath205 gev ) give us hope that these mass differences can be distinguished .
as long as @xmath135 is known to @xmath206-@xmath207 and the sign of @xmath48 can be determined , the chargino mixing uncertainty will not significantly affect the determination of the effective coupling @xmath80 .
we would first like to provide a feel for the sensitivity to the squark mass scale and @xmath49 in the ideal case where we ignore the slepton and gaugino mass uncertainties . in fig .
6 we show the statistical significance of the loop correction by plotting contours of constant cross - section . here
we fix the sneutrino mass and determine @xmath208 and @xmath47 by fixing the one - loop corrected masses @xmath86 , @xmath88 , and @xmath87 . rather than plot the contours according to the value of the production cross - section , we plot the contours corresponding to the number of standard deviations of the fluctuation of the accepted number of events .
the 1-@xmath209 fluctuation corresponds to @xmath210 , where @xmath211 is our nominal value of the number of events at @xmath111 gev and @xmath109 .
the accepted number of events @xmath212 is given by @xmath213 here we took @xmath214 and overall acceptance @xmath215 .
gev , @xmath216 gev and @xmath217 gev .
we further scale our acceptance down , because our calculation of cross - section does not include the effects of initial state radiation , beam energy spread , and beamstrahlung . ]
the number of accepted events at our nominal point @xmath218 is about 12800 for @xmath129 and 11900 for @xmath162 .
we show the contours for @xmath129 in fig .
if @xmath49 is well measured , @xmath22 is constrained to @xmath219 gev at 1-@xmath209 significance . if instead we assume the constraint @xmath220 , the mild @xmath49 dependence yields @xmath221 gev . in the case
@xmath162 ( fig .
6(b ) ) , increasing the squark mass can be compensated for by decreasing @xmath49 , and measuring sneutrino production then determines a region of the @xmath222 plane . as with @xmath129 , @xmath22
is decently constrained if @xmath49 is known ( @xmath223 gev ) . however ,
if we have @xmath220 , only the uninteresting bound @xmath224 tev can be achieved . in fig .
6(c ) we show the contours again for @xmath162 , but now we raise @xmath135 to 500 gev .
larger values of @xmath48 lead to weaker @xmath49 dependence .
however , in this case the cross - section is still sensitive to the heavy chargino mass @xmath225 .
if the heavier chargino is beyond the kinematical reach of the collider , there may be considerable uncertainty in the predicted cross - section .
the cross - section is pretty much independent of @xmath48 once @xmath226 gev , in this case .
if we vary @xmath48 from 450 to 850 gev the cross - section varies by about 2% .
it is good to keep in mind that @xmath48 may be constrained in a situation such as this by measuring the left - right asymmetry of the lighter chargino production cross - section . figs .
6 illustrate that we might have to rely on independent measurements of @xmath49 in order to obtain a nice constraint on @xmath22 . for many regions in parameter space
we expect @xmath49 will indeed be known fairly well .
for example , it was pointed out by ref .
@xcite that the measurement of the chargino forward - backward asymmetry provides a strong constraint on @xmath49 if @xmath227 .
also , for not too large slepton masses and @xmath228 , the measurement of the mass difference between @xmath229 and @xmath200 gives a sensitive measurement of @xmath49 .
other measurements can be instrumental in pinning down @xmath49 . for example
, the higgs sector constraints on @xmath49 are considered in ref .
@xcite , which tend to give decent bounds for @xmath230 .
we now turn to the effect of the mass uncertainties .
the sneutrino production cross - section depends on the masses @xmath197 , @xmath86 , @xmath88 and @xmath87 . the uncertainties in these masses will of course worsen our ability to constrain the squark mass scale . of these massses
the cross - section is most sensitive to the sneutrino mass .
all of the same chirality scalar production cross - sections suffer from the strong @xmath231 kinematic dependence ( @xmath232 ) is proportional to @xmath233 , due to the chiral structure of @xmath234-@xmath235-@xmath236 coupling . ] . near threshold
this results in an especially large sensitivity to the final - state mass .
for example , a 1% uncertainty of @xmath168 results in @xmath237 , for @xmath110 gev , and @xmath112 gev .
this immediately translates into a 5.3% error in the total cross - section .
although a simple statistical scale - up of the results of ref .
@xcite implies a sneutrino mass uncertainty of only 0.3% , this nevertheless leads to a significant degradation in our ability to constrain the squark mass scale .
( systematic errors might be the limiting factor here . )
note , however , that the measurement of the sneutrino mass in ref
. @xcite was obtained by studying a small fraction of the total sneutrino decay modes .
the mode they studied , @xmath238 @xmath239 , amounts to only about 4% of the total sneutrino decays .
using other modes , such as @xmath240 , might reduce the mass error even further . because the @xmath200 production cross - section is significantly larger than the other slepton cross - sections , isolating the various sneutrino signatures is less affected by susy backgrounds such as @xmath241 @xmath242 . in this discussion
we assume the errors are dominantly statistical .
now we show the constraint on the squark mass @xmath81 after taking into account the uncertainty of the masses @xmath243 and @xmath244 .
for this purpose , we define a @xmath166 function as @xmath245 here we consider the @xmath129 case , in which the @xmath246 and @xmath247 mass dependencies are negligible .
we also assume that @xmath49 is well measured . in fig .
7 we plot @xmath248 against @xmath81 , where @xmath249 is a minimum of @xmath166 with respect to variations in @xmath87 and @xmath197 . by construction , the region of @xmath22 where @xmath250 corresponds to @xmath251-@xmath209 error of the squark mass when
the chargino and sneutrino mass uncertainties are taken into account .
the sneutrino mass uncertainty reduces the sensitivity of the production cross - section to @xmath81 considerably , because the effect of increasing @xmath81 can be compensated for by a small increase in @xmath197 .
on the other hand , we do not find any significant effect due to non - zero @xmath244 . from fig .
7 we see that in this case , even with the sneutrino mass uncertainty , we can reasonably constrain the squark mass scale . for example , at the 1-@xmath209 level with @xmath163 tev , we constrain @xmath22 to @xmath252 tev , using the naive scale up ( from 20 fb@xmath198 to 100 fb@xmath198 ) of the statistical errors of ref . @xcite .
this corresponds to the difference between the gauge and gaugino effective couplings , @xmath253 .
this can be compared to the estimate of the constraint @xmath254 from the chargino production measurement @xcite .
such comparisons are sensitive to different choices of parameter space and other assumptions .
the sensitivity of the constraint on the mass errors can be seen by changing the errors .
if we reduce the mass uncertainties by a factor of 2 , we find the interesting constraint @xmath255 gev .
7 illustrates the large dependence on the mass uncertainties .
this sensitivity is greatly reduced in the case of chargino production . as we mentioned previously
, the @xmath0-channel amplitude is very similar to that of sneutrino production ; only its final and intermediate states are exchanged .
the external chargino wave - function renormalization induces the same @xmath73 correction to the @xmath256-@xmath200-@xmath257 coupling .
also , the chargino production cross - section is generally large ; for the set of parameters considered , it is near 1 pb for a left - handed electron beam . because the chargino cross - section is only proportional to a single power of @xmath258 , the uncertainty in the cross - section due to the final state mass is effectively reduced by a factor of 3 .
notice @xmath87 can be measured through the electron energy distribution from the @xmath200 production and decay . also , the chargino production cross - section rises quickly near threshold , which allows for a determination of its mass by a threshold scan . in figs . 6 and 7
, we assumed that the decays into the lightest chargino @xmath259 were detectable as @xmath200 production signal without background .
this may not be true , depending on the decay mode of the chargino .
for example , if both of the charginos decay into @xmath260 , it gives an @xmath261 signal event with missing energy .
pair production of @xmath262 could give the same @xmath261 signal with similar kinematics .
by su(2 ) invariance @xmath229 and @xmath200 are very close in mass , so it is natural to think that this background is necessarily problematic .
additionally , at point 3 ref .
@xcite estimates that the @xmath229 mass uncertainty is larger than that of @xmath200 ; this corresponds to a large uncertainty in the expected @xmath229 production cross - section .
fortunately , in this case , the cross - section of @xmath229 and its branching ratio into @xmath263 are small , and they are constrained by other measurements , so they give a negligible background to the @xmath264 signal .
we also assumed constant @xmath265 throughout figs .
6 and 7 , while the branching ratio is expected to vary with @xmath47 , @xmath48 , @xmath49 , @xmath266 , and @xmath22 .
notice , @xmath47 , @xmath48 , and @xmath266 are constrained from mass measurements , while in fig .
6 @xmath49 and @xmath22 are our fit parameters , so the inclusion of the branching ratio would not introduce additional uncertainty in the @xmath22 constraint unless there is a large correlation .
the chargino branching ratios ( @xmath267 and @xmath268 ) are likewise determined by the measured sparticle masses and their cross - section and fit parameters , or they are directly constrained by the study of chargino and neutralino production at low energy - boson , the branching ratios become somewhat sensitive to squark and slepton masses . ] .
we did not include the expected luminosity and acceptance errors .
needless to say , it is extremely important to achieve a low luminosity error , since we are dealing with a potentially sub - percent measurement .
the effect of the luminosity error can be directly seen in fig .
6 . in the chargino study of ref .
@xcite , the systematic error of acceptance is claimed to be significant in the two jet + lepton final state .
there , the large background from @xmath269 production requires a very tight cut , leading to low acceptance and large systematic error . here
, we consider sneutrino production followed by decays to charginos .
there is very low sm background to the final state @xmath270 missing energy , where the final state electron and positron are hard and isolated .
also , the sm background has very different kinematical structure from the signal . given the precisely measured chargino , neutralino and sneutrino masses , it is reasonable to assume the acceptance error is negligible . in ref .
@xcite , the possibility to measure @xmath271 from chargino production has been studied , resulting in a rather poor constraint @xmath272 . in that analysis
, the assumption was made that only chargino production was accessible .
this leads to a rather large uncertainty due to the unknown sneutrino mass , which could only be determined with a polarized beam through the differential cross - section of the chargino production itself .
that was a reasonable and conservative assumption at the time , before a detailed mc simulation of @xmath200 production was available . including the sneutrino mass measurement will in this case
provide a reasonable constraint on @xmath273 and hence on the squark mass @xcite ; a detailed study will be presented elsewhere @xcite .
supersymmetry is a beautiful symmetry which relates bosons and fermions .
if we wish to determine whether this symmetry is realized in nature , the relations imposed between particles and their superpartners must be confirmed by experiment . of course , discovering a particle with the quantum numbers of a superpartner is the first very important step in this procedure . an equally important test , though , is the confirmation of the hard relations imposed by supersymmetry , for example , the equivalence of the gauge and gaugino couplings . it has been argued that a next generation linear collider would be an excellent tool to verify supersymmetry in this respect .
production cross - sections such as @xmath274 and @xmath275 involve the @xmath0-channel exchange of gauginos or sleptons , so they depend on gaugino couplings .
several papers @xcite have shown that it is possible to verify the equivalence of the gaugino and gauge couplings in a model independent way , by fitting the data to the cross - sections with arbitrary gaugino couplings .
the sensitivity ranges from about 1% to 10% , depending on the assumptions . in this paper
, we approached this problem from a somewhat different direction . because supersymmetry must be badly broken by soft breaking terms , the tree - level relations of the couplings are also broken , by radiative corrections .
the corrections are logarithmically sensitive to the splitting of the supersymmetry multiplets . to quantify this ,
we have calculated the full one - loop correction due to ( s)quark loops of the slepton production cross - sections .
we have explicitly demonstrated that the full one - loop amplitudes can be factorized into a form which corresponds to an effective theory in which the heavy squarks are integrated out .
the difference between the effective lepton - slepton - gaugino couplings @xmath276 and the effective gauge couplings @xmath277 is given by a coupling factor times @xmath278 .
the one - loop formulae for the chargino and neutralino production cross - sections are not given in this paper , but the effective gaugino couplings receive the same corrections to logarithmic accuracy .
we gave an explicit example which illustrates that the statistics at the future linear collider may be enough to constrain the squark mass scale through the measurement of the slepton production cross - section .
we found , with 1-@xmath209 significance , @xmath22 could be constrained to @xmath252 tev by the measurement of the sneutrino production cross - section .
this corresponds to a measurement of the difference between the su(2 ) gauge and gaugino couplings @xmath279 .
we found this constraint in the @xmath129 case where we took into account the errors ( based on existing mc simulation ) of the sneutrino and light chargino masses , but assumed @xmath49 was well constrained by other measurements . in the @xmath162 case
the slepton cross - sections are much more sensitive to @xmath49 and the heavy ( higgsino dominated ) chargino and neutralino masses .
we need to rely on additional measurements in order to precisely measure the gaugino couplings in this case . for either sign of @xmath48
, the mass of the sleptons and gauginos must be measured very precisely in order to successfully constrain the squark mass scale via production cross - section measurements .
the effect of the slepton mass uncertainty on the prediction of the slepton production cross - section is especially significant , as discussed in section 4 . in order to determine the ultimate sensitivity of this procedure ,
a thorough study of the systematic uncertainties in the slepton mass measurements is necessary .
the prediction of the chargino production cross - section is less sensitive to the uncertainties in the masses .
the limiting factor in the chargino mode is sm background and statistics .
it is important to note that the constraint on the squark mass scale can be stronger than the one presented in this paper . here
, we estimated the sensitively to squark mass scale by utilizing sneutrino production followed by its decay into a chargino and an electron . depending on the spectrum and center - of - mass energy , there will typically be many other production processes which involve @xmath0-channel exchange of gauginos or sleptons , and all those amplitudes have @xmath82 corrections .
the constraint on the squark mass we have realized here could be unique in the sense that this information may not be available at the lhc .
even if the lhc squark production rate is large , the gluino could be produced in even larger numbers , creating a large irreducible background to the squark signal .
a large gluino background could make the extraction of the squark mass from kinematical variables difficult .
on the other hand , if information on the squark masses is obtained at the lhc , we would have rather accurate predictions for the gaugino couplings . in this case
, the measurement of the production cross - sections we considered here would constrain new supersymmetry - breaking physics with standard model gauge quantum numbers . in a sense , the study proposed here is similar in nature to studies performed at lep and slc .
the physics of gauge boson two - point functions has been studied extensively at lep and slc , and it has provided strong constraints on new physics .
similarly , a future lc and the lhc might provide precision studies of the gaugino two - point functions , to realize a supersymmetric version of new precision tests .
we thank manuel drees and jonathan feng for useful discussions .
in this section , we show the explicit forms of the quark - squark loop corrections to the amplitudes of the processes @xmath280 where @xmath281 . we adopt the renormalization scheme , and the suffix is implicit throughout this and the next sections . the neutralino mass matrix @xmath282 and the chargino mass matrix @xmath283 are functions of @xmath284 , @xmath285 , @xmath286 , @xmath287 , @xmath288 , and @xmath289 as follows @xcite , @xmath290 here @xmath291 , @xmath292 , @xmath293 and @xmath294 .
the @xmath44 dependence of the right hand side is implicit .
a unitary matrix @xmath212 is defined so that @xmath295 .
the chargino mass matrix is @xmath296 and the chargino mass matrix can be diagonalized by two unitary matrices @xmath297 and @xmath298 as @xmath299 .
we order the charginos and neutralinos according to their masses , so , for example , @xmath300 is the lightest neutralino .
we first show the @xmath0-channel amplitudes . they are expressed in terms of two - point functions @xmath301 , @xmath302 of neutralinos @xmath303 and charginos @xmath304 .
@xmath305 are decomposed as @xmath306 where the momentum @xmath29 flows from @xmath307 to @xmath308 .
@xmath309 are the chiral projectors , @xmath310 . in the following the argument of the @xmath311 functions
is @xmath149 .
the corrected @xmath0-channel amplitude @xmath312 for @xmath313 is then expressed as @xmath314\ , \label{eqa2}\end{aligned}\ ] ] where @xmath315 .
similarly , other @xmath0-channel amplitudes are @xmath316\ , \label{eqa3 } \\ i{\cal m}_{lr}^{(t ) } & = & -i \bar{v}(p_2 ) p_l u(p_1 ) b^{0*}_{\tilde{e}ri}a^0_{\tilde{e}lj } \left [ \frac{m_{\tilde{\chi}^0_i}\delta_{ij}}{p^2-m_{\tilde{\chi}^0_i}^2 } \right .
\nonumber \\ & & \left .
-\frac{1}{(p^2-m_{\tilde{\chi}^0_i}^2)(p^2-m_{\tilde{\chi}^0_j}^2 ) } ( p^2(m_{\tilde{\chi}^0_j } \sigma_{0ij}^l + m_{\tilde{\chi}^0_i } \sigma^r_{0ij})+m_{\tilde{\chi}^0_i}m_{\tilde{\chi}^0_j } \sigma^m_{0ij } + p^2 \sigma^{m*}_{0ij})\right]\ , \label{eqa4 } \\ i{\cal m}_{rl}^{(t ) } & = & -i \bar{v}(p_2 ) p_r u(p_1 ) a^{0*}_{\tilde{e}li}b^0_{\tilde{e}rj } \left [ \frac{m_{\tilde{\chi}^0_i}\delta_{ij}}{p^2-m_{\tilde{\chi}^0_i}^2 } \right .
\nonumber \\ & & \left .
-\frac{1}{(p^2-m_{\tilde{\chi}^0_i}^2)(p^2-m_{\tilde{\chi}^0_j}^2 ) } ( p^2(m_{\tilde{\chi}^0_i } \sigma_{0ij}^l + m_{\tilde{\chi}^0_j } \sigma^r_{0ij } ) + p^2 \sigma^m_{0ij}+m_{\tilde{\chi}^0_i}m_{\tilde{\chi}^0_j } \sigma^{m*}_{0ij})\right]\ , \label{eqa5 } \\ i{\cal m}_{\nu\nu}^{(t ) } & = & -i \bar{v}(p_2 ) \psla p_l u(p_1 ) a^{-*}_{\tilde{\nu}li}a^-_{\tilde{\nu}lj } \left [ \frac{\delta_{ij}}{p^2-m_{\tilde{\chi}^+_i}^2 } \right .
\nonumber \\ & & \left .
- \frac{1}{(p^2-m_{\tilde{\chi}^+_i}^2 ) ( p^2-m_{\tilde{\chi}^+_j}^2 ) } ( m_{\tilde{\chi}^+_i } m_{\tilde{\chi}^+_j } \sigma_{-ij}^l + p^2 \sigma^r_{-ij}+m_{\tilde{\chi}^+_j } \sigma^m_{-ij } + m_{\tilde{\chi}^+_i } \sigma^{m*}_{-ij})\right]\ , \label{eqa6 } \end{aligned}\ ] ] here @xmath317 denote the couplings of neutralinos and charginos to fermions and sfermions , @xmath318 where @xmath319 and @xmath320 are su(2 ) doublets .
explicit forms of @xmath317 for the gauge eigenstates of @xmath321 are written in terms of mixing matrices @xmath322 and yukawa couplings @xmath323 of fermions @xmath324 as @xmath325 we next show the explicit forms of the two - point functions @xmath326 and @xmath327 from quark - squark loops .
for neutralinos , we have @xmath328 for charginos we have @xmath329 here @xmath330 are t hooft - veltman functions in the convention of @xcite . finally , we show the corrected @xmath28-channel amplitudes @xmath331 for @xmath332 . in terms of the transverse part of the gauge - boson self - energies , @xmath333 , they are expressed as @xmath334u(p_1 ) \nonumber\\ & & + \ 2ig_z^2(i_{3\tilde{\ell}}-s_w^2q_{\tilde{\ell } } ) \frac{1}{s - m_z^2 } \left ( 1-\frac{\pi^t_{zz}(s)-\pi^t_{zz}(m_z^2)}{s - m_z^2 } \right ) \nonumber \\ & & \times \bar{v}(p_2 ) \psla ( i_{3e}p_l - s_w^2q_e)u(p_1 ) , \label{eqa11}\end{aligned}\ ] ] where @xmath335 , @xmath336 , and @xmath337 is the @xmath85-boson pole mass .
in this appendix , we show how the full one - loop @xmath0-channel amplitudes factorize into a form corresponding to an effective theory .
the effective theory amplitude depends on effective couplings , masses , and mixing matrices .
we explicitly show what approximations are necessary for this , and justify them . as a by - product , we show the cancellation of uv divergences in the one - loop amplitudes . here , for brevity , only the case of @xmath338 production is discussed .
the generalization to other channels is straightforward .
we first prove that the loop corrected @xmath0-channel amplitude ( eq . ( [ eqa3 ] ) ) can be written in terms of the mass matrix of neutralinos @xmath339 in the gauge eigenbasis , instead of the mixing matrix @xmath212 and @xmath340 . by substituting eq .
( [ eqa8 ] ) , eq .
( [ eqa3 ] ) takes the form ( the @xmath341 s are functions of @xmath149 ) @xmath342 n_{j1}x\ . \label{eqb1}\end{aligned}\ ] ] here we use abbreviations @xmath343 and @xmath344 . at tree - level , @xmath345 , diagonalized masses @xmath346 , and @xmath212 are related by @xmath347 we can transform the tree - level part of eq .
( [ eqb1 ] ) as @xmath348_{\alpha\beta } \nonumber \\ & = & \left [ n^{\dagger}\frac{1 } { p^2-ny^{\dagger}yn^{\dagger}}n \right]_{\alpha\beta } \nonumber \\ & = & \left [ \frac{1}{p^2-y^{\dagger}y } \right]_{\alpha\beta}\ , \label{eqb3}\end{aligned}\ ] ] where @xmath349 in this case .
the loop corrected part can be written by noting that the two - point functions @xmath311 are given in terms of the gauge eigenbasis functions @xmath350 , as @xmath351 next , we absorb the main loop contributions into an effective neutralino mixing matrix and effective couplings .
we define the effective neutralino mass matrix as @xmath353 the one - loop pole masses @xmath354 are obtained from @xmath355 as @xmath356(m_i^2)\ , \end{aligned}\ ] ] the amplitude eq .
( [ eqb5 ] ) becomes , up to @xmath357 , @xmath358 _
\label{eqb7}\end{aligned}\ ] ] the remaining @xmath359 have an important property .
their off - diagonal terms are generated by the breaking of su(2)@xmath360u(1 ) gauge symmetry in the quark - squark sector , and are uv finite .
for sufficiently heavy squarks the contributions of the factors @xmath361 are insignificant and can be dropped .
then , by introducing the effective gaugino coupling @xmath362 the @xmath0-channel amplitude is expressed in a very simple form @xmath363_{11}x \nonumber\\ & & = \left ( \frac{\bar{g}^2_{e\tilde{e}_r\tilde{b } } \overline{n}_{i1}^*\overline{n}_{i1}(p^2 ) } { p^2-\overline{m}_i^2(p^2 ) } \right)x\ , \label{eqb9}\end{aligned}\ ] ] where @xmath364 and @xmath365 are the effective masses and the effective mixing matrix , respectively , obtained by diagonalizing @xmath366 .
( [ eqb9 ] ) takes the same form as the tree - level amplitude eq .
( [ eqb1 ] ) with the replacement of the neutralino masses and couplings with the effective ones .
it is now evident that the corrected amplitude eq .
( [ eqb1 ] ) is uv finite , since both @xmath366 and @xmath367 are finite .
the @xmath0-channel amplitudes for other processes are similarly expressed in terms of effective mass matrices and couplings of neutralinos and charginos .
below we list the forms of other effective lepton - slepton - gaugino couplings which are relevant to our processes : @xmath368 note that the effective couplings are independent of ( s)fermion flavor , and hence the @xmath369 corrections are universal , in analogy with the oblique corrections of the gauge bosons @xcite . the lep collaborations aleph , delphi , l3 , opal , the lep electroweak working group and the sld heavy flavour group , preprint cern - ppe/96 - 183 ; j. ellis , g.l .
fogli and e. lisi , phys . lett . *
b389 * , 321 ( 1996 ) ; w. de boer , a. dabelstein , w. hollik , w. msle , u. schwickerath , hep - ph/9609209 v4(nov , 1996 ) ; k. hagiwara , d. haidt , s. matsumoto , hep - ph/9706331 .
i. hinchliffe , f.e .
paige , m.d .
shapiro , j. soderqvist , w. yao , phys .
d * 55 * , 5520 ( 1997 ) .
jlc group , jlc-1 , kek report no .
9216 ( 1992 ) .
t. tsukamoto et al . ,
d * 51 * , 3153 ( 1995 ) .
feng , m.e .
peskin , h. murayama , x. tata , phys .
d * 52 * , 1418 ( 1995 ) .
nojiri , phys .
d * 51 * , 6281 ( 1995 ) .
h. baer , r. munroe , x. tata , phys .
d * 54 * , 6735 ( 1996 ) .
nojiri , k. fujii , t. tsukamoto , phys .
d * 54 * , 6756 ( 1996 ) .
cheng , j.l .
feng , n. polonsky , hep - ph/9706438 .
cheng , j.l .
feng , n. polonsky , hep - ph/9706476 .
l. randall , e. katz , s. su , hep - ph/9706478 and mit
ctp2646 , inpreparation .
pierce , s. thomas , slac
pub7474 , in preparation .
chankowski , phys .
d * 41 * , 2877 ( 1990 ) .
cohen , d.b .
kaplan , a.e .
nelson , phys . lett .
* b388 * , 588 ( 1996 ) . w. siegel , phys . lett . * 84b * , 193 ( 1979 ) ; d.m .
capper , d.r.t .
jones , p. van nieuwenhuizen , nucl .
b167 * , 479 ( 1980 ) ; i. antoniadis , c. kounnas , k. tamvakis , phys . lett . *
119b * , 377 ( 1982 ) .
grifols , j. sola , nucl .
b253 * , 47 ( 1985 ) .
pierce , j.a .
bagger , k.t .
matchev , r .- j .
zhang , nucl .
b491 * , 3 ( 1997 ) .
d. pierce , a. papadopoulos , nucl .
* b430 * , 278 ( 1994 ) ; d. pierce , a. papadopoulos , phys .
d * 50 * , 565 ( 1995 ) .
gunion , h.e .
haber , nucl .
b272 * , 1 ( 1986 ) ; * b402 * , 567 ( 1993)(e ) .
feng , t. moroi , hep - ph/9612333 .
s. kiyoura , m.m .
nojiri , d.m .
pierce , y. yamada , work in progress . | we investigate an important radiative correction to the slepton and chargino production processes . we consider a supersymmetric spectrum with a large splitting between the squark and slepton masses . in this case , in the effective theory below the squark mass threshold , the supersymmetric slavnov - taylor identities which enforce the equality of the gauge and gaugino couplings are violated .
the gaugino propagators receive potentially large corrections due to virtual quark / squark loop effects .
we compute the full one - loop ( s)quark corrected slepton production cross - sections .
the @xmath0-channel one - loop scattering amplitudes are factorized into an effective chargino / neutralino mass matrix and an effective fermion - sfermion - gaugino coupling .
the difference between the effective gaugino coupling and the gauge coupling is proportional to @xmath1 in the large squark mass limit .
we find that the one - loop corrected slepton production cross - sections can depend on the squark mass strongly , up to @xmath2 .
we investigate the squark mass sensitivity of the slepton cross - section measurements at a future linear collider . for sneutrino production accessible at @xmath3 gev
there can be sensitivity to squark masses at or larger than 1 tev . 8.95 in 6.9 in -0.5truein -0.2truein kek th 520 + slac pub7558 + tu522 + june 1997 + * slepton production as * + 0.5 cm * a probe of the squark mass scale * mihoko m. nojiri,@xmath4 damien m. pierce,@xmath5 and youichi yamada@xmath6 _ @xmath4kek theory group , oho 1 - 1 , tsukuba , ibaraki , 305 japan + @xmath5stanford linear accelerator center , stanford university , stanford , california 94309 , usa + @xmath6department of physics , tohoku university , sendai , 980 - 77 , japan _ _ d.m.p .
is supported by department of energy contract de
ac0376sf00515 .
m.m.n .
is supported in part by grant in aid for science and culture of japan ( 07640428 , 09246232 ) , and jsps japanese - german cooperative science program . _ |
let @xmath1 be a connected , but possibly noncompact , symplectic manifold without boundary .
a subset @xmath2 is said to be _
displaceable _ if there is a compactly supported hamiltonian diffeomorphism @xmath3 such that @xmath4 , and if no such @xmath5 exists , then @xmath6 is said to be _
ever since arnold conjectured that certain lagrangian submanifolds are nondisplaceable , a central theme in symplectic topology has been to determine what subsets are displaceable and what subsets are not .
we will work with symplectic toric manifolds , where a @xmath7-dimensional symplectic manifold @xmath8 is _ toric _ if it is equipped with an effective hamiltonian action of an @xmath9-torus @xmath10 .
associated to a toric manifold is a moment map @xmath11 where the hamiltonian flow of @xmath12 generates the action of the @xmath13-th component circle in @xmath10 , and if @xmath14 is proper the image of @xmath14 is a polytope @xmath15 , called the _
moment polytope_. symplectic toric manifolds come with a natural family of lagrangian tori , namely for each @xmath16 the fiber @xmath17 is a lagrangian torus and an orbit of the @xmath10-action . in this paper
we will restrict ourselves to the study of the displaceability of the lagrangian toric fibers in symplectic toric manifolds , and will assume unless explicit mention is made to contrary that the moment map is proper . determining which lagrangian fibers are nondisplaceable involves two complementary tasks : building and computing invariants that obstruct displaceability , and finding general sufficient criteria for when lagrangian fibers are displaceable .
there are now many well - developed floer - theoretic tools that can be used to prove that certain lagrangian fibers are nondisplaceable : quasi - states from hamiltonian floer homology @xcite , lagrangian floer homology @xcite , and quasi - map floer homology @xcite .
in contrast the only known general method for proving that a lagrangian fiber is displaceable was introduced by mcduff in @xcite and involves the affine geometric notion of a probe in the moment polytope .
schlenk @xcite also used this method of displacement in a slightly different context .
the method of probes was later reinterpreted in @xcite in terms of symplectic reduction .
however the main contribution of that paper was to the other side of the problem in that it allowed one to deduce many nondisplaceability results from a few basic examples .
for very simple examples the method of probes perfectly complements the proven nondisplaceability results .
for instance if @xmath18 is closed and monotone with @xmath19 , then the method of probes displaces everything except an identified fiber @xmath20 ( * ? ? ?
* theorem 1.1 ) , and @xmath20 is known to be nondisplaceable .
however , in general the method of probes does not perfectly complement the proven nondisplaceability results .
the simplest such example is a hirzebruch surface @xmath21 for @xmath22 , which is the projectivization @xmath23 where @xmath24 is a line bundle of chern class @xmath25 . here
these methods leave a line segment of points with unknown displaceability properties .
the next basic example is ( a blow up of ) the weighted projective space @xmath0 , the quotient of @xmath26 by the group action @xmath27 it was pointed out in @xcite that it is possible to resolve the singularities of @xmath0 by small blow ups in such a way that there is an open set of points not displaceable by probes . on the other hand for smooth toric @xmath28-manifolds ,
the nondisplaceable fibers detected by the floer theoretic methods of @xcite lie on a finite number of line segments ( cf .
the proof of proposition [ p : noghost ] ) .
one expects that all fibers with vanishing invariants are displaceable .
our current methods do give better results than standard probes in many cases , but still are not powerful enough to prove this even in four dimensions .
as we show in proposition [ p : noghost ] , the method of quasi - map floer homology , developed by woodward @xcite , gives no more information than standard floer theoretic methods in the closed smooth case .
however , it applies also in the orbifold and noncompact cases and in those cases can give open sets of fibers that are nondisplaceable because they have nonvanishing quasi - map invariants ( called qw invariants , for short ) .
see also the orbifold version of the standard approach by cho and poddar @xcite .
figure [ f:135 ] illustrates the current knowledge about the displaceability of points in @xmath0 . here
the displaceable points are displaced by standard probes .
below , extended probes do not help in triangles . ]
as we explain in [ ss : pot ] the open set of nondisplaceable points comes from varying the position of certain ghost " facets ; cf .
also the proof of theorem [ t : dnd](i ) .
these facets are precisely the ones that can be used to resolve the singular points , and once one has used them for this purpose , so that their position is fixed , the numbers of points with nonvanishing invariants decreases .
thus as one resolves singularities by blowing up , the set of points with nonvanishing qw invariants tends to decrease . in this paper
we will introduce a technique for extending a given probe by deflecting it by an auxiliary probe .
in contrast to the nondisplaceability results explained above , this technique gives no new information for very simple orbifolds such as @xmath0 .
instead it starts to displace more fibers as we resolve the singularities by blow up . because it is a geometric method , it is very sensitive to the exact choice of blow up , i.e. to the choice of support constant that determines where the new facet is in relation to the others . in higher dimensions one could use this technique in more elaborate ways , for example by deflecting a probe several times in different directions .
however , we will restrict to the @xmath29 dimensional case since , even in this simple case , the results are quite complicated to work out precisely .
the next paragraph describes our results rather informally .
more complete definitions and statements are given later .
we say that two integral vectors in @xmath30 are * complementary * if they form a basis for the integral lattice @xmath31 .
let @xmath32 be a rational polygon in @xmath30 , i.e. the direction vectors @xmath33 of the edges @xmath34 are integral .
we call it * smooth * if the direction vectors at each vertex are complementary .
a * probe * @xmath35 in @xmath32 is a line segment in @xmath32 starting at an interior point of some edge @xmath34 ( called its * base facet * ) , whose direction @xmath36 is integral and complementary to @xmath33 .
by @xcite , if @xmath37 lies less than halfway along a probe @xmath35 then the corresponding fiber @xmath38 is displaceable . for short
, we will say that the point @xmath39 itself is displaceable .
a probe @xmath40 is said to be * symmetric * if it is also a probe when its direction is reversed .
that means that its exit point also lies at an interior point of an edge @xmath41 and also that @xmath42 is complementary to @xmath43 .
all points other than the midpoint of a symmetric probe are displaceable .
moreover there is an affine reflection @xmath44 of a neighborhood of @xmath40 in @xmath32 that reverses its direction . in this paper
we show how to lengthen a probe @xmath35 so that it still has the property that points less than halfway along are displaceable .
there are three basic methods whose effects are described in the following theorems .
* theorem [ t : noflag ] : deflecting @xmath35 via a * symmetric probe*. * theorem [ t : pdeflected ] : deflecting @xmath35 by a * parallel probe * @xmath40 , i.e. one whose base facet @xmath45 is parallel to the direction @xmath36 of @xmath35 ; these probes have flags that are parallelograms . *
theorem [ t : deflected ] : deflecting @xmath35 by an arbitrary probe @xmath40 ; these probes have trapezoidal flags . in case ( a ) , the extended probe @xmath46 ( often denoted @xmath47 for clarity ) is a union of line segments , and can be used to displace points less than halfway along it , whether these points lie before or after the intersection with @xmath40 . in cases ( b ) and ( c ) , the extended probe @xmath46 ( or more precisely @xmath48 ) is the union of the initial segment of @xmath35 together with a flag " emanating from @xmath40 , a parallelogram in case ( b ) ( cf .
figure [ f : extendedprobe ] ) or trapezoid in case ( c ) ( cf .
figure [ f : epnp ] ) .
one can only displace points less than halfway along @xmath48 that also lie before the intersection with @xmath40 .
another difficulty with case ( c ) is that the flag may taper to a point , which severely restricts its length .
in particular , if the ray in direction @xmath35 meets the base facet @xmath45 of @xmath40 at @xmath49 then @xmath48 can not be longer than the line segment from the initial point @xmath50 of @xmath35 to its intersection @xmath49 with @xmath45 .
therefore this variant is less useful , though it does displace some new points in certain bounded regions : see proposition [ p : fvanr ] . *
_ symmetric extended probes solve the displaceability problem for hirzebruch surfaces .
previous results show that all hirzebruch surfaces @xmath51 for @xmath52 , with the exception of certain surfaces @xmath53 that are ( small ) blow ups of @xmath54 , have precisely one fiber with nonvanishing floer homology .
proposition [ p : h ] shows that all the other fibers are displaceable by symmetric extended probes .
( the case @xmath55 can be fully understood using standard probes . )
+ .,width=480 ] * _ using parallel extended probes .
_ we next consider open regions of two elementary types .
section [ s : dinc2 ] considers the regions @xmath56 where @xmath57 are relatively prime integers and @xmath58 , with moment polytope @xmath59 as in figure [ fig : x1 ] .
lemmas [ l : dosecp ] and [ l : openep ] specify the points that can be displaced by probes and extended probes . figure [ f : dxnm ] shows that there are open regions of points in the moment polytope of @xmath60 where extended probes are needed . + the second basic region represents the quotient @xmath61 , where the generator @xmath62 of @xmath63 acts via @xmath64 its moment polytope is the sector @xmath65 cf .
figure [ fig : x1 ] and [ s : mnm ] .
here we contrast the set of probe displaceable points with the set of points that are nondisplaceable because they have nonvanishing qw invariants .
theorem [ t : dnd ] states the precise result .
there is an open set of points with unknown behavior whenever the hirzebruch
jung continued fraction @xmath66 has at least one of @xmath67 greater than @xmath29 .
* _ resolving a singular vertex : unbounded case . _ as one resolves the singular point of @xmath68 by blowing up , the set of probe displaceable points increases while that of points with nontrivial invariants decreases .
proposition [ p : asympresolve ] explains what happens after a single blow up , with exceptional divisor corresponding to the horizontal edge @xmath69 .
the moment polytope is then @xmath70 , the closure of @xmath71 .
remark [ r : resolvep ] ( ii ) points out that usually there is an open set of unknown points .
+ in subsection [ s : emr ] we look at some complete minimal resolutions @xmath72 . in this case
all qw invariants vanish , while typically there is still an open subset of points that we do not know how to displace ; cf . figure [ f:5 - 8tr ] .
as we show in corollary [ c : an ] , even in the easy case of an @xmath73 singularity ( which has all @xmath74 ) there are @xmath75 lines of points of unknown status , where @xmath13 is as in equation .
* _ special cases of @xmath68 and @xmath72 .
_ the first , the case @xmath55 , is treated in [ ss : om ] . here
the status of all interior points can be determined by our methods .
in fact , lemmas [ l : dndm ] and [ l : o - m ] show that all points are displaceable except in the case of @xmath76 with @xmath77 odd , in which case there is a ray of points with nontrivial qw invariants : see figure [ f:1 m ] .
+ the second is when @xmath78 in the notation of equation .
then , although there may be an open set of unknown points of @xmath68 , corollary [ c : k1 ] shows that there is at most one line of such points in the resolution @xmath72 , namely the ( affine ) bisector of the angle between the two new edges : see figure [ f : k1 ] .
* _ using general extended probes in regions of finite area . _
the above results only use extended probes of types ( a ) and ( b ) , and it is easy to see that extended probes of type ( c ) would give nothing new .
to demonstrate the use of this kind of extension , proposition [ p : fvanr ] considers the resolution of an open finite volume @xmath73-singularity , i.e. one whose moment polytope has an open upper boundary @xmath79 .
as shown in figure [ f : fvanr ] , there are some points that can be reached only by these new probes ; however there is still an open set of unknown points .
finally , we discuss a simple family of examples with closed moment polytope , namely the weighted projective spaces @xmath80 ( with @xmath81 relatively prime ) and its resolutions .
the moment polytope of @xmath80 is a triangle with vertices at @xmath82 and @xmath83 .
thus it has two singular vertices , the one at @xmath84 modelled on @xmath85 and the one at @xmath83 modelled on @xmath86 where @xmath87 .
if these sectors ( or their resolutions ) have unknown points , and the new edges coming from the resolutions are sufficiently short , one would expect there to be corresponding unknown points in the compactifcation @xmath80 and its resolutions .
this is the case for @xmath80 itself , since there are no symmetric or parallel extended probes and there are not enough edges for there to be any useful probes of type ( c ) : cf .
[ ss : gen ] below .
figure [ f:135 ] shows the situation for @xmath0 : there is an open set of points with nonzero qw invariants , as well as an open set of points with unknown behavior .
one can always displace more points in the resolutions .
some of these newly displaceable points can be displaced by the same parallel extended probes that are used in the resolved sectors @xmath88 .
however , there is another kind of extended probe , constructed using a symmetric probe @xmath40 , that always displaces at least a few more points , though typically there is still an open set of points with unknown behavior ; cf .
proposition [ p : openset ] .
the case @xmath0 is special since in this case one can reverse the direction of the deflected probes formed from @xmath40 . as illustrated in figure
[ f:135r03dand ] , when the singularity of @xmath0 at @xmath89 is resolved , one can displace an open dense set of fibers by probes or extended probes , leaving just a line segment of points together with one more point that can not be determined by our methods .
in particular , all points near @xmath89 can be displaced with the help of the symmetric probes @xmath40 , though they are not all displaceable in the corresponding resolved sector @xmath90 .
the picture does not change significantly when one fully resolves both singular vertices in @xmath0 : cf .
proposition [ p:135trdand ] and figure [ f:135trdand ] .
proposition [ p : noghost ] shows that qw invariants give no more information than standard floer homology in the case where the moment polytope @xmath32 is a smooth closed polytope of @xmath91 .
if in addition @xmath32 is compact and @xmath29-dimensional , it is easy to understand geometrically when the qw invariant @xmath92 does not vanish . for a typical point with @xmath93 ,
the set of facets that are closest to it has at least three elements .
the exception is when there are two closest facets that are parallel .
thus the set of points with nontrivial invariants is the union of a finite set together with at most one line segment . by remarks [ r : nohelp ] and [ r : extprob ]
, one can not lengthen a probe @xmath35 by deflecting it by a probe @xmath40 that starts from the facet at which @xmath35 exits @xmath32 .
thus extended probes do not displace extra points in the sectors @xmath68 , since these have only two edges .
it is also not hard to check that they also do not help in triangles such as @xmath80 , though they do help with the blow up @xmath70 of @xmath68 . after a brief introduction to affine geometry
, we describe probes and symmetric extended probes . theorem [ t : noflag ] explains which points can be displaced by these new probes , and the section ends by illustrating their use in the hirzebruch surfaces @xmath51 .
theorem [ t : pdeflected ] describes the points that can be displaced by parallel extended probes .
the rest of [ s:3 ] illustrates how to use this result to understand the displaceability of points in the open sectors @xmath71 .
next , in
[ s : dintor ] we describe the qw invariants , and use them to prove the results stated above about closed sectors @xmath68 , weighted projective planes @xmath80 , and their resolutions .
finally in [ s : epwfnp ] we describe extended probes with trapezoidal flags , and use them in an open polytope of finite area .
the last section [ s : proofs ] contains the proofs of all the main theorems about probes .
we thank andrew fanoe , yael karshon , egor shelukhin , and chris woodward for useful discussions .
we would also like to thank the referee for their comments and corrections .
the first named author was partially supported by fundao para a cincia e a tecnologia ( fct / portugal ) , the second named author by nsf - grant dms 1006610 , and the third named author by nsf - grant dms 0905191 .
let @xmath94 be a toric symplectic manifold with moment map @xmath95 , where @xmath96 is the dual of the lie algebra @xmath97 of the torus @xmath98 .
we will identify @xmath97 together with its integer lattice @xmath99 with @xmath100 , and , using the natural pairing @xmath101 , will also identify the pair @xmath102 with @xmath100 .
thus we write @xmath103 for clarity , we use greek letters for elements in @xmath97 and latin letters for elements of @xmath104 .
a polytope @xmath105 is _ rational _ if it is the finite intersection of half - spaces @xmath106 where @xmath107 are primitive vectors and are the interior conormals for the half - spaces .
a rational polytope @xmath108 is _ simple _ if each codimension @xmath13 face of @xmath32 meets exactly @xmath13 facets .
a rational simple polytope @xmath108 is _ smooth _
, if at each codimension @xmath13 face , the @xmath13 conormal vectors for the facets meeting at the face can be extended to an integral basis of @xmath109 .
moment polytopes for symplectic toric manifolds are smooth .
delzant @xcite proved that the moment map gives a bijective correspondence between closed symplectic toric manifolds of dimension @xmath7 , up to equivariant symplectomorphism , and smooth compact polytopes in @xmath91 , up to integral affine equivalences by elements of @xmath110 .
[ r : non - compact - orbifolds ] lerman and tolman @xcite generalized delzant s classification result to closed symplectic toric orbifolds , showing that these are uniquely determined by the image of their moment map , a rational simple compact polytope together with a positive integer _ label _ attached to each of its facets .
equivalently , we can allow the conormal vectors @xmath111 to the facets to be nonprimitive , interpreting the label as the g.c.d . of their entries . using the convexity and connectedness results in @xcite , karshon and lerman @xcite
have further extended this classification to the non - compact setting , provided one assumes that the moment map is _ proper _ as a map @xmath112 , where @xmath113 is a convex open set . in this case
@xmath114 is an open polytope as in ( * ? ? ?
* definition 3.1 ) ; cf .
the polytope @xmath115 in figure [ fig : x1 ] above .
an affine hyperplane @xmath116 is _ rational _ if it has a primitive conormal vector @xmath117 , in which case @xmath118 for some @xmath119 .
the _ affine distance _ between a rational affine hyperplane @xmath120 , as above , and a point @xmath121 is @xmath122 for a rational hyperplane @xmath120 , as above , a vector @xmath123 is _ parallel _ to @xmath120 if @xmath124 where @xmath125 is the defining conormal for @xmath120 .
an integral vector @xmath126 is _ integrally transverse _ to @xmath120 if there is an integral basis of @xmath127 consisting of @xmath126 and vectors @xmath128 parallel to @xmath120 , or equivalently if @xmath129 .
note also that if @xmath126 is integrally transverse to @xmath120 , then there is an integral affine equivalence of @xmath130 taking @xmath120 and @xmath126 to @xmath131 and @xmath132 . an affine line @xmath133 in @xmath91 is _ rational _ if the direction vector @xmath126 can be taken to be a primitive integral vector in @xmath134 .
given a rational line @xmath135 with primitive direction vector @xmath136 , the _
affine distance _
@xmath137 between two points @xmath138 is defined by @xmath139 for a primitive vector @xmath136 and a rational hyperplane @xmath120 , the _ affine distance along @xmath126 _ between a point @xmath121 and @xmath120 is defined as @xmath140 if @xmath141 does not meet @xmath120 , then @xmath142 . if @xmath143 is integrally transverse to @xmath120 and the ray @xmath144 meets @xmath120 , then @xmath145 , but otherwise , somewhat paradoxically , we have @xmath146 for example , if @xmath147 , @xmath148 and @xmath149 , then @xmath150 while @xmath151 .
we now recall the definition of a probe in a rational polytope from @xcite and state the method of probes .
a * probe * @xmath35 in a rational polytope @xmath152 , is a directed rational line segment contained in @xmath32 whose initial point @xmath153 lies in the interior of a facet @xmath154 of @xmath32 and whose * direction vector * @xmath155 is primitive and integrally transverse to the * base facet * @xmath154 . if @xmath156 is the endpoint of @xmath35 , then the * length * @xmath157 of @xmath35 is defined as the affine distance @xmath158 . note that if @xmath159 is the interior conormal for the facet @xmath154 , then the definition requires that @xmath160 since @xmath161 needs to be integrally transverse and be inward pointing into @xmath32 .
[ l : probe ] let @xmath35 be a probe in a moment polytope @xmath32 for a toric symplectic orbifold @xmath18 .
if a point @xmath39 on the probe @xmath35 is less than halfway along @xmath35 , meaning that @xmath162 , then the lagrangian fiber @xmath17 is displaceable . since the inverse image of @xmath35 by the moment map is symplectomorphic to the product of a disc @xmath163 of area @xmath157 with @xmath164 , it is easy to construct a proof of this lemma from the remarks after equation below . to generalize this method ,
let us first introduce the notion of a symmetric probe and the associated reflections .
a probe @xmath40 in a rational polytope @xmath152 is * symmetric * if the endpoint @xmath165 lies on the interior of a facet @xmath166 that is integrally transverse to @xmath167 .
associated to a symmetric probe @xmath40 is an affine reflection @xmath168 @xmath169 that swaps the two facets @xmath170 since @xmath171 and @xmath172 .
let @xmath173 be the associated linear reflection @xmath174 observe that @xmath175 preserves @xmath40 set - wise , @xmath176 , and @xmath177 .
note that @xmath175 need not preserve the polytope @xmath32 , but it does preserve a small neighborhood of @xmath40 .
our first generalization of the method of probes deflects a given probe by a symmetric probe .
let @xmath40 be a symmetric probe in a rational polytope @xmath152 , and let @xmath35 be another probe with direction @xmath161 such that @xmath35 ends at the point @xmath178 in the interior of @xmath40 .
a * symmetric extended probe * @xmath47 formed by deflecting @xmath35 with @xmath40 is a union @xmath179 where the extension @xmath180 is a rational line segment in @xmath32 with direction @xmath181 starting at @xmath182 , where @xmath183
we define the * length * of the extended probe @xmath47 to be @xmath184 .
thus the * endpoint * @xmath185 of @xmath180 is @xmath186 .
a visual description of @xmath182 is that it is the unique point on @xmath40 so that @xmath187 cf .
figure [ f : epnf ] . note that
if @xmath161 is parallel to @xmath188 , then @xmath189 is parallel to @xmath166 , being the projection of @xmath161 to the linear hyperplane @xmath190 along @xmath191 . .
left : @xmath39 is on @xmath35 .
right : @xmath39 is on @xmath180.,width=384 ] [ t : noflag ] let @xmath47 be a symmetric extended probe formed by deflecting the probe @xmath35 with the symmetric probe @xmath40 , in a moment polytope @xmath192 for a toric symplectic orbifold @xmath193 .
for a point @xmath39 in the moment polytope @xmath32 : * if @xmath39 is in the interior of @xmath35 and @xmath194 , or * if @xmath39 is in the interior of @xmath180 and @xmath195 , then the lagrangian torus fiber @xmath17 is displaceable in @xmath1 .
see section [ s : proofs ] for the proof .
the idea is to join @xmath35 to @xmath180 using a symplectomorphism of @xmath196 that equals the identity on one boundary component and the lift of the reflection @xmath44 on the other . .
if @xmath35 were extended past @xmath178 it would exit the polytope at the point @xmath197 , so its length is at most @xmath28 .
this is not enough to displace @xmath39 since @xmath198 . using @xmath40 as a deflecting probe does not help ;
the length of the resulting extended probe is equal to @xmath28 .
of course the fiber @xmath199 is nondisplaceable since it is the clifford torus in @xmath200.,width=240 ] [ r : nohelp ] let @xmath201 be a symmetric extended probe . because @xmath185 must stay in @xmath32 , which is convex , it follows that @xmath202 .
also because @xmath175 is an integral affine equivalence in @xmath203 it follows that @xmath204 combining these two we see that @xmath205 the length of the extended probe @xmath206 is less than @xmath207 , which is the maximum length the probe @xmath35 can have before it hits the affine hyperplane @xmath188 .
in particular , this means that one can not make a probe @xmath35 displace more points by deflecting it with a symmetric probe @xmath40 that is based at the facet on which @xmath35 would exit @xmath32 .
see figure [ f : nohelp ] . in the next section
we will generalize the notion of extended probes to the case where @xmath40 is not a symmetric probe .
but before that let us first explain how theorem [ t : noflag ] suffices to settle the question of which lagrangian fibers in hirzebruch surfaces are displaceable .
let @xmath208 be an integer and @xmath209 a real number so that @xmath210 . the moment polytope for the @xmath77th hirzebruch surface is @xmath211 for @xmath212 and @xmath213 .
points not on the line @xmath214 can be displaced by horizontal probes .
when @xmath77 is even , probes with directions @xmath215 based on the two horizontal facets displace everything on @xmath214 except @xmath216 . , width=432 ] when @xmath217 , the only known nondisplaceable fiber is @xmath218 @xcite and when @xmath77 is even , standard probes displace every other fiber .
however for odd @xmath77 and @xmath219 , ( * ? ? ? * lemma 4.1 ) proves that the point @xmath220 which is different from @xmath216 , can not be displaced by probes , so in these cases probes do not perfectly complement the known nondisplaceablility results .
in fact , in these cases one can not use probes to displace any of the points on a line segment that starts at @xmath221 and runs through and a bit past @xmath216 .
we will now show that these unknown fibers can all be displaced using theorem [ t : noflag ] .
[ p : h ] when @xmath217 , then every fiber in the polytope @xmath222 of the @xmath77-th hirzebruch surface is displaceable except for @xmath223 hence @xmath216 is a stem and therefore is nondisplaceable .
recall that a fiber of a moment polytope is called a _ stem _ if every other fiber is displaceable .
it is proven in ( * ? ? ?
* theorem 2.1 ) using the theory of quasi - states that every stem is nondisplaceable .
proposition [ p : h ] stands in contrast to the well - studied case of @xmath224 , which corresponds to a toric blow - up of @xmath200 , where standard probes do complement the known nondisplaceability results .
methods in @xcite prove that when @xmath225 , the fibers @xmath226 in @xmath227 are nondisplaceable , and when @xmath228 only the fiber @xmath216 in @xmath227 is nondisplaceable .
the vector @xmath229 is integrally transverse to @xmath230 and @xmath231 is integrally transverse to @xmath232 .
it is easy to check that probes in these directions displace every point @xmath233 not on the median @xmath234 when @xmath77 is even , one can use a probe with direction @xmath235 based on the facet @xmath236 and a probe with direction @xmath237 based on the facet @xmath238 to displace all the points on @xmath214 except @xmath216 .
when @xmath239 is odd , we will use symmetric extended probes to show that every point on @xmath214 , except @xmath216 , is displaceable .
such points @xmath240 can be written as @xmath241 and we will divide this into two cases : @xmath242 and @xmath243 . in both cases
we will use the symmetric deflecting probe @xmath40 where @xmath244 we will use probes with direction @xmath245 , which are parallel to both @xmath188 and @xmath166 , so it follows that @xmath246 . for @xmath247 , cf .
figure [ f : h3ep ] , let @xmath35 be the probe with @xmath248 and let @xmath201 be the associated extended probe , where @xmath249 that displaces @xmath240 when @xmath247.,width=432 ] the point @xmath240 from lies on @xmath35 and therefore using @xmath250 we have @xmath251 hence @xmath252 is displaceable by theorem [ t : noflag ] . for @xmath243 , cf .
figure [ f : h3epnf ] , let @xmath35 be the probe with @xmath253 and let @xmath201 be the associated extended probe , where @xmath254 and the point @xmath240 is on @xmath180 .
that displaces @xmath240 when @xmath243.,width=432 ] since @xmath255 we have @xmath256 and therefore theorem [ t : noflag ] implies that @xmath257 is displaceable .
the use of symmetric extended probes is fairly restrictive since a symmetric probe @xmath40 represents a torus bundle over @xmath258 . in cases where @xmath40 does not exit the polytope ( or does so non - transversally )
then the following construction can be used with @xmath40 to deflect probes .
[ d : pepwf ] let @xmath35 and @xmath40 be probes in a rational polytope @xmath108 where the probe @xmath35 ends at the point @xmath178 on @xmath40 , and suppose that @xmath161 is parallel to the base facet @xmath188 of @xmath40 . the * parallel extended probe with flag * @xmath48 formed by deflecting @xmath35 with @xmath40 is the subset @xmath259 here the flag @xmath260 is the convex hull of the points @xmath261 , where @xmath262 and @xmath263 lie on @xmath40 , and the vector @xmath264 is parallel to @xmath35 ( see figure [ f : extendedprobe ] ) .
the * length * of the flag @xmath265 is @xmath266 so that @xmath267 the * length * of the extended probe is @xmath268 .
.,width=240 ] the following theorem explains how one can use parallel extended probes to displace lagrangian torus fibers .
[ t : pdeflected ] let @xmath269 be a parallel extended probe in the moment polytope @xmath192 of the toric symplectic orbifold @xmath193 , and let @xmath270
. then the lagrangian fiber @xmath17 is displaceable if the following conditions both hold : * the affine distance from @xmath39 to the facet @xmath154 satisfies @xmath271 * and the flag @xmath260 satisfies the inequality @xmath272 in section [ s : epwfnp ] we will further generalize this to the case where @xmath35 is not parallel to the facet @xmath188 . see section [ s : proofs ] for the proofs . consider the standard toric structure @xmath273 .
the symplectic form is @xmath274 , the torus action is @xmath275 where @xmath276 , the moment map is @xmath277 and the moment polytope is @xmath278 . for ease of notation let us now specialize to the case @xmath279 .
let @xmath57 be relatively prime integers and consider the open subset of @xmath279 @xmath280 where @xmath58 is any positive constant .
its image under the moment map is @xmath281 we will now turn to the investigation of the displaceability of lagrangian toric fibers in @xmath282 . for @xmath283 and @xmath284.,width=211 ] let us first explain what is displaceable in @xmath285 by standard probes .
[ l : dosecp ] the following points @xmath286 can be displaced by probes based on the facet @xmath287 : * points in @xmath288 by probes with direction @xmath289 .
* points in @xmath290 by probes with direction @xmath291 . the following points @xmath286 can be displaced by probes based on @xmath292 : @xmath293 an elementary calculation . and lemma [ l : openep ] .
the points in the gray regions are displaceable , but extended probes are needed for points in the dark gray regions .
see remark [ r : epvsp ] . in this figure and others
the numbers @xmath294 refer to the slope @xmath295 for the line @xmath296 defining the boundary between differently shaded regions . , width=432 ] here is what we can do with extended probes .
[ l : openep ] let @xmath297 .
if @xmath298 is to the left of the line passing through @xmath299 with slope @xmath300 , that is @xmath301 then @xmath298 can be displaced by a parallel extended probe with flag in @xmath302 .
let @xmath303 and let @xmath304 be a point satisfying .
consider the probes @xmath35 and @xmath40 where , for some small @xmath305 , @xmath306 and @xmath35 ends at the point @xmath178 on @xmath40 .
observe that @xmath40 is parallel to the line defined by an equality sign in , so since @xmath240 satisfies it follows that @xmath240 lies in the interior of @xmath35 for sufficiently small @xmath307 .
note that @xmath161 is parallel to the base facet @xmath308 of @xmath40 .
.,width=432 ] for the three parameters @xmath309 , consider the parallel extended probe with flag @xmath310 where @xmath311 see figure [ f : epinc2 ] for an example . first take @xmath312 sufficiently large so that is satisfied as @xmath313 then take @xmath314 to ensure that the endpoints of the flag @xmath315 stay in @xmath316 . finally taking @xmath317
ensures that condition in theorem [ t : pdeflected ] is satisfied .
therefore for these parameter values the extended probe @xmath48 displaces the lagrangian fiber @xmath318 . [
r : epvsp ] if @xmath319 , then there are @xmath320 regions of infinite measure in @xmath316 consisting of points that can be displaced by extended probes but not by standard probes .
if @xmath321 , then the only points in @xmath316 where extended probes are needed are the points on a ray with direction @xmath289 .
see figure [ f : dxnm ] for examples .
wilson and woodward @xcite recently observed that the quasi - map floer homology developed in @xcite , can be used in orbifold and noncompact settings to obtain large families of nondisplaceable lagrangian fibers .
as we will see , after ( partially ) resolving an orbifold singularity , many of these nondisplaceable fibers become displaceable by extended probes . in this section
we investigate this phenomenon . given a presentation of a rational simple polytope @xmath322 where @xmath323 and @xmath107 are integer vectors , one can build a symplectic toric orbifold @xmath324 such that @xmath325 is proper and @xmath326 , which is unique up to equivariant symplectomorphism by @xcite . by (
* proposition 6.8 ) , @xmath32 has at least one vertex exactly if @xmath327 can be represented as a symplectic reduction @xmath328 of the standard @xmath329 , where @xmath330 is a suitable @xmath331 dimensional subtorus and @xmath332 is the moment map for the action of @xmath333 on @xmath327 . if @xmath32 does not have a vertex , then , as was noted in ( * ? ? ?
* corollary 6.9 ) , @xmath334 where @xmath335 is a rational simple polytope with a vertex , @xmath336 , and one can take @xmath337 where @xmath338 . by (
* proposition 6.10 ) , the results of @xcite hold in both cases .
consider the field of generalized laurent series in the variable @xmath339 @xmath340 the field @xmath341 is complete with respect to the norm @xmath342 induced by the non - archimedian valuation @xmath343 with the convention that @xmath344 , and @xmath345 satisfies @xmath346 where the inequality is an equality if @xmath347 .
the subring of elements with only non - negative powers of @xmath339 , @xmath348 is a local ring with maximum ideal @xmath349 .
completeness gives that the exponential function @xmath350 , defined via the standard power series , is surjective onto the units @xmath351 .
associated to a rational simple polytope @xmath32 there is the _ bulk deformed potential _ that for each @xmath352 and @xmath353 is a function @xmath354 where @xmath355 and @xmath356 are from .
the following theorem allows one to prove nondisplaceability results merely by finding critical points of the potential function @xmath357 .
it was proved by fukaya
ono ( * ? ? ?
* theorem 9.6 ) for smooth closed toric manifolds ( and for geometric @xmath358 ) , by woodward ( * ? ? ?
* proposition 6.10 and theorem 7.1 ) for rational simple polytopes , and by wilson woodward ( * ? ? ? * theorem 4.7 ) for rational simple polytopes for open symplectic toric orbifolds .
[ t : potential ] for a toric orbifold @xmath359 as above and @xmath360 , if there exists @xmath361 such that @xmath362 has a critical point , then the lagrangian torus fiber @xmath363 is nondisplaceable .
the basic idea is due to cho
oh @xcite , where the holomorphic disks used to define the @xmath364-structure associated to the lagrangian floer homology and quasi - map floer homology for @xmath365 are explicitly classified .
it turns out that if @xmath366 is a critical point for @xmath357 then , in the smooth case , lagrangian floer homology with differential @xmath367 depending on @xmath366 is defined and nonzero for @xmath368 , so that @xmath365 is nondisplaceable . for general @xmath357
a similar statement holds for the quasi - map floer homology of @xmath365 . note that the parameters @xmath369 and @xmath366 correspond to bulk deformations and weak bounding cochains , in the language of fukaya oh ohta ono .
if @xmath360 is a critical point of @xmath357 we will say that it has * nontrivial ( or nonzero ) qw invariants*. [ r : ww ] ( i ) observe that the potential function @xmath357 depends on the presentation of @xmath32 in as a polytope and not just on @xmath32 as a subset of @xmath91 .
equivalently @xmath357 depends on the presentation of @xmath327 as a reduction of @xmath370 and not just on @xmath327 as a symplectic toric orbifold . in papers such as @xcite that work in the context of lagrangian floer homology on smooth manifolds
it is assumed that the polytope @xmath32 has precisely @xmath371 facets , and one builds the invariant by counting holomorphic discs in @xmath327 that intersect these facets . in this case , we call @xmath357 the geometric `` potential function . however , in the quasi - map approach of woodward @xcite , the invariant is built from holomorphic discs in @xmath372 , that intersect the @xmath371 facets of @xmath373 . since the geometry takes place in @xmath372
there is no need for each of these @xmath371 facets to descend to a geometric facet of @xmath32 ; some of them may be ghosts '' with constants @xmath374 chosen so large that @xmath375 for all @xmath376 where @xmath377 is a ( possibly empty ) face of dimension less than @xmath378 .
\(ii ) it is clear from equation above , that if a ghost facet is parallel to a geometric facet of @xmath32 then we can amalgamate the two corresponding terms in @xmath357 : if the geometric facet has @xmath379 then the ghost facet is @xmath380 where @xmath381 , so that if @xmath358 is the original potential and @xmath382 is the potential with the ghost facet , we have that @xmath383 where @xmath384 thus , the parallel ghost facet affects the terms in @xmath369 with positive @xmath339 weight ; in the language of @xcite it is a bulk deformation .
if @xmath32 is a smooth compact moment polytope with rational support constants , it follows from ( * ? ? ?
* proposition 4.7 ) and ( * ? ? ?
* theorem 4.5 ) that there is always a @xmath385 such that the geometric potential @xmath386 has a critical point for some @xmath387 . in @xcite ,
woodward observed that ghost facets can give new information if @xmath32 has singularities or corresponds to an open symplectic toric orbifold .
lemma [ l : dndm ] and remark [ r : dmdm ] below show precisely how ghost facets may create lines with nontrivial invariants and theorem [ t : dnd](i ) is an example where ghost facets create open sets with nontrivial invariants .
in contrast to this we will now prove that ghost facets give no new information if the polytope is smooth and closed , which explains ( * ? ? ? * remark 6.11 ) .
note that part ( i ) of the next proposition has analogs in all dimensions , but we restrict to dimension @xmath29 for simplicity .
[ p : noghost ] let @xmath32 be a smooth closed polytope in @xmath91 . *
if @xmath32 is compact and @xmath29-dimensional , the set of points in @xmath32 with nontrivial qw invariants is the union of a finite number of points with at most one line segment . * in any dimension , adding ghost facets to the potential does not change the set of points @xmath388 such that @xmath386 has a critical point for some @xmath387 .
we use the notation of .
we first prove ( i ) in the case of the geometric qw potential to explain the idea in a simple case .
we then prove ( ii ) , which implies ( i ) in the general case .
for each point @xmath389 define @xmath390 and denote the set of edges that are closest to @xmath39 by @xmath391 .
if @xmath392 , and the edges in @xmath393 are not parallel then we can choose coordinates so that one edge in @xmath393 has equation @xmath394 , while the other has the form @xmath395 where @xmath396 . then , because for all @xmath397 we have @xmath398 , where @xmath399 , we find that @xmath400 which means that @xmath39 is not a critical point of @xmath386 .
a similar argument shows that @xmath39 is not critical when @xmath401 . on the other hand
if the two edges in @xmath393 are parallel and we choose coordinates so that these have equations @xmath402 , then @xmath403 can be solved to lowest order in @xmath339 .
further the equation @xmath404 starts with terms involving @xmath405 where @xmath406 , and its lowest order terms also have a solution if at least two of these involve the same power of @xmath339 .
equivalently , we need @xmath407 , where @xmath408 consists of those facets not in @xmath393 that are closest to @xmath39 . we may now appeal to ( * ? ? ?
* theorem 4.5 ) which says that if the system of lowest order equations has a solution , then one can choose the higher order terms in @xmath409 to obtain a solution of the full system of equations @xmath410 .
all the other critical points have @xmath411 .
since there are only finitely many such points , it remains to check that there is at most one line segment consisting of points with @xmath392 . to see this ,
note first that for any two parallel edges , the set of points equidistant from them is convex . hence if there are two such line segments , @xmath32 must have two sets of parallel sides , and hence be the blow up of a rectangle .
but in a rectangle only one set of parallel lines can appear as @xmath393 , and if it is a square there are no points with @xmath412 .
this proves ( i ) .
now consider ( ii ) .
remark [ r : ww ] deals with the case when the ghost facet is parallel to some facet of @xmath32 . therefore suppose it is not . without loss of generality
, we may suppose that @xmath413 defines a ghost facet for @xmath414 that intersects @xmath32 when @xmath415 in a codimension @xmath416 face @xmath417 .
then we may choose coordinates so that @xmath418 and @xmath419 where @xmath420 .
so in particular we see that for any interior point @xmath298 and any @xmath414 : @xmath421 the potential with the ghost facet added is given by @xmath422 .
observe for @xmath423 that @xmath424 is unaffected by the ghost term .
for @xmath425 , it follows from that the leading order terms in @xmath424 are unaffected by the ghost term .
therefore the leading order term critical point equation for @xmath426 and @xmath386 are the same , so again by ( * ? ? ?
* theorem 4.5 ) both potentials have the same set of points @xmath385 that give rise to critical points .
this proves ( ii ) . for an integer @xmath428 ,
consider the orbifold @xmath429 whose moment polytope is the sector @xmath430 here @xmath431 where the generator in the cyclic group acts by diagonal multiplication by @xmath432 .
if the torus @xmath433 acts via @xmath434 the moment map is given by @xmath435 the orbifold singularity at the origin can be resolved with the facet @xmath436 in fact , if @xmath437 , then the polytope @xmath438 is smooth and is the moment polytope for the standard toric structure on the line bundle @xmath439 . on the other hand if @xmath440 then @xmath441 as subsets of @xmath442 , so @xmath443 defines a ghost facet in the presentation .
the effect of resolving the orbifold singularity in this case is easy to explain , while the answers become more complicated for the later examples . when @xmath212 .
the displaceable points are in light gray and the nondisplaceable points are on the line @xmath444 .
the nondisplaceability proof uses ghost facets @xmath445 for varying @xmath446.,width=288 ] [ l : dndm ] if @xmath447 is on the line @xmath448 , then the lagrangian fiber @xmath365 is nondisplaceable
. if @xmath298 is not on the line @xmath448 , then @xmath365 is displaceable by a horizontal probe .
the displaceability statement is straightforward . for each @xmath447 on the line @xmath448
, there is some @xmath446 such that @xmath449 for a given @xmath298 and the associated @xmath446 , consider the presentation @xmath450 from , which has the ghost facet @xmath451 .
its potential function is @xmath452 setting @xmath453 and using to cancel out the powers of @xmath339 gives @xmath454 these equations are solved by @xmath455 over @xmath456 .
therefore by theorem [ t : potential ] , the lagrangian fiber @xmath365 is nondisplaceable .
[ r : dmdm ] ( i ) the ghost facet was needed , for without it the potential function is @xmath457 which has no critical points since @xmath458 has just one non - zero term for all @xmath298 and @xmath369 .
\(ii ) the bulk deformation @xmath459 is also needed when @xmath460 . to see this
, note that under the substitution @xmath461 and @xmath462 becomes @xmath463 for @xmath460 this says @xmath464 which has no solution except @xmath465 . if @xmath437 , then @xmath466 is a resolution of @xmath76 .
the next result shows that all the previously nondisplaceable fibers can now be displaced , either by standard probes based on the new facet @xmath467 or by parallel extended probes deflected by a probe @xmath40 that is based on the new facet .
[ l : o - m ] if @xmath437 , then the lagrangian fiber @xmath468 is displaceable by extended probes for all @xmath469 .
just as in the hirzebruch surface case , for even @xmath428 standard probes displace everything .
when @xmath470 is odd , horizontal probes displace everything except the points @xmath471 on the line @xmath448 , which up to translation is identified with @xmath472 so it suffices to prove that points on the line in are displaceable and this follows from lemma [ l : openep ] since @xmath473 . for relatively prime positive integers
@xmath474 consider the complex orbifold @xmath475 where @xmath476 is the cyclic subgroup of @xmath477 generated by the matrix @xmath478 the standard symplectic toric structure on @xmath479 induces a symplectic toric orbifold structure on @xmath480 with an orbifold singularity at the origin , where @xmath481 ) = ( { \left| z_{1 } \right|}^{2},\ , \tfrac{n}{m}{\left| z_{1 } \right|}^{2 } + \tfrac{1}{m}{\left| z_{2 } \right|}^{2})\ ] ] and the moment polytope is @xmath482 @xmath483 note that interior conormals for @xmath68 are @xmath484 and we will call @xmath485 the _ vertical edge _ and @xmath486 the _
slant edge_. note @xmath487 defines the line @xmath488 , which we will call the _ midline_. for @xmath489.,width=288 ] in terms of the polytope , the assumption that @xmath490 is harmless since if @xmath491 , then by applying a shear @xmath492 , which is in @xmath493 , we see that @xmath494 is equivalent to @xmath495 .
associated to @xmath494 is a minimal resolution of the symplectic toric orbifold singularity at the origin which is known in algebraic geometry as a hirzebruch jung resolution .
the version in the symplectic toric setting is due to orlik
raymond @xcite , see also @xcite . to find the resolution ,
one writes @xmath496 as a continued fraction using positive integers @xmath497 : @xmath498 the positive integers @xmath499 are given by the euclidean algorithm where @xmath500 , @xmath501 and @xmath13 is the smallest number such that @xmath502 .
the sequence of integers @xmath503 determine a sequence of @xmath504 interior conormals in @xmath505 , starting with the conormal for the vertical edge @xmath506 and then defined recursively for @xmath507 @xmath508 where the last one is the conormal for the slant edge @xmath509 these conormals are such that if @xmath510 , then @xmath511 for appropriate support constants
@xmath512 the polytope @xmath513 where @xmath514 has @xmath504 edges , is smooth , and corresponds to a symplectic toric manifold @xmath515 that is called a * minimal resolution * of @xmath516 . the class of examples @xmath494 where @xmath474 has the following symmetry , which we will exploit to shorten the proofs below .
let @xmath517 be the integers that solve @xmath518 so that @xmath519 and @xmath520
. then the matrix @xmath521 has @xmath522 and @xmath523 .
furthermore , @xmath524 interchanges the roles of the vertical edge and the slant edge , while mapping the midline to the midline . .
right : @xmath525 .
the matrix @xmath524 in maps @xmath525 to @xmath68 , mapping the light gray region to the light grey region , and likewise for the dark grey regions . here @xmath526 and @xmath527.,width=480 ] therefore we will often only need to prove a result for points to the left of the midline : the properties of the points to the right of the midline will be deduced by applying the matrix @xmath524 .
this symmetry provided by @xmath528 is compatible with the resolution given by the continued fraction expansion .
namely if @xmath503 are associated with the pair @xmath529 , then @xmath530 so that @xmath531 is given by reversing the order of the @xmath499 s .
if @xmath532 are the conormals associated with @xmath88 , and @xmath533 are the conormals associated with @xmath534 , then one can check that @xmath535 where @xmath536 is the transpose .
therefore @xmath524 maps one minimal resolution from to the other @xmath537 in particular we have that the first three and the last three conormals for @xmath88 are @xmath538 for relatively prime positive integers
@xmath474 , consider the symplectic toric orbifold @xmath539 from .
let @xmath540 be given by and let @xmath503 be the sequence of integers from associated to @xmath541 . in this section we will use the notation @xmath542 [ t : dnd ] let @xmath543 . 1 .
the lagrangian fiber @xmath365 is nondisplaceable if @xmath544 2 . if @xmath545 , then all other fibers @xmath365 are displaceable .
3 . if @xmath546 , then all fibers with @xmath547 are displaceable except possibly for those with @xmath548 .
4 . if @xmath549 , then all fibers with @xmath550 are displaceable except possibly for those with @xmath551 . .
the dark grey regions are closed and nondisplaceable , the light grey regions are open and displaceable , and the white regions are unknown.,width=576 ] recall that @xmath552 where @xmath553 for ghost facets , we will use the first two conormals @xmath554 and @xmath555 associated with the hirzebruch
jung resolution .
thus we take @xmath556 which define ghost facets for @xmath68 when @xmath557 are non - negative . observe that if the point @xmath558 satisfies @xmath559
then @xmath560 for suitable @xmath561 .
the potential function of @xmath68 with the added ghost facets @xmath562 and @xmath563 is @xmath564 changing variable so that @xmath461 and @xmath462 and using , the critical point equation becomes @xmath565 then @xmath566 is a critical point of @xmath357 when @xmath567 these equations are solvable by @xmath568 since @xmath569 and neither term in is zero . hence @xmath365 is nondisplaceable by theorem [ t : potential ] .
so far we have proved that for @xmath558 such that @xmath570 the fiber @xmath365 is nondisplaceable .
likewise , for @xmath571 , @xmath572 is nondisplaceable if @xmath573 the image of this region under the symmetry @xmath524 from is the subset of @xmath574 where @xmath575 piecing and together , we have proved that if @xmath558 satisfies , then @xmath365 is nondisplaceable .
the proof of theorem [ t : dnd ] is completed by the following lemma .
[ l : dsec ] probes based on the vertical edge in @xmath68 displace the following points : * points in @xmath288 by probes with direction @xmath289 .
* points in @xmath576 by probes with direction @xmath291 .
probes based on the slant edge in @xmath68 displace the following points : * points in @xmath577 by probes with direction @xmath578 .
* points in @xmath579 by probes with direction @xmath580 .
the first two claims are similar to lemma [ l : dosecp ] , and are straightforward to check .
the last two claims are the transform under the symmetry @xmath524 of the first two claims for the sector @xmath525 .
we next show that the lower bound @xmath581 in theorem [ t : dnd ] ( i ) is optimal with our current methods .
[ l : dnd2 ] the qw invariants vanish for points in @xmath68 with @xmath582 .
if the potential function @xmath583 has a critical point at a point @xmath558 left of the midline , there must be at least one ghost facet @xmath584 such that holds for support constants @xmath446 . if @xmath585 for some non - negative integers , then @xmath586 must hold in order for @xmath587 to define a ghost facet .
since @xmath588 implies @xmath589 , this gives a potentially new lower bound .
the choice @xmath554 is optimal since it gives the bound @xmath590 .
however , adding just this ghost facet by itself is not enough since the second equation in would then have just one term and so have no solution .
any other of @xmath125 must have @xmath591 positive , and we claim that in this case @xmath592 , so that the lower bound is no better than before . to see this , recall that @xmath593 . if @xmath594 , then @xmath595 , so @xmath596 and hence @xmath597 .
suppose @xmath598 , then since @xmath599 , we have that @xmath600 and hence it suffices to prove @xmath601 .
this is equivalent to @xmath602 , which holds since @xmath598 .
observe that in the proof of theorem [ t : dnd](i ) we used the ghost facets with conormal @xmath603 to prove the nondisplaceability of the points in , which are to the left of the midline , and implicitly we used their transforms under the symmetry in with conormal @xmath604 to deal with the points to the right of the midline .
hence , if we partially resolve the orbifold singularity with these two edges , many fibers with previously nonzero qw invariants now have vanishing invariants . at the same time
, since probes with direction @xmath291 based on the vertical edge , are parallel to the new edge with conormal @xmath603 ( and likewise on the right ) , this partial resolution also causes many fibers to become displaceable using parallel extended probes with flags .
[ p : asympresolve ] for the polytope @xmath68 as in , consider a minimal resolution @xmath605 ) given by where @xmath606 are the support constants .
if @xmath607 is not in the region @xmath608 then the lagrangian fiber @xmath609 can be displaced by extended probes in @xmath610 , provided that the terms in @xmath209 are sufficiently close to zero . for @xmath611 , it follows from lemma [ l : openep ] that @xmath365 can be displaced by a parallel extended probe with flag @xmath310 in @xmath610 . here
@xmath35 is based on the vertical edge with direction @xmath612 and @xmath40 is based on the new edge @xmath613 with direction @xmath614 . observe that @xmath615 is the transform of @xmath616 for @xmath525 under the transformation @xmath524 from . in @xmath610 one
builds an extended probe where @xmath35 is based on the slant edge , with direction @xmath617 , and the deflecting probe @xmath40 has direction @xmath618 and is based on the new edge @xmath619 .
[ r : resolvep ] ( i ) unwrapping how proposition [ p : asympresolve ] uses lemma [ l : openep ] gives the following more precise version .
suppose that a ( partial ) resolution @xmath620 of @xmath68 contains an edge @xmath613 with one endpoint on the vertical edge @xmath485 and the other at @xmath621 , where @xmath622 .
then the points @xmath623 such that @xmath624 are displaceable by extended probes using lemma [ l : openep ] .
the analogous statement holds when @xmath619 appears next to the slant edge @xmath486 .
\(ii ) comparing the results of proposition [ p : asympresolve ] to theorem [ t : dnd ] , if @xmath625 then the points @xmath558 in the region @xmath626 can be displaced after we partially resolve @xmath68 with @xmath627 and are nondisplaceable before partially resolving .
if @xmath628 , then this region is empty . in this section
we discuss a few of the minimal resolutions @xmath610 in .
we will use standard probes as well as the extended probes described in proposition [ p : asympresolve ] and remark [ r : resolvep ] . displace no new points . ] in general for minimal resolutions @xmath629 there is no known nondisplaceable fiber . in section [ s:1 m ]
where @xmath630 and hence @xmath631 , we showed in lemma [ l : o - m ] that every fiber is displaceable in a minimal resolution @xmath632 . in every other case
there will be fibers that we can not displace with extended probes .
suppose the continued fraction expansion for @xmath541 has length @xmath634 given by @xmath635 and @xmath636 .
then @xmath637 and the facets for a minimal resolution @xmath638 have interior conormals @xmath639 note that in this case the upper and lower bounds of proposition [ p : asympresolve ] coincide , so we have the following corollary taking into account remark [ r : resolvep ] . .
above : @xmath640 before resolving .
below : the resolution @xmath641 drawn in the case @xmath642 .
the light gray regions are displaceable and the dark grey regions are nondisplaceable .
it is unknown if the points on the ray with slope @xmath643 are displaceable.,width=384 ] [ c : k1 ] let @xmath644 be such that @xmath645 is a minimal resolution from .
a lagrangian fiber @xmath365 in @xmath645 is displaceable by probes provided it is not on the ray given by @xmath646 that is @xmath647 we only need to prove that if @xmath298 is on the line and @xmath648 , then @xmath365 is displaceable .
these points can be displaced by probes with direction @xmath231 based on the edge @xmath649 .
see figure [ f : k1 ] .
if @xmath298 lies on the line in , then @xmath650 and hence the potential function @xmath651 from will not have a critical point .
in particular one can check that @xmath652 for all @xmath653 .
so theorem [ t : potential ] can not be used to prove @xmath365 is nondisplaceable .
the @xmath654-singularity is @xmath655 where @xmath656 is the subgroup of @xmath657 generated by @xmath658 where @xmath659 . comparing with section [ s : mnm ]
, we see that the @xmath654-singularity is given by @xmath660 and note the associated @xmath661 .
the continued fraction expansion for @xmath662 has length @xmath663 and is given by @xmath664 and the facets for a minimal resolution @xmath665 have interior conormals @xmath666 note that in this case the upper and lower bounds of proposition [ p : asympresolve ] coincide , so we have the following corollary taking into account remark [ r : resolvep ] .
the light grey regions are displaceable by standard probes with direction ( 1,1 ) or parallel to one of the added facets .
the medium grey regions are displaceable by extended probes from proposition [ p : asympresolve ] .
displaceability for the white rays is unknown.,width=288 ] [ c : an ] let @xmath667 be such that @xmath668 is a minimal resolution from .
a point @xmath669 is displaceable by probes if it does not belong to one of the @xmath670 rays given by @xmath671 for @xmath672 .
if @xmath503 is the continued fraction decomposition of @xmath541 , suppose that its length is at least @xmath673 and some @xmath674 for @xmath675 .
then in any minimal resolution @xmath610 there will be open regions of points that extended probes do not displace . .
the light grey regions are displaceable by standard probes with direction @xmath289 , @xmath676 , or parallel to one of the added facets .
the medium gray regions are displaceable by extended probes from proposition [ p : asympresolve ] .
displaceability is unknown for points in the white rays and regions.,width=384 ] [ ex : fr5 - 8 ] consider a minimal resolution of @xmath677 .
since @xmath678 we have @xmath679 and the upper and lower bounds in proposition [ p : asympresolve ] do not coincide .
since the continued fraction expansion of @xmath680 is given by @xmath628 , @xmath681 , and @xmath682 , the conormals for @xmath677 are @xmath683 the displaceable fibers in @xmath677 are displayed in figure [ f:5 - 8tr ] ; as we can see there is an open region of unknown fibers .
consider the weighted projective plane @xmath0 , with moment polytope @xmath684 mcduff showed in ( * ? ? ?
* lemma 4.4 ) that @xmath32 has an open subset of points that can not be displaceable by probes ; moreover , this open subset persists even after resolving the orbifold singularities .
woodward on the other hand showed in ( * ? ? ?
* example 4.11 ) that many , but not all , of the fibers that can not be displaced by probes are actually nondisplaceable in @xmath32 .
figure [ f:135 ] summarizes their results . using remarks [ r : nohelp ] and [ r : extprob](ii ) , one can see that one can not do better by using extended probes . in this section we work out
which points can be displaced by extended probes when we resolve the singularities . .
the probe displaceable points in light gray , the nondisplaceable points in dark grey , and the unknown points in white.,width=192 ] observe that @xmath32 near the vertex @xmath89 is locally equivariantly symplectomorphic to a neighborhood of the origin in @xmath685 , and hence locally the results in figure [ f:135 ] match those in figure [ f : dandsectors ] .
the region near @xmath686 in @xmath32 is literally of the form @xmath687 , which by shearing is equivalent to @xmath688 and theorem [ t : dnd ] says that in @xmath688 there is one line of nondisplaceable fibers and everything else is displaceable .
this is what we see in the region near @xmath686 in figure [ f:135 ] .
the normals for @xmath0 are given by @xmath689 and the interior conormals for a minimal resolution of @xmath0 are given by @xmath690 in what follows we include facets @xmath691 for @xmath692 into the presentation of @xmath32 from .
for reference let us note that these half spaces define ghost facets in @xmath32 when @xmath693 consider the resolution of the singularity at @xmath89 given by @xmath694 , which we assume has vertices @xmath695 this resolution corresponds to a minimal resolution of the sector @xmath685 .
since the continued fraction expansion of @xmath696 is given by @xmath697 , it follows from corollary [ c : k1 ] that there is a line of nondisplaceable points near the resolved vertex lying on the bisector of the edges @xmath698 and hence in direction @xmath699 .
these points can not be displaced in the minimal resolution @xmath90 because , although we can deflect a vertical probe @xmath35 starting on the horizontal base facet @xmath700 by a @xmath231 probe @xmath40 starting on @xmath701 , the resulting deflected probe @xmath260 is not parallel .
rather it has a trapezoidal flag and tapers to a point as it reaches this line ; cf .
remark [ r : extprob ] ( ii ) .
however because of the vertical edge @xmath702 , the probe @xmath40 is symmetric in the partial resolution of @xmath0 so that the deflected probe @xmath260 has no flag .
moreover , in the case @xmath0 ( but not in other @xmath703 ) , the deflected probe @xmath260 is symmetric , i.e. it exits the polygon transversally so that its direction can be reversed .
hence this type of probe displaces all but a codimension @xmath704 subset : see figure [ f:135r03dand ] .
proposition [ p:135rdand ] gives the details .
in @xmath0 is resolved as in . *
left * : in grey , points displaceable by symmetric extended probes deflected by a probe @xmath40 , that is based on the edge @xmath701 with conormal @xmath705 . *
middle * : displacing some more points with standard probes that have direction @xmath706 . *
right * : in light grey are the fibers that were displaced in the previous two pictures .
the point @xmath707 and the points on the black line segment connecting @xmath686 and @xmath708 are nondisplaceable .
displaceability is unknown for points on the white line segment , and the points @xmath709 , and @xmath710.,width=576 ] [ p:135rdand ] in the resolution of the singularity at @xmath89 in @xmath32 from where @xmath711 let @xmath712 be such that the vertices of are given by . 1 .
there is an open dense set of points in @xmath694 , whose lagrangian fibers are displaceable by probes or extended probes .
if @xmath298 is on the line segment connecting @xmath713 or the point @xmath714 then the lagrangian fiber @xmath365 is nondisplaceable .
if @xmath298 is on the segment , then for some @xmath715 one can add a ghost facet @xmath716 to so that @xmath717 using and the change of variables @xmath461 and @xmath718 , the potential function with the ghost facet added is @xmath719 hence the critical point equations at @xmath566 are @xmath720 and therefore @xmath566 is a critical point of @xmath357 when @xmath721 it follows from that such @xmath722 exist . hence @xmath365 is nondisplaceable by theorem [ t : potential ] .
points between @xmath708 and @xmath709 , see figure [ f:135r03dand ] are closer to the facet @xmath723 than any other facet , so we can not prove they are nondisplaceable with a potential ; cf . proposition [ p : noghost ] .
if @xmath724 , then @xmath725 by and the change of variables @xmath461 and @xmath462 , the potential function is @xmath726 hence the critical point equations at @xmath566 are @xmath727 and therefore @xmath566 is a critical point of @xmath357 when @xmath728 by , such @xmath722 exist .
hence @xmath365 is nondisplaceable by theorem [ t : potential ] .
* symmetric extended probes : * we will displace everything except the points on the black solid lines on the left in figure [ f:135r03dand ] .
the points are @xmath729 we will use a symmetric probe @xmath40 based arbitrarily close to @xmath730 on @xmath731 , with direction @xmath732 and in particular lies on the line @xmath733 for @xmath734 .
the associated reflection is given by @xmath735 * part 1 : * let @xmath35 be based at @xmath736 for @xmath737 , with direction @xmath738 and form the symmetric extended probe @xmath201 where @xmath739 where the direction @xmath740 .
the assumption on @xmath741 ensures that @xmath185 exists the polytope on @xmath742 .
one can check that @xmath743 so the midpoint of the extended probe always lies on @xmath35 , in fact it lies on the line @xmath744 , which appears as the line connecting @xmath745 and @xmath746 on the left in figure [ f:135r03dand ] . therefore by theorem [ t : noflag ] , every point on @xmath35 before the midpoint is displaceable .
now observe that since @xmath747 is integrally transverse to the facet @xmath748 , on which @xmath180 exits the polytope , we can swap the roles of @xmath35 and @xmath180 .
hence everything on these extended probes past the midpoint are displaceable as well , with the exception of when @xmath749 , since then @xmath185 is not on the interior of a facet .
as @xmath741 and @xmath750 vary , this sweeps out the points in the regions @xmath751 that are not on the line @xmath744 .
these are the light gray regions on the left in figure [ f:135r03dand ] . * part 2 : * now let @xmath35 be based at @xmath736 , with direction @xmath738 for @xmath752 and form the associated symmetric extended probe @xmath201 given by @xmath753 where the direction @xmath740 .
note that the restriction on @xmath741 is to ensure that @xmath754 lies on @xmath755 .
one can check that @xmath743 so the midpoint of the extended probe always lies on @xmath35 .
therefore by theorem [ t : noflag ] , every point on @xmath35 before the midpoint is displaceable .
now observe that since @xmath747 is integrally transverse to the facet @xmath748 , on which @xmath180 exits the polytope , we can swap the roles of @xmath35 and @xmath180 .
hence everything on these extended probes past the midpoint are displaceable as well , with the exception of when @xmath749 , since then @xmath185 is not on the interior of a facet .
as @xmath741 and @xmath750 vary , this sweeps out the points in the regions @xmath756 that are not on the line segment connecting @xmath746 and @xmath757 .
these are the dark gray regions on the left in figure [ f:135r03dand ] .
* standard probes : * it is straightforward to check that standard probes with directions @xmath758 displace everything not already displaced except the points @xmath759 and the points on the line @xmath760 such that @xmath761 , whose endpoint is denoted @xmath710 in figure [ f:135r03dand ] .
see the middle polytope in figure [ f:135r03dand ] .
one can carry out a similar analysis of the points in the full minimal resolution of @xmath0 .
the result is qualitatively the same : there are @xmath28 isolated points that are known to be non displaceable because their qw invariants are nonzero , there are a finite number of line segments of unknown properties ( more than before because there are more vertices ) , and otherwise everything is displaceable .
here are the details .
the polytope is given by @xmath762 where @xmath763 are from , and the support constants satisfy @xmath764 and are such that the vertices of @xmath765 are @xmath766 where @xmath767 are given in .
are resolved as in . *
left * : in grey , points displaceable by symmetric extended probes that use @xmath40 , which is based on @xmath768 , as the deflecting probe . *
middle * : displacing some more points with standard probes that have direction @xmath706 . *
right * : in light grey are the fibers that were displaced in the previous two pictures .
the white line segments are on the lines from .
the point @xmath707 is nondisplaceable just as in proposition [ p:135rdand ] .
the point @xmath769 satisfies @xmath770 , the point @xmath708 satisfies @xmath771 , the point @xmath772 satisfies @xmath773 , and all three points are nondisplaceable .
displaceability is unknown for points on the white line segments and the points marked with white diamonds .
see remark [ r:135tr].,width=576 ] [ p:135trdand ] in a resolution of @xmath0 given by @xmath765 , if @xmath774 does not lie on one of the lines @xmath775 and is not the point @xmath714 , then @xmath365 is displaceable by probes or extended probes .
the proof is the same as the proof of proposition [ p:135rdand](i ) .
see figure [ f:135trdand ] .
[ r:135tr ] the polytope on the right in figure [ f:135trdand ] depicts stronger displaceability results than proposition [ p:135trdand ] .
they are obtained by using the extended probes with trapezoidal flags which are introduced in the next section .
one can reach some points on @xmath776 above @xmath769 by a probe formed by deflecting @xmath35 based on @xmath777 with direction @xmath778 by @xmath40 based on @xmath779 with direction @xmath780 . also , points on @xmath781 above and slightly below @xmath772 , are displaceable by probes formed by deflecting @xmath35 , based on @xmath782 or @xmath783 with direction @xmath784 , with a probe @xmath40 based on @xmath779 with direction @xmath780 .
[ p : openset ] the full minimal resolution of @xmath785 has an open set of points with trivial qw invariants that can not be displaced by extended probes . as illustrated in figure [ f:5 - 8tr ] , the full resolution @xmath786 has an open set of unknown points . in the resolution @xmath787 , these would lie near the vertex @xmath788 in the region above the ray with direction @xmath699 . therefore , when we resolve at @xmath788 they would lie above the symmetric probes @xmath40 starting on the facet @xmath34 with conormal @xmath789 . in the case @xmath0
such points were reached by probes starting on the slant edge with direction @xmath790 and then deflected by @xmath40 to be vertical .
but the corresponding probes do not exist in @xmath787 because @xmath790 is not complementary to @xmath791 .
points in this region do lie on the extensions of vertical probes from the base that are deflected by @xmath40 , but they lie more than halfway along such probes . therefore these points can not be displaced . on the other hand , by proposition [ p : noghost ] , there are at most finitely many points in this region with nonvanishing qw invariants .
similarly , the singularity of @xmath792 at @xmath793 is modelled on @xmath794 and has a nearby open set of points that are not probe - displaceable .
these arguments generalize to show that typically the resolution of @xmath795 has an open set of points with unknown properties .
in this section we will generalize extended probes with flags , definition [ d : pepwf ] and theorem [ t : pdeflected ] , to the case where the probe @xmath35 is not parallel to the base facet @xmath188 of @xmath40 . before diving into the more complicated notation for the non - parallel extended probes ,
let us first demonstrate that theorem [ t : pdeflected ] is not valid as stated when the probe @xmath35 is not parallel to @xmath188 .
we will do this by showing that if it was valid , then we could displace the clifford torus in @xmath200 , which is known to be nondisplaceable @xcite .
if the moment polytope for @xmath200 is given by @xmath796 then the fiber @xmath199 over @xmath797 is the clifford torus .
what follows is similar to remark [ r : nohelp ] for symmetric extended probes . .
why theorem [ t : pdeflected ] is not valid if @xmath35 is not parallel to @xmath188 , the facet on which the deflecting probe is based.,width=240 ] [ e : nohelpf ] let @xmath35 and @xmath40 be the probes where @xmath798 form the ` parallel ' extended probe with flag @xmath310 where @xmath799 we have that @xmath800 $ ] has length @xmath801 and passes through @xmath39 . since @xmath802 if theorem [ t : pdeflected ] applied then @xmath48 would displace the fiber @xmath199 .
of course it does not apply since @xmath35 is not parallel to @xmath188 . despite the above failure , the parallel condition in theorem [ t : pdeflected ] can be restrictive and in trying to relax it we are led to the following general notion of _ extended probes with flags_. , where the flag parameter @xmath803.,width=288 ] [ d : epwf ] let @xmath35 and @xmath40 be probes in a rational polytope @xmath108 where the probe @xmath35 ends at the point @xmath178 on @xmath40 . the * extended probe with flag *
@xmath48 formed by deflecting @xmath35 with @xmath40 is the subset @xmath804 where the flag @xmath805 is the convex hull of the points @xmath261 in @xmath32 .
the points @xmath262 and @xmath263 are on @xmath40 , while @xmath806 the parameter @xmath807 $ ] affects the shape of the flag , and the length of the flag @xmath808 must be small enough so that @xmath809 and @xmath810 stay in @xmath32 .
the * length * of the extended probe with flag @xmath46 is @xmath811 .
we also assume that the line segment @xmath812 $ ] does not cross the line segment @xmath813 $ ] , so that they are boundaries of the flag as in figure [ f : epnp ] .
[ r : flag ] ( i ) if @xmath161 is parallel to the facet @xmath188 where @xmath167 is based , then @xmath814 and hence by , the shape of the flag @xmath260 is independent of the parameter @xmath815 . in this case
we recover the definition of a parallel extended probe with flag , definition [ d : pepwf ] .
\(ii ) when @xmath161 is not parallel to the facet @xmath188 the parameter @xmath815 affects the shape of the flag . for the hyperplane @xmath816 ,
the projection @xmath817 along @xmath167 is @xmath818 and the reflection @xmath819 across @xmath820 via @xmath167 is @xmath821 so as @xmath807 $ ] varies the shape of the flag @xmath805 linearly interpolates between @xmath822 for varying flag parameters , as per remark [ r : flag](ii).,width=480 ] the following theorem explains how one can use extended probes with flags to displace lagrangian torus fibers .
[ t : deflected ] let @xmath823 be an extended probe with a flag constructed from probes @xmath35 and @xmath40 as above , in a moment polytope @xmath192 for a toric symplectic orbifold @xmath193 . for a point @xmath824 , if @xmath39 is in the interior of @xmath35 , the affine distance from @xmath39 to the facet @xmath154 satisfies @xmath825 and the flag @xmath260 satisfies both inequalities @xmath826 then the lagrangian fiber @xmath17 is displaceable .
the second condition in did not appear in theorem [ t : pdeflected ] , for there it is trivially satisfied since if @xmath161 is parallel to @xmath188 , then @xmath827 .
compare the following remark with remark [ r : nohelp ] .
[ r : extprob ] ( i ) if @xmath154 and @xmath188 were the only facets in the polytope , then @xmath207 represents the maximum length @xmath35 could be extended to before it hit the facet @xmath188 . the second condition in implies that this maximum length is an _ a priori _ upper bound @xmath828 on the length @xmath829 for an extended probe with flag @xmath46 formed with probes @xmath35 and @xmath40 .
\(ii ) this a priori upper bound has the following consequence : suppose a probe @xmath35 exits the polytope @xmath32 through the facet @xmath34
. then the displaceability results given by using @xmath35 as a standard probe can not be improved on by using theorem [ t : deflected ] with a deflecting probe @xmath40 based on @xmath34 .
\(iii ) in the counter - example in section [ s : mce ] , the deflecting probe @xmath40 is based on the facet through which the probe @xmath35 would exit the polytope .
so in this example , is violated and hence the second condition in is as well .
let us consider a minimal resolution of an @xmath830-singularity that now has finite volume , so the moment polytope is given by @xmath831 where the finite volume @xmath830-singularity is defined by @xmath832 and the minimal resolution at the origin uses @xmath833 figure [ f : fvanr ] depicts the polytope @xmath834 when @xmath835 in this example there remains an open region of unknown points , even after using extended probes . .
in light gray are points displaceable by probes , in medium gray are points displaceable by extended probes with trapezoidal flags , and in dark grey are nondisplaceable points.,width=384 ] [ p : fvanr ] in the resolution of the finite volume @xmath830-singularity given in , then in the polytope @xmath834 : 1 .
the lagrangian fiber @xmath365 is displaceable by probes if @xmath298 is in one of the regions @xmath836 and not on the line segment @xmath837 .
the lagrangian fiber @xmath365 is displaceable by an extended probe with trapezoidal flag if @xmath298 is in the region @xmath838 3 .
the lagrangian fiber @xmath365 is nondisplaceable if @xmath298 is in the region @xmath839 or on the line segment @xmath840 .
part ( i ) is straightforward with using probes with direction @xmath291 on @xmath485 , direction @xmath289 on @xmath613 and @xmath649 , direction @xmath231 on @xmath649 and direction @xmath841 on @xmath486 .
part ( iii ) is also straightforward . for the points in the region one uses the ghost facets
@xmath842 and @xmath843 for varying @xmath844 and @xmath845 .
displacing the point @xmath39 .
points in the dark grey region can be displaced with these types of extended probes.,width=384 ] for part ( ii ) , let @xmath40 be a probe with direction @xmath846 based on @xmath649 arbitrary close to @xmath847 .
let @xmath35 be based at @xmath848 for @xmath849 , with direction @xmath612 , and form the extended probe with flag @xmath850 where @xmath851 is the flag parameter .
the flag is given by @xmath852 where the upper bound on @xmath265 comes from the second condition in and @xmath262 should be moved slightly closer to @xmath178 so that the first condition in is satisfied . by theorem [ t : deflected ] , this extended probe displaces everything on @xmath35 between @xmath853 and @xmath854 .
as @xmath741 ranges over @xmath849 , these extended probes displace precisely the region in ( ii ) .
consider the symplectic form @xmath855 on @xmath456 .
for this symplectic form , the standard @xmath856 action on @xmath456 by @xmath857 is given by the moment map @xmath858 where @xmath859 .
the symplectic form @xmath860 is also normalized so that @xmath861 where @xmath862 is the disc @xmath863 we will denote its boundary by @xmath864 and the annulus by @xmath865 for each theorem we have an extended probe @xmath46 in a toric symplectic orbifold @xmath193 and our goal is to displace the lagrangian torus fiber @xmath17 . to displace @xmath199 , it suffices to build an embedding @xmath866 such that for some @xmath867 @xmath868 where @xmath869 is the projection .
since @xmath870 there is a hamiltonian isotopy of @xmath871 supported away from the boundary that displaces @xmath872 , and therefore the embedding can be used to extend this to a hamiltonian isotopy of @xmath1 that displaces @xmath199 .
while the precise details for building @xmath873 vary depending on the type of extended probe , the following outline describes the general process . here
@xmath874 are the three parts of the extended probe @xmath46 , where @xmath875 is either @xmath180 , @xmath260 , or @xmath805 depending on the type of extended probe .
stage 1 : : for @xmath876 , produce an embedding @xmath877 except for the second version of theorem [ t : noflag ] , we have @xmath878 and we will show @xmath879 for @xmath880 where by assumption @xmath867 .
stage 2 : : for @xmath881 , produce an embedding @xmath882 in the second version of theorem [ t : noflag ] , we have @xmath883 and we will show @xmath884 for @xmath885 where by assumption @xmath867 .
stage 3 : : use the deflecting probe @xmath40 to build a symplectomorphism @xmath886 of @xmath8 such that @xmath887 and @xmath888 glue together to form an embedding @xmath889 that satisfies . since the fiber @xmath199 is disjoint from @xmath890 , to ensure @xmath873 satisfies the second condition in it suffices to prove that @xmath886 can be built to be supported in any given neighborhood of @xmath891 .
the canonical symplectic form @xmath892 on @xmath893 is @xmath894 if @xmath895 and @xmath896 are dual bases , then in the associated coordinates @xmath897 the canonical symplectic form is @xmath898 and it is clear that the projection @xmath899 is the moment map for the obvious @xmath98-action .
now let @xmath900 be the moment polytope for a symplectic toric orbifold @xmath193 .
then @xmath193 can be modeled by @xmath901 by performing a symplectic cut along @xmath902 for each facet @xmath903 .
if @xmath904 , then this amounts to replacing @xmath905 with @xmath906 where @xmath907 is the circle generated by @xmath34 s primitive interior conormal @xmath908 . in this way we can consider the action - angle coordinates @xmath909 as a global coordinate system on @xmath910 .
for example the disk @xmath911 has action - angle coordinates @xmath912 \times \t^{1}$ ] where the circle @xmath913 is collapsed to a point .
the explicit identification @xmath914 pulls @xmath860 back to @xmath915 .
likewise the annulus @xmath916 has action - angle coordinates @xmath917 \times \t^{1}$ ] with @xmath918 .
let @xmath35 be a probe with direction @xmath919 , length @xmath920 , and based at the point @xmath153 on the interior of a facet @xmath154 , which has primitive interior conormal @xmath921 .
since @xmath161 is integrally transverse to @xmath159 and inward pointing , there is a lattice basis for @xmath505 of the form @xmath922 where @xmath923 using the model @xmath924 for @xmath910 define the embedding @xmath925 where if @xmath926 are coordinates for @xmath927 such that @xmath928 are action - angle coordinates for @xmath929 , then @xmath930 is given by @xmath931 this map is well - defined since for fixed @xmath932 , the image of the map @xmath933 lies in @xmath934 , which in @xmath1 is replaced with @xmath935 . by design @xmath936 as in , where @xmath937 is the point on @xmath35 such that @xmath938 .
this embedding @xmath930 will serve as in stage 1 for all the extended probe theorems .
let @xmath939 be a rational line segment starting at @xmath940 and ending at @xmath941 in @xmath942 , with length @xmath943 and direction @xmath944 .
let @xmath945 be an integral basis for @xmath505 that satisfies with respect to @xmath946 .
then similarly to the case of a probe , using the model @xmath947 for @xmath910 we can define an embedding @xmath948 for any @xmath949 , such that @xmath950 where @xmath951 \times \t^{1}$ ] are action - angle coordinates for @xmath952 .
again we have @xmath953 where @xmath954 is the point on @xmath524 such that @xmath955 .
for @xmath956 , this embedding @xmath957 will serve as in stage 2 for theorem [ t : noflag ] .
the probe @xmath35 has length @xmath920 , direction @xmath919 , starts at @xmath960 , ends at the point @xmath961 . for a choice of lattice basis @xmath962 for @xmath505 that satisfies ,
define the embedding for stage 1 @xmath963 so that for @xmath964 \times \t^{1}$ ] , @xmath965 as in .
the rational line segment @xmath180 has length @xmath966 , direction @xmath967 , starts at the point @xmath968 , and ends at @xmath969 .
since @xmath970 is an element of @xmath971 , the image under @xmath972 , the dual of @xmath970 , of the lattice basis of @xmath505 used for @xmath930 : @xmath973 is still a lattice basis .
this new basis satisfies with respect to @xmath974 since @xmath975 .
define the embedding for stage 2 , where @xmath976 , @xmath977 so that for @xmath978 \times \t^{1}$ ] , @xmath979 as in .
recall that for our symmetric probe @xmath980 , we have the affine reflection @xmath168 and the linear version @xmath981 from and .
observe that since @xmath975 it follows that @xmath982 is a symplectomorphism with respect to the canonical symplectic form .
comparing and we see that to establish stage 3 it suffices to prove the following proposition , which can be seen as a local version of ( * ? ? ?
* proposition 5.5 ) .
[ p : lhi ] let @xmath980 be a symmetric probe in the moment polytope for a symplectic toric orbifold @xmath193 . then for any neighborhood of @xmath983 of @xmath891
, there is a hamiltonian isotopy of @xmath1 supported in @xmath983 with time one map @xmath886 such that @xmath984 for a smaller neighborhood @xmath985 of @xmath890 , where @xmath986 is given by .
consider the special case where @xmath167 is parallel to every facet except @xmath188 and @xmath166 , meaning @xmath987 for all other interior conormals @xmath908 .
we have that @xmath988 where without loss of generality @xmath989 is a lattice basis for @xmath505 and @xmath990 .
let @xmath991 be a dual basis for @xmath992 .
we have that @xmath993 since @xmath994 and @xmath995 otherwise .
we can identify @xmath193 with @xmath996 , which is built by performing symplectic reduction on the standard @xmath329 .
in particular it has the form @xmath997 where the level set @xmath998 for the hamiltonians @xmath999 is symplectically reduced using the action of the @xmath1000 dimensional subtorus @xmath1001 whose action is given by the hamiltonians @xmath1002 on @xmath370 .
the moment map for the action of @xmath98 on @xmath1003 is given by @xmath1004 for a point @xmath339 on the probe @xmath40 , since @xmath1005 and @xmath995 for @xmath1006 , it follows that @xmath1007 and hence @xmath1008 is @xmath1009 now the standard hamiltonian @xmath477 action on @xmath1010 preserves the level set @xmath1011 and commutes with the action of @xmath1012 , so it descends to a hamiltonian @xmath477 action on @xmath1013 .
consider the element @xmath1014 so that @xmath1015 . since @xmath1016 it follows for @xmath1017 that we have @xmath1018 where the second to
last equality uses that @xmath1019 from for points on @xmath1011 . therefore up
to applying a uniform rotation using the toric action , we have that @xmath1014 acts on @xmath1020 as the hamiltonian diffeomorphism @xmath1021 from .
now let @xmath1022 be such that @xmath1023 and let @xmath1024 be the autonomous hamiltonian whose corresponding hamiltonian flow @xmath1025 on @xmath1026 is the action of @xmath1027 .
since @xmath1028 poisson commutes with @xmath1029 for @xmath1006 , it follows that @xmath1025 in @xmath1030 preserves level sets of the form @xmath1031 in particular @xmath1032 is preserved .
now let @xmath1033 be any neighborhood of @xmath1032 and let @xmath1034 be a bump function that is a constant @xmath704 near the level set and @xmath1035 . then the time one flow @xmath1036 for the hamiltonian @xmath1037 is the desired element of @xmath1030 .
let @xmath1038 be a lattice basis satisfying with respect to @xmath167 .
if @xmath339 is a point on @xmath40 , then for @xmath1039 define the rational half - spaces @xmath1040 for @xmath1041 .
since @xmath40 starts and ends at points in the interior of the facets @xmath188 and @xmath166 of @xmath32 , respectively , for @xmath1039 sufficiently small @xmath1042 is a neighborhood of @xmath1043 .
furthermore @xmath1044 is the moment polytope for a symplectic toric manifold @xmath1045 , that satisfies the special condition that all facets except @xmath188 and @xmath166 are parallel to @xmath167 . by the special case of propositon [ p : lhi ]
we can build the desired hamiltonian isotopy @xmath886 in @xmath1046 that is generated by an autonomous hamiltonian supported in @xmath1047 .
since @xmath1048 canonically embeds into @xmath1 , preserving the toric structure , we can see @xmath1049 as our desired hamiltonian isotopy .
[ [ proving - theoremtpdeflected - parallel - extended - probes - with - flags ] ] proving theorem [ t : pdeflected ] : parallel extended probes with flags ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ let @xmath823 be a parallel extended probe with flag .
since the direction @xmath161 of @xmath35 is parallel to the facet @xmath188 , we can pick dual lattice bases @xmath1050 for @xmath505 and @xmath1051 for @xmath992 so that @xmath1052 picking action - angle coordinates @xmath1053 on @xmath1054 with respect to these bases , we can let the points on the probe @xmath1055 have coordinates @xmath1056 where @xmath920 is the length of @xmath35 and @xmath1057 is @xmath1058 .
besides @xmath178 , let the other points on @xmath1059 be @xmath1060 where @xmath1061 are positive and @xmath1062 . by , the end points of @xmath260 are @xmath1063 where @xmath1064 is the length of the extended probe @xmath46 and @xmath312 is the length of the flag @xmath260 . in our coordinates for @xmath96
the flag @xmath260 is given by @xmath1065 it follows from that @xmath1068 and therefore by lemma [ l : apd ] there is a hamiltonian isotopy @xmath1069 of @xmath1070 supported in the interior so that @xmath1071 if @xmath1072 \times \t^{1}$ ] are action - angle coordinates on @xmath1070 , then write this hamiltonian diffeomorphisms as @xmath1073
. now let @xmath978 \times \t^{1}$ ] be action - angle coordinates on @xmath1074 and using the action - angle coordinates @xmath1066 on @xmath1 define the embedding @xmath1075 @xmath1076 observe that the formula for @xmath1077 is just the result of applying @xmath1069 to the @xmath1078 coordinates in the formula for @xmath930 .
it is straightforward to check that this embedding @xmath1077 satisfies the conditions for . in the standard toric structure @xmath1079 ,
consider the probe @xmath1080 given by @xmath1081\}\ ] ] where @xmath1082 are positive , then @xmath1083 for small @xmath1084 we define the subsets of @xmath1085 @xmath1086 \times [ 0,q-\de ] \times \prod_{k\ge 3 } [ b_k-\eps , b_k+\eps],\\ x^\eps : = [ b_1-\eps , b_1+\eps ] \times [ 0,q ] \times ( b_3,\dots , b_n).\end{gathered}\ ] ] [ l : deflecting ] let @xmath1069 be any area preserving diffeomorphism of @xmath1087 that is the identity near @xmath1088 .
if @xmath1089 for @xmath1090 sufficiently small , then there exists a hamiltonian isotopy of @xmath1091 supported in @xmath1092 such that the time one map of the isotopy @xmath886 in a small neighborhood @xmath1093 of @xmath1094 is given by @xmath1095 in terms of the decomposition in and in particular @xmath1096 .
the map @xmath1069 is the time one map of a hamiltonian isotopy , generated by some time - dependent hamiltonian @xmath1028 with support in @xmath1097 .
simply multiply @xmath1028 by a cut - off function @xmath1098 $ ] that is a function of the variable @xmath1099 and @xmath1100 near @xmath1101 .
the time one map @xmath1102 of the hamiltonian isotopy generated by @xmath1103 has the desired properties .
take now @xmath1080 with @xmath1104 and @xmath1105 to be a local model for our probe @xmath980 . by applying lemma [ l : deflecting ] to the hamiltonian diffeomorphism @xmath1069 of @xmath1106 from in stage 2
, the resulting hamiltonian diffeomorphism @xmath1102 can be extended by the identity outside its support to be an element of @xmath1107 .
it follows from that near @xmath890 the hamiltonian diffeomorphism @xmath1102 has the form @xmath1108 in our action - angle coordinates @xmath1066 .
let @xmath1110 be a extended probe with flag .
we can pick dual lattice bases @xmath1050 for @xmath505 and @xmath1051 for @xmath992 so that @xmath1111 where @xmath1112 are relatively prime and without loss of generality @xmath1113 .
picking action - angle coordinates @xmath1053 on @xmath1054 with respect to these bases , we can let the points on the probe @xmath1055 have coordinates @xmath1114 where @xmath920 is the length of the probe @xmath35 .
the points on @xmath1059 are @xmath1115 where @xmath1061 are positive and @xmath1062 . by , the end points of @xmath805 are @xmath1116 where @xmath1064 is the length of the extended probe @xmath46 and @xmath312 is the length of the flag @xmath260 .
assume now that @xmath1118 , so that @xmath161 points towards @xmath188 . by lemma [ l : apd ]
there is a compactly supported hamiltonian isotopy @xmath1069 of @xmath1070 such that @xmath1119 where @xmath1120 by , we have that @xmath1068 and hence @xmath1121 assume that @xmath1122 , so that @xmath161 points away from @xmath188 . by lemma
[ l : apd ] there is a compactly supported hamiltonian isotopy @xmath1069 of @xmath1123 such that @xmath1124 where @xmath1120 by , we have that @xmath1068 and hence @xmath1125 now let @xmath978 \times \t^{1}$ ] be action - angle coordinates on @xmath1074 and using the action - angle coordinates @xmath1066 on @xmath1 define the embedding @xmath1126 observe that the formula for @xmath1127 is just the result of applying @xmath1069 to the @xmath1078 coordinates in the formula for @xmath930 .
it is straightforward to check that this embedding @xmath1127 satisfies the conditions for , in particular @xmath1128 follows from and . when @xmath1129 , note that @xmath1130 .
so the second assumption in , i.e. @xmath1131 , ensures that @xmath1132 in does not collapse to a point .
if it did collapse to a point , then @xmath1127 would no longer be an embedding and this is necessary for our proof .
the rest of the proof is now the same as in the parallel case . by applying lemma [ l : deflecting ] to the hamiltonian diffeomorphism @xmath1069 of @xmath1106 from in stage 2
, the resulting hamiltonian diffeomorphism @xmath1102 can be extended by the identity outside its support to be an element of @xmath1107 .
it follows from that near @xmath890 the hamiltonian diffeomorphism @xmath1102 has the form @xmath1133 in our action - angle coordinates @xmath1066 . for real numbers
@xmath1134 , pick a smooth function @xmath1135 \to [ a , b]$ ] that is non - increasing , is such that @xmath1136 and the function @xmath1137 is non - decreasing for @xmath1138 $ ] .
above we picked @xmath1139 to have the form @xmath1140$}\ ] ] where the parameter @xmath815 corresponds with the flag parameter .
choose a smooth family of disjoint , contractible closed curves @xmath1147\ ] ] that each enclose a region of area @xmath1148 .
this is possible because @xmath305 .
next pick a compactly supported diffeomorphism @xmath1149 of @xmath1150 such that each circle @xmath1151 is mapped to @xmath1152 for all @xmath1153 .
finally isotope @xmath1149 to an area preserving @xmath1069 via moser s method . using that @xmath1152
encloses the same amount of area as @xmath1151 , it is not hard to check that the isotopy is given by flowing along a vector field @xmath1154 that at each time @xmath1155 is tangent to the curves @xmath1152 . | in this paper we introduce a new way of displacing lagrangian fibers in toric symplectic manifolds , a generalization of mcduff s original method of probes .
extended probes are formed by deflecting one probe by another auxiliary probe . using them
, we are able to displace all fibers in hirzebruch surfaces except those already known to be nondisplaceable , and can also displace an open dense set of fibers in the weighted projective space @xmath0 after resolving the singularities .
we also investigate the displaceability question in sectors and their resolutions .
there are still many cases in which there is an open set of fibers whose displaceability status is unknown . |
the addition of oppositely charged polyions to a suspension of charged colloidal particles gives rise to an intriguing and partially unexpected phenomenology @xcite , resulting in the formation of aggregates which play an important role in a wide range of implications @xcite , from membrane biophysics and soft matter physics @xcite to biotechnological processes , such as therapeutic delivery systems @xcite .
these aggregates are governed by a delicate balance between attractive and repulsive interactions resulting in the appearance of an equilibrium ( or a kinetically arrested ) cluster phase , where single particles , stuck together by an electrostatic glue acting between oppositely charge domains , form reversible , relatively large , complexes . in this system , the short - range attraction contribution is promoted by the addition of adsorbing polyions which form a two - dimensional strongly correlated , short - range ordered , structure on the surface of the oppositely charged particle , contrarily to what happens for non - adsorbing polymers , where the attraction contribution is produced by the unbalanced osmotic pressure in the depletion regime .
in fact , correlated adsorption of polyion chains on the surface of oppositely charged particles induces two different phenomena : charge inversion@xcite and reentrant condensation@xcite .
the first one occurs when a colloidal particle binds several oppositely charged multi - valent ions ( polyions ) , so that its effective charge inverts its sign . the second effect , concomitant to the charge inversion , consists in the formation of particle aggregates whose average size increases on increasing the polyion concentration , until it reaches a maximum ( at the point of charge inversion ) , decreasing afterwards to the initial value .
this phenomenology has been investigated in a variety of different colloidal systems both for particles bearing a negative charge ( such as dotap liposomes ) in the presence of anionic polyions @xcite and for positively charged particles ( such as hybrid niosomes ) in the presence of cationic polyions@xcite . )
( a ) and charge inversion ( @xmath0-potential ) ( b ) of positively charged liposomes , in the presence of negatively charged polyions .
right panels : reentrant condensation ( aggregate size @xmath1 ) ( a ) and charge inversion ( @xmath0-potential ) ( b ) of negatively charged hybrid niosomes , in the presence of positively charged polyions .
the data are shown as a function of the charge molar ratio @xmath2 normalized to the value @xmath3 at which the measured @xmath0-potential goes to zero .
@xmath2 is defined as the polyion to the lipid molar ratio .
the experimental values ( hydrodynamic radius measured by means of dynamic light scattering and @xmath0-potential measured by means of doppler laser electrophoretic techniques ) refer to different colloidal systems .
positive charged particles : ( @xmath4 ) : dotap liposomes ( 0.8 mg / ml ) and polyacrylate sodium salt ; ( @xmath5 ) : dotap liposomes ( 1.7 mg / ml ) and polyacrylate sodium salt@xcite .
negative charge particles : ( @xmath6 ) : hybrid niosomes and @xmath7-polylysine ; ( @xmath8 ) : hybrid niosomes and @xmath9-polylysine ; ( @xmath10 ) : hybrid niosomes and pevp ( ionization degree 65% ) ; ( @xmath11 ) : hybrid niosomes and pevp ( ionization degree 95%)@xcite .
hybrid niosomes are built up by tween20 , cholesterol and dicethylphosphate and the cationic polyion pevp is poly[n - ethyl-4-vinyl pyridinium ] bromide.,width=340 ] the occurrence of a sharp reentrant condensation on increasing the adsorption of polyion on charged liposomes has been documented in a series of articles@xcite .
it has been shown that the size of the aggregates goes through a maximum and their overall charge inverts its sign in concomitance to the point of charge inversion , as briefly reviewed in fig .
[ rz ] for some typical cationic or anionic colloidal particle systems .
the formation of aggregates and , consequently , the stability of colloidal charged particles ( solid particles or soft vesicles ) is usually satisfactorily explained in the framework of derjaguin - landau - verwey - overbeek ( dlvo ) theory @xcite , which accounts for electrostatic and van der waals interactions between two approaching particles . in the traditional dlvo theory ,
the particle surface is assumed to be uniformly charged . in the case of polyion - correlated adsorption onto the surface of the particle ,
this assumption is no longer valid and the usual dlvo theory might not account for the effects arising from the surface charge heterogeneity .
indeed , the adsorption of the polymer generates an ordered charged structure , strongly dependent on the polyion size and on the valence and linear charge density @xcite .
some experimental evidences of such a structure can be found in refs.[13 ] and [ 14 ] .
transmission electron microscopy [ tem ] measurements suggest the formation of a local patchy charge distribution . in this respect
, it is reasonable to expect that attractive interactions can arise when `` counterion domains '' on one particle align to a `` counterion - free domains '' on another approaching particle .
recently , velegol and thwar @xcite have developed an analytical model for the effective particle - particle potential @xmath12 in the case of non - uniform charge distributions , based on the derjaguin approximation and on an extention of the hogg - healy - fuerstenau [ hhf]@xcite model for randomly charged surfaces .
the resulting spherically symmetric potential depends on the values of the average surface potential ( here assumed equal to the @xmath0-potential ) and on the standard deviation @xmath13 of the surface potential among different regions on the particle surface .
in particular , velegol and thwar @xcite showed that the value of the second moment of the charge distribution lowers the electrostatic repulsion . in this work , by means of mc simulation
, we analyze the aggregation process of spherical charged particles interacting via the velegol and thwar @xcite dlvo - like potential . starting from isolated monomers ,
we follow the growth of the aggregates until a kinetically arrested state is reached .
we study the final size of the aggregates for several values of the @xmath0-potential and its standard deviation @xmath13 .
the results of the simulations confirm that the final size of the aggregates grows on increasing @xmath13 and on decreasing @xmath0 , providing a microscopic interpretation of the re - entrant condensation phenomenon of polyion - induced charged colloidal particle aggregation ( fig .
according to the hhf model , the electrostatic potential of mean force @xmath14 between two dielectric colloidal spherical particles of radius @xmath15 and @xmath16 and surface potentials @xmath17 and @xmath18 , respectively , is given by @xmath19 @xmath20\ ] ] here , @xmath9 is the dielectric permittivity of dispersing medium , @xmath21 is the debye screening length and @xmath22 is the minimal distance between the two particle surfaces ( see fig .
[ derj ] ) . + and @xmath16 , respectively , width=226 ] eq .
( [ sphere1 ] ) , which is derived under the conditions @xmath23 and @xmath24 ( @xmath7=a , b ) , holds under the following assumption : _ i _ ) @xmath17 and @xmath18
are relatively small ( less than 25 mv at room temperature ) ; _ ii _ ) the particles share the same chemical nature and are charged by similar mechanism ; _ iii _ ) the derjaguin approximation holds@xcite .
velegol and thwar @xcite assume that the surface of a particle @xmath7 is partitioned in @xmath25 regions ( labelled by @xmath26 , with @xmath27 ) of area @xmath28 , each of them characterized by a different value of the surface potential @xmath29 .
these regions must be of sufficient size @xmath30 , so that lateral interactions can be considered negligible ( practically this means that @xmath31 and @xmath32 ) .
the surface potentials @xmath17 and @xmath18 of two approaching particles in the hhf theory are thus replaced by the random value of the potential @xmath33 and @xmath34 , where @xmath26 and @xmath35 label the two domains ( on opposite particles ) facing each other . defining @xmath36 as the average potential , one can write @xmath37 where @xmath38 a random contribution that varies among the different regions .
the @xmath38 are independently distributed and not correlated , i.e. @xmath39 where @xmath40 is the krenecker delta function and @xmath41 the standard deviation of surface potential on spheres of type @xmath7 .
the expression for the resulting potential of mean force provided by velegol and thwar @xcite is @xmath42 @xmath43 where the sum runs over the @xmath44 facing regions and @xmath45 is the gap between _
_ i__th regions .
the ensemble averaging gives : @xmath46 in order to evaluate the sum in eq .
( [ sum3 ] ) , the derjaguin approximation is applied ( fig .
[ derj ] ) , i.e. @xmath47 and , for any function @xmath48 , the following expression are employed @xmath49 where @xmath50 is the flat ring radius ( see fig . [ derj ] ) . + the resulting mean force pair interaction potential is given by @xmath51 @xmath52\ ] ] the mean force potential for typical values of the @xmath0-potential and standard deviation @xmath13
is shown in figs .
[ zvar ] and [ svar ] .
the potential combines a repulsive net charge - dependent monopole term ( @xmath53 ) and an attractive multipole term ( @xmath54 ) arising from the presence of random charge heterogeneity on the particle surface .
+ ) for different values of @xmath55 ( from 11 to 19 mv ) and for a constant value of @xmath56=15 mv .
curves are plotted in units of the thermal energy @xmath57 at room temperature for @xmath58 nm and @xmath59 nm .
solid line : @xmath60 mv ; dashed line : @xmath61 mv ; dotted line : @xmath62 mv ; dot - dashed line : @xmath63 mv ; dot - dot - dashed line : @xmath64 mv.,width=340 ] + ) for different values of @xmath56 ( from 5 to 25 mv ) and for a constant value of @xmath55=15 mv .
curves are plotted in units of the thermal energy @xmath57 at room temperature for @xmath58 nm and @xmath59 nm .
solid line : @xmath65 mv ; dashed line : @xmath66 mv ; dotted line : @xmath67 mv ; dot - dashed line : @xmath68 mv ; dot - dot - dashed line : @xmath69 mv.,width=340 ] these two components lead up to a global maximum and to the presence of a potential barrier . in the case of particles with the same net charge , an attractive component is always present since the first term in eq .
( [ finalform ] ) is negative .
if the two identical particles are uniformly charged ( @xmath70 ) , both the attractive component and the global maximum vanish , recovering the hhf expression for two identical spheres ( eq .
[ sphere1 ] ) . the height of the potential barrier that two approaching particles must overcome in order to aggregate and the distance between the particle surface at which this maximum occurs can be evaluated from eq .
( [ finalform ] ) . for two identical particles ( @xmath71 )
we obtain : @xmath72+\right.\ ] ] @xmath73 \right\}\ ] ] and @xmath74 we point out that : _ i _ ) the height of potential barrier does not depend on debye screening length @xmath21 .
consequently , addition of simple salt in bulk phase ( varying the ionic strength of the solvent ) does not change the strength of the interaction but only modifies the distance at which this maximum interaction occurs ; _ ii _ ) using derjaguin approximation , a linear dependence of the barrier height on the radius @xmath1 of the particles arises , underlying that larger particles interact more intensely than smaller ones and thus aggregation in the presence of large clusters is relatively inhibited ; _ iii _ ) @xmath75 is zero when @xmath76 and @xmath77 , that is consistent with the statement that there is no attractive component for uniformly charged particles and that a lower value of @xmath13 determines a short - ranged attractive interaction . in the framework of the above stated model
, the aggregation process can be viewed as a thermally activated process strictly linked to electrostatic parameters ( the @xmath0-potential and its standard deviation @xmath13 ) and to the clusters dimensions .
colloidal particles with smaller curvature radius , lower @xmath0-potential or higher charge anisotropy will be characterized by a faster aggregation dynamics .
the aggregation process will slow down due to the progressive increase of the repulsive barrier on increasing the size of the aggregates . in the next section ,
we introduce a mc simulation method to analyze the aggregation dynamics of spherical clusters and to evaluate the evolution of the mean cluster radius for several values of @xmath0-potential and standard deviation @xmath13 , chosen to model the corresponding experimental range of values observed for polyion - induced liposome cluster aggregation . in our calculations , the contribution to the attraction associated to the van der waals interaction between two approaching particles is neglected .
the reason why such interaction can be safely neglected arises from the typical range associated to the van der waals interaction , which dyes off before @xmath78 .
indeed , for liposome particles ( aqueous cores of radii @xmath79 and @xmath80 , respectively , covered by phospholipid shell of thickness @xmath81 ) , the van der waals interaction @xmath82 can be written as @xcite @xmath83-\ ] ] @xmath84\ ] ] ( where @xmath85 is the hamaker constant ) . for typical values of the parameters
involved , eq .
[ ham ] results in a range of attraction which becomes negligible beyond @xmath86 nm . as can be seen from figs .
[ zvar ] and [ svar ] , this distance is always smaller than @xmath78 .
consequently , we neglect the van der waals interaction in the irreversible aggregation process .
we study a system composed by @xmath87 colloidal particles of initial diameter @xmath88 @xmath89 in a cubic box of volume @xmath90 with packing fraction @xmath91 where @xmath92 is the number density .
we carry out mc simulation using local metropolis algorithm at @xmath93 @xmath94 .
particles interact via a short range potential defined by eq .
( [ finalform ] ) .
we study five different pairs of @xmath0 and @xmath13 values , comparable to typical values measured in liposome solutions@xcite .
we choose @xmath95 @xmath89 .
we recall that the value of screening length does not modify the barrier height . because of the compactness of the formed clusters , aggregation is modeled as an `` oil drop - like process '' in which the shape of the aggregates retains a spherical form .
more precisely , when two approaching particles ( @xmath85 and @xmath96 ) overcome the potential barrier ( surface gap distance @xmath97 ) , they aggregate forming an unique particle with radius @xmath98 and positioned in the center of mass of two aggregating spheres .
we also assume that @xmath28 ( the size of the uniform potential regions on the particle surface in the velegol - thwar expression ) is independent of the cluster size . to incorporate a brownian dynamics , in the mc algorithm
, the @xmath26th particle is selected with a probability proportional to @xmath99 , where @xmath100 is the initial radius and @xmath101 is the radius of the @xmath26-th aggregate@xcite .
each selected aggregate is moved in each direction by a random quantity uniformly distributed between @xmath102 nm .
simulations have been carried out by varying both the @xmath0-potential and the variance @xmath103 .
the @xmath0-potential values have been chosen similar to the typical ones measured in different colloidal systems ( see fig . 1 and refs .
10 - 11 ) . on the contrary , values for @xmath103 , in absence of any experimental indication , have been chosen arbitrarily , within the validity of the model and in any case within a range of reasonable values .
the aggregation process progressively slows down and simulations are interrupted when a plateau in the time dependence of the aggregate average radius or mass is reached .
in figs . [ logz ] and [ logs ] , we present results of the aggregation process investigated for different values of the characteristic parameters ( @xmath0-potential and standard deviation @xmath13 ) .
first , we will discuss the time evolution of mean cluster radius at constant @xmath13 on varying the @xmath0-potential .
[ logz ] shows that on increasing of the @xmath0-potential , the aggregation process slows down and that the aggregates reach at long time a final limiting size .
the slowing down of the dynamics is already seen at the early stage of the aggregation process , as shown in the inset of fig .
[ logz ] , consistent with the effect of @xmath0-potential on the height of the potential barrier .
[ logs ] shows the complementary case , i.e , the effect of the different values of the standard deviation @xmath13 at a constant value of the @xmath0-potential .
again , data confirm that the increase of the random fluctuations of the charges on the particle surface speeds up the aggregation process . in all the cases
investigated , the growth process slows down at long times and the aggregates appear to reach a long - time limit value , providing a strong evidence for a dynamics slow down and an arrest on the timescale sampled by our simulation .
this finding suggests that a kinetically arrested state can be generated in this class of systems by the coupling between the aggregates size and the electrostatic barrier .
an estimate of the characteristic size reached by the aggregates at long time can be calculated assuming that the arrest is observed on the explored time scales when the interaction energy barrier reaches an _ ad - hoc _ value . indeed , inverting eq .
[ max ] , the relation connecting the radius of interacting aggregates and the potential barrier is easily obtained .
we find that , in all the cases investigated , for all the couples of @xmath0-potential and standard deviation @xmath13 we have considered , the plateau of the average aggregate size corresponds to a characteristic barrier @xmath104 of the order of about @xmath105 @xmath57 ( see figs .
[ logz ] and [ logs ] , where the normalized size @xmath106 - 1 calculated from eq . [ max ] for different values of the @xmath0-potential at fixed @xmath13 or conversely , for different values of @xmath13 at fixed @xmath0-potential is shown ) .
simulations have been carried out for different values of the standard deviation @xmath13 ( in the range from @xmath107 to @xmath108 mv ) with a constant value of the @xmath0-potential @xmath0= @xmath109 mv .
the inset shows , in linear scale , the evolution at short times of the mean size of the aggregates , for different values of the standard deviation @xmath13 ( @xmath13=25 , 20 , 15 , 10 , 5 mv , from top to bottom , respectively).,width=340 ] + -1 .
simulations have been carried out for different values of @xmath0-potential ( in the range from @xmath110 to @xmath111 mv ) with a constant value of the standard deviation @xmath13= @xmath109 mv .
the inset shows , in a linear scale , the evolution at short times of the mean size of the aggregates for different values of the @xmath0-potential ( @xmath0= 11 , 13 , 15 , 17 , 19 mv , from top to bottom , respectively).,width=340 ] it is worth nothing that @xmath0 and @xmath13 values are considered to be both constant during the aggregation process that evolves from the single polyion - decorated liposome to liposome cluster .
this assumption is justified as a consequence of the local character of the interaction , where only a local charge ( potential ) distribution determines the repulsive and attractive components of the pair potential .
the distribution of the local charge of the elementary unit ( single liposome ) surfaces can be considered independent of the size of clusters . in all the simulations , at short time below the slow down , the average size of clusters increases according to a power law @xmath112 with a pre - factor that decreases from @xmath113 to @xmath114 . at the same time , the exponent @xmath115 increases from @xmath116 to @xmath117 , as the @xmath0-potential varies from @xmath60 mv to @xmath64 mv at a constant value of @xmath67 mv .
the same behavior characterizes the @xmath13 dependence at fixed @xmath62 mv in the range from @xmath65 mv down to @xmath69 mv .
as expected , the aggregation process does not follow the diffusion limited cluster aggregation ( dlca model @xcite ) due to the presence of significant potential barriers controlling the rate of aggregation .
deviation from dlca is also evidenced by the not negligible potential barrier height at which the arrested state occurs ( of order of 10 @xmath57 ) .
[ sizez15s15 ] shows the cluster mass distributions at different stages , during the aggregation process for the case of the couple of parameters @xmath118 @xmath109 mv and @xmath119 @xmath109 mv . of the aggregating particles normalized to the mass @xmath120 of the initial particles at different aggregation states for @xmath118 15 mv and @xmath119 15 mv .
( @xmath6 ) : @xmath121 ; ( @xmath8 ) : @xmath122 ; ( @xmath4 ) : @xmath123 ; ( @xmath124 ) : @xmath125 ; ( @xmath126 ) : @xmath127 .
inset : the same distributions in a semi - log scale.,width=340 ] the system evolves from the initial mono - disperse distribution toward a broader distribution characterized by a well defined peak , whose position increases with time .
the distribution freezes when the system kinetically arrests .
the inset in fig .
[ sizez15s15 ] shows the same distribution in a semi - log scale to highlight the presence of a distribution tail compatible with an exponential distribution .
similar behavior is observed for all the cases investigated .
the main differences concern the shape of the distributions in the arrested states , some of which shown in fig .
[ fig : final ] ( see also table [ table : one ] ) . .
[ cols="^,^,^",options="header " , ] ): @xmath0=11 mv , @xmath13=15 mv ; ( @xmath128 ) : @xmath0=15 mv , @xmath13 = 15 mv ; ( @xmath129 ) : @xmath0=15 mv , @xmath13 = 25 mv ; ( @xmath130 ) : @xmath0=15 mv , @xmath13 = 20 mv ; ( @xmath131 ) : @xmath0=13 mv , @xmath13 = 15 mv .,width=340 ] the final size distribution has a smaller standard deviation @xmath132 in the case of higher values of the @xmath0-potential or lower values of @xmath13 . the average mass instead shows an opposite trend , i.e. , it increases for smaller values of the @xmath0-potential or higher values of @xmath13 . a possible explanation of the influence of the electrostatic parameters , @xmath0-potential and standard deviation @xmath13 , on the variance @xmath132 of the cluster size distribution is provided in fig .
[ pot - polydisp ] . here
, we compare the potential profiles ( eq . [ finalform ] ) where a particle interacts with particles of different sizes , for two different values of the @xmath0-potential ( 11 and 15 mv , respectively ) .
the results show that there is a preferential aggregation which favors the increase of smaller cluster rather than the cluster of larger size .
indeed , the inter - particle potential ( eq . ( [ finalform ] ) ) predicts that the test particle ( particle of mass 10@xmath120 in the example sketched in fig .
[ pot - polydisp ] ) preferentially interact with clusters of smaller size ( particle of mass 50@xmath120 ) rather than with clusters of larger size ( particle of mass 200@xmath120 ) .
the graph shows the case of a particle `` 1 '' ( with mass 10@xmath120 ) interacting with particle `` 2 '' ( with mass 50@xmath120 ) and particle `` 3 '' ( with mass 200@xmath120 ) , respectively , for two different values of the @xmath0-potential : @xmath0=@xmath110 mv ( dashed lines ) , @xmath0=@xmath109 mv ( full lines ) .
as can be seen , particle `` 1 '' preferentially aggregates with particle `` 2 '' rather than with particle `` 3 '' , to which corresponds a higher potential barrier.,width=340 ] this kind of preferential interaction , which becomes relatively more and more favorable for larger values of @xmath0 , favors the aggregation of the smaller clusters in the system .
these privileged interactions result in the formation of aggregate size distributions whose width decreases with increasing @xmath0 or decreasing @xmath13 , as shown in fig .
[ fig : final ] .
eq . 8 provides a connection between the experimentally measured values of the radius @xmath1 and the @xmath0-potential of the cluster aggregates and the unknown value of the variance @xmath103 of the surface potential @xmath133 .
this quantity can be evaluated for the different systems shown in fig .
[ rz ] , i.e. , cationic particles in the presence of anionic polyions and , conversely , anionic particles in the presence of cationic polyions . fig .
[ rzs ] shows the dependence of the standard deviation @xmath13 normalized to the value of the @xmath0-potential , @xmath134 , as a function of the normalized molar charge ratio @xmath135 . here
, @xmath2 is defined as the ratio between the polyion and lipid molar concentrations and @xmath3 is the value of @xmath136 at which the @xmath0-potential goes to zero . as can be seen , there is a more or less pronounced decrease of @xmath134 as a function of charge ratio @xmath135 for all the systems investigated , as one can expect when more and more polyion adsorption results in a more uniform charge distribution .
/@xmath137 calculated from eq . 8 on the basis of the experimental values of radius @xmath1 and the @xmath0-potential of the aggregates formed in different polyion - induced particle aggregations .
the marked regions correspond to the experimental instability of the aggregates , where their average size increases with time , until , in long time limit , they flocculate .
a : positive charged particles : ( @xmath4 ) : dotap liposomes ( 0.8 mg / ml ) and polyacrylate sodium salt ; ( @xmath5 ) : dotap liposomes ( 1.7 mg / ml ) and polyacrylate sodium salt@xcite .
b : negative charged particles : ( @xmath6 ) : hybrid niosomes and @xmath7-polylysine ; ( @xmath8 ) : hybrid niosomes and @xmath9-polylysine ; ( @xmath10 ) : hybrid niosomes and pevp ( ionization degree 65% ) ; ( @xmath11 ) : hybrid niosomes and pevp ( ionization degree 95%)@xcite .
hybrid niosomes are built up by tween20 , cholesterol and dicethylphosphate and the cationic polyion pevp is poly[n - ethyl-4-vinyl pyridinium ] bromide.,width=340 ] a final comment is in order .
experimentally , in some of the colloidal systems investigated , it has been observed that , close to the point of charge inversion , the aggregates do not reach an equilibrium radius , but their size continues to increase with time , until , in long time limit , aggregates flocculate .
this window of @xmath135 values is indicated as a dashed area in fig .
interestingly , if the value of the aggregate radius measured just before flocculation is used in eq .
[ max ] in conjunction with the corresponding @xmath0-potential value , no physical solution for @xmath13 is recovered .
more precisely , the region of @xmath1 and @xmath0 values for which no physical solutions for @xmath13 exist is delimited by the curve @xmath138 . along this curve
, the equilibrium radius depends on @xmath0-potential according to @xmath139 . ) for @xmath93 @xmath94 and permittivity @xmath9 of water.,width=340 ] all the points falling in the dashed windows ( instability windows ) of fig .
[ rzs ] are positioned in this area of @xmath140 plane ( dashed area in fig .
[ instability ] ) thus we find particularly convincing the fact that the velegol and thwar potential provides an estimate of this unstable region because of the close agreement with the experimental results .
a large body of experimental @xcite , theoretical @xcite and simulation @xcite investigations have shown that linear flexible or semiflexible polyions induce the aggregation of oppositely charged colloidal particles displaying a large variety of possible structures , thus emphasizing the many facets of charged macroion complex formation . among these structures , clusters of liposomes stuck together by oppositely charged polyions
form a class of model colloids in soft - matter physics which are of the great importance in many biotechnological implications . in these cases ,
the attractive interaction contribution to the inter - particle potential is originated by a correlated adsorption at the particle surface , which causes a non - homogeneous charge distribution at the particle surface .
since the major driving force of these processes is of electrostatic origin , more refine interaction potentials should be used to take into account some peculiar features in the aggregate cluster formation .
we have modeled particle interactions by means of a mean force inter - particle potential recently proposed by velegol and thwar , who developed a closed form analytical model to estimate the effect of a non - uniform charge distribution . within the scenario of polyion - induced charged particle aggregation ,
this potential provides a justification for why an attractive contribution arises in like - charged particle interactions in the presence of heterogeneities on the charged surface .
monte carlo simulations have been carried out for an ensemble of particles interacting via a dlvo - like - velegol and thwar potential , for different values of the parameters characterizing the system .
simulations qualitatively reproduce all experimentally observable trends in a variety of different colloidal systems , from positively charged liposomes in the presence of anionic polyions to negatively charged hybrid niosomes in the presence of cationic polyions . in particular
, simulations clearly evidence the formation of an arrested cluster phase , which is the sign of a kinetic arrested state in low - density colloidal suspensions . for electrostatically highly coupled systems , the use of the velegol and thwar potential offers an interesting promise for the understanding of the role of electrostatic attractive and repulsive interactions between charged colloidal particles and oppositely charged polyions . | polymer chains adsorbed onto oppositely charged spherical colloidal particles can significantly modify the particle - particle interactions . for sufficient amounts of added polymers , the original electrostatic repulsion can even turn into an effective attraction and relatively large kinetically stable aggregates can form . the attractive interaction contribution between two oppositely particles arises from the _ correlated _ adsorption of polyions at the oppositely charged particle surfaces , resulting in a non - homogeneous surface charge distribution . here , we investigate the aggregation kinetics of polyion - induced colloidal complexes through monte carlo simulation , in which the effect of charge anisotropy is taken into account by a dlvo - like intra - particle potential , as recentely proposed by velegol and thwar [ _ d .
velegol and p.k .
thwar , langmuir , 17 , 2001 _ ] .
the results reveal that in the presence of a charge heterogeneity the aggregation process slows down due to the progressive increase of the potential barrier height upon clustering . within this framework
, the experimentally observed cluster phases in polyelectrolyte - liposomes solutions should be considered as a kinetic arrested state . |
how ( and whether ) coherent quasiparticles form in a lightly doped mott insulator is a key question in the physics of strongly correlated electron systems .
a satisfactory theoretical understanding of this issue has been achieved in the limit where magnetic correlations do not play a prominent role , starting with the work of brinkman and rice @xcite . in cuprate superconductors however , the undoped phase is an antiferromagnetic insulator with a rather large exchange coupling @xmath17 ( on the scale of @xmath18 ) , so that we have to face the problem of the interplay between local coherence and magnetic correlations .
furthermore , there is ample experimental evidence that _ carrier localisation _ and _ magnetic frustration _ also play a crucial role in the low to intermediate doping regime .
this is particularly clear in the compound at concentrations just above @xmath19 ( the threshold for the disappearance of the antiferromagnetic long - range order ) , for which true spin - glass ordering of the copper moments has been demonstrated at very low temperature ( with @xmath20 for @xmath21 @xcite ) .
up to which doping concentration does this glassy regime persist when superconductivity is suppressed is not known at this point , but carrier localization is indeed observed at low temperature up to optimal doping in both the @xmath22 and @xmath23 directions when a strong magnetic field is applied @xcite .
it was actually predicted early on @xcite that hole doping induces strong frustration in the system when the holes become localized , replacing locally an antiferromagnetic cu - cu bond with an effectively _ ferromagnetic _ one , with a strength larger than the original @xmath17 .
we observe furthermore that the disappearance of antiferromagnetic long - range order is accompanied by the appearance of new low - energy spin excitations , of a quite different nature than spin waves , as evidenced by inelastic neutron scattering experiments @xcite-@xcite .
it is important to notice that the compounds with a glassy ground - state display , at sufficiently high temperature ( above the onset of localization ) , the same distinctive transport properties as in samples with higher doping , e.g. linear resistivity @xcite .
it is thus tempting to view these low - energy excitations as the source of anomalous scattering in the normal state .
anticipating some of the speculations made at the end of this paper , we shall argue that these low energy excitations are associated with a novel kind of spin state : the `` slushy '' state associated with the disordering of an insulating ( possibly glassy ) ground state by hole motion , quantum fluctuations and thermal effects . in this picture , many distinctive `` anomalous '' properties of the normal state of the cuprate superconductors are associated with the quantum critical regime corresponding to the @xmath24 transition at which the insulating ( glassy ) ground - state melts into a metallic ( fermi - liquid ) ground state
when doping is increased . in this paper , we shall study a highly simplified model of such a slushy state of spins and holes .
our starting point is the work of sachdev and ye @xcite , who showed that in the large-@xmath25 limit of the fully - connected random heisenberg model of @xmath26 spins , quantum fluctuations are strong enough to overcome the tendency to spin - glass ordering . instead
, a gapless spin - liquid state is found down to zero temperature with a large density of low - energy spin excitations @xcite .
remarkably , these excitations are characterized by a local dynamic spin susceptibility which has precisely the form advocated by the `` marginal fermi liquid '' phenomenological description @xcite of the low energy spin excitations in cuprates , namely : @xmath27 this model is one of the few cases in which a response function having the marginal fermi liquid form could be derived explicitly ( see also @xcite ) .
the generalization of eq.([mflt=0 ] ) to finite temperature will be given in sec.[sectcrossovertherm ] ( eq .
( [ mflchi ] ) ) and displays @xmath12 scaling . the physical mechanism for the gaplessness and the high - density of spin excitations in this model is discussed in more detail at the beginning of sec.[physics ] .
it has to do with the large number of transverse components of the spins in the large-@xmath25 limit . in this respect
, it might be a reasonable picture for the disordering of the two - dimensional quantum heisenberg spin - glass due to quantum fluctuations and low dimensionality @xcite .
the main purpose of this paper is to determine whether this marginal fermi liquid spectrum survives the introduction of charge carriers and the associated insulator to metal transition .
the physics of this problem is dominated by the interplay between two competing effects : * the formation of coherent metallic quasiparticles , which can be viewed as a binding of spin and charge degrees of freedom . in the simplest description of a doped mott insulator with @xmath28 , coherent quasiparticles form below a scale of order @xmath29 ( where @xmath30 is the doping and @xmath4 the hopping amplitude ) .
this is a `` naive '' estimate of the effective fermi - energy scale , since it ignores any effect coming from the magnetic exchange ( which will tend to suppress it ) .
* the binding of spin degrees of freedom on neighbouring sites into singlet or triplet states , and the corresponding slow dynamics of the on - site local moment .
this is the phenomenon leading to the formation of the spin - liquid state in the undoped phase , which involves a scale of order @xmath3 ( the characteristic strength of the exchange ) .
it is clear from comparing the scales above that when @xmath3 is larger than the `` naive '' coherence scale @xmath31 , the magnetic exchange prevents the formation of coherent quasiparticles at that scale : in other words , @xmath31 can not possibly be the _ actual quasiparticle coherence scale _ above which free local moments are recovered , since the exchange is still effective at energy scales between @xmath31 and @xmath3 .
it is thus expected that the _ actual _ coherence scale of the system , @xmath32 will be much smaller than @xmath31 , and that a new metallic regime in which spin degrees of freedom form a spin - liquid like state while charge degrees of freedom are incoherent will be found in the intermediate energy and temperature range @xmath33 . from the above estimates , this will be the case at small doping : @xmath34 , while a direct crossover from a coherent metal to an incoherent high temperature state is expected for @xmath1 .
these expectations are entirely borne out from our solution of the doped sachdev - ye model , as evidenced by fig.[crossovers ] , which summarizes the main crossovers found in our analysis .
it should be emphasized that this competition between metallic coherence and magnetic exchange is also essential to the physics of heavy fermion compounds @xcite . in this context , the `` naive '' coherence scale @xmath31 stands for the single - impurity kondo scale ( or rather , any estimate of the lattice kondo scale that ignores rkky interactions ) , while @xmath3 stands for the typical strength of the rkky interaction .
for this reason , the results of the present paper may also have some relevance , with appropriate changes , to the physics of the disordered rare - earth compounds near the quantum critical transition into a spin glass ground state @xcite .
the effect of charge carriers on the sachdev - ye spin - liquid phase will be investigated by generalizing the model of ref .
@xcite to a @xmath35 model , with randomness on the exchange couplings @xmath36 between nearest neighbors sites : @xmath37 in this expression , the @xmath38 spin symmetry of the electrons has been enlarged to @xmath26 @xcite .
@xmath39 is the conduction electron spin density on site @xmath40 and the spin index @xmath41 runs over @xmath42 .
the projection operator @xmath43 enforces the local constraint : @xmath44 in this manner the @xmath45 case exactly coincides with the standard @xmath35 model with the constraint of no double occupancy .
the exchange couplings are quenched random variables with random sign and magnitude , distributed according to a gaussian distribution with : @xmath46 ( throughout this paper the bar will denote an average over the disorder ) . in the following , we shall consider this model on a lattice of connectivity @xmath47 , with a nearest - neighbor hopping amplitude normalized as : @xmath48 and we shall analyze the model in the following double limit : * @xmath49 . in this limit of infinite connectivity ,
a dynamical mean field theory applies which reduces the model to the study of a single - site self - consistent problem @xcite , as detailed in sec .
[ reductionunsite ] .
however this single site model is still a complicated interacting problem . for the sake of simplicity
, the lattice will be taken to be a bethe lattice ( no essential physics is lost in this assumption ) .
* @xmath50 , in which the single site problem becomes tractable . in the absence of a random exchange
, this limit yields the familiar brinkman - rice description of a doped mott insulator @xcite .
the scaling in @xmath47 and @xmath25 in ( [ defjij ] ) and ( [ deftij ] ) are chosen such that this double limit gives non trivial results .
alternatively , one could consider ( as in @xcite ) this model on a fully connected lattice of @xmath51 sites , with _
random _ hopping amplitudes : @xmath52 with @xmath53 , @xmath54 .
this leads to precisely the same equations for single particle green s functions as the @xmath55 bethe lattice @xcite .
we shall use a decomposition of the physical electron operator into a spin - carrying fermion @xmath56 and a slave boson @xmath57 : @xmath58 . the local constraint ( [ contrainte1 ] ) becomes : @xmath59 with this decomposition the hamiltonian ( [ ham1 ] ) can be rewritten as : @xmath60 and the @xmath61 components of the @xmath26 spin operators @xmath62 read : @xmath63 in this section , we explain how the large connectivity limit @xmath64 reduces the problem to the study of a single - site model supplemented by a self - consistency condition .
first we use a path integral representation of the partition function @xmath65 and introduce a lagrange multiplier field @xmath66 on each site in order to handle the constraint ( [ defconstraint ] ) .
we then introduce @xmath67 replicas of the fields ( @xmath68 ) in order to express @xmath69 and average over the disorder .
the action associated with @xmath70 reads : @xmath71 - { 2t\over\sqrt z m } \sum_{<i , j>,\alpha}\int_0^{\beta } d\tau f^{\dagger a}_{i\alpha}(\tau ) b^a_i(\tau ) b^{\dagger a}_j(\tau ) f^a_{j\alpha}(\tau ) \\ & & - { j^2\over 2 zm}\int_0^{\beta}\!\!\!\int_0^{\beta } d\tau d\tau'\ , \sum_{<ij>}\sum_{1\leq\alpha,\beta,\gamma,\delta\leq m } \sum_{1\leq a , b\leq n } \ , s^{a}_{i\alpha \beta } ( \tau ) s^{a}_{j\beta \alpha } ( \tau ) s^{b}_{i\gamma \delta } ( \tau ' ) s^{b}_{j\delta \gamma } ( \tau ' ) \end{aligned}\ ] ] where the action @xmath72 is defined by @xmath73 & \equiv & \int_0^{\beta } d\tau \left ( b^{\dagger } ( \tau)\partial_\tau b(\tau ) + \sum_{\alpha } f^{\dagger}_{\alpha}(\tau)(\partial_\tau-\mu ) f_{\alpha}(\tau ) \right ) \\ & & + i \intbeta d\tau \lambda(\tau ) \left ( \sum_{\alpha } f^{\dagger } _ { \alpha}(\tau ) f_{\alpha}(\tau ) + b^{\dagger } ( \tau ) b(\tau ) - \frac{m}{2 } \right ) \end{aligned}\ ] ] following the `` cavity method '' ( reviewed in @xcite ) , a site of the lattice is singled out , and a trace is performed over all degrees of freedom at the other sites ( concentrating on phases without translational symmetry breaking , so that all sites are equivalent ) . in the @xmath49 limit ,
this can be performed explicitly , and the problem reduces to a single - site effective action which reads : @xmath74 - { j^2 \over 2 m } \sum_{a , b,\alpha , \beta , \gamma , \delta } \intbeta\!\!\!\intbeta d\tau d \tau'\ , s^{a}_{\alpha \beta } ( \tau ) r^{ab}_{\beta \alpha \delta
\gamma } ( \tau-\tau ' ) s^{b}_{\gamma \delta } ( \tau ' ) \\ & & + \left ( \frac{2t}{m } \right)^{2 } \sum_{a,\alpha}\intbeta\!\!\!\intbeta d\tau d \tau'\ , f^{\dagger a}_{\alpha}(\tau ) b^a(\tau ) c^{aa}_{\alpha\alpha}(\tau-\tau ' ) b^{\dagger a}(\tau ' ) f^a_{\alpha}(\tau')\end{aligned}\ ] ] this effective action is supplemented by a self - consistency condition which constrains @xmath75 and @xmath76 to coincide with the local electron green s function and spin correlation function respectively : @xmath77 in each of these equations , the last equality expresses the fact that local correlation functions can be calculated using the single - site action @xmath78 itself .
the limit @xmath79 must eventually be taken in these equations .
we shall study the above self - consistent single - site problem in the large @xmath25 limit , focusing on the paramagnetic phase of the model . in this case , all the above correlators become replica diagonal ( @xmath80 , @xmath81 ) .
furthermore , we shall look for solutions in which the slave boson undergoes a bose condensation .
( solutions with an uncondensed boson when the bosons carry an additional channel index have been investigated by horbach and ruckenstein @xcite ) .
the solutions considered here can be found as a saddle - point of @xmath78 by setting : @xmath82 and looking for solutions in which both @xmath83 and the lagrange multiplier @xmath84 become static at the saddle - point : @xmath85 from the constraint eq.([defconstraint ] ) , the total number of electrons will be related to @xmath30 through : @xmath86 so that @xmath30 measures the number of holes doped into the system .
the saddle point equations then reduce to a non - linear integral equation for the fermion green s function @xmath87 , which reads ( with : @xmath88 the matsubara frequencies ) : [ eqbase ] @xmath89 @xmath90 and to the following relations , which determine the lagrange multiplier @xmath91 and the chemical potential @xmath92 for given values of the doping @xmath30 and the temperature ( given the couplings @xmath3 and @xmath4 ) : @xmath93 @xmath94 the derivation of these saddle point equations from @xmath78 is detailed in appendix [ apxspeq ] .
the _ local _ spin - spin correlation function is directly related to @xmath95 in the @xmath96 limit , as : @xmath97 in the following , we shall often consider the spectral functions associated with the single - particle green s function and the local spin - spin correlation : @xmath98
in this section , we study the nature of the metallic state as a function of the doping level @xmath30 .
let us first recall some of the properties of the spin - liquid insulating state found for @xmath99 , as obtained by sachdev and ye @xcite . in this case ,
our equations ( [ eqbase1]-[eqbase3 ] ) coincide with those of ref.@xcite .
note that eq.([eqbase4 ] ) decouples , being automatically satisfied at @xmath99 , and that particle - hole symmetry imposes @xmath100 .
a low - frequency analysis of the integral equation reveals that the @xmath24 green s function and spectral density have a @xmath101 singularity for @xmath102 .
more precisely @xcite , in the complex frequency plane as @xmath103 : @xmath104 this yields the following behaviour of the local dynamical susceptibility for @xmath105 : @xmath106 fig.[dosspinfluid ] displays a numerical calculation of @xmath107 and @xmath108 at zero doping ( in agreement with the one in ref.@xcite ) .
these results display the above low - frequency behaviour ( but we note that significant corrections to ( [ mflchi ] ) are already sizeable at rather low values of @xmath109 . )
hence the insulator at @xmath99 is a gapless quantum paramagnet ( spin liquid ) , with a rather large density of low - energy spin excitations .
remarkably , ( [ mflchi ] ) is of the same form than the `` marginal fermi liquid '' susceptibility proposed on phenomenological grounds by varma _
@xcite for the normal state of the cuprate superconductors . in the present context , the physical nature of these low - energy excitations is intimately connected to the fact that the exchange couplings @xmath36 are random in sign . in constructing the ground - state of the insulator ,
let us imagine that we first try to satisfy the bonds with the larger exchange constants .
when such a bond is antiferromagnetic , the two spins connected by it will form a non - degenerate singlet . for a ferromagnetic bond
however , the two spins will pair into a state of maximal possible spin ( the generalization to @xmath26 of a triplet state ) .
this state has a degeneracy , which actually becomes very large ( exponential in @xmath25 ) as @xmath25 becomes large . continuing the process in order to accommodate bonds with smaller strengths will tend to remove part of this degeneracy @xcite , but leaves behind a very large density of low - energy spin excitations .
these effects are clearly favored by the ( fermionic ) large-@xmath25 limit considered here , because of the high degeneracies of the `` triplet '' state and because the strength of quantum fluctuations in this limit precludes the appearance of long - range order ( e.g. spin - glass ) which would remove degeneracies in a different manner .
we believe however that this physics is not an _ artefact _ of the large@xmath110 limit . indeed , preliminary theoretical studies @xcite suggest that the local spin correlations near the quantum critical point associated with the @xmath24 transition into a metallic spin - glass phase could be similar to eq.([mflchi ] ) , with a related physics . finally , we note that the single - site action to which the model reduces at zero doping ( _ i.e _ eq .
( [ action2 ] ) with @xmath111 ) has some similarities with the multichannel kondo effect in the _ overscreened _ case . in the present context
however , the `` bath '' seen by the spin is not due to an electronic conduction band , but generated by all the other spins in the lattice .
the spin correlations of both the bath and the spin adjust to the self - consistent long - time behaviour : @xmath112 similar to that of the @xmath26 kondo model with @xmath113 channels @xcite ( 2-channel model in the @xmath38 case ) .
the first question we would like to address is whether the `` marginal fermi liquid '' spin dynamics survives the introduction of charge carriers .
as we shall demonstrate , this depends on the temperature range considered ( fig.[crossovers ] ) .
at low temperature , below some -possibly very low- coherence scale @xmath114 , it turns out that a fermi liquid is recovered .
this is easily seen from a low - frequency analysis of the integral equation for @xmath95 at zero - temperature . at zero doping ,
the green s function and self - energy behave at low - frequency as : @xmath115 . when inserted in eq.([eqbase1 ] ) , this controls the leading low - frequency behaviour of both the r.h.s and l.h.s of the equation taken at @xmath99 , which match each other .
however , for @xmath116 , the term @xmath117 would introduce a @xmath101 singularity and prevent this matching from taking place : this indicates that the low - frequency behaviour of the zero - temperature green s function for arbitrary small doping is no longer @xmath101 . in this respect ,
an infinitesimal doping is a _ singular perturbation _ of the above equations .
this observation directly yields an estimate of the coherence scale @xmath114 such that @xmath118 is recovered for @xmath119 .
indeed , the term @xmath117 becomes comparable to @xmath120 in this regime ( thus providing a cut - off to the singular behaviour ) when @xmath121 .
hence , in the low - doping regime : @xmath122 where @xmath123 will be precised below .
( in the following we shall take eq.([scalesmall ] ) as defining @xmath114 in the low - doping regime , with no additional prefactors ) . in the high - doping regime on the other hand
( or when @xmath124 ) , one should consider first the limit of a vanishing magnetic exchange @xmath125 . in this limit , the usual slave - boson ( large @xmath25 ) description of a doped mott insulator is recovered @xcite .
setting @xmath126 in the equations above yields a semi - circular spectral density : @xmath127 where @xmath128 is given by : @xmath129 the original bandwith @xmath130 of the non - interacting case has been reduced by a factor @xmath30 , and the usual brinkman - rice result for the coherence scale is recovered : @xmath131 turning on @xmath3 as a perturbation from this starting point does not affect the leading low - frequency behaviour of the self - energy , but does lead to a scattering rate @xmath132 characteristic of a fermi liquid ( in contrast the @xmath125 model has infinite quasiparticle lifetime in the large-@xmath25 limit ) . from eqs.([scalelarge],[scalesmall ] ) , it is clear that when the magnetic scattering is strong ( @xmath133 ) , regime ( [ scalesmall ] ) always applies , while for weaker scattering ( @xmath134 ) a crossover between the two regimes is found at a characteristic doping : @xmath135 we thus observe that below some characteristic doping the low - energy coherence scale is strongly affected by the magnetic scattering . when the exchange is large or for doping smaller that @xmath2 , the actual coherence scale @xmath136 is much smaller than the `` naive '' coherence scale @xmath31 ( which holds in the absence of magnetic correlations ) . here
we find @xmath137 and @xmath138 .
this is one of the crucial physical conclusions of this paper .
a numerical solution of the saddle - point equations provide a clear evidence for these two regimes .
the numerical procedure that we have used is explained in appendix [ appendicenum ] .
( [ spectralfunctions ] ) displays the @xmath24 spectral function @xmath139 for three values of @xmath3 .
when @xmath3 is very small , the spectral function is very close to the semi - circular shape ( [ dosj=0 ] ) , while for a larger @xmath3 the @xmath140 divergence is observed over a large frequency range @xmath141 but is cutoff for @xmath142 so that @xmath143 is finite .
anticipating on the results of sec .
[ propunepart ] , we observe that the value of @xmath144 is actually independent of @xmath3 as a consequence of the luttinger theorem . indeed
, the following relation can be established at zero - temperature : @xmath145 where @xmath146 is the non - interacting value of the chemical potential for the tight binding model on the @xmath55 bethe lattice .
this implies : @xmath147 for all values of @xmath3 . at very small doping , a scaling analysis of the saddle - point equations
can be performed in order to characterize more precisely the crossover between the low - frequency and high frequency regimes at @xmath24 . as we now show
, the spectral function ( and the green s function itself ) obeys a scaling form : @xmath148 in order to derive the integral equation satisfied by the scaling function @xmath149 , we rewrite eq.([eqbase1 ] ) at @xmath24 ( using ( [ thmluttinger ] ) ) as : @xmath150 at low doping and low - frequency @xmath95 is of order @xmath151 and @xmath152 is of order @xmath30 .
hence , rescaling frequencies by the coherence scale @xmath153 , we see that the first two terms in the r.h.s of ( [ lowdoping1 ] ) can be neglected . analytically continuing to _ real _ time @xmath4 and frequency @xmath154 , and denoting by @xmath155 the real - frequency @xmath24 green s function ( with the usual feynman prescription ) ,
we define a scaling function @xmath156 associated with @xmath157 by the hilbert transform : @xmath158 we finally obtain from ( [ lowdoping1 ] ) an integral equation satisfied by @xmath159 ( and thus by @xmath149 ) which no longer contains dimensional parameters : @xmath160 ( as explained in appendix .
[ appendicenum ] , a sign change occurs in the expression of the self - energy at @xmath24 ) the universal scaling function @xmath161 can be obtained by solving numerically eq .
( [ equationgf ] ) , and the result is displayed in fig .
[ fntechelletnuldopage ] .
the asymptotic behaviours of @xmath162 for large and small @xmath163 can be obtained analytically and read : @xmath164 where @xmath165 and @xmath166 are two constants .
the low - frequency behaviour reflects the fermi - liquid nature of the low - energy excitation spectrum , while the @xmath101 behaviour characteristic of the undoped spin - liquid is recovered for @xmath167 .
in this section , we focus on the one - particle green s function for the physical electron , which is related to that of the auxiliary fermion by : @xmath168 hence : @xmath169 in this expression , @xmath170 stands for the one - particle energies of a non - interacting tight - binding model on the bethe lattice with hopping @xmath171 between nearest neighbour sites @xcite .
the distribution of these single - particle energies is the semi - circular density of states @xmath172 defined in ( [ defd ] ) . from the large - frequency behaviour of eq.([greenphysique ] )
, we see that the physical electron spectral density in the @xmath96 limit is normalized as @xmath173 ( in contrast @xmath174 ) .
this is expected from the fact that the constraint ( [ defconstraint ] ) on the hilbert space yields a normalisation ( for arbitrary @xmath25 ) @xmath175 ( note that this yields @xmath176 for @xmath45 , as expected for the @xmath28 hubbard model ) .
since our normalisation of the hopping is @xmath177 , the non - interacting conduction electron green s function reads : @xmath178 thus , the physical electron self - energy reads : @xmath179 we observe that it depends solely on frequency , as is generally the case in the limit of large dimensionality @xcite .
we first consider the location of the fermi surface for both the non - interacting and interacting problems , _
i.e _ look for the poles of the electron green s function .
in the non - interacting case , we relate the chemical potential at @xmath24 to the number of particles @xmath180 and find : @xmath181 where the function @xmath146 is defined by the relation : @xmath182 hence the non - interacting fermi surface corresponding to a doping @xmath30 is defined by @xmath183 . in the interacting case ,
we see from eq.([greenphysique ] ) that the fermi surface is located at @xmath184 . in the absence of magnetic scattering ( @xmath125 ) , it can be shown by an explicit calculation @xcite from the saddle - point equations that the r.h.s of this equation is just @xmath146 and thus that the fermi surface is unchanged in the presence of the constraint . when @xmath185 , such an explicit calculation is not possible , since the saddle point equations are coupled non - linear integral equations . however , a proof of luttinger theorem can still be given using the fact that a luttinger - ward functional exists for this problem and is known in explicit form in the large-@xmath25 limit , as detailed in appendix [ lutt ] .
the conclusion of this analysis is that the volume of the fermi surface corresponds to @xmath186 particles per spin flavour and that the zero temperature , zero frequency self energy must obey : @xmath187 we now consider the weight and dispersion of the quasiparticles , that can be read off from eqs.([greenphysique],[sigmac ] ) by expanding around the fermi surface .
we define a renormalisation factor for the auxiliary fermions as : @xmath188 so that the physical electron quasiparticle residue reads : @xmath189 from the low - frequency analysis of the preceding section and the corresponding estimates of the coherence scale , we expect @xmath190 to be of order @xmath191 and thus : @xmath192 in fig.[zfntdelta ] , we display the result of a numerical calculation of @xmath190 as a function of doping , for three values of @xmath193 . these results entirely confirm the above expectations .
we have checked that at small doping @xmath194 scales proportionally to @xmath195 with a universal prefactor . from eq.([greenphysique ] ) , we see that the quasiparticles have a dispersion characterized by an effective hopping @xmath196 ( @xmath197 ) .
hence the effective mass diverges as the mott insulator is reached ( as @xmath198 ) .
the reason for this divergence is the large ( extensive @xcite ) entropy of the insulating spin - liquid ground - state .
this entropy must be released at a temperature of the order of the coherence scale @xmath7 in the doped system .
hence , integrating the specific heat ratio @xmath199 between @xmath24 and @xmath200 leads to @xmath201 , which is the result found above .
this divergence of @xmath202 as @xmath203 is clearly an artefact of the large-@xmath25 and large-@xmath204 limit .
the residual ground - state entropy of the spin - liquid phase should not survive a more realistic treatment of this phase ( whether this happens while preserving @xmath205 in this phase is an open problem at this moment ) .
furthermore , our model does not include a uniform antiferromagnetic exchange constant superimposed on the random part . including this coupling will help locking the spins into singlets and cutoff the divergence of the effective mass ( for a large-@xmath25 treatment of this point , see e.g ref.@xcite ) .
finally , we discuss the shape of the conduction electron spectral density @xmath206 for a _
fixed _ value of the energy @xmath170 as a function of frequency , as relevant for photoemission experiments : @xmath207 numerical results for this quantity are displayed on fig .
[ figphoto ] .
this function is peaked at a frequency @xmath208 with a height of order @xmath209 ( at @xmath24 ) .
moving away from this quasiparticle peak , @xmath210 has the characteristic @xmath211 decay of a fermi liquid only in the limited frequency range @xmath212 , followed ( for @xmath0 ) by a much slower @xmath101 tail corresponding to the spin liquid regime in the frequency range @xmath213 .
( we note that this non - fermi liquid tail is absent in the high - doping regime ) .
these two regimes are clearly apparent on fig .
[ figphoto ] .
if the resolution of a photoemission experiment is not significantly smaller than @xmath7 , the peak will be smeared into a broad feature , and the measured signal will be dominated by the slowly decaying tail .
furthermore , as shown in the next section , temperature has a large effect on the peak , whose height decreases as @xmath214 in the temperature range @xmath215 . the metal - insulator transition at @xmath24 as @xmath216 is a quantum critical point .
the associated crossover regimes at finite temperature can be easily deduced by comparing the coherence scale @xmath114 to the magnetic exchange @xmath3 and to the temperature .
this analysis yields three regimes , as depicted on fig.[crossovers ] : * for @xmath217 , the doped holes form a fermi liquid .
the low - energy degrees of freedom are the fermionic quasiparticles described by the auxiliary fermions @xmath218 , which behave in a coherent manner since their inverse lifetime vanishes at low - frequency as @xmath219 in this regime .
* at low doping @xmath0 , an intermediate temperature regime exists , defined by : @xmath215 . in this regime ,
coherent quasiparticles no longer exist ( as shown below , @xmath220 ) , but the spin degrees of freedom are not free local moments since the temperature is smaller than the magnetic exchange .
hence , the spins behave in this regime as in a spin liquid , with a _ marginal fermi liquid form for the local spin response function_. as shown below , this regime corresponds to the so - called `` quantum - critical '' regime associated with the quantum critical point at @xmath221 . in this regime ,
the low - energy scale @xmath136 drops out from response functions , which obey universal scaling properties as a function of the ratio @xmath222 . *
finally , a high - temperature regime applies , defined by @xmath223 ( for @xmath0 ) or @xmath224 ( for @xmath1 ) in which both spin and charge are incoherent and essentially free .
we note that if @xmath134 and the doping is larger than @xmath2 , the system goes directly from a fermi liquid to this high - temperature regime as temperature is increased , without an intermediate marginal fermi liquid regime .
this qualitative analysis can be established on firmer grounds by generalizing the low - doping scaling analysis of sec.[lowfreqanalysis ] to finite temperature .
assuming that the coherence scale is small as compared to both @xmath3 and the hopping @xmath4 ( i.e that @xmath0 ) , and that @xmath225 , the spectral function takes the following scaling form , generalizing eq.([defphif ] ) : @xmath226 in the following , @xmath227 and @xmath228 stand for @xmath229 and @xmath230 , respectively .
eq.([rhoscalefinitet ] ) yields for the green s function @xmath231 ( with @xmath232 ) .
we assume that the self energy also scales as @xmath233 . from the saddle - point equation ( [ eqbase ] )
we deduce the following equations : @xmath234 in this expression , @xmath235 and @xmath236 are the fermi and bose factor : @xmath237 . with a kramers - kronig transformation one can deduce @xmath238 and @xmath239 from @xmath240 and @xmath241 . from the equation for @xmath242
, we have that @xmath243 where @xmath56 is some scaling function which can in principle be calculated from ( [ fullscaling ] ) .
the function @xmath56 vanishes at small argument ( @xmath244 ) due to luttinger s theorem , and at large argument ( @xmath245 ) due to the auxiliary fermion particle hole symmetry of the undoped model .
we now discuss the solution of this scaled integral equation , and the form taken by @xmath246 in the various regimes described above .
\i ) _ fermi liquid regime _ : @xmath247 . at zero temperature
, the scaling function @xmath246 reduces to that in eq.([defphif ] ) : @xmath248 we can also consider the limit of low - frequency and temperature @xmath249 but with an arbitrary ratio @xmath12 . in this limit , the self - energy term is negligible altogether in eq.([fullscaling ] ) , and one gets simply @xmath250 i.e : @xmath251 note that the r.h.s could a priori be a function of the ratio @xmath12 , but is actually a constant ( as is generically the case in a fermi liquid ) . from this
, we can deduce a scaling form of the scattering rate in the same regime .
indeed ( [ philf ] ) corresponds to the imaginary time green s function : @xmath252 hence , in this limit , the self - energy takes the form : @xmath253 which can be fourier transformed to yield : @xmath254 \ii ) _ spin - liquid regime : _
@xmath255 . in this quantum critical regime
the energy scale @xmath7 drops out from the problem and the spectral density and response functions become functions of the ratio @xmath12 only .
indeed , the scaling function @xmath256 takes the form @xmath257 in the limit @xmath258 . in order to find @xmath259 in explicit form , we divide both side of ( [ fullscaling ] ) by @xmath260 and take the limit @xmath261
, @xmath262 fixed .
then the first term of the r.h.s vanishes and we are left with a scaled equation for @xmath259 in which all dependence on @xmath7 has disappeared .
remarkably , this integral equation can be solved in closed form and yields : @xmath263 some details are provided in appendix [ appfonctionsechelle ] .
this scaling function describes how the @xmath101 singularity associated with the low - energy excitations of the spin - liquid is cutoff ( by the temperature ) at frequencies @xmath264 so that the spectral density is of order @xmath265 at @xmath266 .
( note that if the limit @xmath267 is performed while keeping @xmath227 fixed in ( [ fullscaling ] ) , the same result is obtained as when the limit is taken with @xmath268 .
hence there is no additional crossover in the frequency dependence of the response functions below @xmath269 in this regime ) .
eq.([scalingformrho ] ) corresponds to the following scaling form for the imaginary time green s function : @xmath270 remarkably , ( [ gechelle ] ) has the form which would hold in a model having _ conformal invariance _
, for example a quantum impurity model of a spin interacting with a structureless bath of conduction electrons . in that case , a conformal mapping from the @xmath24 half plane @xmath271 to the finite - temperature strip @xmath272 can be used to show @xcite that if the green s function decays as @xmath273 at @xmath24 , then it takes a scaling form given by ( [ gechelle ] ) at finite temperature ( lower than a high - energy cutoff ) . in the present case ,
the original model is an infinite connectivity _ lattice model _ which does not _ a priori _ satisfy conformal invariance .
it does map onto a single - site quantum impurity model , but with an additional self - consistency condition .
this means that the effective bath for the local spin is given by the local spin - spin correlator itself , and thus does have non - trivial structure at low - energy .
however , this structure appears only as a subdominant correction to the leading low - frequency behaviour @xmath274 . for this reason , our effective single - site model does obey conformal invariance properties in the low - energy limit , which explains the result above .
this remark actually applies in a broader context than the specific model considered here , as will be discussed in more details elsewhere .
let us also consider the scattering rate in this regime , which is obtained by fourier transforming the imaginary time self - energy : @xmath275 which yields : @xmath276 we have calculated numerically the real - frequency , finite - temperature green s function by following the method described in appendix [ appendicenum ] . on fig .
[ dosfnttemp ] , we display results for the spectral density for various temperatures for @xmath277 at a doping of @xmath278 .
these values correspond to a low energy coherence scale @xmath279 .
the crossover from the fermi - liquid regime at low temperature into the quantum critical regime at intermediate temperatures is clearly visible ( in particular , the peak height can be checked to decrease as @xmath214 ) .
note also that @xmath280 remains approximately centered at @xmath281 until @xmath282 and shifts rapidly away from @xmath283 for @xmath284 . in the inset , we also display the thermal scaling function associated with @xmath285 , eq .
( [ scalingformrho ] ) . in this section
, we describe the behaviour of the local spin dynamics in the various temperature regimes . in the large-@xmath25 limit
, the local spin correlation function is given by : @xmath286 in fig .
[ kilocfntt ] , we display @xmath287 for various temperatures and the same choice of parameters as in fig .
[ dosfnttemp ] . in the low doping regime , @xmath288 , @xmath289
, @xmath290 obeys a scaling form which follows from the convolution of ( [ rhoscalefinitet ] ) : @xmath291 let us discuss the limiting forms of this expression as @xmath292 or @xmath255 . * _ at zero temperature , _ @xmath293 has a shape which resembles the undoped spin - liquid case ( fig . [ dosspinfluid ] ) for frequencies @xmath294 . at lower frequency
, the fermi liquid behaviour @xmath295 is recovered .
this results in a peak with a height of order @xmath296 at @xmath24 .
this crossover can be described by a scaling function : @xmath297 where @xmath298 can be obtained by convoluting @xmath149 with itself resulting in the asymptotic behaviours : @xmath299 this can be used to estimate the behaviour of the static local susceptibility at low doping : @xmath300 . in this
integral , the region @xmath301 ( corresponding to spin liquid excitations ) gives the dominant contribution , leading to the logarithmic behaviour for @xmath302 : @xmath303 in contrast , as detailed in appendix [ susceptibilite ] , the _ uniform _ static susceptibility @xmath304 is a constant of order @xmath296 , with no divergence at small doping .
* in the _ spin - liquid regime _
@xmath255 , @xmath305 becomes a function of @xmath12 .
the corresponding scaling function is remarkably simple : from eq .
( [ gechelle ] ) we have @xmath306 which yields : @xmath307 this behaves exactly as the spin response function postulated in the marginal fermi liquid phenomenology @xcite ( @xmath12 for @xmath308 , @xmath309 for @xmath310 ) .
we finally use these results to compute the temperature dependence of the nmr relaxation rate : @xmath311 expanding the scaling form ( [ scalformgenki ] ) to linear order in @xmath154 ( and noting that @xmath312 because @xmath313 is odd ) , we get for @xmath314 : @xmath315 ( with @xmath316 ) . in fig .
[ nmrrate ] , we plot this universal scaling function .
we have also checked the data collapse of our numerical results on this function .
limiting forms are easily obtained from eq . ( [ crudeestki ] ) and ( [ mflchi ] ) : * @xmath247 : @xmath317 hence @xmath318 .
we find a korringa law ( as expected from a fermi liquid ) but with a _ very strong doping dependence_. we also note that in contrast to a non interacting fermi gas , @xmath319 , @xmath320 and @xmath321 obey quite different behaviour as a function of doping . in particular the so - called `` korringa ratio ''
@xmath322 is very large at low doping .
* @xmath323 : @xmath324 , hence @xmath325 as expected in a _ marginal fermi liquid_. we note that @xmath326 is doping - independent in this quantum - critical regime .
this is because the scale @xmath136 no longer appears explicitly . in the limit of large connectivity
, the current - current correlation function has no vertex corrections , due to the odd parity of the current ( see e.g. @xcite ) .
hence the frequency dependent conductivity is given by : @xmath327 where @xmath328 is the single - electron spectral density defined in eq.([defrhoc ] ) .
this expression yields the conductivity in units of @xmath329 where @xmath330 is the lattice spacing and some numerical prefactors have been dropped ( we shall also ignore the prefactor @xmath25 in @xmath331 ) .
we first discuss the behaviour of the dc - conductivity : @xmath332 _ i ) in the fermi liquid regime _
@xmath333 , we have from the behaviour ( [ scattfl ] ) of the scattering rate : @xmath334 and @xmath335 . making the change of variables @xmath336
, we see that the integral over @xmath337 in @xmath338 diverges as @xmath339 .
hence , we find in this regime the expected fermi - liquid behaviour of the resistivity @xmath340 : @xmath341 _ ii ) in the spin liquid regime _
@xmath342 at low doping , @xmath343 is of order @xmath344 ( times a scaling function of @xmath222 ) .
this must be compared to @xmath345 in the denominator of @xmath346 . since @xmath347
, we see that @xmath348 always dominates over @xmath349 which can thus be neglected .
hence one can replace @xmath350 by the _
spectral function @xmath351 . in other words
the limit @xmath352 must be taken before the low temperature limit in this quantum critical regime . using the thermal scaling function eq .
( [ scalingformrho ] ) , we obtain : @xmath353 the integral can calculated explicitly using @xmath354 ( @xcite , eq .
6.412 ) , we finally find : @xmath355 hence the resistivity turns out to have a linear behaviour as a function of temperature in the spin - liquid regime , again as in the marginal fermi liquid phenomenology .
this is rather remarkable in view of the fact that the _ single - particle _ scattering rate behaves as @xmath356 in this regime . as further discussed in the conclusion ,
this is characteristic of a regime of incoherent transport in which the transport scattering rate can not be naively related to the single - particle lifetime .
furthermore , we note that the @xmath14 behaviour of the self - energy is a crucial ingredient in producing a @xmath357-linear resistivity . with a different power law ( @xmath358 )
, the resistivity would behave as @xmath359 in this incoherent regime .
the crossover from @xmath360 to @xmath357 in the resistivity can be captured in a more precise manner in a universal scaling function : @xmath361 we have determined numerically the function @xmath362 , which is depicted in fig .
[ resistivity ] .
we observe that it is linear over a wide temperature range ( with a slope in agreement with ( [ resisspinfluid ] ) ) .
we now turn to the analysis of the frequency - dependent conductivity .
\i ) _ in the fermi liquid regime , _ the conductivity takes the form at @xmath24 : @xmath363 where @xmath128 is the weight of the drude peak and @xmath364 as @xmath365 .
the drude peak is easier to capture by a finite temperature analysis : the delta function is regularised by @xmath357 in the form @xmath366 . performing a low - temperature , low - frequency analysis of ( [ conducbase ] ) leads to the estimation @xmath367 at small doping .
a closed formula can be given for @xmath368 as ( truncated ) convolution of the scaling function @xmath149 .
a low - frequency analysis then shows that @xmath369 , while @xmath370 .
_ ii ) in the regime @xmath371 _ , we have from ( [ conducbase ] ) and ( [ scalingformrho ] ) the scaling form : @xmath372\end{aligned}\ ] ] where @xmath373 . from ( [ defvarphisig ] ) , @xmath374 and
thus we have in this spin - liquid regime : @xmath375 moreover , using the kramers kronig relation @xmath376 we find in the same regime for @xmath377 : @xmath378 hence defining an optical scattering rate from an effective drude formula : @xmath379 we find @xmath380 .
we have also calculated @xmath381 numerically , following the method explained in appendix [ appendicenum ] .
numerical results are displayed for various temperatures on fig.[resigom ] and are in agreement with the previous analysis .
in this paper , we have solved a model of a doped spin - fluid with strong frustration on the exchange constants @xmath36 .
the undoped model is an @xmath26 quantum heisenberg model with random exchange , previously studied by sachdev and ye @xcite in the limit of large-@xmath25 and infinite connectivity .
these authors found that , in this limit , quantum fluctuations are so strong that no spin glass phase forms @xcite .
instead , a gapless spin liquid is found with local spin dynamics identical to the marginal fermi liquid phenomenology @xcite .
we generalised this result to finite temperature and found that the local spin response function displays @xmath12 scaling : @xmath382 ( for @xmath383 ) .
doping this mott insulating phase with holes , we found that a characteristic doping @xmath2 appears separating two quite different doping regimes . in the high - doping regime @xmath1 ,
magnetic effects are weak and a brinkman - rice fermi - liquid description is valid , with a rather large coherence scale of order @xmath5 . in the low doping regime however , the interplay between local coherence and magnetic effects gives rise to a coherence scale @xmath384 , which can be very low . at low temperature @xmath217 , fermi liquid behaviour is recovered , but an incoherent regime is found in a rather wide regime of temperature @xmath8 in which physical properties strongly deviate from fermi liquid theory .
this regime corresponds to the `` quantum critical regime '' associated with the metal - insulator transition which in this model happens at @xmath385 .
we found that both transport properties and response functions in this incoherent regime behave as in the marginal fermi liquid phenomenology , namely @xmath386 , @xmath387 , @xmath388 , and @xmath389 .
remarkably , _ single - particle _ properties deviate much more strongly from fermi liquid theory , with a single - particle scattering rate behaving as @xmath390 ( or @xmath356 ) , in contrast to the @xmath391 behaviour postulated in the marginal fermi liquid phenomenology .
these behaviour result from the solution of the large-@xmath25 saddle point equations , which also yields explicit expressions for the scaling functions of @xmath229 and @xmath12 describing the crossover of the various physical quantities between the fermi liquid and the non - fermi liquid regime .
we also note that in the large-@xmath25 limit , response functions can be calculated from the _ interacting single - particle _ green s function .
hence the behaviour of @xmath392 and of @xmath393 are intimately related .
in contrast , in the marginal fermi liquid phenomenology , the behaviour of @xmath394 is related to a priori unknown higher order vertex functions . in this sense
, the present model yields a solution to the problem of internal consistency of the marginal fermi liquid ansatz , resulting in a more singular form of the single - particle green s function .
we also note that numerical studies of the doped two - dimensional @xmath35 model with uniform antiferromagnetic @xmath3 by imada and coworkers @xcite have some intriguing similarities with the results of the present work .
specifically , a drude weight and coherence temperature scaling as @xmath395 are also found .
the specific heat coefficient is found to scale as @xmath396 in this case , in contrast to the present work .
the reason for this difference is the existence of a residual entropy in the undoped spin - liquid phase of our model .
however , the temperature dependence of the specific heat at the critical point is found to be @xmath356 in both cases .
finally , we briefly discuss the possible instabilities of the metallic paramagnetic phase discussed in this paper .
it can actually be checked that for a given @xmath193 , a low temperature and low doping regime exists in which an instability to phase separation is found , signaled by a negative compressibility .
this is quite easily explained on physical basis for a given realization of the exchange couplings : the holes will tend to cluster in regions with ferromagnetic bonds in order to maximize kinetic energy .
a proper treatment of this phase separated regime should take into account longer range coulomb repulsion . in the infinite connectivity limit ,
an additional term @xmath397 in the hamiltonian reduces to a hartree shift @xmath398 of the chemical potential ( thus the compressibility reads : @xmath399 ) , so that the phase separation boundary can be continuously tuned as a function of @xmath400 . in future work , we are planning to consider other possible instabilities of this model .
the issue of spin glass ordering @xcite does not arise for the @xmath401 fermionic representation considered in this paper @xcite , but spin glass phases are indeed present for @xmath402 for bosonic representations with high enough spin @xcite . even in the fermionic case ,
first order corrections in @xmath403 are likely to restore a regime of spin glass ordering .
finally , an open issue is that of possible pairing instabilities of the metallic phase towards a superconducting state . in this section
, we would like to present arguments suggesting that the problem studied in this paper may be relevant for the understanding of some of the striking aspects of the normal state of cuprate superconductors .
the line of arguments relies on three sets of experimental observations : * the experiments reported in @xcite , in which a @xmath404 magnetic field is used to suppress superconductivity strongly suggest that the ground - state of is actually an _ insulator _ , up to sr - doping of about @xmath405 corresponding to the highest @xmath406 .
this is true even in samples having large values of @xmath407 , making weak localization effects an unlikely explanation of the logarithmic upturn of _ both _ @xmath408 and @xmath331 observed at low temperature . insulating behaviour
is no longer found in overdoped samples .
* at very low doping in the compounds , a low temperature spin glass phase is found for @xmath409 @xcite , in agreement with theoretical arguments @xcite suggesting that localized holes induce locally a strong frustration in the magnetic exchange .
this localization of the carriers induces a strong upturn of @xmath408 at low temperature in these samples , first in a logarithmic manner followed by an activated behaviour . nevertheless , the high - temperature behaviour of the resistivity in these samples is quite similar to that found close to optimal doping . *
inelastic neutron scattering reveals peculiar low - energy spin excitations for all underdoped samples , quite different in nature from spin waves @xcite-@xcite . for very low doping
, these excitations occur in a remarkably low energy range , on the scale of @xmath410 , distinctly smaller than @xmath17 . in a restricted range of frequency and temperature ,
the energy scale for these excitations is actually set by the temperature itself and @xmath12 scaling applies @xcite these excitations , which are present in a wide range of temperature ( much above the freezing transition mentioned above ) and in the whole underdoped regime , correspond to a _
slower spin dynamics _ than in a fermi liquid , as is also clear from the non - korringa behaviour of the copper nmr relaxation time .
similar observations have been made in the compounds @xcite .
this is particularly clear when a small amount of @xmath411 substitution is used to suppress superconductivity @xcite ( we note that this simultaneously opens up again a region of glassy behaviour at low temperature for a rather wide range of oxygen content @xcite ) . in our view , these observations suggest that , in the absence of superconductivity , a @xmath24 metal - insulator transition occurs at some critical value of the doping @xmath412
. this transition might be rather close to optimal doping in @xcite . for @xmath413 ,
the incipient ground - state is a fermi liquid , corresponding to the overdoped regime . for @xmath414 ,
the ground state is a mott - anderson insulator in which holes are localised at @xmath24 . at
very small @xmath415 ( @xmath416 ) , the mechanism for this hole localization has been studied in ref .
@xcite and involves both the freezing of hole motion due to the antiferromagnetic spin background and impurity effects .
this localisation induces strong frustration in the local exchange , in agreement with the arguments of ref.@xcite . as a result
, this insulator will have a glassy nature at @xmath24 for low doping , as indeed found in .
beyond @xmath417 however , the onset of superconductivity has prevented up to now an investigation of the low - temperature properties of this incipient insulating ground - state and the origin of the observed localization is still an open problem .
it may be that the insulator looses its glassy character at some critical doping @xmath418 below @xmath419 , or that the two critical points actually coincide ( @xmath420 ) .
as temperature is raised , the holes become gradually mobile .
this quickly destroys the glassy ordering , leaving the system in a `` slushy '' state of mobile holes and spins . neutron scattering and
nmr experiments show that the spin dynamics in this regime is much _ slower _ than in a fermi liquid state , with local spin correlations decaying ( in some time range ) as @xmath421 ( corresponding to a high density of low - energy spin excitations @xmath422 in some frequency range ) .
we view the model studied in this paper as a simplified description of such a slushy state of spins and holes , valid in the high - temperature quantum critical regime associated with the transition at @xmath24 , @xmath423 ( or @xmath419 ) , as depicted schematically on fig.[phenomenofig ] .
indeed it is a model of a doped mott insulator with strong frustration , in which the effect of quantum disordering the glassy insulating state is mimicked by taking the large-@xmath25 limit .
fluctuations in the transverse components of the spin may actually be an essential ingredient in the disordering process , and this is precisely the effect which is emphasized in the large-@xmath25 limit and produces the high density of low - energy spin excitations . of course the present model
is highly simplified and is meant to retain only the interplay of mott localization with that of frustration in the magnetic exchange constants .
as such , it does not include several important physical aspects of the actual materials , most notably : * \i ) the fact that frustration is a consequence of hole localisation at low temperature @xcite ( in our model frustration is introduced by hand ) .
* \ii ) localization of carriers by disorder ( as a consequence of both i ) and ii ) , the metal - insulator transition occurs at zero - doping in our model ) . *
\iii ) the average antiferromagnetic component @xmath17 of the exchange has not been included ( in that sense we are dealing with a strong frustration limit @xmath424 ) . this could be corrected for by reintroducing @xmath17 in a mean - field manner , leading to : @xmath425 where @xmath426 is the fourier transform of the nearest neighbour connectivity matrix on the lattice .
we note that this formula produces a susceptibility peaked at the antiferromagnetic wave vector , with a correlation length of the order of the lattice spacing , while all the non trivial dynamics comes from local effects .
for these reasons , the present model is unable to address the question of the precise nature of the incipient insulating ground - state of underdoped materials ( or of the low - temperature pseudogap regime associated with it ) , even though the remarks made above point towards a phase separated regime .
various proposals have been made in the literature regarding this issue .
one of the most widely discussed is the `` stripes '' picture , in which there is phase separation between the doped holes and the spins into domain - wall like structures .
we note that as long as the holes remain confined in these structures , the mechanism of ref .
@xcite implies the existence of ferromagnetic bonds in the hole - rich region , as indeed found in numerical calculations @xcite . as temperature or doping
is raised , a melting transition of the stripe structure takes place , and the model of a `` spin - hole slush '' introduced here may become relevant in the associated quantum critical regime . keeping these caveats in mind , we comment on the comparison between our findings and some aspects of the normal state of cuprates in the regime depicted schematically on fig.[phenomenofig ] : * _ low - energy coherence scale _ + the present model yields a remarkable suppression of the low - energy coherence scale of a doped mott insulator in the presence of frustrating exchange couplings .
we find this scale to be of order @xmath427 instead of the naive ( brinkman - rice ) estimate @xmath5 .
we note that , with @xmath428 , and @xmath429 , this scale can be as low as a few hundred degrees .
if relevant for cuprates , this observation suggests that the normal state properties may well be associated , over an extended ( high- ) temperature regime , with _
incoherent behaviour _ characteristic of a quantum critical regime dominated by thermal effects .
we note however that the present model , as any model in which low - energy excitations are local in character , would lead to a large effective mass at low temperature , directly proportional to @xmath430 . in cuprates ,
additional physics sets in at lower temperature ( cf iii ) above ) which quench the corresponding entropy , leading to the experimentally observed moderate effective mass @xcite . *
_ photoemission _ + in the incoherent `` slushy '' regime @xmath431 , we find a single particle green s function decaying as @xmath101 ( and an associated single - particle lifetime @xmath390 ) , leading to a markedly non - fermi liquid tail of the photoemission intensity .
it is worth noting that precisely this form has been recently shown to provide a rather good fit to the high - frequency part of the photoemission lineshape above the pseudogap temperature in underdoped @xmath432 @xcite .
it has been recently argued that the @xmath15 behaviour also holds in the strong coupling limit of antiferromagnetic spin fluctuation theories @xcite . * _ resistivity and optical conductivity _ + we would like to emphasize again the mechanism which yields a linear resistivity in the incoherent regime of our model , starting from a single - particle self energy behaving as @xmath14 .
this holds when scattering is _ local _ and _ incoherent _ so that the effective quasiparticle bandwith ( dispersion ) can be neglected in comparison to lifetime effects . in this limit
, conductivity should be thought of in real space as a tunneling process between neighbouring lattice sites .
this mechanism has a higher degree of generality than the specific model considered in this paper , and should also apply to other models in which the same @xmath433 power - law behaviour of the self - energy holds , such as the model of ref.@xcite .
this model has been proposed in connection with the normal state properties of underdoped cuprates above the pseudo - gap temperature @xcite .
+ we note that the magnitude of the linear resistivity in this incoherent regime is larger or comparable to the mott limit ( @xmath434 ) , as is actually the case over a rather extended high temperature regime in underdoped @xcite and is a quite general feature of ` bad metals ' .
+ regarding optical conductivity , the form we have obtained is quite similar to the marginal fermi liquid one , which has been shown @xcite to provide a very good fit to the data of e.g. ref.@xcite . * _ neutron scattering _
+ neutron scattering experiments on non - superconducting @xcite and with @xmath21 @xcite have revealed spin excitations which are centered at the wavevector @xmath435 with a rather large momentum width .
the frequency dependence of these excitations display @xmath12 scaling and have been successfully fitted by scaling forms very similar to the one found in the present model @xcite . at higher @xmath436 concentration ,
one of the most notable feature of the neutron scattering results is the appearance of sharp peaks at incommensurate wavevectors .
it is likely however that these peaks only carry a small fraction of the total spin fluctuation intensity , as suggested in particular by comparison to nmr data . a broad , weakly @xmath437-dependent contribution most probably persists up to high temperature , carrying a large part of the total weight , and hard to distinguish from background noise in neutron experiments @xcite . in , suppression of superconductivity by @xmath438 doping allow to investigate the spin dynamics of the normal state down to low temperature @xcite .
apart from a very low temperature quasi - elastic peak ( associated with spin freezing into spin - glass like order ) , neutrons scattering results for @xmath439 reveal a strong enhancement of low - frequency spin fluctuations at low temperature , with a distinctively low energy scale and a strong temperature dependence down to very low temperature ( compatible with @xmath12 scaling in a limited range ) .
these features are qualitatively similar to the low - energy excitations found in the present model .
there is furthermore experimental evidence @xcite that these low - energy excitations are associated with the disordering of the spins by transverse fluctuations , as in our model .
we are most grateful to subir sachdev for useful correspondence and remarks .
we acknowledge useful discussions with our experimentalist colleagues : n. bontemps , g. boebinger , h. alloul and particularly with p. bourges , g. colin , y. sidis , s. petit , l.p .
regnault and g. aeppli on their neutron scattering results .
we also thank a. millis , c. varma and particularly e. abrahams for their suggestions and comments .
part of this work was completed during stays of a.g at the i.t.p , santa barbara ( partially supported by nsf grant phy94 - 07194 ) and at the rutgers university physics department .
in this appendix , some further details on the derivation of the saddle - point equations in the large-@xmath25 limit for the single site model defined by eqs.([action2],[defdetr ] ) are provided . in the following
, we will drop the index @xmath440 in @xmath441 . in
( [ defdetr ] ) the brackets denote the average with the action specified in subscript .
as the action @xmath442 is invariant under translations in imaginary time and under the action of @xmath26 ( the rotation invariance for @xmath45 ) , @xmath443 and @xmath444 take the following form : @xmath445 the quartic term in @xmath56 in ( [ action2 ] ) is decoupled using a bi - local field @xmath446 . using the expression of the spin operator eq .
( [ exprsf ] ) and the change of variable : @xmath447 the single - site partition function can be rewritten as : @xmath448 with the actions : @xmath449 in this expression , @xmath450 is defined by @xmath451\equiv \int \gd f^{\dagger } \gd f e^{- s_{00}[\phi , p,\lambda , f]}\ ] ] with : @xmath452 = \sum_{a } \int d\tau f^{\dagger a}(\tau)(\partial_\tau
-\mu ) f^a(\tau ) \\ \nonumber & & + i \int\ ! d\tau \sum_{a}\lambda^{a } ( \tau ) \left ( f^{\dagger a}(\tau ) f^a(\tau ) + \frac{\phi^{\dagger a}(\tau)\phi^a(\tau ) - 1}{2 } \right ) \\ \nonumber & & - j^2 \sum_{a , b } \int\!\!\!\int d\tau d \tau'\ , r^{ab}(\tau-\tau ' ) p^{ab}(\tau,\tau ' ) f^{\dagger b}(\tau ' ) f^{a}(\tau ) \\ & & + t^{2}\sum_{a}\int\!\!\!\int d\tau d \tau'\ , f^{\dagger a}(\tau ) \phi^a(\tau ) g^{a}(\tau-\tau ' ) \phi^{\dagger a}(\tau ' ) f^a(\tau')\end{aligned}\ ] ] in the limit @xmath96 , @xmath453 is controlled by a saddle point with respect to @xmath454 , @xmath455 and @xmath456 .
we assume a _ condensation of the boson _ : after the change of variable ( [ condensationboson ] ) @xmath457 is taken to be a finite constant at the saddle - point : @xmath458 and @xmath459 is static : @xmath460 . moreover , in this limit the correlation functions of @xmath56 are given by the average with the action @xmath461 taken for these values of @xmath462 .
as @xmath461 is quadratic in @xmath56 ( and the boson is condensed ) , the model is completely solved in this limit as soon as @xmath463 has been calculated .
the saddle point equations are given by the minimisation of @xmath464 with respect to @xmath454 , @xmath455 and @xmath456 respectively , which leads to : @xmath465 and finally gives eq.([eqbase1],[eqbase2],[eqbase3],[eqbase4 ] ) given in the text .
in order to find the volume of the fermi surface in the interacting system , we proceed along the lines of ref.@xcite and we observe that the auxiliary fermion self - energy can be obtained as the functional derivative of the following functional : @xmath466 the number of particles reads : @xmath467 and we use the identity : @xmath468 using the invariance of the luttinger - ward functional under a shift of all frequencies ( @xmath469 ) , the integral of the last term vanishes : @xmath470 and the integral of the first term can be explicitly calculated by transforming to retarded green s functions ( denoted by @xmath471 ) in the following manner : @xmath472\ ] ] as @xmath471 has no pole nor zeros in the upper half plane , the first integral can be closed there and vanishes .
hence we have : @xmath473 from eq.([lutt2 ] ) and the definition of @xmath152 , we finally obtain : @xmath474 which is the desired relation and insures that luttinger theorem holds in the presence of both the constraint and the magnetic scattering .
we also checked that this property is verified in our numerical calculations at @xmath24 .
in this appendix , we explain the main steps that we followed in solving numerically the saddle point equations eqs.([eqbase ] ) .
the calculation of the green s function is divided in two steps .
first a matsubara frequency / imaginary time algorithm is used in an iterative manner in order to find the value of the chemical potential @xmath92 and lagrange multiplier @xmath476 for a given doping @xmath30 , interaction strength @xmath193 and temperature .
convolutions are calculated using a fft algorithm and a simple iteration is used : starting from a given @xmath477 , a self - energy is obtained which is then reinjected into the expression for @xmath477 until a converged set @xmath478 is reached ( for given values of @xmath479 ) . a second routine uses the previous one to adjust @xmath480 and @xmath481 in order ( [ eqbase3],[eqbase4 ] ) to be satisfied .
once the imaginary - time green s function and values of @xmath479 have been found using this imaginary time algorithm , a different algorithm is used to obtain _ real frequency _
green s functions and spectral densities .
this is done in the following manner .
we consider a _
finite - temperature _ generalisation of the green s function with the feynman prescription : @xmath482 which reduces to the usual green s function @xmath483 at @xmath24 .
we now define : @xmath484 where @xmath4 is the real time . expressing both @xmath485 and @xmath486 as integrals of the spectral density with the spectral representation of @xmath487 ,
we obtain , after some calculations ( r is a superscript denoting _ retarded _ quantities ) : @xmath488 where @xmath489 and @xmath235 is the fermi factor . at @xmath24 ,
( [ relfondnum ] ) shows that @xmath490 simply coincides with @xmath491 , the usual self - energy at @xmath24 , with the feynman prescription and that eq .
( [ eqbase ] ) can be rewritten as , at @xmath24 : @xmath492 together with the equation corresponding to ( [ eqbase3],[eqbase4 ] ) .
note the change of the sign in front of @xmath493 .
this form was used in ( [ equationgf ] ) . from ( [ eqbasetnul ] )
one can write an algorithm for the computation of @xmath155 in the @xmath24 formalism , similar to the computation in imaginary time .
of course , as our large-@xmath25 limit performs a resummation of the perturbation theory , one can also obtain these equations with the diagrammatic rules , but these rules do not apply at finite temperature to the green s function @xmath477 in a systematic manner . a finite temperature , we use the following iterative algorithm : starting from @xmath487 , we get @xmath477 .
we then obtain @xmath490 by direct convolution in real time and the second term of the r.h.s of ( [ relfondnum ] ) by a double convolution of @xmath494 .
hence we obtain @xmath486 and go back to @xmath487 with eq .
( [ eqbase ] ) . as a starting point of the iteration , in order to speed up convergence , we take an analytic continuation of the solution in imaginary time , obtained by a standard pade approximation . note that the parameter @xmath92 and @xmath481 are fixed in this iteration on the real axis , since they have been calculated before in the matsubara formalism .
as soon as the green function has been obtained , some other quantities are straightforwardly calculated from the formula given in the text .
in particular , @xmath290 is expressed as a convolution .
however the computation of the uniform susceptibility @xmath495 ( considered in appendix [ susceptibilite ] ) is more involved : we solve eq.([suscepunif ] ) for @xmath496 by another iterative loop analogous to the previous ones .
the scaling function describing the effect of the doping at @xmath24 , is computed from ( [ equationgf ] ) by an iterative algorithm similar to the previous ones .
we also give some useful details on the numerical calculation of the frequency - dependent resistivity .
there , it is very convenient to integrate analytically over @xmath497 in ( [ conducbase ] ) using ( [ defrhoc ] ) and the relation : @xmath498 we thus obtain ( the green function is the retarded one ) : @xmath499 \frac{n_{f}(\nu ) - n_{f}(\nu + \omega)}{\omega}\ ] ] for the dc - conductivity , eq.([conducreduit ] ) simplifies to : @xmath500 \frac{1}{\cosh ^{2 } \frac{\beta \omega } { 2}}\ ] ] in both case the integrals are computed simply by transforming them into a riemann sum .
in this appendix , some details about the thermal scaling analysis of section [ sectcrossovertherm ] in the spinfluid regime are provided .
as explained in the text , in this regime @xmath7 disappears from the thermal scaling functions and thus the calculation can be performed in the undoped model @xmath501 . in this appendix
, @xmath463 and @xmath502 will denote the thermal scaling function of these quantities .
they satisfy the scaled saddle point equation @xmath503 .
the calculation is very similar to the low temperature , low frequency analysis made in @xcite so we here just give the main steps of the analysis .
we first check that the scaling behaviour ( [ gechelle ] ) in imaginary time solves the saddle point equation ( for @xmath501 ) using the fourier formulas ( which follows from @xcite 3.631 ) : [ fourier ] @xmath504 @xmath505 where @xmath506 are the matsubara frequencies .
one can then show that ( [ scalingformrho ] ) and ( [ scalingformsigma ] ) are the scaling function on the real axis using the following method .
first we use @xmath507 and @xmath508 ( see formula 3.313.2 of @xcite ) .
we then deduce the full green function ( and thus @xmath502 ) by performing the hilbert transform of @xmath285 using ( [ fourier2 ] ) again and @xmath509 ( see formula 3.112.1 of @xcite ) .
in this appendix , we briefly explain how to calculate the _ uniform susceptibility @xmath510 _ and analyse its behaviour at small temperature in the undoped model @xmath501 .
the magnetic field is introduced in the @xmath26 hamiltonian ( [ ham1 ] ) in the following way : @xmath511 this formula clearly reduces to the usual one for @xmath45 . here ,
only color 1 and 2 are coupled to this field but note that this choice is not unique though convenient for our calculation ( more generally the magnetic field must be coupled to an element of a cartan subalgebra of su(m ) ) .
the large @xmath204 and large @xmath25 limit computation is similar to the zero field one explained previously , although it is more involved .
a simplification occurs in this double limit : due to the fact that only 2 colors over @xmath25 are coupled to @xmath512 , the green function @xmath513 for colors @xmath514 are solutions of the zero - field equations ( [ eqbase ] ) .
moreover we find @xmath515 and @xmath516 where @xmath517 is given by : ( @xmath92 , @xmath481 and @xmath520 are always determined by eqs.([eqbase ] ) ) .
the magnetisation is given by @xmath521 .
let us define @xmath496 by @xmath522 . from eq.([eqbase5],[eqbase6 ] ) , @xmath496 satisfies : @xmath523 with these notations , the uniform susceptibility is given by @xmath524
. in the undoped case , eq.([suscepunif ] ) reduces to @xmath525 and the susceptibility at @xmath24 is formally given by : @xmath526 from the low frequency behaviour eq.([lowombehavg ] ) we have @xmath527 .
thus we have to investigate the low - frequency behaviour of @xmath528 . generalising the luttinger theorem ( as expressed by eq.([luttfinal ] ) ) for the colors 1 and 2 , we obtain ( at @xmath24 ) : @xmath529 in this equation , @xmath530 is given by : @xmath531 where @xmath532 is the number of particles of color 1 .
hence we obtain : @xmath533 taking first the limit @xmath216 and then @xmath534 we obtain finally ( as @xmath535 is bounded by definition ) : @xmath536 thus the leading low - frequency singularity in @xmath496 cancels so from eq.([chitnulle ] ) we see that @xmath495 is smaller than @xmath537 at small temperature . strictly speaking we can not prove from the previous argument that @xmath495 reaches a finite value at zero temperature , but it is a very natural guess which is moreover very well supported by our numerical calculation as displayed in fig.[chiunif ] . for reviews on the large-@xmath25 approach ,
e.g _ d.m .
newns and n. read , _ adv .
phys . _ * 36 * , 799 ( 1987 ) and g. kotliar , in _
`` strongly interacting fermions and high @xmath406 superconductivity '' _ , b.doucot and j. zinn - justin eds . , ( elsevier science pub . ) ( 1994 ) . for a review of neutron experiments on ybco , see p. bourges , to appear in _ the gap symmetry and fluctuations in high - temperature superconductors _ , proceedings of the nato cargese summer school , j. bok , g. deutscher , d. pavuna and s.a .
wolf eds ( plenum , 1998 ) .
p. mendels , h. alloul , j.h .
brewer , g.d .
morris , t.l .
duty , s. johnston , e.j .
ansaldo , g. collin , j.f .
marucco , c. niedermayer , d.r .
noakes and c.e .
stronach phys .
b. rapid . comm . * 49 * 10035 ( 1994 ) we note that the two - dimensional quantum xy model with bond disorder has been suggested to have a spin - liquid ground state for a large enough concentration of frustrating bonds : see p.gawiec and d. grempel , _ phys . rev .
b _ * 54 * , 3343 ( 1996 ) . see e.g. : f. steglich , b. buschinger , p. gegenwart , m. lohmann , r. helfrich , c. langhammer , p. hellmann , l. donnevert , s. thomas , a. link , c. geibel , m. lang , g. sparn and w. assmus _
j. phys.:condens matter _ 8 9909 ( 1996 ) ; m.b .
maple , m.c .
de andrade , j. herrmann , y. dalichaouch , d.a .
gajewski , c.l .
seaman , r. chau , r. movshovich , m.c . aronson and r. osborn _ j. low temp .
phys . _ * 99 * 223 ( 1995 ) for reviews , see _
e.g _ d.m .
newns and n. read , _ adv .
phys . _ * 36 * , 799 ( 1987 ) and g. kotliar , in _
`` strongly interacting fermions and high @xmath406 superconductivity '' _ , b.doucot and j. zinn - justin eds . , ( elsevier science pub . ) ( 1994 ) .
only part of the degeneracy is actually removed : the spin - liquid ground - state turns out to have an extensive entropy in the large-@xmath25 limit , as studied in detail in @xcite .
b a. m. sengupta _ preprint cond - mat 9707316 _ ; a. georges and a.m. sengupta , _ | we study the doping of a mott insulator in the presence of quenched frustrating disorder in the magnitude and sign of the magnetic exchange .
two quite different doping regimes @xmath0 and @xmath1 are found , with @xmath2 ( @xmath3 is the characteristic magnitude of the exchange , and @xmath4 the hopping amplitude ) . in the high - doping regime , a ( brinkman - rice ) fermi liquid description applies with a coherence scale of order @xmath5 . in the low doping regime , local magnetic correlations strongly affect the formation of quasiparticles , resulting in a very low coherence scale @xmath6 .
fermi liquid behaviour does apply below @xmath7 , but a `` quantum critical regime '' @xmath8 holds , in which _ marginal fermi liquid _
behaviour of several physical properties is found : nmr relaxation time @xmath9 , resistivity @xmath10 , optical lifetime @xmath11 together with @xmath12 scaling of response functions , e.g. @xmath13 .
in contrast , _ single - electron _ properties display stronger deviations from fermi liquid theory in this regime with a @xmath14 dependence of the inverse single - particle lifetime and a @xmath15 decay of the photoemission intensity . on the basis of this model and of various experimental evidence
, it is argued that the proximity of a quantum critical point separating a glassy mott - anderson insulator from a metallic ground - state is an important ingredient in the physics of the normal state of cuprate superconductors . in this picture
the corresponding quantum critical regime is a `` slushy '' state of spins and holes with slow spin and charge dynamics responsible for the anomalous properties of the normal state
. this picture may be particularly relevant to @xmath16doped materials . |
many materials with open @xmath0- or @xmath1-shells show metal - insulator transitions which are commonly classified to be of mott - hubbard type due to strong electron - electron correlations @xcite .
the conventional model to study this type of transition is the one - band hubbard model @xcite , which in standard notation reads @xmath2 major progress in understanding the physics of the mott - hubbard metal - insulator transition ( mhmit ) of the model ( [ equ : hubbard ] ) has been achieved in the last decade through the development of the dynamical mean - field theory ( dmft ) @xcite . at @xmath3 the mhmit occurs at a critical value of the coulomb parameter @xmath4 @xcite , where @xmath5 denotes the bandwidth of the density of states at @xmath6 .
interestingly , the transition is of first order @xcite for @xmath7 with a second order end point at a @xmath8 and @xmath9 .
such a second order end point is also seen in the phase diagram of typical mott - hubbard systems like v@xmath10o@xmath11 @xcite .
therefore , the one - band hubbard model has been frequently used as a microscopic model for these materials ( see refs . and , in particular , ref . which focusses on the critical regime close to the second order end point ) .
on the other hand , the justification to base the microscopic description on a one - band model ( see ref . ) has been questioned recently@xcite . for a proper description of materials such as transition metal oxides ,
the orbital structure of the relevant electronic degrees of freedom has to be taken into account .
this can lead to a fairly complicated form of the underlying tight - binding bandstructure ( the kinetic energy term acquires a matrix structure ) .
furthermore , additional local coulomb matrix elements arise which describe the interactions between electrons in different orbitals .
the simplest possible extension of the model ( [ equ : hubbard ] ) to the case of orbital degeneracy is the two - band hubbard model .
it is a relevant model whenever the electronic degrees of freedom close to the fermi level are two - fold degenerate , as for the @xmath12 states in materials like lamno@xmath11 or kcuf@xmath11 @xcite .
here we investigate the two - band hubbard model in the following form : @xmath13{l } \displaystyle-\sum_{ij}\sum_{mm'\sigma}t_{ij}^{mm ' } c^{\dagger}_{im\sigma}c^{\phantom{\dagger}}_{jm'\sigma}\\[5 mm ] \displaystyle+\frac{u}{2}\sum_{i}\sum_{m\sigma}n_{im\sigma}n_{im\bar{\sigma}}\\[5 mm ] \displaystyle+\frac{2u'-j}{4}\sum_{i}\sum_{m\ne m'}\sum_{\sigma\sigma'}n_{im\sigma}n_{im'\sigma'}\\[5 mm ] \displaystyle - j\sum_{i}\sum_{m\ne m'}\vec s_m\cdot\vec s_{m'}\\[5 mm ] \displaystyle-\frac{j}{2 } \sum_{i}\sum_{m\ne m'}\sum_\sigma c_{im\sigma}^\dagger c_{im\bar\sigma}^\dagger c_{im'\bar\sigma}^{\phantom{\dagger } } c_{im'\sigma}^{\phantom{\dagger } } \ \ , \end{array}\;\;\ ] ] with the orbital index @xmath14 .
the coulomb parameters @xmath15 and @xmath16 describe the inter - orbital coulomb interaction and hund s exchange coupling , respectively . the last term in ( [ equ : tbhubbard ] )
is necessary to ensure rotational invariance of the interaction . by virtue of this rotational invariance
the coulomb parameters are related by @xmath17 .
generally , the hierarchy of interactions is @xmath18 . a multi - band hubbard model as in eq .
( [ equ : tbhubbard ] ) displays a mott transition at all integer fillings ( not only at half filling as in the single - band case ) .
the additional coulomb interactions also modify the value of @xmath19 and , possibly , the character of the transition , as has already been investigated within the dmft framework @xcite .
various theoretical and numerical techniques have been employed to investigate multi - band hubbard models within dmft .
the quantum monte carlo method @xcite , which is very successful in treating the multi - band hubbard models in present applications of the lda+dmft approach @xcite , can not , however , handle the rotationally invariant form of the interaction in ( [ equ : tbhubbard ] ) due to the sign problem ( for recent progress in reducing this sign problem , see ref . ) . on the other hand , exact diagonalization @xcite , linearized dmft @xcite and exact treatments in the limit of infinite orbital degeneracy @xcite are not able to reliably calculate dynamical properties . in this paper , we use wilson s numerical renormalization group ( nrg ) @xcite to solve the effective two - orbital quantum impurity problem which appears in the dmft for the two - band hubbard model . in sec .
ii we start with some technical issues related to the nrg in the two - band case .
section iii shows thermodynamic and dynamic quantities for the two - orbital single - impurity anderson model , with the focus on the role of the hund s coupling . in sec .
iv we discuss the two - band hubbard model ; here we concentrate on the simplest case , i.e. degenerate orbitals and intra - orbital hopping only : @xmath20 .
the model eq .
( [ equ : tbhubbard ] ) is studied on a bethe lattice ( mainly to compare our results with those from other approaches ) ; the generalization to other lattices ( other densities of states ) is straightforward .
the paper is summarized in sec .
the use of the nrg to solve the effective quantum impurity model appearing in the dmft self - consistency has been extensively discussed in the literature @xcite . here
we focus on the additional problems arising in the two - orbital impurity anderson model , which constitutes the effective local model arising in the dmft for the two - orbital hubbard model .
the two - orbital anderson impurity hamiltonian ( in standard notation ) is given by @xmath21{l } \displaystyle\sum\limits_{\vec km\sigma}\epsilon_{\vec km\sigma } c^{\dagger}_{\vec km\sigma}c^{\phantom{\dagger}}_{\vec km\sigma } + \sum_{m\sigma } \epsilon_d d^{\dagger}_{m\sigma}d^{\phantom{\dagger}}_{m\sigma}\\[5 mm ] \displaystyle+\frac{u}{2}\sum_{m\sigma}n_{m\sigma}^dn_{m\bar{\sigma}}^d\\[5 mm ] \displaystyle+\frac{2u'-j}{4}\sum_{m\ne m'}\sum_{\sigma\sigma'}n_{m\sigma}^dn_{m'\sigma'}^d\\[5 mm ] \displaystyle - j\sum_{m\ne m'}\vec s_m\cdot\vec s_{m'}\\[5 mm ] \displaystyle-\frac{j}{2 } \sum_{m\ne m'}\sum_\sigma d_{m\sigma}^\dagger d_{m\bar\sigma}^\dagger d_{m'\bar\sigma}^{\phantom{\dagger } } d_{m'\sigma}^{\phantom{\dagger}}\\[5 mm ] \displaystyle\frac{v}{\sqrt{n}}\sum_{\vec km\sigma } c^{\dagger}_{\vec km\sigma}d^{\phantom{\dagger}}_{m\sigma}+\mbox{h.c . }
\end{array}\ ] ] within the nrg approach , the quantum impurity problem is mapped onto a semi - infinite chain form @xcite with the impurity at the first site of the chain and the conduction band written in a one - dimensional tight - binding form .
the model in the semi - infinite chain form is then solved by iterative diagonalization . therefore , the major obstacle in applying the nrg to multi - band models is the dramatic increase of the hilbert space with each nrg step ( in each step , one additional site of the semi - infinite chain is included ) .
a possible solution to this problem is to use very large values of the discretization parameter @xmath22 so that the number of states can be reduced significantly .
averaging over many discretizations ( the so - called `` z - trick '' ) has been shown to give reliable results for thermodynamic quantities ( see ref . ) , even for large values of @xmath22 . however , this approach becomes at least cumbersome for the calculation of dynamic quantities which are required for the dmft self - consistency .
here we adopt two different strategies which allow to use a small value of the nrg discretization parameter @xmath22 and to keep enough states in each step so that dynamic quantities can be calculated reliably .
the first one is to explicitely include the orbital quantum number in the iterative construction of the basis states .
this additional quantum number significantly reduces the typical matrix size , so that o@xmath23 states can be kept in each nrg iteration with reasonable computation time and memory consumption on modern smp high - performance computers like e.g. the ibm regatta .
the price to pay is that one has to omit the last term in the coulomb interaction , i.e. @xmath24 because it explicitely breaks the orbital symmetry .
however , it turns out that this term does not influence the thermodynamics , and dynamic quantities are only slightly affected via the multiplett structure of the hubbard bands .
the second strategy is an asymmetric truncation scheme : instead of adding both orbital degrees of freedom simultaneously , the hilbert space is truncated after adding each orbital individually .
this also leads to a significant reduction of the hilbert space in the iterative diagonalization .
however , the asymmetric truncation scheme does not guarantee that orbital symmetry is preserved during the nrg iterations .
in fact , a slight violation of orbital symmetry is observed for very small energies , typically much lower than the kondo temperature .
it turns out that for the dmft - calculations presented in sec .
[ sec : tohm ] , both methods give almost identical results , at least on the scale shown in the figures . , scaledwidth=45.0% ] figure [ fig : compare ] shows a comparison of the local single - particle density of states ( dos ) for the impurity anderson model eq .
( [ equ : tbsiam ] ) with twofold degeneracy , particle - hole symmetry , and calculated with both schemes ( symmetric truncation with orbital quantum number and asymmetric truncation ) in the kondo limit . here
, the nrg discretization parameter is @xmath25 and @xmath26 states were kept in each nrg step .
the model parameters are @xmath27 , @xmath28 and @xmath29 . as usual , @xmath30 denotes the bare hybridization width ( @xmath31 is the conduction electron dos at the fermi energy ) . obviously , the asymmetric truncation ( dashed line ) leads to accurate results for both the value of the low - energy scale and the form of the kondo resonance .
in addition , it does not violate particle - hole symmetry and orbital symmetry , at least on the scale shown in fig .
[ fig : compare ] .
the aymmetric truncation scheme introduced here might be of advantage in cases where the orbital symmetry is violated from the outset so that the orbital quantum number can not be used in the calculation to reduce the matrix size .
before we turn to the application of the nrg to the two - orbital hubbard model let us first discuss the effects of hund s coupling for the single impurity model ( [ equ : tbsiam ] ) . for simplicity
we consider a conduction band with constant dos , @xmath32 , in the interval @xmath33 $ ] and choose @xmath34 as unit of energy .
the local energy @xmath35 is chosen such that the model is particle - hole symmetric , i.e. @xmath36 .
as we will see , there is a profound difference between the cases with rotationally invariant and ising - like exchange .
the two - orbital impurity anderson model has already been investigated with nrg by several groups @xcite . , @xmath37 , and @xmath38 ;
dashed lines : entropy , solid lines : effective squared moments .
the coulomb parameter @xmath39 and the hybridization @xmath40 were chosen such that the system is in the kondo limit .
the inter - orbital coulomb parameter is fixed to @xmath17 by rotational invariance . for comparison
the results of a single - orbital siam are included.[fig : th_of_j],scaledwidth=45.0% ] here we want to concentrate on the influence of hund s coupling on low energy scales and the possibility of a quantum phase transition . to this end
we present results for thermodynamic and dynamic properties . to improve the accuracy of thermodynamic properties we employed oliveira s `` z - trick '' @xcite , which allows to use a larger discretization @xmath22 ( we used @xmath41 ) and
reduce the number of states kept ( @xmath42 after truncation here ) .
let us begin with thermodynamic properties of the particle - hole symmetric two - orbital single impurity model in the kondo limit , that is for a hybridization width @xmath43 much smaller than the other bare energy scales @xcite .
the temperature evolution of the effective squared moment introduced by wilson@xcite should not be confused with the isothermal quantity @xmath44 .
kondo screening can be seen only in the former ; the latter goes to a finite constant value as @xmath45 . ]
@xmath46 and entropy @xmath47 for @xmath48 , @xmath37 and @xmath38 is shown in fig .
[ fig : th_of_j ] .
the calculations were done with the rotationally invariant exchange coupling and @xmath17 , but neglecting the term breaking orbital symmetry . for comparison
we also include results for a single - orbital siam
marked @xmath49 in fig .
[ fig : th_of_j ] with the same values of @xmath39 and @xmath40 . for @xmath48 ,
i.e. @xmath50 , we observe an intermediate `` local - moment regime '' with entropy @xmath51 and effective moment @xmath52 corresponding to the six degenerate states , two of them magnetic , in the atomic limit .
as expected from general @xmath53 arguments @xcite this local moment is eventually kondo screened with an _ enhanced _ kondo scale @xmath54 . as a technical sidemark
let us point out that for @xmath48 , unlike finite @xmath55 , the asymmetric truncation does _ not _ work properly , leading to wrong results for @xmath56 and @xmath57 in the local moment regime .
in the atomic limit , any finite @xmath58 leads to a spin triplet @xmath59 as ground state with moment @xmath60 and entropy @xmath61 .
apparently , this situation is realized for intermediate temperatures for both @xmath28 and @xmath62 . at low temperatures ,
this local triplet is again quenched by the conduction electrons like for an ordinary @xmath63 kondo effect . obviously , the two - orbital system has a considerably reduced low - energy scale @xmath64 , which in addition decreases strongly with increasing @xmath55 . at present
we do not have a satisfactory explanation for this observation , but believe that it is related to the problem how spin-@xmath65 electrons screen a true @xmath59 object .
a direct consequence of this substantial reduction of the low - energy scale for the application to the hubbard model is that critical interactions for an mit are also strongly reduced for finite @xmath55 ( see sec .
[ sec : tohm ] below ) . for a particle - hole symmetric two - orbital impurity model with hund s rule coupling @xmath48 , @xmath37 and @xmath38 .
other parameters are the same as in fig .
[ fig : th_of_j].[fig : dos_full_j],scaledwidth=45.0% ] the local dos at @xmath66 for @xmath28 and @xmath62 is shown in fig .
[ fig : dos_full_j ] .
the calculations were done with a discretization parameter @xmath25 and @xmath67 states kept after truncation . on the scale used in the main panel of fig . [
fig : dos_full_j ] the kondo resonance for @xmath62 appears to be a vertical line , pointing to a strongly reduced kondo temperature , too .
in addition , new structures on the scale of @xmath55 appear as shoulders in the dos .
if one zooms into the region @xmath68 $ ] around the fermi energy ( see inset to fig . [
fig : dos_full_j ] ) , only the resonance for @xmath62 remains visible with an energy scale well below @xmath69 .
this again confirms the result from the thermodynamic quantities , viz that with increasing @xmath55 an exponential reduction of the kondo temperature occurs .
it is quite obvious , that such a reduction in @xmath64 will later manifest itself in a corresponding reduction of the critical @xmath39 for the mott - hubbard metal - insulator transition within the dmft .
let us point out that friedel s sum rule implies as usual the constraint @xmath70 .
this constraint is fulfilled with high precision due to the calculation of the dos via the self - energy according to ref . .
a completely different picture is obtained for an ising - like exchange interaction in model ( [ equ : tbsiam ] ) , which is realized by replacing @xmath71 and neglecting the term @xmath72 in this case , the atomic ground state is doubly degenerate and consists of and two values of the hybridization parameter @xmath40 .
apparently , there exists a critical @xmath73 which separates a strong coupling fixed point ( @xmath74 ) from a local moment like behaviour ( @xmath75).[fig : compare_th],scaledwidth=45.0% ] the two states where two electrons with the same spin occupy different orbitals .
in contrast to the full exchange we find that the properties change quite dramatically with the ratio @xmath76 , where @xmath40 denotes the hybridization width . in fig . [
fig : compare_th ] we compare calculations for @xmath28 and two different values of @xmath40 . for large @xmath40
we find the expected screening and corresponding formation of a fermi liquid at low temperatures .
however , for small @xmath40 , this behaviour is replaced by the formation of a state reminiscent of a local moment with entropy @xmath77 and effective local moment @xmath78 .
obviously , the neglect of the spin - flip terms in hund s exchange leads to a `` critical '' ratio @xmath79 separating strong - coupling from local moment behavior .
we interpret this feature in the following way .
for the full exchange interaction , spin - flip scattering as in the conventional @xmath63 case is presumably leading to a scenario similar to the standard kondo effect . for ising - like hund coupling @xmath55 , on the other hand , the atomic ground state consists , as already mentioned , of the two states where two electrons with the same spin occupy different orbitals .
quite apparently , these two states can not be connected by low - energy processes like spin - flips , i.e. the mechanism leading to the kondo effect is not present here . however , if the coupling to the band states is large enough such that the @xmath63 kondo temperature is larger than @xmath55 , the system can screen the spins for each orbital individually before the coupling @xmath55 locks the system into the states with @xmath80 , leading to the observed strong - coupling behavior at large @xmath40 . presently , it is not clear whether the change in the impurity properties is connected to some quantum critical behavior like in the pseudo - gap model @xcite .
the clarification of this question is of course of some interest in its own right and will be discussed in detail in a forthcoming publication . .
parameters are the same as in fig .
[ fig : compare_th].[fig : compare_dos],scaledwidth=45.0% ] the impurity dos corresponding to the two different regimes is shown in fig .
[ fig : compare_dos ] . for @xmath74 ( full line )
the typical structure is obtained .
decreasing @xmath40 below @xmath73 completely changes the structure of the dos . instead of a kondo peak , we now find a structure with a pseudo - gap at the fermi energy .
it is quite evident , that this `` criticality '' in the impurity model has profound effects on the mit in the dmft calculations for the hubbard model .
theoretical investigations of multi - orbital hubbard models within dmft have already led to a better understanding of various issues such as the nature of the mott - transition as a function of orbital degeneracy @xcite and the structure of the spectral function in realistic treatments within the lda+dmft approach @xcite .
detailed results have been obtained for the dependence of the critical interaction strength @xmath81 on the number of orbitals @xmath82 and different integer fillings @xmath83 .
numerical dmft - qmc calculations have been performed for @xmath84 @xcite .
remarkably , in the limit of large orbital degeneracy @xmath85 an analytical treatment of the dmft becomes possible for the mit @xcite . for @xmath66 , a scaling @xmath86 for the actual transition
is found while @xmath87 is obtained for the critical interaction where the insulating solution breaks down @xcite .
this is consistent with the linear dependence for large @xmath82 found in refs . and with the square - root dependence reported in ref .
the inclusion of the hund s rule exchange coupling @xmath55 has been shown to significantly reduce the value of @xmath81 @xcite . in particular , a qualitative change from continuous for @xmath48 to discontinuous for any finite @xmath55 has been observed in ref .
( see also the gutzwiller results in ref . ) . a significant quantitative change of @xmath81 when excluding the spin - dependent part from the exchange coupling in eq .
[ equ : tbhubbard ] has already been mentioned in ref . , but detailed results have not been published yet .
recently , the issue of possible orbital - selective mott transitions has been investigated in refs . .
let us now discuss the results from the nrg for the particle - hole symmetric case . to allow a direct comparison with earlier results , we use as non - interacting dos the semielliptic form @xmath88 ( bethe lattice ) with the same bandwidth for both orbitals .
as nrg discretization paramter we choose @xmath25 and keep @xmath67 states after truncation . except for @xmath48 ,
all the following results were obtained with the asymmetric truncation scheme introduced in sec .
ii including all coulomb interactions from the model ( [ equ : tbhubbard ] ) .
we have observed that the symmetric truncation scheme ( taking into account the orbital quantum number and negelcting the last term in the interaction ) leads to identical results for spectral functions , apart from some weak redistribution of spectral weight in the hubbard bands due to a different atomic multiplett structure .
however , the latter effect is only barely noticable due to the broadening introduced in calculating continous spectra from the nrg .
more important , the critical values @xmath19 for all two - band calculations within dmft are not affected ; possible problems due to the breaking of orbital symmetry by the asymmetric truncation scheme do not play a role here , except again for @xmath48 .
we begin by comparing results for the one - band hubbard model ( [ equ : hubbard ] ) ( fig .
[ fig:2b_1]a ) and a two - orbital hubbard model with @xmath48 and @xmath50 ( fig . [ fig:2b_1]b ) . for ( a ) one - band model eq .
( [ equ : hubbard ] ) and ( b ) two - orbital model eq .
( [ equ : tbhubbard ] ) with @xmath48 and @xmath50.[fig:2b_1],scaledwidth=45.0% ] as is well - known from earlier dmft calculations @xcite , the critical coulomb parameter , @xmath89 , for the model with @xmath48 increases strongly with the orbital degeneracy @xmath82 , @xmath90 @xcite .
this is also apparent from the results in fig .
[ fig:2b_1]b .
the system stays metallic up to the largest @xmath39 shown .
the actual mit occurs for a value @xmath91 . the results in figs .
[ fig:2b_1 ] and [ fig:2b_2 ] are calculated with a fairly large value of @xmath25 and broadening parameter @xmath92 ( see eq .
( 8) in ref . ) for both the single - band case in fig .
[ fig:2b_1]a and the two band case in figs .
[ fig:2b_1]b and [ fig:2b_2 ] .
we therefore expect that our critical values @xmath19 for the mott transition differ from more precise calculations , and indeed we find them to be somewhat overestimated . in the single - band case , for example , our result for @xmath93 is slightly ( 10% ) larger than the well established value @xmath94 @xcite .
a similar overestimation is also present for @xmath95 , where @xmath96 is reported in literature @xcite .
however , the qualitative features of the transition are not altered by this overestimation .
the influence of hund s coupling on the development of the spectra and the occurence of the mit can be seen in fig .
[ fig:2b_2]a , where results for different values of @xmath39 and a full hund s exchange @xmath97 are presented@xcite .
as has been noted before @xcite , finite @xmath55 substantially reduces for the two - orbital model with ( a ) full @xmath97 and ( b ) two - band model with ising - like @xmath97 . in both cases @xmath17 was used .
the assignment between the different line styles and values of @xmath39 is the same as in fig .
[ fig:2b_1 ] .
[ fig:2b_2],scaledwidth=45.0% ] @xmath19 .
such a behavior is also seen in fig .
[ fig:2b_2]a .
the critical coulomb parameter is reduced from @xmath98 for @xmath48 to @xmath99 .
presently , the standard technique to solve quantum impurities with orbital degeneracy is quantum monte - carlo ( qmc ) .
however , due to the minus sign problem , one has to restrict the coulomb interaction to density - density type only , i.e. an ising - like hund s exchange .
the last term in ( [ equ : tbhubbard ] ) has to be neglected completely .
this raises the following questions : what are the consequences of this approximation for the dynamics and in particular the mit ? to answer this question ( at least partially ) we performed calculations with ising - like exchange interaction as defined in the previous section ( see ref . )
the results are shown in fig .
[ fig:2b_2]b . at a first glance ,
the results are not very different , except that the critical @xmath39 is further reduced to @xmath100 . on the other hand , the results for the impurity calculation in figs .
[ fig : compare_th ] and [ fig : compare_dos ] already indicate that the replacement of hund s exchange by an ising - like term has more severe consequences than a mere quantitative change of energy scales . in the following
we show that this approximation indeed leads to a qualitative change in the physical properties of the mott - hubbard mit .
let us now turn to the nature of the mott transition .
for a one - band model , it is now commonly accepted that the transition is of second order at @xmath66 with a quasiparticle weight that vanishes smoothly as one approaches @xmath19 .
there is , however , a substantial region below @xmath19 , where the insulator is metastable @xcite .
previous work using the so - called linearized dmft ( l - dmft ) suggests that for orbitally degenerate systems with finite @xmath55 this may be different @xcite .
the authors of ref .
found a first order transition for small to intermediate @xmath55 signalled by a jump in the quasiparticle weight at @xmath19 . and @xmath97 for the full ( circles ) and ising - like exchange interaction ( squares ) .
the latter shows a strong jump in @xmath101 at @xmath19 , leading to a first order transition , while the former vanishes continuously .
the triangles represent a calculation with fixed @xmath102 .
the lines are meant as guide to the eyes .
[ fig : mstar],scaledwidth=45.0% ] the nrg results for the inverse effective mass ( @xmath103 quasiparticle weight ) for the case @xmath97 are shown in fig .
[ fig : mstar ] .
the circles were obtained from calculations using the full interaction , while the squares represent calculations with ising - like interactions .
apparently , the latter signal a strong first order transition at @xmath104 , while the former lead to a continuously vanishing quasiparticle weight . for a fixed @xmath102 ( triangles in fig . [ fig : mstar ] ) the quasi - particle weight near @xmath19 also shows a jump at @xmath19 as predicted by l - dmft @xcite .
however , the magnitude of this jump comes out much smaller in our calculations . for the two - band hubbard model and different values of @xmath55 .
the lines are meant as guide to the eye .
note that for @xmath102 and ising - like @xmath97 a discontinuity in @xmath44 occurs at the critical @xmath39.[fig : sz_sqr],scaledwidth=45.0% ] the differences between the different calculations ( @xmath48 , @xmath102 and @xmath97 ) become more apparent when one looks at the local squared moment @xmath44 . for @xmath105
this quantity has the value @xmath106 ( for @xmath102 it is actually slightly larger ) , while deep in the mott insulator it acquires the atomic value enforced by hund s coupling , i.e. @xmath107 for @xmath48 , @xmath60 for finite full @xmath55 and @xmath108 for ising - like @xmath55 .
this behavior is readily found in the calculated values of @xmath44 in fig .
[ fig : sz_sqr ] . in accordance with the results presented in ref .
, the limiting value for @xmath44 is approached smoothly for @xmath48 .
the same holds for @xmath97 , consistent with the results for the quasi - particle weight in fig .
[ fig : mstar ] .
the slope , however , strongly increases when one approaches @xmath19 from below .
for constant @xmath102 , the numerical results are not decisive , and could be interpreted as both a small discontinuity at @xmath19 and a continuous approach with diverging slope . on the other hand ,
a quite strong discontinuity at @xmath19 appears for an ising - like hund s coupling @xmath97 , signalling a rather strong first order transition in this case .
the above results are in rough agreement with the l - dmft predictions @xcite , although there one observes a first - order transition also for smaller values of @xmath55 .
the appearance of an unambiguous and rather strong first order transition for an ising - like exchange coupling shows that in this case the physics underlying the mott - hubbard transition is very different from the one for the rotationally invariant exchange interaction . as for the single impurity model
, we believe that a transition between individually screened orbitals on the metallic side to a local moment regime enforced by the ising coupling on the insulating side occurs as soon as @xmath55 becomes of the order of the fermi liquid scale . depending on the details of the non - interacting dos ( its value at the fermi level and the band width ) and the values of @xmath39 and @xmath55 , this can lead in the worst case to a serious underestimation of @xmath19 and possibly an incorrect description of the behavior of physical quantities close to the transition .
in this paper we presented first studies of the mott - hubbard transition in a two - orbital hubbard model within the dmft at @xmath66 using wilson s nrg .
in addition to a standard nrg implementation using the orbital quantum number , we proposed an asymmetric truncation scheme which turns out to work rather well in both the two - orbital single impurity anderson model and the two - band hubbard model .
as a first interesting result , we observed that for the particle - hole symmetric case a finite hund s exchange @xmath58 leads to a tremendous reduction in the low - energy scale @xmath64 .
this is in striking contrast to the result for @xmath48 , i.e. @xmath50 , where the behavior conventionally expected for an @xmath53 kondo model , viz @xmath109{t_{\rm k}^{n=1}}$ ] , is found@xcite . at present , the precise theoretical reason for this rather unexpected strong influence of @xmath55 on @xmath64 is not clear .
interestingly , it is also rather different from the case @xmath110 where we find a mild _ increase _ of @xmath64 with increasing @xmath55 .
obviously , a detailed study of the physics of multi - orbital quantum impurity models has to be an important future aim .
a completely different behavior occurs if one replaces the rotationally invariant hund exchange by an ising - like one . in this case
, the levels of the atomic doublet with @xmath80 enforced by the ising coupling can not be connected by schrieffer - wolff type spin - flip processes .
thus , the kondo effect can only occur for @xmath111 , while for larger @xmath55 the systems is locked into a local moment enforced by the exchange coupling .
note that this interpretation also implies that in the strong - coupling phase the spins on each individual orbital will be screened separately , while for the rotationally invariant case a full @xmath59 system must be screened .
the details of the quantum phase transition between strong - coupling and local moment phases have not yet been analyzed .
however , in view of a possible relevance of an ising - anisotropy in the presence of crystal fields , a further investigation of this problem is certainly interesting .
we also applied the nrg to the two - orbital hubbard model in the framework of dmft to investigate the mott - hubbard metal - insulator transition at @xmath66 for the half - filled case .
the major goal was here to eludicate the influence of hund s coupling on the mit and to investigate how the restriction to an ising - like exchange changes the nature of the mit .
our results are in general agreement with previous ones.@xcite in particular , for finite @xmath55 quantities like the effective mass or @xmath44 show diverging slopes as @xmath112 , possibly even discontinuities as proposed by the l - dmft.@xcite for an ising - like hund s exchange coupling the situation becomes qualitatively different .
as can be anticipated from the behavior of the impurity model , the mit is strongly first order with clear jumps in the effective mass and @xmath44 .
note that the former also implies a discontinuous vanishing of the quasi - particle peak at @xmath66 as one reaches @xmath19 .
also the physics underlying this transition is quite different compared to the rotationally invariant case , reflecting the lack of spin - flip scattering processes connecting the two states @xmath80 .
thus , the metallic phase with ising - like interaction will be characterized by individual screening of the spins on the two orbitals rather than a kondo screening of a total spin @xmath59 . note that this subtlety will most likely influence low - energy properties on the metallic side close to @xmath19 , but be less important for `` high - energy '' properties like magnetic or orbital ordering .
it is clear , that the investigations presented here are merely a starting point to systematically study properties of multi - orbital impurity models or correlated lattice models within the dmft at @xmath45 using wilson s nrg .
the major advantage of this method is obviously its unmatched ability to handle exponentially small energy scales and nevertheless provide reliable information on dynamics and thermodynamics even on high - energy scales .
thus , at least for two - orbital models we are now in a position to systematically study their physical properties and address questions that are of fundamental interest for a realistic description of , for example , transition metal oxides but require local degrees of freedom beyond a simple one - band hubbard model .
we acknowledge useful conversations with f. anders , a. lichtenstein , m. vojta , and d. vollhardt .
this work was supported by the dfg through the collaborative research center sfb 484 , the leibniz computer center , the computer center of the max - planck - gesellschaft in garching and the norddeutsche verbund fr hoch- und hchstleistungsrechnen .
g. kotliar and d. vollhardt , physics today * 57 * , no . 3
( march ) , 53 ( 2004 ) ; k. held , i. a. nekrasov , g. keller , v. eyert , n. blmer , a. k. mcmahan , r. t. scalettar , th .
pruschke , v. i. anisimov , and d. vollhardt , psi - k newsletter # 56 , 65 ( 2003 ) [ psi-k.dl.ac.uk/newsletters/news_56/highlight_56.pdf ] .
we have chosen a fixed ration for @xmath113 , because in typical experiments we expect the bandwidth to be the control parameter ( through the variation of pressure ) ; this means it is more natural to change @xmath114 and @xmath115 , but keep @xmath113 fixed . | the mott - hubbard metal - insulator transition is investigated in a two - band hubbard model within dynamical mean - field theory . to this end
, we use a suitable extension of wilson s numerical renormalization group for the solution of the effective two - band single - impurity anderson model .
this method is non - perturbative and , in particular , allows to take into account the full exchange part of the hund s rule coupling between the two orbitals .
we discuss in detail the influence of the various coulomb interactions on thermodynamic and dynamic properties , for both the impurity and the lattice model . the exchange part of the hund s rule coupling turns out to play an important role for the physics of the two - band hubbard model and for the nature of the mott - transition . |
the study of galaxy clusters provides important insight into the formation of structure in the universe and allows tight constraints to be placed on cosmological parameters .
observations at different wavelengths probe complementary aspects of clusters properties . in optical bands , the properties of individual cluster galaxies are studied and the measurement of their velocities enable the estimation of the total binding gravitational mass , via the application of the virial theorem ( e.g. * ? ? ?
observations at optical wavelengths also allow cluster mass estimates via strong and weak lensing analyses ( e.g. * ? ? ?
* ; * ? ? ?
at longer wavelengths , radio observations of clusters have important applications .
the inverse compton scattering of cosmic microwave background photons by the energetic electrons in the intra - cluster medium ( icm ) gives rise to the sunyaev zeldovich effect ( sze ; e.g. * ? ? ?
observations of this effect can be used to detect clusters and probe the properties of the icm .
radio - emitting jets from active galactic nuclei ( agn ) interact with the icm , inflating cavities and driving shocks in the x - ray gas , and are likely to have an important impact on icm properties ( e.g. * ? ? ?
the most sensitive measurements of the icm properties are currently made at x - ray wavelengths .
the ionised gas of the icm is the dominant baryonic component of galaxy clusters , and x - ray observations yield fundamental properties of the icm such as x - ray luminosity and temperature , and permit mass estimates ( assuming the gas is in hydrostatic equilibrium ) even for clusters at high redshifts ( see * ? ? ?
* for a review ) .
in the local universe it has been found that most ( @xmath8 ) clusters host cool cores @xcite .
the high gas densities in these core regions give rise to efficient radiative cooling with corresponding bright peaks in the x - ray emission and suppressed temperatures .
early models predicted that large amounts of gas should cool out of the icm in this process , although the fate of the condensing material was unknown ( e.g. * ? ? ? * ) .
more recent _ xmm - newton _ observations have found that the majority of the core gas does not cool out of the x - ray emitting regime , requiring some form of heating to regulate the process @xcite .
agn feedback is a strong candidate for providing this energy input ( e.g. * ? ? ?
the gas cooling in cool cores is still significant however ; core temperatures of @xmath9 of the global value are typical , with detectable cooling extending to @xmath10 ( e.g. * ? ? ?
* ; * ? ? ?
* ) is the radius enclosing a mean overdensity of 500 with respect to the critical density at the cluster s redshift . ] .
while cool cores are common features in relaxed systems , clusters are observed in a variety of dynamical states , and x - ray observations allow cluster morphologies to be studied .
early work categorised clusters into different classes based on their appearance @xcite , while more quantitative methods have been developed enabling cluster morpholgies to be correlated with other properties ( e.g. * ? ? ?
* ; * ? ? ?
using such methods , evidence has been found that clusters are generally less relaxed at @xmath1 @xcite .
x - ray imaging analysis has also been used recently to demonstrate that the fraction of cool core clusters in the population drops from @xmath11 locally to @xmath12 at @xmath1 @xcite .
these two observational findings are consistent with the higher merger rates expected at high - redshifts .
further insight into the icm properties can be gained from the emission lines in the x - ray spectra of clusters .
these are due to significant amounts of heavy elements in the gas . the observed metal abundances are consistent with models of the enrichment of the icm by supernovae based on observed star formation rates @xcite .
x - ray observations have enabled spatially resolved abundance measurements in clusters , showing sharply central peaks in cool - core clusters , and only mild abundance gradients in non cool - core clusters ( e.g. * ? ? ?
recent work has also found lower average icm metal abundances at high redshifts than locally @xcite .
the _ chandra _ x - ray observatory , with its high spatial and spectral resolution , is well suited to the study of galaxy clusters .
imaging analyses and spatially resolved spectroscopy of the icm can be performed without the complications of point source contamination and deconvolution of the telescope point spread function ( psf ) .
an important contribution of _ chandra _ to the study of galaxy clusters is its large public data archive .
this enables the construction of large samples of clusters whose properties are determined in a consistent manner , allowing powerful statistical studies of cluster properties , their correlations , and their evolution . in this
_ paper _ we present a catalogue of 115 galaxy clusters observed by _
chandra_. the construction of the sample and the data reduction methods are detailed in [ s.cat ] and [ s.reduction ] . our image analysis and x - ray images of the clusters
are presented in [ s.img ] and the spectral analysis and spectral properties of the clusters are given in [ s.spec ] .
notes on individual clusters are given in [ s.notes ] and the results are presented and discussed in [ s.results ] . a @xmath13cdm cosmology of @xmath14 ( @xmath15 ) , and @xmath16 ( @xmath17 )
is adopted throughout and all errors are quoted at the @xmath18 level .
the sample was assembled from all publically available _
chandra _ data as of november 2006 .
the positions of all _ chandra _ pointings were correlated with the nasa / ipac extragalactic database ( ned ) , and observations with a galaxy cluster or group listed in ned within @xmath19 of the _ chandra _ aimpoint were kept .
this list of observations was then refined by selecting only those for which the detector was acis - i , a galaxy group or cluster was the target of the observation , and the object had a published redshift greater than @xmath20 .
no lower mass limit was applied to the target objects , but the lower redshift cutoff and the choice of the acis - i detector limits the number of galaxy group size objects in the sample .
this combination of detector and lower redshift cutoff was chosen to ensure that the cluster emission within the radius @xmath21 ( our measurement of @xmath21 is discussed in [ s.spec ] ) was within the field of view , allowing properties to be measured to that radius without extrapolation .
quantitatively , we required that @xmath22 of the area in the annulus @xmath23 intersected with the active acis - i ccds .
all clusters at @xmath24 met this requirement , with the exception of @xmath25 ( obsid 6103 ) at @xmath26 , which was rejected from the sample accordingly .
observations were then examined individually and rejected if insufficient useful data were left after lightcurve filtering ( described in [ s.reduction ] ) , if they were affected by low level background flares ( identified by inspection of the spectrum of the local background , as described in [ s.spec ] ) , or if no significant extended emission was detected at the cluster position .
the latter was the case for a few candidate high - redshift clusters for which detections in earlier data ( generally from _ rosat _ ) were resolved into distinct point sources by _
the final sample then consisted of 149 observations of 115 galaxy clusters .
the individual observations are summarised in table [ t.cat ] .
the clusters are listed in order of right ascension , and the positions given are the icrs equatorial coordinates of the x - ray centroids .
column 2 gives the _ chandra _ observation identifier for each observation and column 5 gives the redshift of each cluster ( redshifts are taken from ned or bax , the x - rays clusters database ) . in columns 6 and 7
the date of each observation and corresponding `` blank - sky '' background period are given . finally , column 8 lists the cleaned exposure time of each observation ( the data cleaning process is described in [ s.reduction ] ) .
the redshift histogram of the final sample is shown in fig .
[ f.zhisto ] .
+ the clusters form a heterogeneous collection , having been observed for different purposes .
nearly half of the clusters ( 55 ) were observed in the guaranteed time observations of leon van speybroeck as part of a program to combine x - ray data with sze observations in order to place constraints on @xmath27 and @xmath28 @xcite .
one of the main goals in the current _ paper _ is to present a catalogue of the properties of this rich sample of clusters .
the standardised methods that were used to analyse the data ensure that the results presented are self - consistent and can be duplicated in , for instance , comparative analyses of `` mock observations '' derived from numerical simulations . in follow - up papers
we will investigate the x - ray scaling relations for this sample , including their evolution and scatter , and look in detail at clusters that deviate strongly from the trends measured for the sample as a whole .
initially , the level 1 event files were reprocessed with the latest calibration ( as of september 2006 ) .
if the focal plane temperature of the observation was cooler than @xmath29 then the charge - transfer inefficiency and time - dependent gain corrections were applied . for early observations with warmer focal plane temperatures , only the standard gain correction was used .
background flare filtering was then performed with _
lc_clean _ in the @xmath30 band for data from the acis - i ccds , excluding the target ccd and all sources detected in the standard _
chandra _
x - ray centre pipeline processing .
lightcurves were cleaned by @xmath31 clipping and periods with count rates @xmath32 above the mean rate were rejected .
this is consistent with the cleaning applied to the blank - sky background datasets .
the background dataset appropriate for the date of each observation was chosen , and if background period d was used , and the observation was taken in vfaint telemetry mode , then the additional vfaint cleaning procedure was applied to the source and background datasets . finally ,
if the observation was not taken in the same detector configuration as the background files ( acis-01236 ) , then ccds other than acis-0123 were removed from both source and background datasets .
otherwise ccd 6 was left in both the source and background files .
the use of blank - sky background data means that background data is available for the same detector regions as the source data .
this eliminates systematic differences between the source and background data in both imaging and spectral analysis .
however , the background consists of a particle - induced component which dominates at @xmath33 and varies with time , and a soft galactic x - ray component which varies spatially over the sky .
the contributions of these components in the blank - sky background will be different from those in the source dataset .
these differences are accounted for in the imaging and spectral analysis methods described below .
an additional background dataset was produced for each observation to enable the subtraction of the acis readout artifact . while a ccd is being read out , source photons are still detected , creating a streak of source photons along the readout direction .
the ratio of the readout time to the exposure time of one frame is @xmath34 , so @xmath34 of the source photons will be affected . following the method described by @xcite , we account for this effect by generating an additional background dataset from the source dataset , with the photon chip - y positions randomised and coordinates and energies recalculated .
this normalisation of this readout background is set to @xmath34 of that of the source , and the blank - sky background normalisation is reduced accordingly .
we note that the readout artifact was only significant for the clusters with very bright central peaks in their surface brightness distribution , but this correction procedure was applied to all data for consistency .
exposure maps were generated at an energy of @xmath35 , chosen to correspond to the peak acis effective area .
point and extended sources in the field were detected using the wavelet detection algorithm of @xcite . a broad band ( @xmath36 ) exposure - corrected image was used for point source detection , while the @xmath37 band was used for the detection of extended emission .
the latter energy band was chosen to optimise signal to noise for typical cluster spectra and was used for all of the following imaging analysis .
the centroid of the cluster emission was then located by refining its position with 5 iterations within @xmath38 and then 5 iterations within @xmath39 centered on the previous centroid ( with point sources masked out ) .
the target coordinates for each observation were used as the starting point for the iterations .
this centroid was then used to define the position of the cluster throughout the analysis .
extended sources that were detected as distinct separate sources compared to the target cluster were excluded , but any substructure that was detected in the cluster was left untouched .
the division between separate extended sources and cluster substructure is not always simple , and in many cases redshifts were not available for different emission components .
the thirteen clusters with excluded extended sources within @xmath21 ( in projection ) are noted in [ s.notes ] .
exposure - corrected radial profiles of the source and background images were then produced , with point sources excluded .
an iterative process was followed to define the outer detection radius of the cluster emission and correctly normalise the background image in the imaging energy band .
we defined the detection radius ( @xmath40 ) as the outer radius of the last bin with a width between @xmath41 and @xmath38 in which emission was detected at @xmath42 .
while this is a weak criterion for detection , our purpose here is to be conservative in defining a source - free region in each observation , and the detection radii were confirmed by visual inspection of the source and background radial profiles . in each step of the iteration ,
the detection radius was determined , and a background scaling factor was defined as the ratio of source and background count rates in the imaging band , with the emission within @xmath43 excluded .
the background image was renormalised by this factor , and a new detection radius was computed .
the process was repeated until @xmath43 converged to within @xmath44 .
this method provides a robust normalisation of the background in the imaging band , and definition of the background emission in the source image .
the key assumption of this method is that some of the field of view is free of cluster emission .
visual inspection of the source and background radial profiles showed that this was the case even for the brightest clusters in the sample .
if the cluster emission were not fully excluded in this process , then the remaining cluster emission would cause the soft background level to be overestimated .
this would result in a correlation between the derived background scaling and the cluster flux .
this is plotted in fig .
[ f.bgfx ] , with the background scaling defined as the ratio of the source to blank - sky count rates in the background region in the imaging band ( @xmath37 ) .
the cluster fluxes were measured from the spectral analysis , using a different background correction ( a scale factor derived at high energies ; see [ s.spec ] ) and so should be independent of the soft background scaling factor used here .
a spearman s rank test gives a @xmath45 correlation between background scaling and flux .
however , this can be caused by a selection effect whereby fainter clusters are preferentially observed in fields with lower soft x - ray background levels .
+ this was tested by correlating the cluster fluxes and background scaling factors with the @xmath46 x - ray background flux measured in the _ rosat_all - sky survey @xcite in an annulus of @xmath47 centered on each cluster .
we found significant positive correlations between the soft x - ray background and both the cluster flux ( at @xmath48 probability ) and our background scaling factor ( @xmath49 ) .
these two relations were then combined to predict the expected relation between background scaling and cluster flux under the hypothesis that any relation between those two variables is only due to the association of fainter clusters with lower background regions .
this predicted relation is plotted in fig .
[ f.bgfx ] and reproduces the weak trend found in the data . when this artifical correlation is factored out , any remaining trend is extremely weak , placing a limit of @xmath50 on the level of contamination of the background region by cluster emission .
note also that the clusters with large scale factors ( @xmath51 ) in fig .
[ f.bgfx ] are those with very high levels of galactic foreground emission , and are not affected residual cluster emission .
this is illustrated in [ s.spec ] and is corrected for in the imaging analysis by the simple scaling of the background image .
the background - subtracted , exposure - corrected x - ray surface brightness ( @xmath52 ) profiles were then binned so that all radial bins were significant at the @xmath31 level and were fit with a simple one dimensional ( 1d ) @xmath3-model of core radius @xmath53 and outer slope @xmath3 @xcite , @xmath54 images of the clusters were then reduced in size to a square of side @xmath55 ( @xmath56 is defined in [ s.rv ] ) centred on the cluster centroid and fit with a two - dimensional ( 2d ) elliptical @xmath3-model .
the ellipticity of the model was defined as @xmath57 , where @xmath58 and @xmath59 are the semi - major and -minor axes respectively .
the background was modeled with a constant level , which was fit to the normalised background image .
all models were convolved with an image of the telescope psf generated for the detector position of each cluster centroid at an energy of @xmath35 , and multiplied by the exposure map .
it is well known that the @xmath3-model is not generally a good description of cluster surface brightness profiles , with additional components required to match the sharp surface brightness peaks found in the cores of many clusters ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ) and evidence for steepening at large radii @xcite .
the model retains some appeal , however , as it is simple to deproject analytically to estimate the gas density distribution .
as discussed in [ s.mgas ] , we use a more complex model to fit the cluster emission and derive gas density profiles .
the @xmath3-model is used solely to allow us to investigate the steepening of the profiles at large radii , and provide a first measurement of the cluster morphologies via their ellipticity . in order to measure the local surface brightness slope at @xmath21(@xmath60 ) , the background - subtracted , exposure corrected profiles were grouped into logarithmic bins .
a straight line was fit in log space to the data in the radial range @xmath61 giving the best fit slope and its uncertainty .
this slope is independent of the core properties of the clusters . for 8 of the faintest clusters ,
the emission was not detected significantly at @xmath62 , so @xmath60 was not be measured for these clusters .
the logarithmic slope is related to the slope of the standard @xmath3-model by @xmath63 ( for @xmath64 ) .
we thus defined @xmath60 as @xmath65 the same procedure was used to measure @xmath66 , the slope at @xmath21 of the 1d @xmath3-model that was fit to the entire profile .
the measured ellipticities and surface - brightness profile slopes are are given in table [ t.morpho ] .
the x - ray emissivity of the intra - cluster medium depends strongly on the gas density and only weakly on its temperature .
this means that the observed projected emissivity profile can be used to accurately measure the gas density profile . for each annular bin in the surface brightness profile ,
the observed net count rate was corrected for area lost to chip gaps , bad pixels and excluded regions , and converted to an integrated emission measure .
the conversion factor was calculated for each bin assuming an absorbed mekal @xcite plasma model folded through an arf generated for that region , and an on - axis rmf .
the absorption of the spectral model was set at the galactic value and the metal abundance was set at @xmath67 . as the data are not sufficient to allow temperature profiles to be measured for most of the clusters in the sample , the temperature of the spectral model in each radial bin was obtained from the mean temperature profile found by @xcite , normalised to the global temperature measured for each cluster ( see [ s.rv ] ) . as noted above
, the dependence of the conversion from count rate to emission measure on the assumed temperature is weak .
the conversion factor decreases by @xmath12 when the model temperature is increased from @xmath68 to @xmath69 , with most of the decrease occurring below @xmath70 .
our analysis was repeated assuming the cluster was isothermal for the purpose of deriving the emission measures , and the results were not significantly changed .
the resulting profiles were fit with a modified version of the standard 1d @xmath3-model as proposed by ( * ? ? ?
* herafter v06 ) ( see * ? ? ?
* for other variations on the @xmath3-model ) ; @xmath71 the first term modifies a @xmath3-model to include a power - law cusp with slope @xmath72 at @xmath73 .
the second term allows a change of slope by @xmath74 at radius @xmath75 ( @xmath76 ) across a transition region whose width is controlled by @xmath77 .
following v06 , @xmath77 was fixed at 3 for all fits . in order to simplify the model slightly , so that it could be used to fit the range of high and lower quality data in the sample we ignore a second @xmath3-model component included in the v06 models for some of their clusters .
this model was then projected along the line of sight and fit to the observed projected emission measure profile .
the effect of the small _ chandra _ psf was neglected in this analysis .
there are strong correlations between the parameters in the emission measure model so the individual parameters are not reported here . however , the gas masses derived from the best fitting models are given in table [ t.ap ] , with uncertainties determined from monte - carlo randomisations of the observed emission measure profiles .
various methods have been used to quantify the amount of substructure in cluster images .
power ratios derived from different moments of the surface brightness distribution have been found to be a useful diagnostic of cluster morphology @xcite .
centroid shifts measured from the variation in separation between the peak and centroid of cluster emission with aperture size have similarly been used to measure cluster morphologies ( _ e.g. _ * ? ? ? * ; * ? ? ?
@xcite used numerical simulation of cluster mergers to test the effectiveness of these methods and found that the centroid shift method was the most sensitive to the dynamical state of clusters and the least sensitive to noise in the cluster images .
we thus adopt the centroid shift as our preferred statistic .
centroid shifts were measured following the method of @xcite .
the centroid of the cluster emission was determined in a series of circular apertures centred on the cluster x - ray peak .
the radius of the apertures were decreased in steps of @xmath78 from @xmath21 to @xmath79 and the centroid shift , @xmath80 , was defined as the standard deviation of the projected separations between the peak and centroid in units of @xmath21 . to increase the sensitivity of this statistic to faint structure , the central @xmath81 was excluded for the centroid ( but not the x - ray peak ) measurements .
the measured centroid shifts are summarised in table [ t.morpho ] .
the errors quoted for the @xmath80 are the statistical uncertainties on a standard deviation calculated from @xmath82 measurements ( @xmath83 ) .
centroid shifts were measured from exposure - corrected images in order to eliminate the effects of chip gaps , and vignetting .
the high spatial resolution of _ chandra _ means that point sources could be excluded without significantly biasing the measured centroid position .
as the purpose of the statistic is to quantify substructure , all extended sources were left untouched .
the redshifts of many of the extended sources detected around the target clusters are unknown , so those sources known to be foreground or background objects were not excluded in order to treat all clusters consistently . in fact , as detailed in [ s.notes ] , only @xmath84 of the 115 clusters had distinct extended sources that are projected within @xmath21 of the cluster core and just one of those ( v1701 + 6414 ) is known to be at a redshift different from the target cluster . in order to examine the morphologies of the clusters , contour maps of the x - ray emission
were generated using images that were adaptively smoothed to @xmath31 significance with the smoothing algorithm _
* we modified the original algorithm slightly to include exposure correction ) .
the raw images were produced in the @xmath85 band and were binned by a factor of 2 , giving approximately @xmath86 pixels .
the images were then cropped to a box @xmath87 on a side centred on the cluster s x - ray centroid . to correct for vignetting ,
ccd gaps and exposure variations in combined observations , exposure maps were provided to the smoothing algorithm . to enable the statistical significance of morphological features to be determined from the contour plots
, we selected contour levels to bound regions that were detected at a significant level above the local emission .
the methodology is explained in detail in @xcite , but briefly , the emission enclosed between a contour and the one above it is detected at @xmath88 significance ( after exposure correction ) with respect to the emission between the contour and the one below it . in the case of the lowest contour ,
the emission is detected at @xmath88 above the background emission in the same region ( measured in the blank - sky images ) .
note that this contouring scheme means that contour levels can not be interpreted in the same way as in standard contour plots ; separate contours that are the same number of levels above the background are unlikely to correspond to the same flux level in the smoothed image .
[ f.contour ] shows example contour plots for a sample of the clusters .
the plots for the full set of clusters are available in the electronic version of this paper .
in addition , for each cluster , a region file containing the contours is available as online data .
the intention is that these provide a useful resource enabling contours of the x - ray emission to be easily overlaid on images obtained from other detectors .
the region files also contain the flux levels of each contour to aid in their interpretation .
+ + +
in order to monitor the spectrum of the background emission in each observation , a spectrum was extracted from the source - free region and from the same region in the blank - sky background .
these spectra were compared to test the validity of using the blank - sky background spectrum for the analysis of that dataset . in order to account for differences in the particle - induced hard background ,
the background spectra were normalised to match the count rate in the @xmath89 band in the source datasets . after this normalisation ,
an excellent agreement was generally found , except below @xmath90 due to differences in the soft galactic foreground emission between the target and blank - sky fields . in rare cases ,
significant differences were found in the level of the continuum emission at @xmath91 due to low level background flares that were not detected by the standard lightcurve filtering . if there were other , unaffected observations of the same target , then the data with low flares were rejected . for two clusters ( a1682 and v1701 + 6414 ) there was no alternative data , and so a local background spectrum was used for the spectral analysis .
background spectra are shown in fig .
[ f.bgspec ] for v1701 + 6414 and for a1942 , a more typical example where the blank - sky and local background spectra agree well except for the galactic foreground emission at @xmath92 .
+ + in order to account for the differences in the soft background , the blank sky and readout background spectra were then subtracted from the local background spectrum to give a `` soft residual '' spectrum .
these residuals were generally positive , indicating a higher soft background flux in the target pointings than the blank - sky fields , but in some cases were negative .
the soft x - ray background is thermal in origin , and while the residual spectrum is the difference between thermal spectra , it is dominated by the oxygen lines at @xmath93 and is adequately described by a @xmath94 apec @xcite model with zero redshift and positive or negative normalisation @xcite .
the best fitting model to the soft residuals was included as an additional background component in all fits to cluster spectra , with the normalisation scaled for the differences in extraction area . for all spectral fits ,
the source emission was modelled with an absorbed apec model in the @xmath95 band .
the temperature , metal abundance and normalisation of the hot component were free to vary , while the absorbing column was fixed at the galactic value inferred from @xmath96 observations @xcite .
the model was refit with the soft galactic component normalisation set to @xmath97 and again with the overall blank - sky background renormalised by @xmath98 , and the resulting systematic temperature uncertainties were added in quadrature to the statistical uncertainties . for each source spectrum that was fit ,
the amount of flux lost to any excluded point and extended sources was estimated using the azimuthal average of the exposure - corrected , background subtracted flux at that cluster - centric radius .
while this assumes azimuthal symmetry for the clusters , it is an improvement over a simple geometric flux correction that would not take into account the radially varying surface brightness distribution of the cluster .
cluster properties were measured within the radius @xmath21 .
this radius generally matches the detected extent of the clusters emission , and corresponds to approximately half of the virial radius .
@xmath21 was measured following the approach recommended by @xcite to estimate @xmath21 from @xmath99 , where @xmath99 is defined as the product of the cluster temperature and gas mass . @xmath99
provides a reliable method for estimating cluster masses ( and hence @xmath21 ) , as the relation has low scatter , and the shape and scatter in the relation is insensitive to merging clusters @xcite .
this makes it ideal for our sample , which contains clusters with a wide range of dynamical states .
the low - scatter of the relation and its self - similar evolution have been verified for a subset of the clusters in this sample @xcite .
@xmath21 and @xmath99 were computed iteratively , measuring the temperature in the aperture @xmath100 and the gas mass within @xmath21 , computing a new @xmath99 , and hence estimating a new value of @xmath21 .
the process was repeated until @xmath21 converged to within @xmath44 .
an initial temperature estimate of @xmath101 was used to estimate the first value of @xmath21 to begin the iteration . in order to estimate @xmath21 from @xmath99
we used the relation measured for the @xcite sample of clusters @xmath102 with @xmath103 , @xmath104 and @xmath105 ( a. vikhlinin , priv .
@xmath106 describes the redshift evolution of the hubble parameter , and is given by @xmath107 by definition , @xmath21 is then given by @xmath108 .
this observed relation has a normalisation @xmath12 lower than that found in the simulations of kvn06 , which may be due to the effect of turbulent pressure support which is neglected in the mass derivations for the observed clusters ( kvn06 ) .
adopting the normalisation from the simulations would have the effect of increasing @xmath21 by @xmath109 ( @xmath110{15}\%$ ] ) , increasing @xmath111 by @xmath12 with a negligible effect on other properties .
the aperture of @xmath100 within which temperatures were measure in defining @xmath21 was chosen to exclude any effects of cool core emission in the clusters , and is consistent with the relation used .
once @xmath21 was determined , the cluster properties were also measured with the core emission included , and the values obtained for both apertures are given in table [ t.cat ] .
we found that exclusion of the core ( @xmath112 ) removed @xmath9 of the flux in non cool - core clusters ( and a larger fraction from clusters with cool cores ) .
this meant that for 4 of the faintest clusters ( rxj0910 + 5422 , clj1216 + 2633 , clj1334 + 5031 and rxj1350.0 + 6007 ) , we were unable to measure a temperature in that aperture .
for those clusters , all of the cluster emission ( @xmath113 ) was used for the temperature measurements in estimating @xmath21 .
none of these clusters showed evidence for the presence of cool cores , so the effect of including the core emission on biasing the temperature should be negligible .
twenty nine of the clusters in this sample were observed more than once by _
chandra_. in these cases , the observations were first analysed separately as described above , and then the data were combined for certain stages of the analysis .
source , background and exposure map images were projected onto a common coordinate system and summed .
the source lists from the individual observations were combined and those sources were excluded .
the summed images were then used for all imaging analysis .
multiple observations were also combined for the spectral analysis .
source and background spectra were extracted as above for each individual observation .
the spectra were then fit simultaneously as before with the temperature and metal abundance of the hot apec components tied together .
this method was followed when determining @xmath21 and the global temperature .
in this section we note any peculiarities or points of interest about individual observations or clusters that in some cases required a departure from the analysis procedure described above .
* a115 ( 3c28.0 ) : * this cluster is undergoing a major off - axis merger , with the two subclusters separated by @xmath114 ( @xmath115 ) in projection .
the fainter southern subcluster was manually excluded from our analysis .
the _ chandra _
observation of this system is discussed in detail by @xcite .
* clj0152.7 - 1357 : * this system is dominated by a north - south merger between equal mass clusters whose cores are separated by @xmath116 ( @xmath117 ) in projection .
the cluster is also associated with a network of large scale structure @xcite .
the northern and southern subclusters ( hereafter clj0152.7 - 1357n and clj0152.7 - 1357s ) were analysed separately with the emission from each excluded during the analysis of the other . * macsj0329.7 - 0212 :
* obsids 3257 and 6108 were rejected due to background flaring , leaving obsid 3582 . *
macsj0404.6 + 1109 : * an extended source at @xmath118=4^{\rm h}4^{\rm m}42.85^{\rm s}$ ] @xmath119=+11^{\circ}09{\ensuremath{\mathrm{^\prime}}}39.6{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath120 ( @xmath121 ) from the the cluster core , was manually excluded .
* rxj0439.0 + 0715 : * obsid 526 was rejected due to a very high background level , leaving obsid 3583 .
* 1e0657 - 56 : * this cluster , a spectacular merger known as the `` bullet cluster '' @xcite , has been observed 10 times with _ chandra _
acis - i for a total of @xmath122 .
for simplicity we use only two of the observations in our analysis which total @xmath123 , which is more than sufficient for our purposes .
* macsj0717.5 + 3745 : * obsid 1655 was rejected due to a very high background level , leaving obsid 4200 . * ms0906.5 + 1110 : * this cluster is part of a structure that includes a750 ( @xmath118=09^{\rm h}08^{\rm m}59.25^{\rm s}$ ] @xmath119=+11^{\circ}02{\ensuremath{\mathrm{^\prime}}}39.9{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] ) at the same redshift .
the centre of a750 is @xmath114 ( @xmath121 ) from the core of ms0906.5 + 1110 .
the emission from a750 falls partially within the _ chandra _ field of view in this observation and also within the @xmath21 of ms0906.5 + 1110 and was manually excluded .
* rxj1234.2 + 0947 : * this system ( also referred to as z5247 in * ? ? ?
* ) apparently consists of a binary merger between two clusters of similar mass .
however , only one redshift is available in @xcite for this system , and the position given is that of the southern subcluster .
the northern subcluster was thus manually excluded from the current analysis .
* a1682 : * this cluster is part of a large scale filament comprising at least two other extended x - ray sources at the cluster redshift .
these sources are at @xmath118=13^{\rm h}06^{\rm m}58.52^{\rm s}$ ] @xmath119=+46^{\circ}31{\ensuremath{\mathrm{^\prime}}}37.5{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , and @xmath118=13^{\rm h}07^{\rm m}13.33^{\rm s}$ ] @xmath119=+46^{\circ}29{\ensuremath{\mathrm{^\prime}}}08.4{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , and are @xmath124 ( @xmath125 ) and @xmath126 ( @xmath127 ) south east of the cluster core , respectively .
these sources were manually excluded for the current analysis .
furthermore , the only acis - i observation of this cluster was affected by long , low level flares causing the background spectrum to differ significantly from the blank - sky background . for this reason ,
a local background was used for the spectral analysis .
* clj1334 + 5031 : * the only acis - i observation of this cluster was affected by long , low level flares causing the background spectrum to differ significantly from the blank - sky background . for this reason ,
a local background was used for the spectral analysis . *
a1763 : * an extended source at @xmath118=13^{\rm h}34^{\rm m}53.21^{\rm s}$ ] @xmath119=+40^{\circ}56{\ensuremath{\mathrm{^\prime}}}55.34{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath128 ( @xmath129 ; @xmath130 ) from the the cluster core , was manually excluded .
* a2069 : * an extended source at @xmath118=15^{\rm h}24^{\rm m}27.70^{\rm s}$ ] @xmath119=+30^{\circ}01{\ensuremath{\mathrm{^\prime}}}44.29{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath131 ( @xmath132 ; @xmath130 ) from the the cluster core , was manually excluded .
* rxj1524.6 + 0957 : * this cluster is in a region of high galactic x - ray emission at the base of the north polar spur , and thus had an extremely strong background residual spectrum . a cool apec model at @xmath133 was required to fit this soft foreground component . *
a2111 : * an extended source at @xmath118=15^{\rm h}39^{\rm m}32.70^{\rm s}$ ] @xmath119=+34^{\circ}28{\ensuremath{\mathrm{^\prime}}}07.37{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath134 ( @xmath127 ) from the the cluster core , was manually excluded .
* a2163 : * the lightcurve of obsid 1653 was cleaned by hand to remove a long , low - level background flare .
* macsj1621.3 + 3810 : * the lightcurve of obsid 3594 was cleaned by hand to remove several periods of very high background . *
ms1621.5 + 2640 : * an extended source at @xmath118=16^{\rm h}23^{\rm m}48.25^{\rm s}$ ] @xmath119=+26^{\circ}34{\ensuremath{\mathrm{^\prime}}}22.11{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath135 ( @xmath132 ; @xmath130 ) from the the cluster core , was manually excluded .
* rxj1701 + 6414 : * this cluster ( at @xmath136 ) is separated by @xmath137 in projection from the foreground cluster a2246 at @xmath138 .
a2246 was manually excluded for this analysis .
the observation was also affected by long , low level flares causing the background spectrum to differ significantly from the blank - sky background . for this reason ,
a local background was used for the spectral analysis . *
a2261 : * an extended source at @xmath118=17^{\rm h}22^{\rm m}12.15^{\rm s}$ ] @xmath119=+32^{\circ}06{\ensuremath{\mathrm{^\prime}}}54.0{{\ensuremath{\mathrm{^\prime}}}\hskip -0.1em{\ensuremath{\mathrm{^\prime}}}}$ ] , @xmath139 ( @xmath140 ) from the the cluster core , was manually excluded .
we have used two simple quantities ( ellipticity and centroid shift ) to measure the morphology of the clusters in the sample . in fig .
[ f.morpho ] we plot the clusters on the plane of these two quantities .
the dashed lines indicate the median values for the entire sample .
for some nearby clusters , @xmath21 was not completely contained within the field of view , so @xmath141 ( the maximum radius within which @xmath142 could be measured ) was less than @xmath21 .
the @xmath143 clusters for which @xmath144 were rejected from this morphological study to ensure uniformity . broadly speaking , clusters to the bottom left of fig .
[ f.morpho ] are the most relaxed , with undisturbed , circular morphologies ( _ e.g . _
a383 ) while those to the top left are elliptical clusters without significant substructure ( _ e.g . _
the top right quadrant is home to clusters that are both elliptical and disturbed ( _ e.g . _
clj0956 + 4107 ) and finally the bottom left quadrant contains disturbed clusters with broadly circular geometries ( _ e.g . _
adaptively smoothed images of these four example clusters are plotted in fig .
[ f.morphoimg ] the sample was split into low and high - redshift subsets using a redshift cutoff of @xmath145 and the two subsets are marked separately on fig .
[ f.morpho ] .
there is a clear absence of very relaxed clusters at @xmath1 , although the distribution of ellipticities does not vary significantly with redshift .
the distribution of @xmath80 alone is plotted in fig . [ f.whisto ] for the low and high - redshift subsets .
again the absence of very relaxed clusters in the high - redshift subset is clear .
a kolmogorov - smirnov test gave a probability that the two @xmath80 subsets came from the same parent distribution of @xmath146 ( @xmath147 ) .
this reinforces , at greater significance , the results of @xcite who found more substructure in distant clusters than local clusters by using power ratios on _ chandra _ images ( see also * ? ? ?
the abundance of metals in the icm is the signature of star formation activity ( and the ensuing supernovae ) in the member galaxies , and the observed evolution of metal abundances can be used to trace the history of these processes ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ) . measuring metal abundances in distant clusters is challenging , and requires a higher quality of data than measuring temperature alone ( see [ sec : depend - spectr - uncert ] ) . in fig .
[ f.z500 ] we plot metal abundances measured within @xmath21 as a function of redshift and lookback time .
there are large uncertainties on the individual measurements , and for 9 clusters , the abundance measured was an upper limit .
the abundance values are iron abundances relative to the solar values of @xcite . while these have been superceded by more recent measurements ( e.g *
* ) , they allow straightforward comparison with other works . a simple scaling of @xmath148 converts from the @xcite iron abundance to values relative to the @xcite abundances . to improve the precision of the measurements , the clusters were grouped by redshift , and the spectra were fit simultaneously with the galactic absorption , redshifts and soft foreground component fixed at their measured values for each cluster .
the model temperatures were free to fit independently for each cluster ( but tied together for multiple observations of the same cluster ) and the metal abundances were tied for all clusters in the group . the metal abundance measured for each group of clusters is then effectively a weighted mean , with the higher signal - to - noise ( s / n ) spectra naturally having a stronger weighting due to the @xmath149 fitting method .
the clusters were binned into groups of 15 clusters ( with the highest redshift bin containing 11 clusters ) and the abundances measured for each group are plotted with the unbinned data in fig .
[ f.z500 ] .
the locus of each bin is at the median redshift of the clusters in that bin .
the grouped abundances decrease with redshift from @xmath150 locally to @xmath151 at @xmath6 , with most of the decrease occurring at @xmath1 .
the significance of the decrease was investigated by fitting a constant level to the grouped abundances .
this model was rejected at the @xmath152 level .
note that fitting a model to the unbinned data is problematic because of the upper limits at higher redshifts ; a survival analysis that can include the upper limits correctly is not appropriate here , as it is not possible to include measurement errors in such analyses .
clusters with cooling cores have been observed to have central peaks in their abundance profiles within @xmath153 , with abundances approaching solar values @xcite .
outside of the core the abundance profiles flatten significantly to values around @xmath154 .
@xcite also found significant ( although much weaker ) abundance gradients in non cool - core clusters .
as the global abundances measured here are emission weighted , the sharp surface brightness peaks in cool - core clusters cause their central abundance peaks to be overrepresented in the global value , biasing it high . a similar , but much milder bias will also be present in the non cool - core clusters .
it is therefore instructive to remove this effect by excising the core regions from the metal abundance measurements . to this end
, abundances were measured from spectra extracted in the annulus @xmath155 , although this leads to significantly lower signal to noise and hence larger uncertainties .
the clusters were grouped as before and the spectra were fit simultaneously . the abundance evolution measured in this way
is plotted in fig .
[ f.zcf ] along with the binned data from fig .
[ f.z500 ] ( which show the abundances measured when the core region was not excluded ) .
when the core regions are excluded , the data show the same trend of decreasing abundance with redshift , but at lower significance ; the constant abundance model could only be rejected at the @xmath156 level .
the metal abundances are lower on average at all redshifts when the core was excluded , but the difference becomes less significant at @xmath1 .
this can be partially explained by the decrease with redshift of the fraction of cool core clusters in the population @xcite .
however , the data suggest that weak abundance gradients such as those found by @xcite in non cool - core clusters are still present at @xmath157 . our results are consistent with those of @xcite based on a sample of 56 clusters observed in with _
chandra _
( including many clusters in common with our sample ) .
key differences between these studies are the sizes of our samples , and the choice of spectral aperture . @xcite used circular apertures in the approximate range @xmath158 to maximise the s / n for each cluster , and included the core emission .
we used the apertures @xmath159 and @xmath155 for all clusters .
the @xcite method has the advantage of yielding the most precise measurements for each cluster , while our method is consistent for all clusters , and allows us to address the effects of cool cores on the abundance measurements .
the overall decline in abundance with redshift that we observe is in qualitative agreement with the theoretical predictions of @xcite ; the line in fig .
[ f.zcf ] is the solid line taken from their fig .
5 . the abundances measured with the cluster cores excluded agree well with the theoretical prediction , while those measured from the entire cluster are systematically higher than the model .
this is likely to be due to the cool - core bias on our emission weighted global abundances . on the other hand
, the abundances measured with the cores excluded will somewhat under - represent the true global value , because the higher metallicity gas in the cores is then ignored completely .
it would be interesting to compare the evolution of mass - weighted abundances or total iron mass with theoretical predictions , but such measurements are not possible for the distant clusters with the current data .
the slope of the x - ray surface brightness profile is an important quantity because hydrostatic cluster masses derived within some radius depend on the slope of the gas density profile at that radius , which is measured from the surface brightness profile .
the slope is often measured by fitting a simple @xmath3-model to the data , but this has been found to underestimate the slope of the data at large radii as the fit is driven by the higher s / n data in the inner parts of the profile @xcite .
more recently , v06 demonstrated that much of the discrepancy between the normalisation of the mass - temperature relation from observational studies and simulations can be explained by underestimates of the gas density profile slope at large radii .
the ratio @xmath160 was calculated for each cluster , and a kernel density plot of the population is show in fig .
[ f.bratio ] . to form this plot
a gaussian kernel was computed for each measurement with the central position and @xmath161 given by the measured @xmath160 and uncertainty .
the kernels were normalised to have an area of unity and summed .
the advantage of this plot over a histogram is that the uncertainties in the variable are included ; well measured values contribute a high , narrow kernel while those with large uncertainties contribute a broad , low kernel .
the resulting curve peaks at @xmath162 and the weighted mean of the values is @xmath163 .
thus , on average , the slope of the cluster surface brightness profiles at @xmath21 is steeper by @xmath12 than the best - fitting @xmath3-model .
this is consistent with the findings of @xcite and v06 and further reinforces the need for careful modelling of surface brightness profiles at large radii .
a simple and useful model of the structure of galaxy clusters is that they are self - similar .
that is to say that clusters are identical when scaled appropriately by their mass
. this would be the case if gravity were the only important factor in the formation , growth , and evolution of clusters .
there are many examples of cluster properties deviating from self similarity ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but it remains a useful baseline model , and departures from self - similarity provide useful clues to non - gravitational processes that should be incorporated into cluster models .
in the self - similar model , the surface brightness profiles of clusters should be identical , once scaled by mass .
however , it has been established that there is a trend for cooler systems to exhibit shallower surface brightness profiles ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this demonstrates the increasing relative contribution of non - gravitational processes in clusters of lower masses .
the @xmath164 relation was investigated for our sample and is plotted in fig .
the best - fitting relations were measured with an orthogonal , weighted `` bces '' regression ( as described by * ? ? ?
* ) , on the data in log space .
the sample was divided at @xmath145 into low and high - redshift subsets , and these were fit separately and are plotted in fig .
there was a significant trend for cooler clusters to have lower values of @xmath60 in the low - z subset ; the best fitting relation had a slope of @xmath165 .
however , for the high - resdhift subset , there was no significant correlation , the best - fitting slope was @xmath166 .
the low - redshift slope is thus steeper than the high - redshift slope at the @xmath167 level .
this is not simply due to the absence of cooler systems in the distant subset ; the significance was unchanged when the systems cooler than @xmath168 were excluded from the local subset .
this result , while not strongly significant , could indicate that the self - similarity breaking is due to processes that become more important at lower redshifts .
alternatively , the possible flattening of the @xmath164 relation with redshift could be related to the morphological evolution of the clusters .
however , we find no correlation between @xmath60 and @xmath80 to support this hypothesis . finally , the large size of the sample enabled us to examine how the uncertainties on measured spectral properties depend on the signal to noise ratio of the spectra .
[ f.sn ] shows the fractional uncertainties on the temperature and metal abundance measured for the clusters plotted against the signal to noise ( s / n ) ratio in the spectral fitting band .
there are certainly factors other than data quality that influence the measurement uncertainties , such as multiple temperature and abundance components and different levels of absorption .
however , fairly tight power law correlations exist in both cases allowing us to define approximate relations that can be used for a `` rule of thumb '' assessment of data quality .
we find @xmath169 and @xmath170 .
we have performed a systematic analysis of the largest available sample of galaxy clusters at @xmath24 observed with _
chandra _
acis - i . images and contour files are made available for each cluster . the structural properties and metal abundances of the clusters
were investigated , and the main results can be summarised as follows : * there is a significant absence of relaxed clusters at high redshift as measured by the centroid shifts of the cluster emission . *
the metal abundance of the icm decreases with redshift , in line with theoretical predictions based on supernova rates . *
the abundance evolution is still present ( although less significant ) if the core regions of the clusters are excluded .
this indicates that the observed evolution is not a result of the disappearance of cool - core clusters from the population at high redshifts .
* metal abundances are lower at all redshifts when the cluster cores were excluded .
this suggests that weak abundance gradients are present in the icm even at high redshifts where the influence of cool - cores is absent .
* the surface brightness slope at @xmath21 ( @xmath60 ) is steeper on average by @xmath12 than a simple @xmath3-model fit to the entire profile .
* there is a significant correlation between @xmath60 and cluster temperature for local clusters , and an indication that the slope of this relation is shallower at high redshifts .
an important factor in these results is likely to be the higher cluster merger rate at high redshifts .
this is predicted by cosmological simulations ( _ e.g. _ * ? ? ?
* ) and explains the absence of relaxed , cool - core clusters at @xmath7 @xcite .
the changing cool - core population in turn has an effect on the metal abundance evolution , both because the metal abundances in the cool cores is higher and because the high emissivity of those regions then biases the the global emission - weighted abundances towards higher values .
the fact that the evolution is also present when the cores are excluded suggests that the influence of cool cores is not the only effect . indeed ,
the observed evolution agrees with that predicted by @xcite based on models of the rates star formation and subsequent supernova feedback with redshift .
furthermore , enhanced star formation has been found in the central galaxies of some cool core clusters ( _ e.g. _ * ? ? ?
this suggests that cool cores could be directly influencing the rate of enrichment of the icm , although the amount of cool - core induced star formation is not well constrained . finally ,
if the slope of the @xmath164 relation were confirmed to be shallower at @xmath7 , then this would suggest that the breaking of self - similarity in the surface - brightness profiles was due to processes that become important at lower redshifts .
this could be related to the onset at lower redshifts of cooling cores and their regulation by agn outbursts ( _ e.g. _ * ? ? ?
* ) . as a final point
, we remind the reader that the sample is not statistically complete , and is subject to the different selection processes by which the target clusters were originally chosen to be observed . as such , unquantifiable biases could be present , and so some caution is required in the interpretation of the results based on this sample . however , the agreement between the results presented here and other studies suggest that the impact of the heterogenous selection on our results is small . in subsequent papers
we will use this large sample to study the cluster scaling relations , their evolution , departures from self - similarity , and the sources of scatter in the relations ( _ e.g. _ * ? ? ?
this research has made use of the x - rays clusters database ( bax ) which is operated by the laboratoire dastrophysique de tarbes - toulouse ( latt ) , under contract with the centre national detudes spatiales ( cnes ) .
we also used the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
we thank alexey vikhlinin for providing some of the software used in this work and diab jerius for help assembling the list of clusters .
bjm is supported by nasa through chandra postdoctoral fellowship award number pf4 - 50034 issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 03060 .
lccccccc ms0015.9 + 1609 & 520 & 00:18:33.63 & + 16:26:14.6 & 0.541 & 2000 - 08 - 18 & c & 55.7 + rxj0027.6 + 2616 & 3249 & 00:27:45.92 & + 26:16:19.9 & 0.367 & 2002 - 06 - 26 & d & 9.1 + clj0030 + 2618 & 5762 & 00:30:33.77 & + 26:18:09.7 & 0.500 & 2005 - 05 - 28 & d & 16.1 + a68 & 3250 & 00:37:06.11 & + 09:09:33.6 & 0.255 & 2002 - 09 - 07 & d & 7.3 + a115 & 3233 & 00:55:50.69 & + 26:24:37.4 & 0.197 & 2002 - 10 - 07 & d & 43.6 + a209 & 3579 & 01:31:53.47 & -13:36:46.1 & 0.206 & 2003 - 08 - 03 & d & 8.8 + a209 & 522 & 01:31:53.59 & -13:36:46.1 & 0.206 & 2000 - 09 - 09 & c & 7.5 + clj0152.7 - 1357 & 913 & 01:52:40.03 & -13:58:24.2 & 0.831 & 2000 - 09 - 08 & c & 30.6 + a267 & 1448 & 01:52:42.12 & + 01:00:41.4 & 0.230 & 1999 - 10 - 16 & b & 6.5 + macsj0159.8 - 0849 & 6106 & 01:59:49.32 & -08:50:00.7 & 0.405 & 2004 - 12 - 04 & d & 30.3 + macsj0159.8 - 0849 & 3265 & 01:59:49.44 & -08:50:00.6 & 0.405 & 2002 - 10 - 02 & d & 15.0 + clj0216 - 1747 & 6393 & 02:16:32.93 & -17:47:30.5 & 0.578 & 2005 - 10 - 04 & d & 20.0 + rxj0232.2 - 4420 & 4993 & 02:32:18.34 & -44:20:50.3 & 0.284 & 2004 - 06 - 08 & d & 15.0 + macsj0242.5 - 2132 & 3266 & 02:42:35.86 & -21:32:25.8 & 0.314 & 2002 - 02 - 07 & d & 8.8 + a383 & 2320 & 02:48:03.43 & -03:31:46.2 & 0.187 & 2000 - 11 - 16 & c & 15.8 + a383 & 524 & 02:48:03.65 & -03:31:44.9 & 0.187 & 2000 - 09 - 08 & c & 8.8 + macsj0257.6 - 2209 & 3267 & 02:57:41.33 & -22:09:14.4 & 0.322 & 2001 - 11 - 12 & d & 16.6 + ms0302.7 + 1658 & 525 & 03:05:31.63 & + 17:10:10.6 & 0.424 & 2000 - 10 - 03 & c & 9.1 + clj0318 - 0302 & 5775 & 03:18:33.60 & -03:02:55.3 & 0.370 & 2005 - 03 - 15 & d & 12.7 + macsj0329.6 - 0211 & 3582 & 03:29:41.54 & -02:11:45.9 & 0.450 & 2002 - 12 - 24 & d & 14.5 + macsj0404.6 + 1109 & 3269 & 04:04:32.66 & + 11:08:16.8 & 0.355 & 2002 - 02 - 20 & d & 18.7 + macsj0429.6 - 0253 & 3271 & 04:29:35.95 & -02:53:06.1 & 0.399 & 2002 - 02 - 07 & d & 20.0 + rxj0439.0 + 0715 & 3583 & 04:39:00.67 & + 07:16:03.8 & 0.230 & 2003 - 01 - 04 & d & 16.1 + rxj0439 + 0520 & 527 & 04:39:02.35 & + 05:20:43.7 & 0.208 & 2000 - 08 - 29 & c & 8.0 + macsj0451.9 + 0006 & 5815 & 04:51:54.41 & + 00:06:19.2 & 0.430 & 2005 - 01 - 08 & d & 8.0 + a521 & 901 & 04:54:06.58 & -10:13:15.2 & 0.253 & 1999 - 12 - 23 & b & 35.8 + a520 & 528 & 04:54:09.60 & + 02:55:20.8 & 0.199 & 2000 - 10 - 10 & c & 8.5 + a520 & 4215 & 04:54:09.72 & + 02:55:17.8 & 0.199 & 2003 - 12 - 04 & d & 48.7 + ms0451.6 - 0305 & 529 & 04:54:11.26 & -03:00:52.6 & 0.550 & 2000 - 01 - 14 & b & 12.7 + clj0522 - 3625 & 4926 & 05:22:14.66 & -36:24:58.7 & 0.472 & 2004 - 06 - 17 & d & 15.8 + clj0542.8 - 4100 & 914 & 05:42:50.14 & -41:00:02.2 & 0.634 & 2000 - 07 - 26 & c & 44.0 + macsj0647.7 + 7015 & 3584 & 06:47:49.92 & + 70:14:56.0 & 0.584 & 2003 - 10 - 07 & d & 17.9 + macsj0647.7 + 7015 & 3196 & 06:47:50.16 & + 70:14:55.7 & 0.584 & 2002 - 10 - 31 & d & 15.3 + 1e0657 - 56 & 3184 & 06:58:29.52 & -55:56:39.1 & 0.296 & 2002 - 07 - 12 & d & 69.5 + 1e0657 - 56 & 554 & 06:58:30.00 & -55:56:39.5 & 0.296 & 2000 - 10 - 16 & c & 22.3 + macsj0717.5 + 3745 & 4200 & 07:17:31.68 & + 37:45:32.0 & 0.546 & 2003 - 01 - 08 & d & 51.8 + a586 & 530 & 07:32:20.16 & + 31:37:54.5 & 0.171 & 2000 - 09 - 05 & c & 8.6 + macsj0744.9 + 3927 & 3585 & 07:44:52.32 & + 39:27:25.2 & 0.697 & 2003 - 01 - 04 & d & 14.8 + macsj0744.9 + 3927 & 6111 & 07:44:52.32 & + 39:27:27.0 & 0.697 & 2004 - 12 - 03 & d & 39.9 + macsj0744.9 + 3927 & 3197 & 07:44:52.56 & + 39:27:26.3 & 0.697 & 2001 - 11 - 12 & d & 16.6 + a665 & 531 & 08:30:53.04 & + 65:49:55.2 & 0.182 & 1999 - 12 - 29 & b & 7.8 + a665 & 3586 & 08:30:59.04 & + 65:50:44.2 & 0.182 & 2002 - 12 - 28 & d & 26.2 + a697 & 4217 & 08:42:57.84 & + 36:21:57.2 & 0.282 & 2002 - 12 - 15 & d & 16.1 + clj0848.7 + 4456 & 1708 & 08:48:47.76 & + 44:56:12.5 & 0.574 & 2000 - 05 - 03 & c & 44.3 + clj0848.7 + 4456 & 927 & 08:48:50.88 & + 44:55:27.8 & 0.574 & 2000 - 05 - 04 & c & 95.1 + zwclj1953 & 1659 & 08:50:06.96 & + 36:04:18.1 & 0.320 & 2000 - 10 - 22 & c & 17.9 + clj0853 + 5759 & 5765 & 08:53:14.64 & + 57:59:47.8 & 0.475 & 2005 - 02 - 19 & d & 22.0 + clj0853 + 5759 & 4925 & 08:53:14.88 & + 58:00:02.9 & 0.475 & 2004 - 09 - 19 & d & 14.3 + ms0906.5 + 1110 & 924 & 09:09:12.72 & + 10:58:33.6 & 0.180 & 2000 - 10 - 02 & c & 26.4 + rxj0910 + 5422 & 2452 & 09:10:44.40 & + 54:22:04.4 & 1.110 & 2001 - 04 - 24 & d & 54.4 + rxj0910 + 5422 & 2227 & 09:10:45.36 & + 54:22:07.3 & 1.110 & 2001 - 04 - 29 & d & 84.2 + a773 & 5006 & 09:17:52.80 & + 51:43:40.4 & 0.217 & 2004 - 01 - 21 & d & 15.8 + a773 & 3588 & 09:17:53.04 & + 51:43:37.9 & 0.217 & 2003 - 01 - 25 & d & 8.8 + a773 & 533 & 09:17:53.04 & + 51:43:39.4 & 0.217 & 2000 - 09 - 05 & c & 9.8 + a781 & 534 & 09:20:26.16 & + 30:30:04.7 & 0.298 & 2000 - 10 - 03 & c & 8.6 + clj0926 + 1242 & 4929 & 09:26:36.48 & + 12:43:04.8 & 0.489 & 2004 - 02 - 06 & d & 16.3 + clj0926 + 1242 & 5838 & 09:26:36.48 & + 12:43:02.3 & 0.489 & 2005 - 02 - 21 & d & 27.2 + rbs797 & 2202 & 09:47:13.20 & + 76:23:13.6 & 0.354 & 2000 - 10 - 20 & c & 9.8 + macsj0949.8 + 1708 & 3274 & 09:49:51.84 & + 17:07:08.0 & 0.384 & 2002 - 11 - 06 & d & 12.9 + clj0956 + 4107 & 5759 & 09:56:03.12 & + 41:07:14.2 & 0.587 & 2005 - 01 - 28 & d & 34.7 + clj0956 + 4107 & 5294 & 09:56:03.36 & + 41:07:13.1 & 0.587 & 2003 - 12 - 30 & d & 13.5 + a907 & 535 & 09:58:21.84 & -11:03:49.0 & 0.153 & 2000 - 06 - 29 & c & 9.1 + a907 & 3205 & 09:58:22.08 & -11:03:49.3 & 0.153 & 2002 - 10 - 30 & d & 37.1 + a907 & 3185 & 09:58:22.08 & -11:03:50.0 & 0.153 & 2002 - 06 - 14 & d & 38.4 + ms1006.0 + 1202 & 925 &
10:08:47.52 & + 11:47:40.6 & 0.221 & 2000 - 06 - 22 & c & 23.3 + ms1008.1 - 1224 & 926 & 10:10:32.16 & -12:39:30.2 & 0.301 & 2000 - 06 - 11 & c & 14.5 + zw3146 & 909 & 10:23:39.60 & + 04:11:11.3 & 0.291 & 2000 - 05 - 10 & c & 40.4 + clj1113.1 - 2615 & 915 & 11:13:05.28 & -26:15:41.0 & 0.725 & 2000 - 08 - 13 & c & 44.6 + a1204 & 2205 & 11:13:20.40 & + 17:35:39.1 & 0.171 & 2001 - 06 - 01 & d & 19.4 + clj1117 + 1745 & 4933 & 11:17:30.00 & + 17:44:52.1 & 0.305 & 2004 - 06 - 18 & d & 16.6 + clj1117 + 1745 & 5836 & 11:17:30.00 & + 17:44:52.8 & 0.548 & 2005 - 02 - 15 & d & 39.9 + clj1120 + 4318 & 5771 & 11:20:07.20 & + 43:18:06.1 & 0.600 & 2005 - 01 - 11 & d & 16.6 + rxj1121 + 2327 & 1660 & 11:20:57.36 & + 23:26:29.0 & 0.562 & 2001 - 04 - 23 & d & 62.7 + a1240 & 4961 & 11:23:37.68 & + 43:05:44.5 & 0.159 & 2005 - 02 - 05 & d & 43.5 + macsj1131.8 - 1955 & 3276 & 11:31:55.20 & -19:55:52.7 & 0.307 & 2002 - 06 - 14 & d & 11.1 + ms1137.5 + 6625 & 536 & 11:40:22.32 & + 66:08:16.1 & 0.782 & 1999 - 09 - 30 & b & 103.1 + macsj1149.5 + 2223 & 3589 & 11:49:35.28 & + 22:24:09.0 & 0.545 & 2003 - 02 - 07 & d & 15.0 + a1413 & 5003 & 11:55:18.00 & + 23:24:17.6 & 0.143 & 2004 - 03 - 06 & d & 60.9 + a1413 & 1661 & 11:55:18.00 & + 23:24:14.4 & 0.143 & 2001 - 05 - 16 & d & 9.3 + clj1213 + 0253 & 4934 & 12:13:35.04 & + 02:53:46.4 & 0.409 & 2004 - 07 - 17 & d & 16.1 + clj1216 + 2633 & 4931 & 12:16:19.68 & + 26:33:13.7 & 0.428 & 2004 - 05 - 12 & d & 15.3 + rxj1221 + 4918 & 1662 & 12:21:26.40 & + 49:18:28.1 & 0.700 & 2001 - 08 - 05 & d & 66.4 + clj1226.9 + 3332 & 5014 & 12:26:57.84 & + 33:32:47.8 & 0.890 & 2004 - 08 - 07 & d & 25.1 + clj1226.9 + 3332 & 3180 & 12:26:57.84 & + 33:32:47.8 & 0.890 & 2003 - 01 - 27 & d & 25.4 + rxj1234.2 + 0947 & 539 & 12:34:18.00 & + 09:46:18.8 & 0.229 & 2000 - 03 - 23 & c & 8.6 + rdcs1252 - 29 & 4198 & 12:52:54.96 & -29:27:20.5 & 1.237 & 2003 - 03 - 20 & d & 133.2 + a1682 & 3244 & 13:06:51.12 & + 46:33:29.5 & 0.234 & 2002 - 10 - 19 & d & 7.3 + macsj1311.0 - 0310 & 6110 & 13:11:01.68 & -03:10:37.6 & 0.494 & 2005 - 04 - 20 & d & 52.6 + macsj1311.0 - 0310 & 3258 & 13:11:01.92 & -03:10:36.0 & 0.494 & 2002 - 12 - 15 & d & 13.0 + a1689 & 5004 & 13:11:29.52 & -01:20:29.8 & 0.183 & 2004 - 02 - 28 & d & 16.6 + a1689 & 1663 & 13:11:29.52 & -01:20:28.7 & 0.183 & 2001 - 01 - 07 & d & 8.8 + a1689 & 540 & 13:11:29.52 & -01:20:30.4 & 0.183 & 2000 - 04 - 15 & c & 8.8 + rxj1317.4 + 2911 & 2228 & 13:17:20.88 & + 29:11:15.0 & 0.805 & 2001 - 05 - 04 & d & 88.4 + clj1334 + 5031 & 5772 & 13:34:18.72 & + 50:31:02.3 & 0.620 & 2005 - 08 - 05 & d & 15.3 + a1763 & 3591 & 13:35:18.24 & + 40:59:59.3 & 0.223 & 2003 - 08 - 28 & d & 17.4 + rxj1347.5 - 1145 & 3592 & 13:47:30.72 & -11:45:10.4 & 0.451 & 2003 - 09 - 03 & d & 45.9 + rxj1350.0 + 6007 & 2229 & 13:50:48.72 & + 60:07:02.6 & 0.804 & 2001 - 08 - 29 & d & 47.7 + clj1354 - 0221 & 5835 & 13:54:17.04 & -02:21:52.3 & 0.546 & 2005 - 05 - 17 & d & 29.8 + clj1354 - 0221 & 4932 & 13:54:17.28 & -02:21:44.8 & 0.546 & 2004 - 12 - 07 & d & 15.3 + clj1415.1 + 3612 & 4163 & 14:15:11.04 & + 36:12:03.6 & 1.030 & 2003 - 09 - 16 & d & 72.9 + rxj1416 + 4446 & 541 & 14:16:28.08 & + 44:46:42.6 & 0.400 & 1999 - 12 - 02 & b & 27.7 + macsj1423.8 + 2404 & 1657 & 14:23:47.76 & + 24:04:40.8 & 0.543 & 2001 - 06 - 01 & d & 16.9 + a1914 & 3593 & 14:26:01.20 & + 37:49:34.0 & 0.171 & 2003 - 09 - 03 & d & 16.1 + a1914 & 542 & 14:26:01.20 & + 37:49:35.4 & 0.171 & 1999 - 11 - 21 & b & 7.2 + a1942 & 3290 & 14:38:22.08 & + 03:40:06.2 & 0.224 & 2002 - 03 - 13 & d & 52.6 + ms1455.0 + 2232 & 543 & 14:57:15.12 & + 22:20:34.4 & 0.258 & 2000 - 05 - 19 & c & 9.6 + ms1455.0 + 2232 & 4192 & 14:57:15.12 & + 22:20:34.8 & 0.258 & 2003 - 09 - 05 & d & 73.1 + rxj1504 - 0248 & 4935 & 15:04:07.44 & -02:48:18.4 & 0.215 & 2004 - 01 - 07 & d & 12.2 + a2034 & 2204 & 15:10:12.48 & + 33:30:28.4 & 0.113 & 2001 - 05 - 05 & d & 51.6 + a2069 & 4965 & 15:24:09.36 & + 29:53:10.0 & 0.116 & 2004 - 05 - 31 & d & 53.4 + rxj1525 + 0958 & 1664 & 15:24:39.84 & + 09:57:42.6 & 0.516 & 2002 - 04 - 01 & d & 43.6 + rxj1532.9 + 3021 & 1665 & 15:32:53.76 & + 30:20:59.3 & 0.345 & 2001 - 09 - 06 & d & 8.8 + a2111 & 544 & 15:39:41.28 & + 34:25:10.2 & 0.229 & 2000 - 03 - 22 & c & 9.1 + a2125 & 2207 & 15:41:08.64 & + 66:15:51.8 & 0.246 & 2001 - 08 - 24 & d & 71.3 + a2163 & 545 & 16:15:45.60 & -06:08:56.5 & 0.203 & 2000 - 07 - 29 & c & 9.1 + macsj1621.3 + 3810 & 3254 & 16:21:24.72 & + 38:10:07.3 & 0.463 & 2002 - 10 - 18 & d & 8.3 + macsj1621.3 + 3810 & 6109 & 16:21:24.72 & + 38:10:10.2 & 0.463 & 2004 - 12 - 11 & d & 32.9 + macsj1621.3 + 3810 & 6172 & 16:21:24.72 & + 38:10:09.5 & 0.463 & 2004 - 12 - 25 & d & 23.1 + ms1621.5 + 2640 & 546 & 16:23:35.28 & + 26:34:19.9 & 0.426 & 2000 - 04 - 24 & c & 26.2 + a2204 & 6104 & 16:32:47.04 & + 05:34:32.5 & 0.152 & 2004 - 09 - 20 & d & 9.1 + a2218 & 1666 & 16:35:52.08 & + 66:12:34.6 & 0.176 & 2001 - 08 - 30 & d & 37.3 + a2218 & 1454 & 16:35:52.08 & + 66:12:34.2 & 0.176 & 1999 - 10 - 19 & b & 10.1 + clj1641 + 4001 & 3575 & 16:41:52.80 & + 40:01:40.4 & 0.464 & 2003 - 09 - 24 & d & 36.0
+ rxj1701 + 6414 & 547 & 17:01:23.04 & + 64:14:11.4 & 0.453 & 2000 - 10 - 31 & c & 38.4 + rxj1716.9 + 6708 & 548 & 17:16:49.44 & + 67:08:26.9 & 0.813 & 2000 - 02 - 27 &
c & 42.5 + a2259 & 3245 & 17:20:08.64 & + 27:40:09.8 & 0.164 & 2002 - 09 - 16 & d & 7.8 + rxj1720.1 + 2638 & 1453 & 17:20:10.08 & + 26:37:30.0 & 0.164 & 1999 - 10 - 19 & b & 7.0 + rxj1720.1 + 2638 & 3224 & 17:20:10.08 & + 26:37:29.3 & 0.164 & 2002 - 10 - 03 & d & 18.4 + rxj1720.1 + 2638 & 4361 & 17:20:10.08 & + 26:37:29.3 & 0.164 & 2002 - 08 - 19 & d & 20.2 + macsj1720.2 + 3536 & 3280 & 17:20:16.56 & + 35:36:21.6 & 0.387 & 2002 - 11 - 03 & d & 17.4 + a2261 & 5007 & 17:22:27.12 & + 32:07:56.6 & 0.224 & 2004 - 01 - 14 & d & 19.7 + a2261 & 550 & 17:22:27.12 & + 32:07:57.4 & 0.224 & 1999 - 12 - 11 & b & 8.3 + a2294 & 3246 & 17:24:09.60 & + 85:53:11.0 & 0.178 & 2001 - 12 - 24 & d & 8.3
+ macsj1824.3 + 4309 & 3255 & 18:24:18.96 & + 43:09:49.0 & 0.487 & 2002 - 09 - 14 & d & 12.4 + macsj1931.8 - 2634 & 3282 & 19:31:49.68 & -26:34:32.9 & 0.352 & 2002 - 10 - 20 & d & 12.4 + rxj2011.3 - 5725 & 4995 & 20:11:27.12 & -57:25:09.8 & 0.279 & 2004 - 06 - 08 & d & 21.8 + ms2053.7 - 0449 & 1667 & 20:56:21.12 & -04:37:47.2 & 0.583 & 2001 - 10 - 07 & d & 37.1 + ms2053.7 - 0449 & 551 & 20:56:21.12 & -04:37:46.6 & 0.583 & 2000 - 05 - 13 & c & 38.4 + macsj2129.4 - 0741 & 3595 & 21:29:25.92 & -07:41:30.0 & 0.594 & 2003 - 10 - 18 & d & 16.9 + rxj2129.6 + 0005 & 552 & 21:29:40.08 & + 00:05:19.6 & 0.235 & 2000 - 10 - 21 & c & 8.5 + a2409 & 3247 & 22:00:53.04 & + 20:58:27.8 & 0.148 & 2002 - 10 - 08 & d & 9.1 + macsj2228.5 + 2036 & 3285 & 22:28:32.88 & + 20:37:11.6 & 0.412 & 2003 - 01 - 22 & d & 16.1 + macsj2229.7 - 2755 & 3286 & 22:29:45.36 & -27:55:36.5 & 0.324 & 2002 - 11 - 13 & d & 12.7 + macsj2245.0 + 2637 & 3287 & 22:45:04.80 & + 26:38:03.1 & 0.301 & 2002 - 11 - 24 & d & 12.7 + rxj2247 + 0337 & 911 & 22:47:28.08 & + 03:36:59.5 & 0.200 & 2000 - 05 - 11 & c & 41.5 + as1063 & 4966 & 22:48:44.88 & -44:31:44.4 & 0.252 & 2004 - 05 - 17 & d & 22.3 + clj2302.8 + 0844 & 918 & 23:02:48.00 & + 08:43:53.0 & 0.722 & 2000 - 08 - 05 & c & 87.9 + a2631 & 3248 & 23:37:37.92 & + 00:16:07.5 & 0.273 & 2002 - 07 - 08 & d & 8.0 + lcccc ms0015.9 + 1609 & @xmath171 & @xmath172 & 1.00 & @xmath173 + rxj0027.6 + 2616 & @xmath174 & @xmath175 & 1.00 & @xmath176 + clj0030 + 2618 & @xmath177 & @xmath178 & 1.00 & @xmath179 + a68 & @xmath180 & @xmath181 & 1.00 & @xmath182 + a115 & @xmath183 & @xmath184 & 1.00 & @xmath185 + a209 & @xmath186 & @xmath187 & 1.00 & @xmath188 + clj0152.7 - 1357s & @xmath189 & @xmath190 & 1.00 & @xmath191 + a267 & @xmath192 & @xmath193 & 1.00 & @xmath194 + clj0152.7 - 1357n & @xmath195 & @xmath196 & 1.00 & @xmath197 + macsj0159.8 - 0849 & @xmath198 & @xmath199 & 1.00 & @xmath200 + clj0216 - 1747@xmath201 & @xmath202 & @xmath203 & 1.00 & @xmath204 + rxj0232.2 - 4420 & @xmath205 & @xmath206 & 1.00 & @xmath207 + macsj0242.5 - 2132 & @xmath208 & @xmath209 & 1.00 & @xmath210 + a383 & @xmath211 & @xmath212 & 1.00 & @xmath213 + macsj0257.6 - 2209 & @xmath214 & @xmath215 & 1.00 & @xmath216 + ms0302.7 + 1658 & @xmath217 & @xmath218 & 1.00 & @xmath219 + clj0318 - 0302 & @xmath220 & @xmath221 & 1.00 & @xmath222 + macsj0329.6 - 0211 & @xmath223 & @xmath224 & 1.00 & @xmath225 + macsj0404.6 + 1109 & @xmath226 & @xmath227 & 1.00 & @xmath228 + macsj0429.6 - 0253 & @xmath229 & @xmath230 & 1.00 & @xmath231 + rxj0439.0 + 0715 & @xmath232 & @xmath233 & 1.00 & @xmath234 + rxj0439 + 0520 & @xmath235 & @xmath236 & 1.00 & @xmath237 + macsj0451.9 + 0006 & @xmath238 & @xmath239 & 1.00 & @xmath240 + a521 & @xmath232 & @xmath241 & 1.00 & @xmath242 + a520 & @xmath243 & @xmath244 & 1.00 & @xmath245 + ms0451.6 - 0305 & @xmath246 & @xmath247 & 1.00 & @xmath248 + clj0522 - 3625 & @xmath249 & @xmath250 & 1.00 & @xmath251 + clj0542.8 - 4100 & @xmath252 & @xmath253 & 1.00 & @xmath254 + macsj0647.7 + 7015 & @xmath255 & @xmath256 & 1.00 & @xmath257 + 1e0657 - 56 & @xmath258 & @xmath259 & 1.00 & @xmath260 + macsj0717.5 + 3745 & @xmath261 & @xmath262 & 1.00 & @xmath263 + a586 & @xmath264 & @xmath265 & 0.81 & @xmath266 + macsj0744.9 + 3927 & @xmath267 & @xmath268 & 1.00 & @xmath269 + a665 & @xmath243 & @xmath270 & 0.77 & @xmath271 + a697 & @xmath272 & @xmath273 & 1.00 & @xmath274 + clj0848.7 + 4456 & @xmath275 & @xmath276 & 1.00 & @xmath277 + zwclj1953 & @xmath278 & @xmath279 & 1.00 & @xmath280 + clj0853 + 5759 & @xmath217 & @xmath281 & 1.00 & @xmath282 + ms0906.5 + 1110 & @xmath283 & @xmath284 & 0.95 & @xmath285 + rxj0910 + 5422@xmath201 & @xmath286 & @xmath287 & 1.00 & @xmath204 + a773 & @xmath186 & @xmath288 & 1.00 & @xmath289 + a781 & @xmath290 & @xmath291 & 1.00 & @xmath292 + clj0926 + 1242 & @xmath293 & @xmath294 & 1.00 & @xmath295 + rbs797 & @xmath296 & @xmath297 & 1.00 & @xmath298 + macsj0949.8 + 1708 & @xmath299 & @xmath300 & 1.00 & @xmath301 + clj0956 + 4107 & @xmath302 & @xmath303 & 1.00 & @xmath304 + a907 & @xmath305 & @xmath306 & 1.00 & @xmath307 + ms1006.0 + 1202 & @xmath308 & @xmath309 & 1.00 & @xmath310 + ms1008.1 - 1224 & @xmath311 & @xmath312 & 1.00 & @xmath313 + zw3146 & @xmath171 & @xmath314 & 1.00 & @xmath315 + clj1113.1 - 2615@xmath201 & @xmath316 & @xmath317 & 1.00 & @xmath204 + a1204 & @xmath318 & @xmath199 & 1.00 & @xmath319 + clj1117 + 1745 & @xmath320 & @xmath321 & 1.00 & @xmath322 + clj1120 + 4318 & @xmath323 & @xmath324 & 1.00 & @xmath325 + rxj1121 + 2327 & @xmath326 & @xmath327 & 1.00 & @xmath328 + a1240 & @xmath192 & @xmath329 & 1.00 & @xmath330 + macsj1131.8 - 1955 & @xmath180 & @xmath331 & 1.00 & @xmath332 + ms1137.5 + 6625 & @xmath333 & @xmath199 & 1.00 & @xmath334 + macsj1149.5 + 2223 & @xmath246 & @xmath335 & 1.00 & @xmath336 + a1413 & @xmath337 & @xmath199 & 0.90 & @xmath338 + clj1213 + 0253@xmath201 & @xmath339 & @xmath340 & 1.00 & @xmath204 + clj1216 + 2633@xmath201 & @xmath341 & @xmath342 & 1.00 & @xmath204 + rxj1221 + 4918 & @xmath311 & @xmath343 & 1.00 & @xmath344 + clj1226.9 + 3332@xmath201 & @xmath345 & @xmath346 & 1.00 & @xmath204 + rxj1234.2 + 0947 & @xmath347 & @xmath348 & 0.95 & @xmath349 + rdcs1252 - 29 & @xmath350 & @xmath351 & 1.00 & @xmath352 + a1682 & @xmath353 & @xmath354 & 0.70 & @xmath355 + macsj1311.0 - 0310 & @xmath264 & @xmath356 & 1.00 & @xmath357 + a1689 & @xmath358 & @xmath294 & 1.00 & @xmath359 + rxj1317.4 + 2911 & @xmath360 & @xmath361 & 1.00 & @xmath362 + clj1334 + 5031 & @xmath363 & @xmath364 & 1.00 & @xmath365 + a1763 & @xmath366 & @xmath367 & 0.95 & @xmath368 + rxj1347.5 - 1145 & @xmath369 & @xmath370 & 1.00 & @xmath371 + rxj1350.0 + 6007 & @xmath372 & @xmath373 & 1.00 & @xmath374 + clj1354 - 0221 & @xmath375 & @xmath376 & 1.00 & @xmath377 + clj1415.1 + 3612 & @xmath378 & @xmath379 & 1.00 & @xmath380 + rxj1416 + 4446 & @xmath381 & @xmath382 & 1.00 & @xmath383 + macsj1423.8 + 2404 & @xmath384 & @xmath385 & 1.00 & @xmath386 + a1914 & @xmath318 & @xmath387 & 0.81 & @xmath388 + a1942 & @xmath267 & @xmath324 & 1.00 & @xmath389 + ms1455.0 + 2232 & @xmath186 & @xmath230 & 1.00 & @xmath390 + rxj1504 - 0248 & @xmath272 & @xmath391 & 1.00 & @xmath392 + a2034 & @xmath393 & @xmath394 & 0.81 & @xmath395 + a2069 & @xmath396 & @xmath397 & 0.74 & @xmath398 + rxj1525 + 0958 & @xmath399 & @xmath400 & 1.00 & @xmath401 + rxj1532.9 + 3021 & @xmath278 & @xmath402 & 1.00 & @xmath403 + a2111 & @xmath192 & @xmath404 & 1.00 & @xmath405 + a2125 & @xmath406 & @xmath407 & 1.00 & @xmath408 + a2163 & @xmath205 & @xmath409 & 0.63 & @xmath410 + macsj1621.3 + 3810 & @xmath278 & @xmath385 & 1.00 & @xmath411 + ms1621.5 + 2640 & @xmath345 & @xmath412 & 1.00 & @xmath413 + a2204 & @xmath414 & @xmath385 & 0.86 & @xmath415 + a2218 & @xmath186 & @xmath416 & 1.00 & @xmath417 + clj1641 + 4001 & @xmath418 & @xmath419 & 1.00 & @xmath420 + rxj1701 + 6414 & @xmath252 & @xmath327 & 1.00 & @xmath421 + rxj1716.9
+ 6708 & @xmath422 & @xmath423 & 1.00 & @xmath424 + a2259 & @xmath425 & @xmath426 & 0.95 & @xmath427 + rxj1720.1 + 2638 & @xmath428 & @xmath429 & 1.00 & @xmath274 + macsj1720.2 + 3536 & @xmath430 & @xmath233 & 1.00 & @xmath431 + a2261 & @xmath432 & @xmath433 & 1.00 & @xmath434 + a2294 & @xmath435 & @xmath436 & 0.77 & @xmath437 + macsj1824.3 + 4309 & @xmath438 & @xmath439 & 1.00 & @xmath440 + macsj1931.8 - 2634 & @xmath261 & @xmath441 & 1.00 & @xmath442 + rxj2011.3 - 5725 & @xmath443 & @xmath444 & 1.00 & @xmath445 + ms2053.7 - 0449 & @xmath446 & @xmath447 & 1.00 & @xmath448 + macsj2129.4 - 0741 & @xmath449 & @xmath450 & 1.00 & @xmath451 + rxj2129.6 + 0005 & @xmath452 & @xmath187 & 1.00 & @xmath453 + a2409 & @xmath454 & @xmath455 & 0.70 & @xmath456 + macsj2228.5 + 2036 & @xmath457 & @xmath458 & 1.00 & @xmath459 + macsj2229.7 - 2755 & @xmath460 & @xmath461 & 1.00 & @xmath462 + macsj2245.0 + 2637 & @xmath463 & @xmath294 & 1.00 & @xmath464 + rxj2247 + 0337 & @xmath465 & @xmath466 & 1.00 & @xmath467 + as1063 & @xmath232 & @xmath468 & 1.00 & @xmath469 + clj2302.8 + 0844 & @xmath470 & @xmath471 & 1.00 & @xmath472 + a2631 & @xmath473 & @xmath474 & 1.00 & @xmath475 + lp0.4cmcp1.5cmccccp1.5cmcc ms0015.9 + 1609 & 1.24 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & & @xmath480 & @xmath481 & @xmath482 & @xmath483 + rxj0027.6 + 2616 & 0.93 & @xmath484 & @xmath485 & @xmath486 & @xmath487 & & @xmath488 & @xmath489 & @xmath490 & @xmath491 + clj0030 + 2618 & 0.93 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & & @xmath496 & @xmath497 & @xmath498 & @xmath499 + a68 & 1.19 & @xmath500 & @xmath501 & @xmath502 & @xmath503 & & @xmath504 & @xmath505 & @xmath506 & @xmath507 + a115 & 1.25 & @xmath508 & @xmath509 & @xmath510 & @xmath511 & & @xmath512 & @xmath513 & @xmath514 & @xmath515 + a209 & 1.30 & @xmath512 & @xmath516 & @xmath517 & @xmath518 & & @xmath519 & @xmath520 & @xmath521 & @xmath522 + clj0152.7 - 1357s & 0.72 & @xmath492 & @xmath523 & @xmath524 & @xmath525 & & @xmath526 & @xmath527 & @xmath528 & @xmath529 + a267 & 1.04 & @xmath530 & @xmath531 & @xmath532 & @xmath533 & & @xmath534 & @xmath535 & @xmath536 & @xmath537 + clj0152.7 - 1357n & 0.78 & @xmath538 & @xmath539 & @xmath540 & @xmath541 & & @xmath542 & @xmath543 & @xmath544 & @xmath545 + macsj0159.8 - 0849 & 1.28 & @xmath546 & @xmath547 & @xmath548 & @xmath549 & & @xmath550 & @xmath551 & @xmath552 & @xmath553 + clj0216 - 1747 & 0.71 & @xmath554 & @xmath555 & @xmath556 & @xmath557 & & @xmath558 & @xmath559 & @xmath560 & @xmath561 + rxj0232.2 - 4420 & 1.34 & @xmath562 & @xmath563 & @xmath564 & @xmath565 & & @xmath566 & @xmath567 & @xmath568 & @xmath569 + macsj0242.5 - 2132 & 1.09 & @xmath570 & @xmath571 & @xmath572 & @xmath573 & & @xmath574 & @xmath575 & @xmath576 & @xmath577 + a383 & 0.98 & @xmath578 & @xmath579 & @xmath580 & @xmath581 & & @xmath582 & @xmath583 & @xmath584 & @xmath585 + macsj0257.6 - 2209 & 1.14 & @xmath586 & @xmath587 & @xmath588 & @xmath589 & & @xmath590 & @xmath591 & @xmath592 & @xmath593 + ms0302.7 + 1658 & 0.79 & @xmath594 & @xmath595 & @xmath596 & @xmath597 & & @xmath598 & @xmath599 & @xmath600 & @xmath601 + clj0318 - 0302 & 0.88 & @xmath602 & @xmath603 & @xmath604 & @xmath605 & & @xmath606 & @xmath607 & @xmath608 & @xmath609 + macsj0329.6 - 0211 & 0.99 & @xmath610 & @xmath611 & @xmath612 & @xmath613 & & @xmath614 & @xmath615 & @xmath616 & @xmath617 + macsj0404.6 + 1109 & 1.07 & @xmath618 & @xmath619 & @xmath620 & @xmath621 & & @xmath622 & @xmath623 & @xmath624 & @xmath625 + macsj0429.6 - 0253 & 1.12 & @xmath626 & @xmath627 & @xmath628 & @xmath629 & & @xmath630 & @xmath631 & @xmath632 & @xmath633 + rxj0439.0 + 0715 & 1.14 & @xmath634 & @xmath635 & @xmath636 & @xmath637 & & @xmath638 & @xmath639 & @xmath640 & @xmath641 + rxj0439 + 0520 & 0.94 & @xmath642 & @xmath643 & @xmath644 & @xmath645 & & @xmath646 & @xmath647 & @xmath648 & @xmath649 + macsj0451.9 + 0006 & 0.97 & @xmath650 & @xmath651 & @xmath652 & @xmath653 & & @xmath654 & @xmath655 & @xmath656 & @xmath657 + a521 & 1.18 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & & @xmath662 & @xmath663 & @xmath664 & @xmath665 + a520 & 1.31 & @xmath666 & @xmath667 & @xmath302 & @xmath668 & & @xmath669 & @xmath670 & @xmath671 & @xmath672 + ms0451.6 - 0305 & 1.11 & @xmath673 & @xmath674 & @xmath675 & @xmath676 & & @xmath677 & @xmath678 & @xmath679 & @xmath680 + clj0522 - 3625 & 0.77 & @xmath681 & @xmath682 & @xmath683 & @xmath684 & & @xmath685 & @xmath686 & @xmath687 & @xmath688 + clj0542.8 - 4100 & 0.90 & @xmath689 & @xmath690 & @xmath691 & @xmath692 & & @xmath693 & @xmath694 & @xmath695 & @xmath696 + macsj0647.7 + 7015 & 1.18 & @xmath697 & @xmath698 & @xmath699 & @xmath700 & & @xmath701 & @xmath702 & @xmath687 & @xmath703 + 1e0657 - 56 & 1.58 & @xmath704 & @xmath705 & @xmath706 & @xmath707 & & @xmath708 & @xmath709 & @xmath710 & @xmath711 + macsj0717.5 + 3745 & 1.36 & @xmath712 & @xmath713 & @xmath714 & @xmath715 & & @xmath716 & @xmath717 & @xmath478 & @xmath718 + a586 & 1.16 & @xmath719 & @xmath720 & @xmath721 & @xmath722 & & @xmath723 & @xmath724 & @xmath725 & @xmath726 + macsj0744.9 + 3927 & 1.04 & @xmath727 & @xmath728 & @xmath729 & @xmath730 & & @xmath731 & @xmath732 & @xmath733 & @xmath734 + a665 & 1.37 & @xmath735 & @xmath736 & @xmath737 & @xmath738 & & @xmath739 & @xmath740 & @xmath741 & @xmath742 + a697 & 1.41 & @xmath743 & @xmath744 & @xmath664 & @xmath745 & & @xmath746 & @xmath747 & @xmath748 & @xmath749 + clj0848.7 + 4456 & 0.58 & @xmath750 & @xmath751 & @xmath752 & @xmath753 & & @xmath754 & @xmath755 & @xmath624 & @xmath756 + zwclj1953 & 1.12 & @xmath757 & @xmath758 & @xmath759 & @xmath760 & & @xmath761 & @xmath762 & @xmath763 & @xmath764 + clj0853 + 5759 & 0.85 & @xmath765 & @xmath766 & @xmath767 & @xmath768 & & @xmath769 & @xmath647 & @xmath770 & @xmath771 + ms0906.5 + 1110 & 1.06 & @xmath772 & @xmath773 & @xmath774 & @xmath775 & & @xmath776 & @xmath777 & @xmath778 & @xmath779 + rxj0910 + 5422@xmath201 & 0.53 & @xmath780 & @xmath781 & @xmath540 & @xmath782 & & @xmath204 & @xmath783 & @xmath204 & @xmath784 + a773 & 1.25 & @xmath785 & @xmath786 & @xmath580 & @xmath787 & & @xmath788 & @xmath789 & @xmath790 & @xmath791 + a781 & 1.10 & @xmath792 & @xmath793 & @xmath794 & @xmath795 & & @xmath796 & @xmath797 & @xmath798 & @xmath799 + clj0926 + 1242 & 0.81 & @xmath800 & @xmath801 & @xmath802 & @xmath803 & & @xmath804 & @xmath805 & @xmath806 & @xmath807 + rbs797 & 1.13 & @xmath808 & @xmath809 & @xmath810 & @xmath811 & & @xmath812 & @xmath813 & @xmath814 & @xmath815 + macsj0949.8 + 1708 & 1.23 & @xmath816 & @xmath817 & @xmath818 & @xmath819 & & @xmath820 & @xmath821 & @xmath822 & @xmath823 + clj0956 + 4107 & 0.78 & @xmath824 & @xmath825 & @xmath826 & @xmath827 & & @xmath828 & @xmath829 & @xmath830 & @xmath831 + a907 & 1.12 & @xmath508 & @xmath832 & @xmath833 & @xmath834 & & @xmath634 & @xmath835 & @xmath217 & @xmath836 + ms1006.0 + 1202 & 1.11 & @xmath837 & @xmath838 & @xmath839 & @xmath840 & & @xmath841 & @xmath842 & @xmath843 & @xmath844 + ms1008.1 - 1224 & 1.02 & @xmath845 & @xmath619 & @xmath846 & @xmath847 & & @xmath534 & @xmath848 & @xmath849 & @xmath850 + zw3146 & 1.30 & @xmath851 & @xmath852 & @xmath853 & @xmath854 & & @xmath855 & @xmath856 & @xmath857 & @xmath858 + clj1113.1 - 2615 & 0.66 & @xmath859 & @xmath860 & @xmath861 & @xmath862 & & @xmath863 & @xmath864 & @xmath865 & @xmath866 + a1204 & 0.92 & @xmath867 & @xmath868 & @xmath372 & @xmath869 & & @xmath870 & @xmath871 & @xmath872 & @xmath873 + clj1117 + 1745 & 0.73 & @xmath874 & @xmath875 & @xmath876 & @xmath877 & & @xmath878 & @xmath879 & @xmath880 & @xmath881 + clj1120 + 4318 & 0.88 & @xmath882 & @xmath883 & @xmath884 & @xmath885 & & @xmath886 & @xmath887 & @xmath888 & @xmath889 + rxj1121 + 2327 & 0.77 & @xmath890 & @xmath891 & @xmath892 & @xmath893 & & @xmath894 & @xmath895 & @xmath896 & @xmath897 + a1240 & 0.92 & @xmath898 & @xmath899 & @xmath900 & @xmath901 & & @xmath898 & @xmath902 & @xmath903 & @xmath904 + macsj1131.8 - 1955 & 1.35 & @xmath905 & @xmath906 & @xmath907 & @xmath908 & & @xmath909 & @xmath910 & @xmath911 & @xmath912 + ms1137.5 + 6625 & 0.79 & @xmath913 & @xmath914 & @xmath915 & @xmath916 & & @xmath917 & @xmath918 & @xmath919 & @xmath920 + macsj1149.5 + 2223 & 1.22 & @xmath921 & @xmath922 & @xmath923 & @xmath924 & & @xmath925 & @xmath926 & @xmath927 & @xmath928 + a1413 & 1.26 & @xmath929 & @xmath930 & @xmath931 & @xmath932 & & @xmath933 & @xmath934 & @xmath275 & @xmath935 + clj1213 + 0253 & 0.76 & @xmath936 & @xmath937 & @xmath938 & @xmath939 & & @xmath940 & @xmath941 & @xmath942 & @xmath943 + clj1216 + 2633@xmath201 & 0.87 & @xmath944 & @xmath945 & @xmath946 & @xmath947 & & @xmath204 & @xmath948 & @xmath204 & @xmath949 + rxj1221 + 4918 & 0.86 & @xmath950 & @xmath951 & @xmath952 & @xmath953 & & @xmath954 & @xmath955 & @xmath956 & @xmath957 + clj1226.9 + 3332 & 0.94 & @xmath958 & @xmath959 & @xmath960 & @xmath961 & & @xmath962 & @xmath963 & @xmath964 & @xmath965 + rxj1234.2 + 0947 & 1.17 & @xmath966 & @xmath967 & @xmath968 & @xmath969 & & @xmath970 & @xmath971 & @xmath972 & @xmath973 + rdcs1252 - 29 & 0.52 & @xmath974 & @xmath975 & @xmath976 & @xmath977 & & @xmath978 & @xmath979 & @xmath980 & @xmath981 + a1682 & 1.13 & @xmath982 & @xmath983 & @xmath984 & @xmath985 & & @xmath986 & @xmath987 & @xmath988 & @xmath989 + macsj1311.0 - 0310 & 0.95 & @xmath634 & @xmath990 & @xmath991 & @xmath992 & & @xmath993 & @xmath994 & @xmath995 & @xmath996 + a1689 & 1.37 & @xmath997 & @xmath998 & @xmath302 & @xmath999 & & @xmath1000 & @xmath1001 & @xmath1002 & @xmath1003 + rxj1317.4 + 2911 & 0.50 & @xmath1004 & @xmath1005 & @xmath1006 & @xmath1007 & & @xmath1008 & @xmath1009 & @xmath1010 & @xmath1011 + clj1334 + 5031@xmath201 & 1.00 & @xmath1012 & @xmath1013 & @xmath1014 & @xmath1015 & & @xmath204 & @xmath1016 & @xmath204 & @xmath1017 + a1763 & 1.32 & @xmath1018 & @xmath1019 & @xmath1020 & @xmath1021 & & @xmath1022 & @xmath1023 & @xmath1024 & @xmath1025 + rxj1347.5 - 1145 & 1.45 & @xmath1026 & @xmath1027 & @xmath1028 & @xmath1029 & & @xmath1030 & @xmath1031 & @xmath1032 & @xmath1033 + rxj1350.0 + 6007@xmath201 & 0.70 & @xmath1034 & @xmath1035 & @xmath1036 & @xmath1037 & & @xmath204 & @xmath1038 & @xmath204 & @xmath1039 + clj1354 - 0221 & 0.76 & @xmath1040 & @xmath1041 & @xmath1042 & @xmath1043 & & @xmath1044 & @xmath781 & @xmath1045 & @xmath1046 + clj1415.1 + 3612 & 0.64 & @xmath1047 & @xmath1048 & @xmath1049 & @xmath1050 & & @xmath1051 & @xmath1052 & @xmath1053 & @xmath1054 + rxj1416 + 4446 & 0.86 & @xmath1055 & @xmath1056 & @xmath1057 & @xmath1058 & & @xmath1059 & @xmath1060 & @xmath1061 & @xmath1062 + macsj1423.8 + 2404 & 0.99 & @xmath1063 & @xmath1064 & @xmath1065 & @xmath1066 & & @xmath1067 & @xmath1068 & @xmath1069 & @xmath1070 + a1914 & 1.37 & @xmath1071 & @xmath1072 & @xmath275 & @xmath1073 & & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 + a1942 & 0.94 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & & @xmath1082 & @xmath1083 & @xmath1084 & @xmath1085 + ms1455.0 + 2232 & 1.04 & @xmath1086 & @xmath1087 & @xmath1088 & @xmath1089 & & @xmath1090 & @xmath1091 & @xmath1092 & @xmath1093 + rxj1504 - 0248 & 1.34 & @xmath1094 & @xmath1095 & @xmath1096 & @xmath1097 & & @xmath1098 & @xmath1099 & @xmath1100 & @xmath1101 + a2034 & 1.22 & @xmath1102 & @xmath1103 & @xmath1104 & @xmath1105 & & @xmath1106 & @xmath1107 & @xmath217 & @xmath1108 + a2069 & 1.20 & @xmath1109 & @xmath1110 & @xmath1111 & @xmath1112 & & @xmath1113 & @xmath1114 & @xmath1115 & @xmath1116 + rxj1525 + 0958 & 0.80 & @xmath1117 & @xmath1118 & @xmath1119 & @xmath1120 & & @xmath1121 & @xmath1122 & @xmath1123 & @xmath1124 + rxj1532.9 + 3021 & 1.12 & @xmath658 & @xmath1125 & @xmath421 & @xmath1126 & & @xmath1127 & @xmath1128 & @xmath1129 & @xmath1130 + a2111 & 1.18 & @xmath1131 & @xmath1132 & @xmath1133 & @xmath1134 & & @xmath1135 & @xmath1136 & @xmath1137 & @xmath1138 + a2125 & 0.77 & @xmath1139 & @xmath899 & @xmath1140 & @xmath1141 & & @xmath1142 & @xmath1143 & @xmath1144 & @xmath1145 + a2163 & 1.86 & @xmath1146 & @xmath1147 & @xmath1148 & @xmath1149 & & @xmath1150 & @xmath1151 & @xmath1152 & @xmath1153 + macsj1621.3 + 3810 & 1.01 & @xmath1154 & @xmath1155 & @xmath1156 & @xmath1157 & & @xmath1158 & @xmath1159 & @xmath1160 & @xmath1161 + ms1621.5 + 2640 & 1.03 & @xmath1158 & @xmath1162 & @xmath1163 & @xmath1164 & & @xmath1127 & @xmath1165 & @xmath1166 & @xmath1167 + a2204 & 1.37 & @xmath1168 & @xmath1169 & @xmath1170 & @xmath1171 & & @xmath586 & @xmath1172 & @xmath1173 & @xmath1174 + a2218 & 1.21 & @xmath1168 & @xmath1175 & @xmath1176 & @xmath1177 & & @xmath1109 & @xmath1178 & @xmath1179 & @xmath1180 + clj1641 + 4001 & 0.74 & @xmath1181 & @xmath1182 & @xmath1183 & @xmath1184 & & @xmath1185 & @xmath1186 & @xmath1187 & @xmath1188 + rxj1701 + 6414 & 0.93 & @xmath1189 & @xmath694 & @xmath1190 & @xmath1191 & & @xmath1192 & @xmath1122 & @xmath1193 & @xmath1194 + rxj1716.9 + 6708 & 0.78 & @xmath1195 & @xmath1196 & @xmath1197 & @xmath1198 & & @xmath1199 & @xmath1200 & @xmath1201 & @xmath1202 + a2259 & 1.09 & @xmath638 & @xmath1203 & @xmath640 & @xmath1204 & & @xmath1189 & @xmath1205 & @xmath1206 & @xmath1207 + rxj1720.1 + 2638 & 1.24 & @xmath1208 & @xmath1209 & @xmath1088 & @xmath1210 & & @xmath1211 & @xmath1212 & @xmath1213 & @xmath1214 + macsj1720.2 + 3536 & 1.16 & @xmath1215 & @xmath1216 & @xmath1217 & @xmath1218 & & @xmath1219 & @xmath1220 & @xmath1221 & @xmath1222 + a2261 & 1.31 & @xmath669 & @xmath1223 & @xmath1224 & @xmath1225 & & @xmath1226 & @xmath1227 & @xmath1228 & @xmath1229 + a2294 & 1.30 & @xmath1230 & @xmath1231 & @xmath1232 & @xmath1233 & & @xmath1234 & @xmath762 & @xmath1235 & @xmath1236 + macsj1824.3 + 4309 & 0.84 & @xmath1237 & @xmath493 & @xmath1238 & @xmath1239 & & @xmath1240 & @xmath1241 & @xmath1242 & @xmath1243 + macsj1931.8 - 2634 & 1.13 & @xmath1244 & @xmath1245 & @xmath1179 & @xmath1246 & & @xmath1247 & @xmath1248 & @xmath1249 & @xmath1250 + rxj2011.3 - 5725 & 0.85 & @xmath642 & @xmath848 & @xmath1251 & @xmath1252 & & @xmath1253 & @xmath1254 & @xmath1255 & @xmath1256 + ms2053.7 - 0449 & 0.75 & @xmath1257 & @xmath1258 & @xmath1259 & @xmath1260 & & @xmath1261 & @xmath647 & @xmath1262 & @xmath1263 + macsj2129.4 - 0741 & 1.12 & @xmath1264 & @xmath1265 & @xmath1266 & @xmath1267 & & @xmath1268 & @xmath1269 & @xmath1270 & @xmath1271 + rxj2129.6 + 0005 & 1.20 & @xmath634 & @xmath1272 & @xmath1273 & @xmath1274 & & @xmath1275 & @xmath619 & @xmath1276 & @xmath1277 + a2409 & 1.16 & @xmath1278 & @xmath1279 & @xmath1280 & @xmath1281 & & @xmath1282 & @xmath1283 & @xmath1173 & @xmath1284 + macsj2228.5 + 2036 & 1.23 & @xmath731 & @xmath1285 & @xmath1286 & @xmath1287 & & @xmath1288 & @xmath1289 & @xmath1290 & @xmath1291 + macsj2229.7 - 2755 & 1.06 & @xmath1292 & @xmath1293 & @xmath1294 & @xmath1295 & & @xmath1296 & @xmath1297 & @xmath1298 & @xmath1299 + macsj2245.0 + 2637 & 1.04 & @xmath1300 & @xmath1301 & @xmath1302 & @xmath1303 & & @xmath1304 & @xmath1305 & @xmath1306 & @xmath1307 + rxj2247 + 0337 & 0.62 & @xmath1308 & @xmath1309 & @xmath1310 & @xmath1311 & & @xmath1312 & @xmath1313 & @xmath1314 & @xmath1315 + as1063 & 1.42 & @xmath1316 & @xmath1317 & @xmath1318 & @xmath1319 & & @xmath1320 & @xmath1321 & @xmath1322 & @xmath1323 + clj2302.8 + 0844 & 0.73 & @xmath1324 & @xmath891 & @xmath1325 & @xmath1326 & & @xmath1327 & @xmath1328 & @xmath1329 & @xmath1330 + a2631 & 1.22 & @xmath1331 & @xmath1332 & @xmath1333 & @xmath1334 & & @xmath1335 & @xmath1336 & @xmath1337 & @xmath1338 + | we have assembled a sample of 115 galaxy clusters at @xmath0 with archived _ chandra _ acis - i observations .
we present x - ray images of the clusters and make available region files containing contours of the smoothed x - ray emission .
the structural properties of the clusters were investigated and we found a significant absence of relaxed clusters ( as determined by centroid shift measurements ) at @xmath1 .
the slope of the surface brightness profiles at large radii were steeper on average by @xmath2 than the slope obtained by fitting a simple @xmath3-model to the emission .
this slope was also found to be correlated with cluster temperature , with some indication that the correlation is weaker for the clusters at @xmath1 .
we measured the mean metal abundance of the cluster gas as a function of redshift and found significant evolution , with the abundances dropping by @xmath4 between @xmath5 and @xmath6 .
this evolution was still present ( although less significant ) when the cluster cores were excluded from the abundance measurements , indicating that the evolution is not solely due to the disappearance of relaxed , cool core clusters ( which are known to have enhanced core metal abundances ) from the population at @xmath7 . |
mobile devices can establish opportunistic networks a flavour of delay - tolerant networks ( dtns ) @xcite among each other in a self - organising manner and facilitate communication without the need for network infrastructure .
the capabilities of today s smart mobile devices yield substantial computing and storage resources and means for creating , viewing , archiving , manipulating , and sharing content .
in addition , content downloaded from internet is recommended to be cached at the handset @xcite .
this yields a huge reserve of data stored in mobile devices , which users may be willing to share .
while this is typically done using internet - based services , opportunistic networks enable content sharing among users in close proximity of each other . in this paper , we focus on a human - centric dtn , where nodes search for information stored in some of the nodes .
the nodes lack a global view of the network , i.e. , there is no service that could index the stored content and assist a searching node in finding content ( or indicate that the sought information does not exist ) .
this means that operations are decentralized and , as mobile nodes have resource constraints ( e.g. , energy ) , we need to control the spread of the messages when searching .
one such control mechanism is imposing the maximum number of hops a message can travel .
we call a search scheme that limits the message s path to maximum @xmath0 hops as @xmath0-hop search .
_ hop - limited search _
@xcite is of our interest as it is a lightweight scheme that does not require intensive information collection about the nodes or the content items .
however , determining the optimal @xmath0 is not straightforward .
although a large @xmath0 tends to increase replication , it also increases the chance of finding the desirable target ( content or searching node ) , thereby decreasing replication .
while many works in the literature apply hop limitations @xcite , mostly two - hop such as @xcite , to the designed routing protocols , the motivation behind setting a particular hop value is not clear . in @xcite
, we analyze the effect of hop limit on the routing performance by modelling the optimal hop limited routing as _ all hops optimal path _ problem @xcite .
however , opportunistic search necessitates considering the availability of the sought content as well as routing of the response to the searching node . in our previous work
@xcite , we analytically modelled the search utility and derived the optimal hop count for a linear network , e.g. , search flows through a single path and response messages follow the same path . in this paper , we provide an elaborate analysis of hop - limited search in a mobile opportunistic network considering the search success , completion time , and the cost .
search is a two - phase process .
we refer to the first phase in which query is routed towards the content providers as _ forward path _ and the second phase in which response is routed towards the searching node as _
return path_. first , we present an analysis on the forward path via an analytical model .
we show the interplay between tolerated waiting time ( how long the searching node can wait for the forward path ) , content availability , and the hop count providing the maximum search success ratio for a specific setting .
next , we verify our analysis of the forward path and elaborate also on the return path via simulations .
our results suggest the following : * generally speaking , search performance increases with increasing @xmath0 especially for scarcely available content .
however , the highest improvement is often achieved at the second hop .
after that , the improvements become smaller and practically diminish when @xmath1 . *
we observe that return path requires on average longer hops / time compared to the forward path , which may imply that optimal settings for the forward path does not yield the optimal performance on the return path .
* we show that the search cost first increases and after several hops ( @xmath2 ) it tends to stabilise .
this is due to the small diameter of the network ; even if @xmath0 is large letting the nodes replicate a message to other nodes , each node gets informed about the search status quickly so that replication of obsolete messages is stopped .
the rest of the paper is organised as follows .
section [ sec : sysmodel ] introduces the considered system model followed by section [ sec : hoplimited ] introducing the hop - limited search .
section [ sec : numericalanalysis ] numerically analyzes the forward path , while section [ sec : sims ] focuses on the whole search process .
section [ sec : related ] overviews the related work and highlights the points that distinguish our work from the others .
finally , section [ sec : conclusions ] concludes the paper .
we consider a mobile opportunistic network of @xmath3 nodes as in fig .
[ fig : model ] .
nodes move according to a mobility model which results in i.i.d .
meeting rates between any two nodes .
we use the following terminology : * searching node ( @xmath4 ) * is a node that initiates the search for a content item . we assume that content items are equally likely to be sought , i.e. , uniform _
content popularity_. in reality , content items have diverse popularities ; e.g. , youtube video popularity follows a zipf distribution @xcite .
we choose uniform distribution to avoid a bias toward the `` easy '' searches .
* tagged node * is a node that holds a copy of the sought content .
only a fraction @xmath5 of the nodes are tagged , where @xmath5 is referred to as the _ content availability_. although it is expected that search dynamics such as caching changes @xmath6 , we assume that it does not change over time .
the sets of all and tagged nodes are denoted by @xmath7 and @xmath8 , and their sizes @xmath9 and @xmath10 , respectively . the number of tagged nodes is @xmath11 .
every node is equally likely to be a tagged node .
* forward path and return path : * in the first step of the search , a query is disseminated in the network to find a tagged node . in case
the first step is completed , a response generated by a content provider is routed back towards @xmath4 in the second step .
we refer to the former as the _ forward _ or _ query path _ and to the latter as the _ return _ or _ response path_. * tolerated waiting time ( @xmath12 ) * is the maximum duration to find the content ( excluding the return path ) . note that total tolerated waiting time is @xmath13 if we assume the same delay restriction for the return path .
to describe the basic operation of hop - limited search , assume that @xmath4 creates a query that includes information about its search at time @xmath14 .
when @xmath4 encounters another node that does not have this query , @xmath4 replicates the query to it to reach a tagged node faster .
a node acquiring a copy of the message becomes a _
discovered node_. the discovered node also starts to search and replicate the query to _
undiscovered nodes_. each message contains a header , referred to as _ hop count _ , representing the number of nodes that forwarded this message so far .
messages at @xmath4 are 0-hop messages ; those received directly from @xmath4 are @xmath15-hop messages .
we refer to a search scheme as @xmath0-hop search if a search message can travel at most @xmath0 hops .
hence , when a node has @xmath0-hop message , it does not forward it further .
we also call a node with a @xmath16-hop message as the @xmath16-hop node for that particular message . in @xmath0-hop search , when an @xmath16-hop and @xmath17-hop node meet ( without loss of generality we assume @xmath18 ) , the state of @xmath17-hop node is updated to @xmath19-hop if @xmath20 .
the forward path completes when a copy of the query reaches a tagged node .
let us first consider a search scheme that does not put any hop limitation but instead limits the total number of replications on the forward path and the return path to @xmath21 and @xmath22 , respectively . for a content item with availability @xmath6 , we can calculate the forward success ratio of this search scheme as : @xmath23 similarly , we calculate the search success ratio , i.e. , both steps are completed , as : @xmath24 we can expand the above formulation which leads to : @xmath25 let @xmath26 , i.e. , the probability that a response reaches @xmath4 . after replacing @xmath27 into ( [ eq : ps ] ) ,
we apply some manipulations by the help of binomial theorem : that is @xmath28 then , we find the search success as : @xmath29 please refer to appendix [ appendix ] for the details of the above derivation .
[ fig : search_success_m ] plots the success ratio with increasing fraction of nodes that receive the message .
we plot the success of both the forward path and the total search under various content availability values .
the results are for equal number of replications for the forward and return path , i.e. , @xmath30 .
the figure shows that to ensure a desirable level of success , search has to cover certain fraction of nodes , which depends on the content availability . in hop - limited search
, there is no explicit restriction on number of replications @xmath21 , but rather it is implicitly set by @xmath0 and @xmath12 .
this result brings us to the question of how search coverage in the number of nodes changes with @xmath0 and @xmath12 .
let @xmath31 be the set of discovered nodes excluding @xmath4 at time @xmath32 and @xmath33 its size .
et al . _
@xcite define @xmath34 for a static graph as node neighborhood within @xmath0 hops .
similarly , we define @xmath35 as the number of nodes that can be reached from a source node less than or equal to @xmath0 hops under time limitation @xmath12 . for @xmath36.,scaledwidth=39.0%
] let @xmath37 denote the _ forward path success ratio _ defined as the probability that a query reaches one of the tagged nodes in a given time period @xmath12 under @xmath0-hop limitation . for a search that seeks a content item with availability @xmath5 , we approximate @xmath38 as follows : @xmath39
\nonumber\\ & \approx 1-(1-\alpha)^{e[{n}_{h}(t ) ] } \label{eq : pcompletion}.\end{aligned}\ ] ] note that ( [ eq : pcompletion ] ) provides an upper bound for @xmath37 and will only be used to understand the effect of @xmath0 and @xmath12 . in numerical evaluations , we relax all the simplifications ( e.g. , i.i.d meeting rates ) and experiment using real mobility traces
. given ( [ eq : pcompletion ] ) , our problem reduces to discovering how @xmath40 $ ] changes with @xmath0 and @xmath12 .
however , as observed in @xcite , each meeting may not lead to a new node being discovered , i.e. , some nodes may meet again or a node may be met by several nodes .
therefore , total meetings in time period @xmath12 results in a very optimistic estimate for @xmath40 $ ] .
in addition to the overlaps of opportunistic contacts , the hop restriction may result in lower @xmath40 $ ] .
in contrast to static networks @xcite , modelling @xmath40 $ ] for general time - evolving networks is not straightforward .
therefore , we derive @xmath40 $ ] from mobility traces and plug it into ( [ eq : pcompletion ] ) . see
also @xcite for an analysis of how mobility and node density affect the communication capacity of a mobile opportunistic network .
we define the effect of increasing hop count from @xmath0 to @xmath41 on the search performance as the _ added benefit _ and calculate it as the difference between @xmath37 and @xmath42 : @xmath43 , where @xmath38 values are either approximated as above or obtained from experiments . given that users can perceive only significant improvements rather than minimal changes in performance , we are more interested in larger values @xmath44 .
let @xmath45 denote the hop count beyond which the added benefit of one additional hop is lower than @xmath46 .
more formally , @xmath45 is defined as : @xmath47 while @xmath48 yields the optimal hop count @xmath49 achieving the highest performance , i.e. , @xmath50 , we search for @xmath0 values that lead to higher @xmath46 .
setting @xmath48 , we assess until which hop there is still some gain , albeit minimal , whereas higher @xmath46 values help discovering the last hop with substantial benefit .
we refer to these @xmath46 values as _ any - benefit _ and _ fair - benefit _ hop , respectively . in this section , we derive the _ forward path completion time _ which is the time required for an initiated query to reach one of the content providers . let us denote by @xmath51 the total number of @xmath16-hop nodes at time @xmath32 .
these nodes have received the message via ( @xmath52 ) relays .
subsequently , we denote the state of a continuous time markov chain ( ctmc ) by : @xmath53 where * @xmath54 is the number of nodes with a copy of the query ( @xmath4 and all discovered nodes ) and * @xmath51 nodes are @xmath16-hop nodes and @xmath55 .
states that have @xmath56 nodes holding the search query are represented as @xmath57 .
the state space consists of all @xmath58 transient states where @xmath59 and the absorbing state @xmath60 , which represents discovery of a tagged node .
hereafter , we drop the time parameter for clarity . from state @xmath57 ,
three types of events trigger state transitions : * _ meeting a tagged node _ : an arbitrary @xmath16-hop node in @xmath61 where @xmath62 meets with one of the tagged nodes in @xmath8 .
the resulting state is @xmath60 and the search ends .
there is only one such transition from every state . *
_ meeting a node that is undiscovered and not tagged _ : an @xmath16-hop node meets an undiscovered untagged node .
there are @xmath63}$ ] such transitions from this state where indicator function @xmath64}$ ] yields 1
if @xmath65 evaluates to _ true _ , and the resulting state is @xmath66 .
* _ meeting among discovered nodes _ : we call a meeting between @xmath16-hop and @xmath17-hop node an _ @xmath67-meeting _ where @xmath68 . if @xmath20 , @xmath34 does not change but @xmath69 and @xmath70 change . the new state is @xmath71 .
the number of such transitions from a state @xmath57 is @xmath72}\left(\sum_{j = i+2}^{h}1_{[m_j>0]}\right).$ ] we call these three events _ type - tagged _ , _ type-1 _ , and _ type-0 _ events ; the corresponding transition rates are denoted by @xmath73 , @xmath74 , and @xmath75 , respectively .
if the type-1 event is a meeting with an @xmath16-hop node , we denote the respective rate as @xmath76 .
similarly , if a type-0 event is due to an @xmath67-meeting , the rate is @xmath77 . for state
@xmath78 the transitions leading to a state change are : @xmath79 and the corresponding transition rates are : @xmath80 where @xmath81 is the pairwise meeting rate . since solving the given markov model may not be practical due to the state space explosion for large @xmath0 and @xmath9 , we approximate @xmath82 and show its high accuracy in section [ sec : numericalanalysis ] by comparing it with the results of markov model .
let @xmath83 denote the remaining time to search completion under @xmath0-hop search and when @xmath56 nodes hold the search query . under @xmath0-hop search
, only @xmath84 nodes are actively searching and can forward the query to their encounters .
assume that @xmath85 are identical for all @xmath86 .
then , the number of searching nodes @xmath87 is : @xmath88 we calculate @xmath83 as : @xmath89 we expand @xmath90 similarly and substitute in ( [ eq : tmh ] ) . after a certain number of nodes
are searching , the remaining time to search completion converges to zero , i.e. , @xmath91 .
denote @xmath92 .
then , we find : @xmath93 solving for @xmath94 gives an approximation for @xmath82 : @xmath95
in this section , we evaluate the performance of hop - limited search ( referred to as hop ) while varying ( i ) content availability , ( ii ) tolerated waiting time , ( iii ) network density , and ( iv ) mobility scenarios .
we use @xmath96 for _ high _ , _ medium _ , and _ low _ content availability , respectively . in our analysis , we use real traces of both humans and vehicles . for the former
, we use the infocom06 dataset @xcite which represents the traces of human contacts during infocom 2006 conference . for the latter ,
we use the cabspotting dataset @xcite that stores the gps records of the cabs in san francisco . to gain insights about more general network settings ,
we also analyze a synthetic mobility model that reflects realistic movement patterns in an urban scenario .
below , we overview the basic properties of each trace : * infocom06 : * this data set records opportunistic bluetooth contacts of 78 conference participants who were carrying imotes and 20 static imotes for the duration of the conference , i.e. , approximately four days . in our analysis
, we treated all devices as identical nodes which host content items and initiates search queries .
this trace represents a small network in which people move in a closed region .
* cabspotting : * this data set records the latitude and longitude information of a cab as well as its occupancy state and time stamp .
the trace consists of updates of the 536 cabs moving in a large city area for a duration of 30 days . for our analysis
, we focused only on the first three days and a small region of approximately 10km@xmath9710 km area .
496 cabs appear in this region during the specific time period .
the cab information is not updated at regular intervals .
hence , we interpolated the gps data so as to have an update at every 10 s for each cab
. next , we set transmission range to 40 m to generate contacts among cabs .
this trace represents an urban mobility scenario @xcite .
* helsinki city scenario ( hcs ) : * this setting represents an opportunistic human network in which the walking speed is uniformly distributed in [ 0.5,1.5]m / s .
the nodes move in a closed area of [email protected] km .
hcs @xcite uses the downtown helsinki map populated with _
points - of - interests _ ( e.g. , shops ) , between which pedestrians move along shortest path .
we derive the neighborhood size @xmath35 as an average of 500 samples where each sample represents an independent observation of the network starting at some random time , from an arbitrary node and spanning an observation window equal to the tolerated waiting time .
next , we calculate @xmath37 using ( [ eq : pcompletion ] ) .
we use r @xcite for these analysis . in the following ,
we mostly focus on the more challenging cases such as short @xmath12 or low @xmath6 , and report the representative results due to the space limitations .
[ fig : neighborhood ] illustrates the change in @xmath98 represented as fraction of the network size for infocom06 for various tolerated waiting time @xmath12 and content availability @xmath6 . for each @xmath12 , we plot @xmath98 and corresponding @xmath37 . from fig .
[ fig : infocom06nodeunionbyhop_98 ] , we can see the significant growth of @xmath98 at the second hop for all settings .
while further hops introduce some improvements , we observe the existence of a _ saturation point _ @xcite .
after this point , the change in @xmath99 is marginal either because all nodes are already covered in the neighborhood or higher @xmath0 can not help anymore without increasing @xmath12 .
we present the resulting @xmath100 for low content availability in fig.[fig : infocom06_ptagged_98_t05 ] . as expected
, the second hop provides the highest performance gain compared to the previous hop ( @xmath44 ) as a reflection of highest @xmath101 .
as infocom06 has good connectivity , almost all queries reach one of the tagged nodes after @xmath2 .
we present the effect of content availability for short @xmath12 in fig .
[ fig : ptagged_98_t600 ] . regarding the effect of content availability ,
_ we observe that search for a rare content item , i.e. , low @xmath6 , benefits more from increasing @xmath0 compared to highly available content items_. when many nodes are tagged , @xmath4 meets one of these after some time .
however , if the probability of meeting a tagged node is fairly low , using additional hops exploiting the mobility of the encountered nodes and spreading the query further is a better way of searching .
a smart search algorithm can keep track of the content availability via message exchanges during encounters and can adjust the hop count depending on the observed availability of the content . for low and medium content availability , _
the highest benefit is obtained at the second hop_. for a content item which is stored by a significant fraction of nodes , even a single hop search may retrieve the sought content .
[ fig : density_vs_neighborhood ] illustrates the growth of @xmath102 under various network size @xmath9 and @xmath12 .
we use our synthetic model hcs by setting @xmath103 nodes to observe the effect of network density on @xmath102 . for this particular
setting , the clustering of lines based on @xmath12 shows that time restriction is more dominant factor in determining @xmath102 compared to @xmath9 . as expected
, @xmath102 is higher for higher @xmath9 .
however , all settings exhibit the same growth trend with increasing @xmath0 . fig . [
fig : allbeta ] illustrates the change in @xmath45 with increasing @xmath12 for @xmath104 which represent any - benefit and fair - benefit hops , respectively . as stated before
, infocom06 has good connectivity among nodes as conference takes place in a closed area and nodes share the same schedule ( e.g. , coffee breaks and sessions ) . as a result of this good connectivity ,
first one or two hops provide almost all the benefits of multi - hop search ( fig .
[ fig : infocomnobenefitpoints_98_benefit5e-04fair ] ) .
however , fig.[fig : infocomnobenefitpoints_98_benefit5e-04any ] shows that the optimal hop in terms of the highest @xmath100 is achieved at higher @xmath105 . for hcs scenario that represents a network of urban scale ,
the effect of increasing @xmath12 is a bit different .
[ fig : helsinkifair ] shows that short and longer @xmath12 may have the same operating point in terms of @xmath45 . for low @xmath12 , the benefit of increasing @xmath0 diminishes due to the limited number of encounters and the resulting small set of discovered nodes , whereas for long @xmath12 almost all nodes are discovered without requiring any further hops . the lower @xmath45 with increasing @xmath12
can also be explained by `` _ _ shrinking diameter _ _ '' phenomenon , which states that the average distance between two nodes decreases over time as networks grow @xcite .
indeed , the diameter of the message search tree gets shorter over time and leads to a smaller hop distance between the searching node and a tagged node .
note the decreasing _ any - benefit _ hop in fig .
[ fig : helsinkiany ] .
for example , searching for a content item with medium availability achieves the highest performance at @xmath106 for @xmath107 s , while it decreases gradually to @xmath108 with increasing @xmath12 .
table [ tab : searchtime_markovchain ] shows the search time @xmath82 which is obtained by solving ctmc introduced in section [ sec : ttsearchcompletion ] and @xmath109 given by ( [ eq : tmhapproxm1 ] ) .
we normalize every value by the maximum search time of each setting . in the table
, we also list the approximation errors .
first thing to note is the drastic decrease in search time at the second hop . in agreement with our conclusions from @xmath100
, we see that second hop speeds the search significantly resulting in approximately @xmath110 shorter search time for the low availability , @xmath111 for the medium , and @xmath112 decrease for the high availability scenario . as @xmath100 ,
the most significant improvement in search time occurs for low content availability .
our approximation also exhibits exactly the same behaviour in terms of the change in the search time .
as the error row shows , it deviates from the expected search time to some degree : @xmath113 under - estimation to @xmath114 overestimation . [ ! htb ] .search time and its approximation . [ cols="<,<,^,^,^,^,^",options="header " , ] [ tab : correlationanalysis ] table [ tab : correlationanalysis ] summarises this analysis for infocom06 , which also agrees with the analysis for cabspotting scenario .
first , we do not observe any strong correlation between return and forward path lengths for any of the settings ( i.e. , @xmath115 and @xmath116 are around 0.10.4 ) .
this result may be conflicting with the intuition that the information found in nearby / far - away will also be routed back quickly / slowly . however , search process is more complicated due to the intertwined effects of mobility and restrictions on the total search time .
for example , consider a forward path with a very large number of hops . due to the remaining short time before the tolerated waiting time expires , search can only go a few hops towards the searching node .
this leads to a long forward path with a very short return path which challenges the above - mentioned intuition . with higher content availability
, the required number of forward hops decreases whereas the return path length seems to be barely affected .
hence , the corresponding @xmath117 and @xmath118 increase . for all scenarios ,
return path is on the average longer than the forward path .
using this observation , a tagged node receiving a query can set the time - to - live field of its response message longer than the received query s forward path time . obviously , increasing hop count increases the neighborhood which in turn increases the chance of finding the sought content .
however , the larger neighborhood should also be interpreted as larger number of replication , i.e. , higher search cost . in fact , the neighborhood size represents the upper bound of number of replications for a search message if no other search stopping algorithm is in effect . in this section
, we evaluate the cost of search considering two simple mechanisms that aim to keep replication much below the upper bound set by @xmath102 .
we consider three cases : ( i ) oracle : there is a central entity from which a node retrieves the global state of a message in its buffer and can ignore it if outdated , e.g. a completed search or a query already reaching a tagged node , ( ii ) exch : upon encounter , nodes exchange their local knowledge about search activities in the network , ( iii ) local : nodes exchange their knowledge _ only _ on the shared messages .
note that an oracle can be an entity in the cellular network and nodes can access it via a control channel .
in fact , this communication with the cellular network is more costly ( e.g. , energy ) compared to the opportunistic communication .
nevertheless , this scenario serves as an optimal benchmark to assess the performance of the other scenarios .
exch pro - actively spreads information about existing queries to all nodes opportunistically , which may be considered as leaking information to nodes not involved in search .
local circumvents this by only using its local knowledge and sharing information with peers only about queries that the other node has already seen .
[ fig : costoraclevsnooracle ] shows _ query spread ratio _ which is defined as i.e. , the fraction of nodes that have seen this query , and search time .
first , we should note that the resulting search success ( not plotted ) is almost the same under all schemes .
second , note the non - increasing query spread ratio for @xmath1 .
this result confirms that the network is a small - world network where all nodes get informed quickly about the search status and drop the outdated messages timely .
hence , even for larger @xmath0 , nodes detect the outdated messages via local and shared knowledge . in fig .
[ fig : queryspreadratio15_600_015 ] , we observe that exch maintains the same performance as oracle , whereas local results in more replication as fewer nodes are informed about the completed queries / responses . nevertheless , because of the small network diameter , a higher @xmath0 does not result in an explosion of query spread in the network .
regarding search time , we observe substantial decrease in search time also for @xmath119 and @xmath120 in contrast to the vanishing benefits after second hop derived from our analytical model ( table [ tab : searchtime_markovchain ] ) .
search time tends to stabilise after @xmath121 .
hence , although several hops are sufficient in terms of search success , further hops ( e.g. , @xmath2 ) can be considered for faster search .
our analysis shows that the two factors affecting the optimal hop count are the content availability ( @xmath5 ) and the product of meeting rate and tolerated waiting time ( @xmath122 ) .
the latter is the number of meetings before tolerated waiting time expires where @xmath81 depends on network density and mobility model .
if both @xmath5 and @xmath122 are low ( scarce content and very few contacts due to network sparsity or short search time ) , search performance is expected to be low .
however , it increases with increase in either of these factors . if one of these factors is large , it is sufficient to have a low hop count limit ( e.g. , two or three ) to obtain good performance .
that is , when @xmath123 is large relative to the expected number of tagged nodes met during the search time , limiting the search to a few hops still achieves good results . as @xmath123 decreases , the required hop limit to maintain good search performance is larger .
this allows devising adaptation when issuing search queries .
nodes can monitor the request and response rate for certain ( types of ) content items and thus infer popularity and availability in their area .
they can also assess the regional node density and meeting rate @xmath81 and thus determine the required hop count @xmath0 given @xmath12 or vice versa .
moreover , nodes can monitor search performance and determine how well their ( region in a ) network operates . to improve performance
, nodes can decide to increase the availability of selected contents via active replication of the scarce content , obviously trading off storage capacity and link capacity for availability .
such decision could be based entirely upon local observations , but could also consider limited information exchange with other nodes .
we believe that with the guidance of our analysis , a node can decide on the best hop count depending on the network density and the content availability that can both be derived from past observations by the node .
although we show that search cost stops increasing after a few hops , keeping @xmath0 low may be desirable if we interpret it as a measure of the social relation between two nodes ( e.g. , one hop as friends ; two hop as friend - of - friend ) . in other words , lower
@xmath0 can be interpreted as more _ trustworthy _ operation .
moreover , protocols involving lower number of relays are more scalable and energy - efficient @xcite .
while the dtn literature has many proposals for message dissemination , which exploit the information about the network such as ( estimated ) pairwise node contact rates , node centrality , communities , and social ties , efficient mechanisms for content search remain largely unexplored . in a sense , this is reasonable as search can be considered as a two - step message delivery : query routing on the forward path and the response routing on the return path .
the forward path is less certain as the content providers are unknown . in this regard ,
the return path is less challenging as the target node ( and a recent path to reach it ) is already known .
however , routing the response looks for a particular node , whereas the forward search is for a _ subset of nodes _ whose cardinality is proportional to the content availability .
thus , search requires special treatment rather than being an extension of the message dissemination .
two questions for an efficient search are ( i ) which nodes may have the content @xcite and ( ii ) when to stop a search @xcite
. the former question requires assessment of each node in terms of its potential of being a provider for the sought content .
for example , _ seeker assisted search _ ( sas ) @xcite estimates the nodes in the same community to have higher likelihood of holding the content as people in the same community might have already retrieved the content .
given that people sharing a common interest come together at a certain space " ( e.g. , office , gyms ) , @xcite defines _ geo - community _ concept and matches each query with a particular geo - community .
hence , the first question boils down to selecting relays with high probability of visiting the target geo - community . as @xcite aims to keep the search cost minimal , searching node employs two - hop routing and determines the relays , which thereby reduces the issue of when to stop the search .
deciding when to stop search is nontrivial as the search follows several paths and whether the searching node has already discovered a response is not known by the relaying nodes .
et al . _
@xcite model the expected search utility with increasing hop count and then finds the optimal hop , while @xcite estimates the number of nodes having received a query and possible responses by using the node degrees . in our work
, we showed that simple schemes via information sharing can stop search timely and maintain similar performance to that of an oracle due to the small world nature of the studied networks .
different from all these above works , main focus of our paper is more on a fundamental question : _ how does the hop limitation affect the search performance ? _
while this question has been explored in general networking context , content - centricity and the time constraints require a better understanding of flooding in the context of search .
therefore , we first provided insights on search on a simplified setting and next analyzed the effect of various parameters , e.g. , time and real mobility traces , via extensive simulations .
in the literature , several works focus on two - hop forwarding in which the source node replicates the message to any relay and the relays can deliver the message only to the final destination @xcite .
et al . _ in their seminal work @xcite show that two hops are sufficient to achieve the maximum throughput capacity in an ideal network with nodes moving randomly .
the capacity increase is facilitated by the reduced interference on the links from source to relay and relay to the destination .
et al . _
@xcite assess this two - hop forwarding scheme in a dtn scenario with power - law inter - contact times and employ an _ oblivious forwarding algorithm _
( e.g. , memoryless routers that do not use any context information such as contact history ) .
similarly , @xcite focuses on a dtn with power - law distributed inter - contact times and derives the conditions ( i.e. range of pareto shape parameter ) under which message delivery has a finite delay bound for both two - hop and multi - hop oblivious algorithms .
the authors show that `` _ _ as long as the convergence of message delivery delay is the only concern , paths longer than two - hops do not help convergence _ _ '' as two hops are sufficient to explore the relaying diversity of the network @xcite .
another work supporting two - hop schemes is @xcite which shows that two - hop search is favourable for opportunistic networks a resource - scarce setting , as
_ one - hop neighbors are able to cover the most of the network in a reasonable time _ " in a network with _
sufficiently many _ mobile nodes . in our work
, we include those cases when longer paths still yield ( some ) performance improvement .
finally , our work supports the conclusion of @xcite , which theoretically proves the _ small - world _ in human mobile networks .
our work is closely related to the @xmath124-hop flooding @xcite which models the spread of flooded messages in a random graph .
unlike @xcite , we focus on content search in a realistic setting , and using real mobility traces ( both a human contact network and a vehicular network ) we provide insights on the effect of increasing hop count on the search success , delay , and cost under various content availability and tolerated waiting time settings . despite the differences , it is worthwhile noting that our results agree with the basic conclusions of @xcite .
given the volume of the content created , downloaded , and stored in the mobile devices , efficient opportunistic search is paramount to make the remote content accessible to a mobile user .
current schemes mostly rely on routing based on hop limitations . to provide insights about the basics of such a generic search scheme
, we focused on a hop - limited search in mobile opportunistic networks . first , by modelling only the forward path , we showed that ( i ) the second hop and following few hops bring the highest gains in terms of forward path success ratio and ( ii ) compared to single - hop delivery , increase in hop count leads to shorter search time and after a few hops search time tends to stabilize .
next , we revisited these findings via simulations of the entire search process .
while simulations validated our claim for the forward path , we observed that return path on average requires longer time and more hops .
moreover , our results do not indicate strong correlation between the return and forward paths .
finally , we showed that search completes in less than five hops in most cases .
this is attributed to the small diameter of the human contact network which has also a positive impact on search cost ; nodes are informed about search state quickly and stop propagation of obsolete messages .
our simulations validated that increasing hop count to several hops accelerates the search and later search completion time stabilizes .
as future work , we will design a search scheme that adapts the hop count of query and response paths based on the observed content availability and popularity .
moreover , we believe that content - centric approaches should be paid more attention to implement efficient search schemes in mobile opportunistic networks .
this work was supported by the academy of finland in the pdp project ( grant no .
260014 ) .
10 l. a. adamic , r. m. lukose , and b. a. huberman .
local search in unstructured networks . , 2006 .
a. al hanbali , p. nain , and e. altman . performance of ad hoc networks with two - hop relay routing and limited packet lifetime .
, 65(6):463483 , 2008 .
s. bayhan , e. hyyti , j. kangasharju , and j. ott .
seeker - assisted information search in mobile clouds . in _
acm sigcomm workshop on mobile cloud computing _ , 2013 .
s. bayhan , e. hyyti , j. kangasharju , and j. ott .
analysis of hop limit in opportunistic networks by static and time - aggregated graphs . in _ ieee icc _ , 2015 .
c. boldrini , m. conti , and a. passarella .
less is more : long paths do not help the convergence of social - oblivious forwarding in opportunistic networks . in _ acm int .
workshop on mobile opportunistic networks _ , 2012 .
a. chaintreau , p. hui , j. crowcroft , c. diot , r. gass , and j. scott .
impact of human mobility on opportunistic forwarding algorithms .
, 6(6):606620 , 2007 .
a. chaintreau , a. mtibaa , l. massoulie , and c. diot .
the diameter of opportunistic mobile networks . in _
acm conext _ , 2007 .
k. fall . a delay - tolerant network architecture for challenged internets . in _ acm sigcomm _ , 2003 .
m. faloutsos , p. faloutsos , and c. faloutsos . on power - law relationships of the internet topology . in _
acm sigcomm computer communication review _ , volume 29 , pages 251262 , 1999 .
j. fan , j. chen , y. du , p. wang , and y. sun .
delque : a socially aware delegation query scheme in delay - tolerant networks . ,
60(5):21812193 , jun 2011 .
w. gao , q. li , and g. cao .
forwarding redundancy in opportunistic mobile networks : investigation , elimination , and exploitation . , 2014 .
p. gill , m. arlitt , z. li , and a. mahanti .
youtube traffic characterization : a view from the edge . in _
acm sigcomm conf . on internet measurement _ , 2007 .
m. grossglauser and d. tse .
mobility increases the capacity of ad hoc wireless networks .
, 10(4):477486 , aug 2002 .
r. gurin and a. orda .
computing shortest paths for any number of hops .
, 10(5):613620 , 2002 .
m. a. hoque , x. hong , and b. dixon .
efficient multi - hop connectivity analysis in urban vehicular networks .
, 1(2):7890 , april 2014 .
e. hyyti , s. bayhan , j. ott , and j. kangasharju .
searching a needle in ( linear ) opportunistic networks . in _
acm mswim _ , 2014 .
e. hyytia and j. ott .
criticality of large delay tolerant networks via directed continuum percolation in space - time . in _
ieee infocom _ , 2013 .
a. kernen , j. ott , and t. krkkinen . .
in _ proc . of int .
conf . on simulation tools and techniques _ , 2009 .
j. leskovec , j. kleinberg , and c. faloutsos .
graphs over time : densification laws , shrinking diameters and possible explanations . in _
acm kdd _ ,
j. ott and m. pitknen . .
in _ the first ieee wowmom workshop on autonomic and opportunistic communications ( aoc ) _ , june 2007 .
m. piorkowski , n. sarafijanovic - djukic , and m. grossglauser .
data set epfl / mobility ( v. 2009 - 02 - 24 ) , feb . 2009 .
downloaded from http://crawdad.org / epfl / mobility/. m. pitknen , t. karkkainen , j. greifenberg , and j. ott .
searching for content in mobile dtns . in _
ieee percom _ , 2009 .
f. qian , k. s. quah , j. huang , j. erman , a. gerber , z. mao , s. sen , and o. spatscheck .
web caching on smartphones : ideal vs. reality . in _
acm mobisys _ , 2012 . .
r foundation for statistical computing , vienna , austria , 2013 .
j. scott , r. gass , j. crowcroft , p. hui , c. diot , and a. chaintreau .
data set cambridge / haggle ( v. 2006 - 01 - 31 ) .
http://crawdad.org/cambridge/haggle/ , jan .
m. vojnovic and a. proutiere .
hop limited flooding over dynamic networks . in _
ieee infocom _ , 2011 .
given that the message is received by @xmath21 nodes and each node has probability @xmath6 of holding the requested content , the probability that an initiated query reaches one of the content provider(s ) equals to the probability that at least one of the @xmath21 nodes has the content .
we denote then the forward success ratio as : @xmath125 if we consider the whole search path with the assumption that @xmath21 nodes at maximum holds the query and only @xmath126 copies are allowed for the response message , we calculate the search success ratio , i.e. , both steps are completed , as : @xmath127 the first factor in ( [ eq : ps_appendix ] ) , corresponding to the event that @xmath124 content providers are discovered , obeys binomial distribution with parameters @xmath128 .
the second factor is conditioned on @xmath124 , the number of content providers reached by the query replicas . out of these @xmath124 responses
, we calculate the probability that at least one of them reaches the destination , i.e. , the searching node . with the assumption that @xmath126 copies of the response is allowed on the return path , to calculate the probability of finding @xmath4 , we find the probability that none of the @xmath124 responses reaches @xmath4 .
since there are @xmath129 nodes a response message created by a particular content provider can reach and @xmath126 selection without replication , the probability that a response reaches @xmath4 equals to : @xmath130 substituting these formulations into ( [ eq : ps_appendix ] ) , we find : @xmath131 separating ( [ eq : app_0 ] ) into two parts , we find : @xmath132 next , we notice that each summation term above can be compactly written by the help of the binomial theorem , which is : @xmath133 taking into account that the summation in ( [ eq : app_1 ] ) starts from 1 , we simplify the first term in ( [ eq : app_1 ] ) as follows : @xmath134 applying the same expansion for the second term in ( [ eq : app_1 ] ) , we find : @xmath135 substituting ( [ eq : app_first_term ] ) and ( [ eq : app_second_term ] ) into ( [ eq : app_1 ] ) , we find : @xmath136 which gives us the search success probability : @xmath137 | while there is a drastic shift from host - centric networking to content - centric networking , how to locate and retrieve the relevant content efficiently , especially in a mobile network , is still an open question .
mobile devices host increasing volume of data which could be shared with the nearby nodes in a multi - hop fashion .
however , searching for content in this resource - restricted setting is not trivial due to the lack of a content index , as well as , desire for keeping the search cost low . in this paper , we analyze a lightweight search scheme , _ hop - limited search _ , that forwards the search messages only till a maximum number of hops , and requires no prior knowledge about the network .
we highlight the effect of the hop limit on both search performance ( i.e. , success ratio and delay ) and associated cost along with the interplay between content availability , tolerated waiting time , network density , and mobility .
our analysis , using the real mobility traces , as well as synthetic models , shows that the most substantial benefit is achieved at the first few hops and that after several hops the extra gain diminishes as a function of content availability and tolerated delay .
we also observe that the _ return path _ taken by a response is on average longer than the _ forward path _ of the query and that the search cost increases only marginally after several hops due to the small network diameter . |
the cosmic microwave background ( cmb ) has been used as a window through which to study not only the early universe but also its evolutionary history .
the challenge of such observations arises from the fact that the cmb carries information spanning a period of about 14 billion years , so that specially - designed instruments and analysis methods are required to separate the effects originated from different physical processes .
one recent focus is the measurement of the sunyaev - zeldovich effects ( sze ) resulting from the hot gas in galaxy clusters @xcite .
several cmb experiments are dedicated to this purpose , such as act @xcite , ami @xcite , apex - sz @xcite , ovro / bima - sze @xcite , spt @xcite , sza @xcite , and ocra - p @xcite . here
we report the first results of sze observations with the y. t. lee array for microwave background anisotropy ( amiba ; * ? ? ?
amiba is devoted to cmb observation with particular emphasis on detecting sze clusters , the cmb anisotropy , and possibly cosmic defects @xcite .
the instrument is described in @xcite , @xcite , and @xcite . in our initial configuration with seven close - packed antennas of 60-cm diameter , we concentrated on pointed observations of the sze and the measurement of the cmb temperature power spectrum @xcite . at the amiba operating frequency ( 86 - 102 ghz ; @xmath2-mm wavelength )
the sze is expected to induce a decrement in cmb intensity as compared with the undistorted planckian spectrum ( see sec .
[ sec - tel - sze ] ) .
this provides a powerful tool for discriminating between galaxy clusters and other astronomical sources , because the latter normally emit photons characteristic of much harder spectra than the cmb , and therefore always induce an increment in intensity rather than a decrement . by measuring the sze intensity deficit and its profile , we hope to probe not only the cluster physics but also the related cosmic origins
this paper is organized as follows . in section 2
we describe the amiba telescope and its relation to the sze . in section 3 , we summarize our initial observations in 2007 with emphasis on the observing strategy and target selection . in section 4 , we describe our analysis formalism and procedures for obtaining the calibrated visibilities from the raw time - lag data , and the sze images and profiles from the calibrated visibilities .
the main results are also presented here . in section 5
we discuss possible systematic errors . in section 6
, we summarize the scientific implications of these results , including the estimation of hubble constant @xcite , the scaling relationship between the sze and x - ary derived properties @xcite , and the baryonic fraction when our sze data are jointly analyzed with the subaru lensing data @xcite .
brief conclusions are given in section 7 .
we have other companion papers investigating the system performance @xcite , the foreground @xcite , and data integrity @xcite .
amiba is an interferometric experiment initiated in taiwan in 2000 and dedicated on mauna - loa ( 3400 m in elevation ) , big island , hawaii on october 3 , 2006 .
it has dual - channel receivers operating at 86102 ghz , designed to have full polarization capabilities @xcite .
currently it has seven close - packed cassegrain antennas of 60-cm diameter @xcite giving a synthesized resolution of 6 arcminutes ( see sec .
[ sec - res ] ) , expandable to a total of 19 elements with a synthesized resolution of about 2 arcminutes .
the project has been funded for an expansion to 13 elements with dishes of 1.2-m diameter .
the 13-element system is expected to start operating in the early 2009 .
its capability in studying the sze science is investigated by @xcite .
the receiver - antenna elements are reconfigurable and co - mounted on a six - meter platform , which is driven by a hexapod mount @xcite .
each element has a cooled heterodyne receiver , consisting of hemt amplifiers of 46 db in amplification , subharmonic mixers , and 218 ghz if amplifiers . for each baseline , the signals from two dual - channel receivers are cross - correlated in an analogue 4-lag correlator , whose time - lag outputs are convertible to two complex visibilities corresponding to the upper and lower frequency bands ( see sec . [ formalism ] ) .
the cross correlation between the l and r polarization modes ( the dual channels ) of a pair of receivers enables the measurement of the four stokes parameters , @xmath3 , @xmath4 , @xmath5 , and @xmath6 .
currently amiba operates with only two cross - polarization modes of ll and rr , focusing on the measurement of @xmath3 .
the typical receiver noise temperatures are between 80 and 110 k @xcite . in 2007 ,
the seven - element amiba focuses on targeted sze observations .
the theoretically expected sze in flux density is @xmath7 where @xmath8 and @xmath9 are the thermal and kinematic sze respectively , defined as @xmath10 , \\
\delta i_{\rm ksze}(f,\tau , v_{\rm p } ) & = & - i_0 \beta\tau h(x).\end{aligned}\ ] ] here @xmath11 is the observing frequency , @xmath12 jy sr@xmath13 for a cmb temperature of @xmath14 , @xmath15 is the compton parameter , @xmath16 $ ] with @xmath17 , @xmath18 is a relativistic correction , @xmath19 is the peculiar radial velocity in units of speed of light , @xmath20 is the optical depth , @xmath21 , @xmath22 is the speed of light , @xmath23 is the electron mass , @xmath24 is the electron number density in the cluster , @xmath25 is the electron temperature , @xmath26 is the boltzmann constant , @xmath27 is the planck constant , and @xmath28 is the thomson cross section . for the thermal effects , the @xmath8 has a maximum decrement with respect to the cmb planckian spectrum @xmath29 at @xmath30 ghz . this optimal frequency for sze observations
is well covered by the amiba operating frequency of @xmath31 $ ] ghz . for amiba @xmath32 at the center frequency @xmath33 ghz , with a variation of less than @xmath34 for the range of @xmath35 $ ] ghz .
therefore for a typical massive cluster with @xmath36 , @xmath37 kev , @xmath38 , @xmath39 km@xmath40s , and an angular size of @xmath41 , we expect the sze signals to be @xmath42 mjy and @xmath43 mjy for amiba , indicating @xmath44 % .
the fractional contribution from the relativistic correction @xmath18 in such case is about 7 % .
thus we expect the amiba sze signals to be dominated by the thermal effects , with a feature of decrement in flux density .
we also note that @xmath45 jy sr@xmath13 , so that @xmath46
. the operating frequency of @xmath35 $ ] ghz also has the advantage that the sze signals are less affected by the point sources when compared with the observations at a lower frequency , which is more typical for interferometric experiments such as ami ( 15 ghz ; * ? ? ?
* ) , sza ( 2735 ghz ; * ? ? ?
* ) , and ovro / bima ( 30 ghz ; * ? ? ?
this is because typically the point sources have power - law spectra of flux density with negative spectral indices , so that they contribute less towards higher frequencies .
we note that although sza also has instruments at 9098 ghz , their baselines are about ten times longer than ours providing information at much smaller angular scales that is complementary to what amiba can supply .
another advantage of the amiba frequency range is that it also suppresses the galactic synchrotron radiation and dust emission .
detailed investigation about the amiba foreground is presented in @xcite .
these features in frequency and resolution make amiba a unique cmb telescope , complementing the capability of other cmb projects .
before the 60-cm antennas were mounted onto the platform , their beam patterns were individually measured in the far field ( @xmath47 m ) by scanning a wide - band white source with a system retrofitted from a commercial equatorial mount @xcite .
figure [ fig - beam](a ) shows an example .
an azimuthal average of the beam pattern ( figure [ fig - beam](b ) ) reveals a full - width - at - half - maximum ( fwhm ) of about 23 arcminutes .
the first side lobe is about 20 db below the primary peak and the beam profile is as per designed .
the index of asymmetry @xcite is calculated showing that the discrepancy of the beam pattern from its azimuthal average is below 10% within the fwhm .
all beam patterns of the antennas are well fitted by a gaussian beam of 23 arcminutes within the 2-@xmath48 region for @xmath49 error . in 2007 ,
amiba was used with the seven 57.6-cm diameter antennas close - packed on the platform , providing 21 baselines of three different lengths @xmath50 cm , 105.0 cm , and 121.2 cm .
each baseline operates at two frequency channels centered at @xmath51 and 98 ghz , each with a band width of 8 ghz .
thus the three baseline lengths correspond to six multipole bands @xmath52 $ ] , @xmath53 $ ] , @xmath54 $ ] , @xmath55 $ ] , @xmath56 $ ] , @xmath57 $ ] .
these baselines have a six - fold symmetry @xcite . to achieve better @xmath58-@xmath59 coverage and thus a better imaging capability , we observe each target at eight different polarization angles of the platform with an angular step of 7.5 degrees , uniformly covering azimuthal angles in the @xmath58-@xmath59 plane with 48 discrete samples for each of the three different baseline lengths at each of the two frequency bands .
this makes a total of 288 discrete @xmath58-@xmath59 samples .
figure [ fig - uv_coverage ] shows such typical @xmath58-@xmath59 coverage , with the corresponding noise - weighted point spread function from a typical observation , the so - called ` dirty beam ' .
the fwhm of the dirty beam is about 6 arcminutes , defining the synthesized resolution of the seven - element amiba .
when using an equal - weighted ( rather than the current noise - weighted ) scheme to construct the dirty beam , we still obtain a similar resolution of about 6 arcminutes because the noise level for each baseline is about the same . to minimize , if not completely remove , the effects from ground pickup and from a dc component in the electronics that leaks into the data , we adopt a ` two - patch ' observing strategy that alternates between the target and a region of nominally blank sky .
this strategy is similar to that used by cbi for the measurement of cmb polarization @xcite . in this mode , a leading patch that contains our target
is tracked for three minutes , and then a trailing patch of the presumably blank sky is tracked for the same period of time .
the trailing patch is located at an ra @xmath60 greater than the ra of the target .
since it takes 10 seconds for the telescope to slew between the fields , the two patches are observed over identical azimuth and elevation tracks . during both tracks the platform polarization angle is controlled so that each baseline corresponds to a fixed @xmath58-@xmath59 mode .
the recorded data during the two tracks should then contain the same contribution from ground pickup , which is removed by a subsequent differencing of the two tracks .
this strategy requires that both the electronic offset and the ground emission are stable within 6 minutes , which is far shorter as compared with the measured time dependence @xcite .
the penalty for adopting such a two - patch strategy is the loss of efficiency by a factor of @xmath61 , where a factor of 2 comes from doubling the observing time and the other factor of 2 from doubling the noise variance when differencing the two patches .
we also note that there is a cmb component as well as possible combinations from galactic foreground and extragalactic point sources in the differencing map though at a level much lower than the expected sze of massive clusters @xcite . nevertheless before observations we did check our fields for such contaminations .
figure [ fig-2-patch ] demonstrates the application of the two - patch observing strategy for sze mapping .
the left panel is an image constructed from the leading patch , containing cluster abell 2142 ( see sec .
[ formalism ] for the formalism and procedure of data analysis ) .
the middle panel shows the corresponding image from the trailing patch of blank sky .
it is clear that both patches are contaminated by signals much stronger than the sze ( mainly from the ground pickup ) , and that the images are closely similar .
the difference of the two patches reveals the cluster signal ( right panel ) , whose signal strength is only about 1% of the contamination signal in individual patches . to calibrate these data , we observe planets roughly every 3 hours using the same 2-patch observing strategy .
this time the two patches are each observed for four minutes and separated by @xmath62 in ra .
such a separation in ra is equivalent to an angular separation of at least 57 arcminutes , assuring that leakage of the planetary signal into the second patch is small , less than @xmath63 of the planetary flux density according to the primary beam pattern measured in sec .
[ primary - beam ] . the detailed formalism on how the data are converted to calibrated visibilities
is presented in sec .
[ calibration ] , and the determination of the planetary flux densities is discussed by @xcite .
long - duration observations of the planets show that the system is stable at night ( between about 8 pm and 8 am ) so that a planetary calibration every three hours controls the calibration error to be within 5% in gain and 0.1 radian in phase @xcite . to test the overall performance of our observing , calibration , and analysis strategy , we first used the sun to successfully phase - calibrate an observation of jupiter , and then jupiter to phase and amplitude calibrate an observation of saturn .
the left plot of figure [ fig - saturn - quasar ] shows a dirty image of saturn calibrated by an observation of jupiter made four hours later .
the fwhm of this image is identical to that of the dirty beam shown in figure [ fig - uv_coverage ] , and the index of asymmetry @xcite of the cleaned image is well below 5% . after first light of the seven - element amiba in september 2006 , which was a drift scan on jupiter and
whose image was quickly constructed , several months of start - up tests were used to study the instrumental properties , to fine tune our system , to study the ground pickup , and to find an optimal observing strategy . during this period , saturn ,
mars , venus , and the crab nebula were also observed for system tests . in february 2007 ,
when testing the two - patch observing strategy , we detected a first extragalactic object , quasar j0325 + 2224 with a total ( unresolved ) flux density of 550 mjy ( right plot of figure [ fig - saturn - quasar ] ) .
after further fine tuning of the system , in april 2007 we detected our first sze cluster , abell 2142 with a central sze decrement of around -320 mjy per beam .
during 2007 we observed a total of six s - z clusters with @xmath1 : a1689 , a1995 , a2142 , a2163 , a2261 , and a2390 .
the observations were always at night time , when the system is stable .
due to the weather , system engineering , and observability of the targets , a total of roughly 150 observing hours was spent in observing the clusters or their trailing fields over a span of 5 months , from april to august in 2007 . during this period a lot of efforts
were dedicated to the system tests and tuning @xcite .
table [ tab - sz - obs ] summarizes the pointing directions , the observing and integration times , and the synthesized beam sizes , for each of the six clusters . due to the two - patch observing strategy , which modulates between on- and off - source positions , the total time spent during cluster observations is double the on - cluster time listed in the table .
the integration time on source ( @xmath64 ) is less than the actual observing time ( @xmath65 ) because of the data trimming that will be discussed in sec .
[ time - domain - processing ] .
c | c c c c c c c cluster & & @xmath65 & @xmath64 & syn .
+ & ra & dec & ( hr ) & ( hr ) & res .
+ a1689 & @xmath66 & @xmath67 & 8.75 & 7.11 & @xmath68 + a1995 & @xmath69 & @xmath70 & 15.90 & 5.56 & @xmath71 + a2142 & @xmath72 & @xmath73 & 7.05 & 5.18 & @xmath74 + a2163 & @xmath75 & @xmath76 & 7.80 & 6.49 & @xmath71 + a2261 & @xmath77 & @xmath78 & 19.15 & 8.87 & @xmath79 + a2390 & @xmath80 & @xmath81 & 16.05 & 11.02 & @xmath79 + the six clusters were chosen based on observational convenience , the expected flux density based on sze observations at other frequencies , and the existence of extensive published x - ray data .
the observational convenience requires that the targets can be seen at the mauna - loa latitude of @xmath82 n , at the night time when the system is stable ( between about 8 pm and 8 am ) , and during the main observing period between april and august in 2007 .
several cluster catalogs are compiled and studied for the expected sze flux at amiba frequency , including ovro ( 30 ghz ; * ? ? ?
* ) , ovro / bima ( 30 ghz ; * ? ? ?
* ; * ? ? ?
* ) , vsa ( 34 ghz ; * ? ? ?
* ) , and suzie ii ( 145 , 221 , and 355 ghz ; * ? ? ?
the redshift range of @xmath83 is so chosen that massive clusters of a typical virial radius of few mpc is resolvable by amiba , whose resolution ( @xmath0 ) corresponds to a physical scale of @xmath84 mpc at @xmath85 in a flat @xmath86cdm cosmology with matter density parameter @xmath87 , cosmological constant density parameter @xmath88 , and hubble constant @xmath89 . among these six clusters , only a1689 and
a2390 are known to have relaxed with a nearly isothermal profile .
a1689 has a regular morphology in x - ray while a2390 is elongated .
a2163 , by contrast , is a merger with shocks and high - temperature regions , and a2142 is a merger with a strong frontal structure . due to the missing - flux problem
, most studies for the scientific implications of our results require the assistance of x - ray derived properties of the clusters . in table
[ tab - x - properties ] , we summarize the critical x - ray derived parameters for the amiba clusters , which will be used by our companion science papers @xcite .
the x - ray temperatures are from @xcite for a2163 , a2261 , a1689 , and a1995 , @xcite for a2142 , and @xcite for a2390 . the power - law index @xmath90 in the isothermal @xmath90-model and the core radius @xmath91 are from @xcite for a2163 , a2261 , a1689 , and a1995 , @xcite for a2142 , and @xcite for a2390 .
the x - ray core radii are mostly below one arcminute except for a2142 and a2163 .
these parameters are mostly x - emission weighted but not corrected for any cooling flows , which we further discuss in some science analyses such as the estimation of the hubble constant @xmath92 @xcite .
c | c c c c cluster & @xmath93 & @xmath94 & @xmath90 & @xmath95 + & & ( kev ) & & ( arcsec ) + a1689 & 0.183 & @xmath96 & @xmath97 & @xmath98 + a1995 & 0.322 & @xmath99 & @xmath100 & @xmath101 + a2142 & 0.091 & @xmath102 & @xmath103 & @xmath104 + a2163 & 0.202 & @xmath105 & @xmath106 & @xmath107 + a2261 & 0.224 & @xmath108 & @xmath109 & @xmath110 + a2390 & 0.232 & @xmath111 & @xmath112 & @xmath113 +
we describe the formalism that we use to analyze amiba data from their raw format to the calibrated visibilities , images , and cluster profiles .
amiba operates with a bandwidth of 16 ghz centered at 94 ghz .
its four - lag correlators provide two pass - bands of 8 ghz processing dual linear polarizations .
the challenge for the analysis of such data arises because the measured output of the instrument is not visibilities and because there are only two channels of wide bands causing considerable band - smearing effect in frequency . with the formalism presented here we successfully used the sun to calibrate the image of jupiter , and then used several planets ( jupiter , saturn , mars , venus ) to cross - calibrate .
jupiter was chosen as the primary calibrator for our science results .
when we observe a general source field @xmath114 with a synthesized primary beam @xmath115 , the visibility along a baseline @xmath116 of length @xmath117 is @xmath118 where @xmath119 is the corresponding @xmath58-@xmath59 mode , @xmath120 is the observing frequency , a tilde denotes the fourier transform of a quantity , and @xmath121 denotes a convolution .
the four lag - outputs of a correlator for the source field can thus be modeled as @xmath122df \right\ } , \label{eq : vis2lag}\end{aligned}\ ] ] where @xmath123 ( @xmath124 ) is the time delay for each of the four lags , @xmath125 and @xmath126 are the instrumental frequency - dependent gain- and phase - responses respectively , @xmath127 is the frequency of the local oscillator ( 86 ghz ) , and @xmath128 and @xmath129 indicate the upper and lower ends of the response frequency range i.e. @xmath130=[86,102]$ ] ghz for amiba .
our analysis must invert equation ( [ eq : vis2lag ] ) to obtain the calibrated visibilities @xmath131 .
figure [ fig - lag ] shows our first - light lag data @xmath132 of a drift scan on jupiter .
the fringing rate is determined by the angle between a baseline and the drift direction .
initially the cluster data must be processed to remove baselines for which the system is malfunctioning .
in addition to checking the instrumental logs of receivers and correlators , nightly drift - scan data for jupiter or saturn are used to identify obvious problems .
we then apply kolmogorov - smirnov ( k - s ) test to filter out datasets for which the noise appears non - gaussian @xcite . in the k - s test
, a @xmath133 significance level is used throughout and shown to be efficient in identifying data sets with known hardware problems .
finally , we remove periods of data where occasional mount control problems appear . after this flagging
, we difference the tracking data @xmath134 in the two - patch observation in order to remove ground pickup and the electronic dc offset .
4@xmath48 outliers in the temporal domain are clipped before the differenced tracking data are integrated over the 3-minute ( 4-minute for the calibration ) tracking period to yield the lag data @xmath135 .
thus each two - patch observation yields a set of four lags @xmath132 ( @xmath124 ) .
we carefully verified that the noise of the lag data is white within the time scale of 10 minutes , assuring the scaling of the noise level with the two - patch integration time @xcite .
further flagging of 4@xmath48 outliers is applied in the visibility space ( see sec . [ calibration ] ) .
table [ tab - sz - flag ] shows the percentages of data that are flagged out by each step . on average the fraction of good data
is about 60% .
c | c c c c | c cluster & baseline & non - gau .
& mount & 4@xmath48 & good + & prob . & noise & prob . & outliers & data + a1689 & 2 & 11 & 4 & 2 & 81 + a1995 & 35 & 4 & 5 & 21 & 35 + a2142 & 8 & 15 & 1 & 2 & 74 + a2163 & 7 & 8 & 1 & 1 & 83 + a2261 & 25 & 1 & 1 & 27 & 46 + a2390 & 5 & 16 & 3 & 7 & 69 + we must invert equation ( [ eq : vis2lag ] ) to convert the four lags @xmath132 to two visibilities @xmath131 . ideally ,
if @xmath136 is continuous , then we need only perform an inverse fourier transform from the @xmath137 domain to the @xmath11 domain .
here is sampled discretely , and conservation of the degrees of freedom provided by four lag measurements implies that we can obtain at best two uncorrelated complex visibilities , with no redundancy .
we perform the inversion first by dividing the 16-ghz frequency band into two , @xmath138 $ ] and @xmath139 $ ] , each associated with one ` band - visibility ' , and define a ` visibility vector ' as @xmath140)\}\\ \im\{v^b([f_1,f_d])\}\\ \re\{v^b([f_d , f_2])\}\\ \im\{v^b([f_d , f_2])\ } \end{array } \right ] .
\label{eq : vb}\ ] ] if we also write the four lags in column - vector form @xmath141 ( @xmath124 ) , then the visibility - to - lag transform resulted from equation ( [ eq : vis2lag ] ) is simply @xmath142 , where @xmath143 is a @xmath144 matrix .
the four columns in @xmath143 can be easily obtained from equation ( [ eq : vis2lag ] ) as the four vectors @xmath145 when setting each component in @xmath146 to unity in turn while zeroing the other three components .
subsequently the inversion is @xmath147^{-1 } { \bf c}^b = { \bf v}^b { \bf c}^b .
\label{eq : lag2vis0}\ ] ] once given any lag data @xmath145 , we can use equation ( [ eq : lag2vis0 ] ) to construct the band - visibilities @xmath148)$ ] and @xmath149)$ ] . in principle , if @xmath125 and @xmath126 are accurately known , then @xmath150 can be accurately determined so that there is no need for further calibration .
however , in reality @xmath125 and @xmath126 are time - dependent so that @xmath150 is also time - dependent and thus needs to be calibrated for every observation .
because the transform ( [ eq : lag2vis0 ] ) is linear , any reasonable choice of @xmath151 , @xmath125 and @xmath126 should yield the same @xmath146 after calibration .
hence we make the simplest choice @xmath152 ghz , @xmath153 , and @xmath154 , and then calibrate the data using the formalism described in section [ calibration ] .
this choice makes the inversion ( [ eq : lag2vis0 ] ) exactly a discrete inverse fourier transform that accounts for the band - smearing effect .
an alternative , and equally natural , choice might be to have equal power in the two frequency bands , with @xmath151 so chosen that @xmath155 , but this is difficult to implement since @xmath125 is time - dependent . for calibration ,
we observe planets using the same 2-patch observing strategy ( sec .
[ sec - cal - eve ] ) , the same time - domain pre - processing ( sec .
[ time - domain - processing ] ) , and the same lag - to - visibility transform ( eq . ( [ eq : lag2vis0 ] ) , sec .
[ lag - to - visibility - transform ] ) .
the underlying band - visibilities of these calibrators have the form @xmath156,\ ] ] where @xmath157 and @xmath158 are real constants , corresponding to the fluxes at different frequencies .
however , the band - visibilities constructed from observational data using equation ( [ eq : lag2vis0 ] ) , where we have chosen @xmath153 and @xmath154 , will take the form @xmath159 .
\label{eq : vis - obs}\ ] ] since the involved operations are linear , there must exist a calibration matrix @xmath160 that corrects this discrepancy , i.e.@xmath161 we may model @xmath160 as @xmath162 , \label{eq : cb}\ ] ] where the parameters @xmath163 and @xmath164 account for the gain- and phase - corrections relevant to the effects from @xmath125 and @xmath126 in equation ( [ eq : vis2lag ] ) respectively .
by comparing the moduli and phases of @xmath165 and @xmath166 , we obtain @xmath167 finally , for a set of observed lag data @xmath145 , the calibrated band - visibility vector is @xmath168 we emphasize that this is a linear transform , so different initial choices for @xmath125 and @xmath126 that generate different @xmath150 should lead to different @xmath160 but the same final calibrated band - visibilities @xmath146 .
for each two - patch observation , we obtain one @xmath146 for each cross - polarization mode and each baseline .
for each baseline and each day , 4@xmath48 outliers in @xmath169 are flagged .
in further analyses such as image making and model fitting for the cluster profile , a reliable estimate for the error in the band - visibilities derived above will be needed . first
the 2-patch differenced lag data @xmath170 are modeled as the linear sum of the signal and noise , @xmath171 . since such data for clusters
are dominated by electronic noise ( i.e. @xmath172 ) and the noise is white @xcite , the noise variance associated with each time - integrated @xmath173 can be directly estimated as @xmath174 ^ 2 \rangle \equiv \langle \langle { \bf n}^b_{c}(t ) \rangle_t^2 \rangle \approx \frac{\langle [ { \bf c}^b ( t)]^2 \rangle_t}{\#_t},\ ] ] where @xmath175 is the number of temporal data in the time integration .
then @xmath176 constitute the diagonal elements of the lag - lag noise correlation matrix @xmath177 .
the off - diagonal elements are found to be less than 10% of the diagonal elements @xcite so we approximate them as zeros . in a similar fashion ,
the band - visibility obtained from equation ( [ eq : lag2vis ] ) can be modeled as a linear sum of the signal and noise , @xmath178 .
finally the noise correlation matrix for the calibrated band - visibilities is @xmath179 we construct the ` dirty ' image of a cluster by making a continuous inverse fourier transform of all the band - visibilities from all days and all baselines , with noise weighting based on the noise variance .
the instantaneous @xmath58-@xmath59 coverage is improved in amiba operation by rotating the platform to eight polarization angles , giving a uniform angular interval of @xmath180 between sampled @xmath58-@xmath59 modes @xmath119 ( see sec .
[ sec - res ] ) .
figure [ fig - sz_images ] shows the dirty images of these clusters .
the flux decrement at the center of each target is evident , as expected for the sze signal at the amiba center frequency of 94 ghz ( see sec .
[ sec - tel - sze ] ) .
the left part of table [ tab - images ] summarizes the angular sizes ( azimuthally averaged fwhm ) and the peak fluxes directly measured from these images .
c | c c | c c c c & & + cluster & size & flux & size & flux & s / n & scale + & & ( mjy ) & & ( mjy ) & ratio & ( mpc ) + a1689 & ( @xmath181 ) & -217 & ( @xmath182 & -168 & 6.0 & ( 1.05 ) + a1995 & ( @xmath183 ) & -167 & ( @xmath184 & -161 & 6.4 & ( 1.91 ) + a2142 & @xmath185 & -320 & @xmath186 & -316 & 13.7 & 0.92 + a2163 & @xmath187 & -347 & @xmath188 & -346 & 11.7 & 2.24 + a2261 & ( @xmath189 ) & -115 & ( @xmath190 ) & -90 & 5.2 & ( 1.25 ) + a2390 & @xmath191 & -180 & @xmath192 & -158 & 6.6 & 1.78 + a dirty beam for each cluster data set is constructed using the same method , and the dirty image and dirty beam are then processed by a clean procedure @xcite in miriad @xcite to yield a cleaned image .
figure [ fig - sz_images_clean ] shows the cleaned sze images of the six amiba clusters , where the cleaned regions are indicated by the white circles , which are the fwhm contour of the primary beam .
this process significantly reduces the convolution effects from the finite @xmath58-@xmath59 coverage .
the basic properties measured from the cleaned images are summarized in the right part of table [ tab - images ] .
the apparent angular sizes , as compared with the synthesized resolution of about @xmath0 , indicate that we have partially resolved clusters a2142 , a2163 , and a2390 , while a1689 , a1995 , and a2261 appear unresolved in these images .
the sze signals observed here are at the level of few hundred mjy , indeed consistent with the theoretical expectation @xmath193 mjy as discussed in sec .
[ sec - tel - sze ] . in table [ tab - images ] , the s / n ratios are computed as the peak flux of the cleaned model ( uncorrected for the primary - beam attenuation ) divided by the rms of the noise residual map .
we note that the noise level in these results are dominated by the instrumental noise , leading to the fact that the scaling among the peak flux , s / n ratios , and integration time is consistent with the noise equivalent flux ( point - source sensitivity ) of 63 mjy@xmath194 for the on - source integration time in a 2-patch observation @xcite , i.e. for a 2-hour observation , one hour per patch , the noise rms is 63 mjy .
simulations also show that our analysis method ( described in sec . [ lag - to - visibility - transform][noise - estimation ] ) does not bias the sze amplitude but induces a statistical error at the few - percent level .
the calibration error is controlled within 5 % @xcite . the systematic effect from cmb anisotropy
is estimated to be at a similar level as the instrumental noise , and the point - source contamination causes an underestimate of the sze amplitude by about 10 % @xcite .
it is useful for science purposes to estimate parameters describing the structures of the sze clusters from the visibilities .
we employ a maximum - likelihood analysis to estimate the band - visibilities taking the priors that all the visibilities represent the same circularly - symmetric source and that the visibility moduli within a frequency band are constant .
thus there are only six non - redundant band - visibilities @xmath195 of six multipole bands corresponding to the two frequency bands @xmath196 $ ] ( @xmath197 ) of three baselines lengths ( see sec .
[ sec - res ] and table [ tab - mcmc ] ) , i.e. the observed band - visibilities are modeled as @xmath198| \{v_p\ } , { \bf x}_0 ) & = & b({\bf x}_0-{\bf x}_1 ) v_{p(\in b , n ) } \times \nonumber\\ & & \frac{1}{k^n_2-k^n_1 } \int_{{\bf k}^n_1}^{{\bf k}^n_2 } e^{i{\bf k}({\bf x}_0-{\bf x}_1 ) } d{\bf k},\end{aligned}\ ] ] where @xmath199 is the sze cluster center , @xmath200 is the pointing direction , @xmath115 is the primary beam , and the integration accounts for the frequency band - smearing effect .
these @xmath201 are equivalent to the band - visibilities at @xmath202 and therefore are real numbers .
they indicate the cluster profile in the visibility space .
markov chain monte - carlo ( mcmc ) approach with metropolis - hastings sampling is used for this eight - dimensional likelihood analysis . for each cluster
three mcmc chains of 200,000 samples are used .
the results are given in table [ tab - mcmc ] and figure [ fig - vis - profile ] .
for all clusters the expected sze decrement in flux density is evident .
we note that although the primary cmb makes a non - negligible contribution to our observed visibilities @xcite , it appears with random phase in @xmath58-@xmath59 space and thus will not bias our analysis here but rather contribute as part of the error bars .
c | c c | c c c c c c cluster & & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 + & ra & dec & ( mjy ) & ( mjy ) & ( mjy ) & ( mjy ) & ( mjy ) & ( mjy ) + a1689 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 + a1995 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + a2142 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 + a2163 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 & @xmath239 & @xmath240 + a2261 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + a2390 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 + we emphasize that these results for @xmath201 and @xmath199 are independent of cluster model .
they can be further compared or combined with other experimental results , or fitted with a specific cluster model , for further study .
the approach here is different from that in @xcite , where an isothermal @xmath90-model is directly fitted with the visibilities to estimate the central brightness @xmath257 , with the core radius @xmath95 and the power index @xmath90 taken from x - ray analysis .
to verify that our detections of sze clusters are real rather than from the instrument or from the foreground , we implement several tests .
a commonly used powerful test is the so - called sum - and - difference test , where the temporal data are divided into two subsets of equal size and then processed separately .
figure [ fig - a2142-diff ] shows an example for a2142 , where the two half - data images ( left and middle ) show a clear signal while their difference ( right ) reveals no signal but noise .
their average is very close to the overall dirty image shown in figure [ fig - sz_images ] .
we have verified that such feature for the existence of signal is independent of the scheme for dividing the data into two halves .
all data of the 6 clusters have passed this test .
we carried out several other tests for systematics .
long - duration ( 12-hour ) observations of planets were used to check the stability of the system ; an independent analysis path produced results for a2142 consistent to high accuracy with those presented here ; and several two - patch blank - sky observations showed no signal above the expected cmb confusion @xcite .
the noise properties of the lag data were investigated in detail indicating no systematics and no significant non - gaussianity @xcite .
the asymmetry of the antenna beams @xcite was shown to be negligible .
the pointing error is small enough that it has negligible effects on our results @xcite , and the radio alignment was tuned to improve efficiency and tested for stability @xcite .
the scientific implications of the amiba sze results presented here are further studied by companion papers .
the hubble constant is estimated by first deriving the angular diameter distances of the clusters , and found to be @xmath258 @xcite in a flat @xmath86cdm cosmology with @xmath87 and @xmath88 .
we also estimate the integrated compton parameter @xmath259 @xcite , which is the compton @xmath260 parameter integrated out to the angular radius at which the mean overdensity of the cluster is equal to 2500 times the critical density of the universe at that redshift .
table [ tab - sz - properties ] summarizes the results , as compared with the results from observations at 30 ghz ( ovro / bima , deduced from * ? ? ?
* relativistic correction considered ) and 145 ghz ( suzie ii , deduced from * ? ? ?
* relativistic correction considered ) .
our results are consistent with those from ovro / bima except for a1995 , but seem to be systematically lower than those from suzie ii .
we also investigate the scaling relations between the @xmath259 from amiba data and various x - ray derived properties such as the gas temperature @xmath25 , total mass @xmath261 and luminosity @xmath262 .
the scaling powers of these relations are consistent with the predictions of the self - similar model @xcite . in the above studies , due to the missing flux problem
the spherical isothermal @xmath90 model is used with the spectral index @xmath90 and the core radius estimated by the x - ray data ( see table [ tab - x - properties ] ) , and the normalization calibrated by the amiba sze visibilities analyzed here @xcite .
c | c c c cluster & + & ovro / bima & amiba & suzie ii + a1689 & @xmath263 & @xmath264 & @xmath265 + a1995 & @xmath266 & @xmath267 & - + a2142 & - & @xmath268 & - + a2163 & @xmath269 & @xmath270 & @xmath271 + a2261 & @xmath272 & @xmath273 & @xmath274 + a2390 & - & @xmath275 & @xmath276 + we also perform a joint analysis of our sze data with the weak gravitational lensing data from subaru observations @xcite . for the four clusters of a1689 , a2142 , a2261 , and a2390 ,
the two data sets are found to be in great agreement in morphology .
quantitative analysis even yields an estimation for the baryonic fraction of @xmath277 and concludes that when compared with the cosmic baryonic fraction @xmath278 @xcite , @xmath279 % of the baryons are missing from the hot phase of clusters @xcite . the morphological agreement between our data and the lensing data and the consistency of @xmath280 , @xmath92 , @xmath259 and scaling relations with literature
are encouraging , indicating that amiba is a reliable cmb telescope .
we successfully detected six sze clusters with the seven - element amiba in its compact configuration .
the analysis method and results presented here mark a milestone for the amiba project and provides the first successful results for sze clusters in the 3-mm band .
these results are consistent with the published results based on data from other observations , and provide complementary information at the same time .
a number of tests show that the system performs as expected from its design , and we anticipate that the current expansion to a 13-element system @xcite will boost its capability for the study of cmb cosmology .
we thank the ministry of education , the national science council , the academia sinica , and national taiwan university for their support of this project .
we thank the smithsonian astrophysical observatory for hosting the amiba project staff at the sma hilo base facility .
we thank the noaa for locating the amiba project on their site on mauna loa .
we thank the hawaiian people for allowing astronomers to work on their mountains in order to study the universe .
this work is also supported by national center for theoretical science , and center for theoretical sciences , national taiwan university for j.h.p .
we are grateful for computing support from the national center for high - performance computing , taiwan .
support from the stfc for m. birkinshaw and k. lancaster is also acknowledged .
allen , s.w . , 2000 ,
, 315,269 benson , b.a . , _
_ 2004 , , 617 , 829 birkinshaw , m. , lancaster , k. 2007 , newar , 51 , 346 behringer , h. , _ et al .
_ 1998 , a&a , 334 , 789 chen , m .-
et al . _
2008 , s , submitted dunkley , j. , _ et al . _
2008 , s , submitted ( arxiv : astro - ph/0803.0586 ) grego , l. , _ et al . _
2001 , , 552 , 2 halverson , n.w . , _
et al . _
2008 , , submitted ho , p.t.p . , _
_ 2008a , , submitted ho , p.t.p . , _
et al . _ 2008b , mod.phys.lett.a , 23 , 1243 hogbom , j.a
, 1974 , a&as , 15 , 417 huang , c.w . , _
et al . _
2008 , , submitted kneissl , r. , _ et al . _
2001 , , 328 , 783 koch , p. , raffin , p.a . ,
wu , j .- h.p . , _
et al . _
2006 , proc .
eucap ( esa sp-626 ) , p.668.1 koch , p. , _ et al . _
2008a , s , submitted koch , p. , _ et al . _ 2008b , ,
submitted kosowsky , a. , _ et al . _ 2003 ,
new astronomy reviews , 47 , 939 laroque , s. , _ et al . _
2006 , , 652 , 917 lancaster , k. , _ et al . _
2005 , , 359 , 16 lancaster , k. , _ et al . _ 2007 , , 378 , 673 leitch , e.m
et al . _
2005 , , 624 , 10 lin , k.y . , _
et al . _
2004 , , 608 , l1 lin , k.y . , _
et al . _
2008 , , submitted liu , g.c . , _ et al . _
2008 , , submitted markevitch , m. , forman , w. r. , sarazin c. l. , vikhlinin a. , 1998 , , 503 , 77 mason , b.s . , myers , s.t . , &
readhead , a.c.s . , 2001 , , 555 , l11 miriad , australia telescope national facility , + http://www.atnf.csiro.au/computing/software/miriad/ molnar , s. , _ et al . _ 2008 , ,
submitted morandi , a. , _ et al . _ 2008 , , in press muchovej , s. , _ et al . _
2007 , , 663 , 708 nishioka , h. , _ et al . _
2008 , , submitted readhead , a.c.s . , _ et al . _
2004 , science , 306 , 836 reese , e. d. , carlstrom , j. e. , joy , m. , mohr , j. j. , grego , l. , holzapfel , w. l. , 2002 , , 581 , 53 ruhl , j. , 2004 , the south pole telescope , 11 sievers , j.l . , _
et al . _
2005 , , submitted ( astro - ph/0509203 ) sanderson a. j. r. , ponman t. j. , 2003 , , 345 , 1241 sunyaev , r. a. , zeldovich , y. b. , 1972 , comments on astrophysics and space physics , 4 , 173 umetsu , k. , _ et al .
_ , 2004 , mod.phys.lett.a , 19 , 1027 umetsu , k. , _ et al .
_ 2008 , , submitted wu , j .- h.p . , _
et al . _ 2001 , , 132 , 1 wu , j .- h.p . , _ et al . _
2004 , mod.phys.lett.a , 19 , 1019 wu , j .- h.p . , _
et al . _ 2008a , mod.phys.lett.a , 23 , 1675 wu , j .- h.p
et al . _ , 2008b , in preparation | we present observations , analysis and results for the first - year operation of amiba , an interferometric experiment designed to study cosmology via the measurement of cosmic microwave background ( cmb ) .
amiba is the first cmb interferometer operating at 3 mm to have reported successful results , currently with seven close - packed antennas of 60-cm diameter giving a synthesized resolution of around @xmath0 . during 2007
amiba detected the sunyaev - zeldovich effects ( sze ) of six galaxy clusters at redshift @xmath1 .
an observing strategy with on - off - source switching is used to minimize the effects from electronic offset and ground pickup .
planets were used to test the observational capability of amiba and to calibrate the conversion from correlator time - lag data to visibilities . the detailed formalism for data analysis is given .
we summarize our early tests including observations of planets and quasars , and present images , visibility profiles , the estimated central coordinates , sizes , and sze amplitudes of the galaxy clusters .
scientific implications are summarized .
we also discuss possible systematic effects in the results . |
it is possible to consider that cognitive psychology appeared as a reaction to behaviorist approaches , where the mental content plays the role of a black box @xcite . in contrast
, this content constitutes a central issue in cognitive sciences .
consequently , the use of computers to implement or imitate human intellectual tasks naturally emerged as a methodological tool , and even as a powerful metaphor , in the investigation of mental processes such as intelligence , learning , memorization , among others .
it is along these lines that we prepared a sequence of experiments with humans , to be later on compared with similar experiments , or simulations , done with computers provided with appropriate algorithms , in particular simple perceptrons .
it should of course be clear that such comparisons have philosophical implications ( see , for instance , @xcite ) , which we address in the present work only in the conclusions ( section v ) . in the present study ,
two different memorization tasks were implemented .
the first of them was relatively simple , namely the memorization of simple binary codes in @xmath0 and @xmath6 matrices .
most of the present effort is dedicated to this analysis .
the second task was , intellectually speaking , sensibly more complex .
it consisted of learning ambiguous images , where the figure - background reversal is crucial .
although semantic and strategic aspects are present in both learning tasks , the second one is by far more delicate and reveals higher level cognitive phenomena .
consistently , our study leads to quantitative information concerning the first task , whereas only some qualitative features were determined for the latter . at the level of the comparison with computational simulations ,
our emphasis is put on whether learning occurs in an _ extensive _ or a _ nonextensive _ manner .
these terms will be mathematically defined and further analyzed in section iii .
they are currently used in statistical mechanics and thermodynamics , branches of physics dedicated to the study of the connections between the microscopic and the macroscopic worlds . at the present stage ,
it is enough to think of extensivity as a form of _ nonglobality _ or _ locality _
@xcite , as opposed to nonextensivity or _
globality _ or _
nonlocality_. in a loose manner , they respectively correspond to the _ molecular _ and _ molar _ approaches in psychology , i.e. , the system is perceived as the sum of its parts , or as different from the sum of its parts . in section
ii we present the experimental study with humans . in section iii
we describe the entropic concepts that we use with regard to the computational simulations that have been implemented . in section
iv we compare both approaches , namely with humans and computers . finally , we conclude in section v.
the experiment consists in individually learning fixed binary codes ( represented by the signs * @xmath7 * and * @xmath8 * ) in @xmath9 matrices .
it was presented to each person , initially one @xmath0 matrix ( see fig . 1 ) and , then , two @xmath10 binary matrices ( see figs . 2 and 3 ) . in order to avoid any kind of uncontrolled cultural effect , the location of the symbols @xmath7 and @xmath8 in all the matrices was randomly generated by computer and fixed once for ever .
it was carried out with approximatively 150 university students whose specialty was not directly related to geometry , mathematics or images , in order to avoid professional bias .
for instance , students of physics , mathematics , engineering , architecture , visual publicity were excluded .
the experiment population was mainly constituted by students of psychology , administration , social service and social comunication of the federal university of rio de janeiro .
approximately 30 students were used for preliminary tests in order to fix an optimal experimental protocol ( matrix sizes , binary code of each matrix , exhibition and hiddening times , among others ) .
then , precisely 120 ( 92 female and 28 male ) students were exposed to the same protocol , and only their results were taken into account for quantitative purposes in the statistical processing ( averaging , in particular ) .
their ages ranged between 17 and 27 years old , the majority of them being around 22 years old .
each of them was sequentially isolated in a peaceful room and tested , for about 40 minutes , by one of us .
information about the scope of the research was given to each one of the 120 individuals .
it was told that the experiment was not measuring their intelligence ( so that they would feel relaxed ) , and that it was important for understanding how learning occurs ( so that they would seriously try to perform satisfactorily ) .
after these instructions were given to each one of the 120 students , three matrices were shown . in all cases ,
a @xmath0 matrix was first shown , one of the two complementary matrices ( i.e , obtained by interchanging the symbols * + * and * o * ) .
then , one of the four @xmath10 matrices was shown ( matrix in fig .
2 to 30 students , matrix in fig . 3 to other 30 , and so on ) . then the other noncomplementary @xmath10 matrix was shown .
the sequence of each individual test was as follows : + _ ( 1 ) _ after some general explanations , an empty @xmath0 matrix and the @xmath8 and @xmath7 symbols were shown , and the subjects were asked to randomly fill the matrix with the two symbols .
this step aimed to provide to the individuals some familiarity with the experiment ; + _ ( 2 ) _ then fig . 1 was exhibited during 8 seconds and then hidden . the student had to try to reproduce it on the spot on an empty matrix .
when this was done , the operation was repeated after a rest interval of 10 seconds .
the @xmath0 matrix was never shown more than 10 times ( the individuals started feeling tired after 10 times ) .
the matrix was considered to be learnt if no error was made after two successive exhibitions ; + _ ( 3 ) _ then the learning test was repeated by successively using two of the four @xmath10 matrix , one at a time ; + _ ( 4 ) _ finally , the student was asked to briefly describe how he(she ) proceeded to learn . the _ error _
@xmath11 of the @xmath12-th individual ( @xmath13 ; @xmath14 corresponds to the initial random filling ) is defined as the number of wrong elements of the filled @xmath9 matrix ( @xmath15 ) ; @xmath16 .
typical results are presented in figs . 4 , 5 and 6 for the @xmath0 and in figs . 7 and 8 for one of the four @xmath10 matrices ( the figures associated with the other three @xmath10 matrices are quite similar in fact ) .
although in a quite different context , results that have some connection with the present ones have been exhibited in @xcite .
the averages @xmath17 ( with @xmath18 if the entire set of students is used for averaging ) are shown in fig .
these curves already achieved the aspect presented in this figure when the averages were performed with approximately 80 individuals . using the entire set of 120 results just improved the precision but incorporated no new qualitative elements in the curves .
if we define the _ learning time _ as the number of times shown before learning , excluding those who did not suceed learning that particular matrix untill the end of the experiment , we can check that it is of the order of 6 .
incidentally we verified an interesting ( and indeed unexpected ) cultural phenomenon .
for the @xmath0 matrix we were expecting @xmath19 to be close to @xmath20 since , before starting to show the codified matrix , we asked to _ randomly _ fill the empty matrix with symbols * o * and * + * , indicated in _ this _ order above the empty matrix , if we look at them from left to right . in variance with this reasonable expectation , we found , for the first 60 students , @xmath21 .
after some hesitation about what could be the cause of this asymmetry ( e.g. , could it be the different semantics humans associate with a circle or a cross ? ) , we speculated that it could be the fact that portuguese language ( within which the brazilian population is educated ) is read _ from left to right_. then , to the second and last set of 60 students , an empty fig .
1 was presented with the symbols in the ordering * + o * instead of * o + * .
very symptomatically , we then obtained @xmath22 , the overall average being consequently @xmath23 , reasonably close to 12.5 as initially expected ! to confirm this cultural cause of the observed asymmetry , it would be interesting to repeat the experiment with say arabic students ( educated within _ right - to - left _ reading ) .
another interesting feature that we observed is that , for many individuals , @xmath24 , thus systematically contradicting the overall monotonic tendency of @xmath25 to decrease with time .
the reason for this kind of a priori unexpected behavior appeared to be ( as commented by the individuals themselves during the free final conversation ) that , after seeing for the first time the code to be learnt , the student dedicated a good part of his ( her ) attention to establish astrategy " for learning rather to properly learn the matrix .
let us mention , by the way , that in an experiment like the present one it is quite hard to differentiate between learning the strategy and memorizing the matrix within that strategy .
this kind of modelization is supported by the fact that , several months after conclusion of the experiment , quite a few students still remembered the strategy , while they had completely forgotten the particular matrix code itself .
let us now briefly comment the second experiment we developed .
we chose several ambiguous images , all of them being susceptible of two mutually excluding interpretations on the basis of figure - background reversal .
for example , if one sees in fig .
15 @xcite a young woman , one does not simultaneously see the old woman , and reciprocally .
we implemented this more complex learning task by showing the image and then asking to the individual what he sees .
then we tried to count the time needed by the person in order to recognize the other image interpretation .
this perception mechanism is sometimes referred to as _ reversal _ of the figure - background .
it turned out that the times involved in this type of experiment , and very specifically the slowness of learning , if any , how to recognize the reversal , were so ill - defined that we decided not to proceed with this protocol .
the question of how to conveniently quantify such learning remains , therefore , an open question ( see also @xcite ) .
a great variety of computer learning algorithms are available in the literature .
it is clear that all of them process , in one way or another , information .
entropy is well known to be a convenient tool for quantifying ( lack of ) information .
it can therefore be used in the context of any learning algorithm , at least in principle .
we shall use it here in connection with the specific perceptron we shall describe later on . for convenience ,
let us briefly review at this point some basic notions about the entropic forms we are referring to in the present paper .
the boltzmann - gibbs - shannon ( bgs ) entropy is the basis of standard statistical mechanics and thermodynamics .
it is defined ( in its discrete version ) by @xcite @xmath26 where @xmath27 is the number of microscopic possibilities accessible to the system , and @xmath28 are the associated probabilities ( @xmath29 ) ; for simplicity , we have taken boltzmann constant equal to unity . this entropy becomes maximal at equiprobability , i.e. , @xmath30 for all @xmath12 , and achieves the value @xmath31 which is the celebrated boltzmann formula . if we consider a composite system made of two ( probabilistically ) independent systems @xmath32 and @xmath33 , i.e. , if we assume that @xmath34 , and replace this into eq .
( 1 ) , we straightforwardly obtain @xmath35 which can be phrased as _ the entropy of the whole is the sum of the entropies of the parts_. the entropic form ( 1 ) is the basis of standard , boltzmann - gibbs ( bg ) , statistical mechanics and thermodynamics , and property ( 3 ) is known as _
extensivity _ or _ additivity_. this entropy , as well as others , are in some sense ubiquitous .
indeed , they emerge in a great variety of discussions .
for example , they have often been used concerning complex phenomena such as the organization of living matter ( see , for instance , @xcite ) , as well as other types of organization , including that of knowledge ( e.g. , memorization and learning ) , economics , linguistics , to mention but a few ( see , for instance , @xcite ) . before addressing the generalization of @xmath36 we are interested in here ,
let us mention that optimization of eq .
( 1 ) in the presence of a constraint of the type @xmath37 ( where @xmath38 might be say microscopic energy levels ) , leads to @xmath39 where @xmath40 is a parameter to be determined through the value of the constraint . in statistical mechanics ,
( 4 ) is in fact the celebrated boltzmann - gibbs weight . in 1988 ,
one of us ( ct ) proposed @xcite the generalization of bg statistical mechanics on the basis of a more general entropic form , namely @xmath41 @xmath1 being any real number .
we can verify that @xmath42 , in other words , the bg formalism becomes now the @xmath2 particular case of this more general formalism .
if we assume , once again , two independent systems @xmath32 and @xmath33 , we can prove that @xmath43 which can be phrased as _ the generalized entropy of the whole is different from the sum of the generalized entropies of the parts_. this property is referred to as _
nonextensivity _ or _
nonadditivity_. to be more precise , if we take into account that @xmath44 is always zero or positive , eq .
( 6 ) implies that @xmath45 if @xmath46 and @xmath47 if @xmath48 .
only if @xmath2 we have that @xmath49 .
it is from property ( 6 ) that the terms _ nonextensive _ statistical mechanics and thermodynamics have been coined ( for reviews see @xcite ) .
analogously to what we did before , if we optimize @xmath44 in the presence of the constraint @xmath50 , we obtain @xcite @xmath51^{1/(1-q ) } \;,\ ] ] where , as before , @xmath52 is a parameter to be determined through the value of the constraint .
notice that for the normal regime for @xmath52 , i.e. , @xmath53 , we have , for large values of @xmath38 , a long _ power - law _
tail for @xmath48 , whereas we have a short _ exponential _
tail for @xmath2 ; finally , for @xmath46 , we have a cutoff .
( 7 ) can be re - written in the boltzmann - gibbs form , namely @xmath54 where @xmath55 \;.\ ] ] in other words , in what concerns the optimizing distribution , @xmath56 plays the role of an effective energy which replaces @xmath38 ( in the @xmath57 limit , of course we recover @xmath58 ) .
this effective energy can be used to @xmath59generalize a variety of microscopic and mesoscopic equations .
one such example is the langevin equation , on which the perceptron that we use here has been constructed .
matrix.,width=226 ] matrix .
this matrix is referred to as the @xmath7diagonal matrix .
its dual matrix is obtained by permutating the @xmath7 and the @xmath8 symbols , and is referred to as the @xmath8diagonal matrix.,width=207 ] matrix .
this matrix is referred to as the @xmath8column matrix .
its dual matrix is obtained by permutating the @xmath7 and the @xmath8 symbols , and is referred to as the @xmath7column matrix .
, width=207 ] matrix .
the three left correspond to three different individuals ( the abscissa is the number of presentations of the matrix ; the ordinate is the number of matrix elements incorrectly reproduced ) .
the three right ( see [ 16 ] for further details ) correspond to three different initial conditions for the perceptron ( the abscissa is the number of iterations ; the ordinate is the percentual error).,width=340 ] matrix as a function of the number of presentations : circles , squares and triangles respectively correspond to averaging over @xmath60 , @xmath61 and @xmath62 individuals.,width=302 ] matrix .
the abscissa is the ordinal of the presentation at which the individual learnt the matrix : 82 ( out of 120 ) individuals learnt the matrix before or at the 9th presentation ( we recall that the 10th presentation was used to confirm the learning at the 9th one ) ; 38 individuals did not succeed ( and are not computed in the histogram).,width=264 ] @xmath7diagonal matrix as a function of the number of presentations : the circles correspond to averaging the results of @xmath63 individuals to whom the @xmath7diagonal matrix was shown in _ first _ place ; the squares correspond to averaging the results of @xmath63 individuals to whom the @xmath7diagonal matrix was shown in _
second _ place ( after having seen , in _ first _ place , the @xmath8column matrix ) .
, width=264 ] matrix , either the @xmath7diagonal or the @xmath8diagonal ones .
the abscissa is the ordinal of the presentation at which the individual learnt the matrix : 50 ( out of 60 ) individuals learnt the matrix before or at the 9th presentation ( we recall that the 10th presentation was used to confirm the learning at the 9th one ) ; 10 individuals did not succeed ( and are not computed in the histogram).,width=264 ] matrix ) as a function of the number of iterations ( @xmath64 is a parameter of the perceptron ) .
the perceptron has been chosen with @xmath65 binary inputs , in order to simulate the human task ( dots ) as closely as possible on the average ( taken on 92 individuals in this example ) .
see [ 16 ] for further details.,width=321 ] matrix ) as a function of the number of iterations ( @xmath64 is a parameter of the perceptron ) .
we notice the extreme sensitivity to the value of @xmath1 in the neighborhood of @xmath2 .
see [ 16 ] for further details.,width=321 ] matrix , where @xmath66 is the value of @xmath67 at which the error becomes half of its value at @xmath14 .
the dots correspond to averaging the experimental data with @xmath62 humans .
the continuous curve has been obtained averaging a large number of initial conditions for the perceptron with the gain parameter @xmath68 , the temperature - like parameter @xmath69 and the entropic index @xmath70 .
it is with these values that optimal fitting was obtained for the experimental data .
the parabolic extrapolation of the experimental data corresponding to @xmath71 and @xmath72 provides , for @xmath14 , the value of @xmath73 for the percentual error , which corresponds to @xmath74 for the absolute error .
therefore the value of @xmath66 corresponds to the value of @xmath67 at which the absolute error equals @xmath75 . since no integer value of @xmath67 corresponds exactly to this value , a linear interpolation has been performed , yielding @xmath76 for the experimental data with 120 individuals .
see [ 16 ] for further details.,width=321 ] as we see , the entropic property ( 6 ) has a kind of _ gestalt_-like flavor .
its use constitutes a natural choice if we desire to deal with informational phenomena involving global or nonlocal aspects .
since this might well be the case of human learning , we have adopted this formalism in order to have the possibility of comparing human and machine learnings .
to do so , a nonextensive perceptron has been implemented @xcite which performs a task similar to the learning of the @xmath77 and @xmath78 matrices that were exposed to the students , according to the experimental protocol we described earlier . in order to perform calculations
, the perceptron needs an internal dynamical equation . to fulfill this requirement , the @xmath1-generalized langevin equation previously introduced by stariolo @xcite
was implemented in the perceptron .
some typical runs of the perceptron are shown in fig .
4 , and typical averages are shown in figs .
16 , 17 and 18 ( from @xcite ) .
the purpose of the present section is to compare the results obtained with humans and those obtained with the nonextensive perceptron .
the comparison will be illustrated on the learning / memorizing of the @xmath0 matrix .
we shall verify that , for this specific task , the human and perceptron results can be amazingly similar .
this can be checked on fig .
4 . the three individual results ( on the left ) are indeed similar to the perceptron realizations with three different initial conditions ( on the right ) .
the three human examples have been chosen as to exhibit typical cases .
the three perceptron examples have been chosen in order to have an overall aspect similar to the human ones .
averages over many realizations of the data just presented are shown ( with dots ) in figs .
16 and 18 .
we have rescaled the time variable in such a way that comparison becomes possible on the same graph .
more precisely , we have expressed time in units of the corresponding half - time @xmath66 , defined as the value of time at which the average error curve decays to its half value .
a rescaling such as this one is clearly necessary in order to quantitatively compare the results . indeed ,
time " is here represented as the number of presentations , whereas perceptron time " essentially corresponds to the number of computer iterations .
these two numbers being of a completely different nature ( see also @xcite ) , it is clear that rescaling becomes necessary .
we may consider fig .
18 as the central result of the present work .
we verify that , for the specific task of learning / memorizing 25 binary states ( on a matrix for the humans ) , humans and machines are remarkably similar .
of course , the parameters of the perceptron have been chosen in such a way as to optimize the overall fitting to the human data .
the number of individuals that have been averaged is 120 , and we have verified that no sensible variation is obtained under increase of that number . in the language of statistical mechanics , we may say that we have practically attained the thermodynamic limit .
there is a little bump in the human results at @xmath79 : we have not identified its origin ( perhaps fatigue of the tested individuals , perhaps something else ) .
the perceptron does not exhibit such bump . as we see
, the perceptron that fits the data has some degree of nonextensivity , as was conjectured in the beginning of the present work .
although @xmath1 is quite close to unity , we must take into consideration the fact that the error curves have been shown to be ( see fig .
17 ) extremely sensitive to the degree of nonextensivity in the neighborhood of @xmath2 ( extensive case ) .
in summary , we have implemented an experimental study aiming to measure the learning / memorizing performance of humans on simple codes on matrices .
it is on purpose that we simultaneously use the words learning " and memorization " .
indeed , the experiments clearly showed that the improvement of correct answers was due to a mixture of memorization of the specific codified example and devising learning _ strategies _ ( symmetry rules and other mnemonic tricks ) in order to efficiently implement the memorization effort .
in fact , after several months , we informally verified that the individuals had forgotten the codes , but still remembered the strategy they used for memorizing them . through comparison with machine results
, we verified that this particular human task was executed with clear indication of a slight , though very efficient nonextensivity ( or globality ) quantified by the entropic index @xmath1 .
this index appeared to be slightly _ above _ unity , which characterizes _ slower _ learning / memorizing ( the error curves takes longer to become basically zero ) , but perhaps higher ability for devising strategies .
it is then allowed to conjecture that human nature evolved , during successive generations , not so much to strongly improve the speed associated with such kind of memorization , but rather to improve the capacity of spontaneously and quickly generating intellectual strategies for performing tasks such as memorization .
we also applied a similar experimental protocol for learning / memorizing how to analyze complex figure - background images in order to quickly realize alternative ( typically two ) interpretations of the figures .
we verified that , unless much more sophisticated experimental protocols and computational algorithms are deviced , cognitive tasks with important semantic content are by all means nontrivial to measure and compare .
for such complex tasks , even more than for the simple binary learning / memorization addressed here , the role of strategies might well be fundamental , although this remains to be proved .
on more general grounds , the scenario which emerges is that nonextensivity seems to serve to humans for achieving _ abduction _ , one of charles sanders pierce three basic forms of inference ( see @xcite and references therein ) . in other words ,
given its intrinsic _ nonlocal _ nature ( strong collective correlations are necessary for making the entropic index @xmath1 to differ from unity ) , it is plausible that nonextensivity constitutes the structure necessary to make _ metaphors_. given the very high intellectual level attributed , since aristotle , to metaphors @xcite , it is allowed to think that it has some specific relation with the nature of the one that we might consider as the _ animal who makes metaphors_. we may use _ homo metaphoricus " _ to express this concept .
further developments , on both philosophical and cognitive - psychological grounds , within the frame that we have outlined here would naturally be very welcome .
they could reinforce or exclude the interpretation that our present human - machine comparison suggests .
we thank a.b . lima and k.b .
miziara for assistance during the early stages of the present work , as well as s.a .
cannas and d.a .
stariolo for making available to us their perceptron curves . computational assistance and useful remarks by l. silva , c. anteneodo , f. baldovin and m.p .
albuquerque are acknowledged as well .
we finally thank capes , cnpq , pronex and faperj ( brazilian agencies ) for partial financial support .
h. atlan , _ application of information theory to the study of the stimulating effects of ionizing radiation , thermal energy , and other environmental factors _ , j. theoret
* 21 * , 45 ( 1968 ) ; _ on a formal definition of organization _ , j. theoret
. biol . * 45 * , 295 ( 1974 ) ; _ self - organizing networks : weak , strong and intentional , the role of their underdeter mination _ , in _ functional models of cognition _ , ed .
a. carsetti ( kluwer academic , amsterdam , 1999 ) , p. 127 . c. tsallis , j. stat .
_ possible generalization of boltzmann - gibbs statistics _ , j. stat .
52 * , 479 ( 1988 ) ; e.m.f .
curado and c. tsallis , _ generalized statistical mechanics : connection with thermodynamics _ , j. phys .
a * 24 * , l69 ( 1991 ) [ corrigenda : * 24 * , 3187 ( 1991 ) and * 25 * , 1019 ( 1992 ) ] ; c. tsallis , r.s .
mendes and a.r .
plastino , _ the role of constraints within generalized nonextensive statistics _ , physica a * 261 * , 534 ( 1998 ) .
a regularly updated bibliography can be accessed at http://tsallis.cat.cbpf.br/biblio.htm s.r.a .
salinas and c. tsallis , eds.,_nonextensive statistical mechanics and thermodynamics _
* 29 * ( 1999 ) ; s. abe and y. okamoto , eds . , _ nonextensive statistical mechanics and its applications _ , series _ lecture notes in physics _
* 560 * ( springer - verlag , heidelberg , 2001 ) [ isbn 3 - 540 - 41208 - 5 ] ; p. grigolini , c. tsallis and b.j . west , eds . ,
_ classical and quantum complexity and nonextensive thermodynamics _ , chaos , solitons and fractals * 13 * , number 3 , 371 ( pergamon - elsevier , amsterdam , 2002 ) ; c. tsallis , _ nonextensive statistical mechanics : a brief review of its present status _ , annals of the brazilian academy of sciences * 74 * , 393 ( 2002 ) ; g. kaniadakis , m. lissia and a. rapisarda , eds . , _ non extensive statistical mechanics and physical applications _ , physica a * 305 * , 129 ( 2002 ) ; m. gell - mann and c. tsallis , eds . , _ nonextensive entropy - interdisciplinary applications _ , ( oxford university press , new york , 2004 ) ; h.l .
swinney and c. tsallis , eds .
, _ anomalous distributions , nonlinear dynamics and nonextensivity _ , physica d * 193 * ( 2004 ) ; c. tsallis , _ algumas reflexoes sobre a natureza das teorias fisicas em geral e da mecanica estatistica em particular _ , in _ tendencias da fisica estatistica no brasil _ , ed . t. tome , volume honoring s.r.a .
salinas ( editora livraria da fisica , sao paulo , 2003 ) , page 10 .
aristotle , _ ars poetica _ [ the greatest thing by far is to be a master of metaphor .
it is the one thing that can not be learned from others ; it is also a sign of genius , since a good metaphor implies an eye for resemblance " ] . | simple memorizing tasks have been chosen such as a binary code on a @xmath0 matrix .
after the establishment of an appropriate protocol , the codified matrices were individually presented to 150 university students ( conveniently pre - selected ) who had to memorize them .
multiple presentations were offered seeking perfect performance verified through the correct reproduction of the code .
we measured the individual percentual error as a function of the number of successive presentations , and then averaged over the examined population .
the _ learning curve _ thus obtained decreases ( almost monotonically ) until becoming virtually zero when the number of presentations attains six .
a computer simulation for a similar task is available which uses a two - level perceptron on which an algorithm was implemented allowing for some degree of _ globality _ or _ nonlocality _ ( technically referred to as entropic _ nonextensivity _ within a current generalization of the usual , boltzmann - gibbs , statistical mechanics ) .
the degree of nonextensivity is characterized by an index @xmath1 , such that @xmath2 recovers the usual , extensive , statistical mechanics , whereas @xmath3 implies some degree of nonextensivity . in other words , @xmath4 is a ( very sensitive ) measure of globality ( gestalt perception or learning ) .
the computer curves fit well the human result for @xmath5 .
it has been verified that even extremely small departures of @xmath1 from unity lead to strong differences in the learning curve .
our main observation is that , for the very specific learning task on which we focus here , humans perform similarly to slightly nonextensive perceptrons .
in addition to this experiment , some preliminary studies were done concerning the human learning of ambiguous images ( based on figure - background perception ) . in spite of the complexity of drawing conclusions from such a comparison
, some generic trends can be established .
moreover , the enormous and well known difficulty for computationally defining semantic , hierarchic and strategic structures reveals clear - cut differences between human and machine learning . . |
nonhomogeneous media , multilayered structures and functionally graded materials ( fgms ) are of considerable technical and engineering importance as well as of mathematical interest ( see , e.g. , @xcite ) .
generally , such materials refer to heterogenous composite materials in which the material moduli vary smoothly and continously from point to point .
typical examples of fgms are ceramic / ceramic and metal / ceramic systems .
compressive reviews on aspects of fgms can be found in @xcite and @xcite .
however , only few investigations have been made to assess the role and the importance of dislocations in fgms .
for the first time , @xcite found the displacement and stress fields of a screw dislocation in an isotropic medium which is arbitrary graded in the @xmath0-direction .
one important example for a fgm are exponentially graded materials . for such media
the elasticity moduli are exponential functions depending on the space coordinates and the new material parameters describing the gradation ( see , e.g. , @xcite ) .
the green function for a two dimensional exponentially graded elastic medium has been found by @xcite and the three - dimensional green function is given by @xcite .
classical continuum theories possess no intrinsic length scales and , therefore , they are scale - free continuum theories which are not able to describe size effects being of importance at micro- and nanoscales and near defects . in order to describe such effects
generalized continuum theories are needed .
such generalized continuum theories are , for instance , strain gradient theories @xcite and the dislocation gauge theory @xcite which enrich the classical theories with additional material lengths . using strain gradient elasticity ,
cracks in fgms have been investigated by @xcite and @xcite .
for the first time ever , the solution of a screw dislocation in fgms , using the theory of gradient elasticity , was given by @xcite .
gradient elasticity is a theory which is very popular in engineering sciences .
but not everything is well - understood about the physical importance of the higher - order stresses ( hyperstresses ) .
gradient elasticity should be understood as an effective theory since it was invented for compatible deformations . on the other hand , a more physically motivated ` gradient - like '
theory is the so - called dislocation gauge theory . in the dislocation gauge theory both the incompatible elastic distortion tensor and the dislocation density tensor
are physical state quantities giving contribution to the elastic stored energy density .
also for nonhomogeneous media these physical state quantities are gauge - invariant . in this approach ,
the higher - order stress is realized as pseudomoment stress which is the response to the presence of dislocations .
this pseudomoment stress is related to the moment stress known from generalized elasticities like cosserat elasticity ( see , e.g. , @xcite ) .
thus , the force stress tensor is asymmetric like in cosserat elasticity . in general , in gradient elasticity the hyperstress ( double stress ) possesses five material moduli @xcite . in a simplified version of gradient
elasticity @xcite under ad - hoc assumptions the double stress can be simplified . in the dislocation gauge theory
such assumptions are not necessary . for that reason we use the gauge theory of dislocations in this paper .
in this paper we study a screw dislocation in fgm in the framework of the translation gauge theory of dislocations . in section 2
we start with the general framework of the translational gauge theory of dislocations and we derive the isotropic three - dimensional field equations of dislocations for nonhomogeneous media as well as for exponentially graded materials . in section 3
we specialize in the anti - plane problem and we give the gauge - theoretical solution of a screw dislocation in exponentially graded materials .
we find rigorous and ` simple ' analytical solutions for the force stresses , elastic distortion , torsion and pseudomoment stresses produced by a screw dislocation in fgms .
our hope is that the gauge solution will provide new physical insight to possible improvements of designing fgms .
this section introduces the notation and constitutive equations of the translational gauge theory of dislocations @xcite , which will be used to investigate a screw dislocation in fgms . in the dislocation gauge theory ,
the elastic distortion tensor @xmath2 is incompatible due to the occurrence of a so - called translational gauge field @xmath3 @xmath4 where @xmath5 is the displacement vector .
it is important to note that @xmath5 and @xmath3 are just the canonical field quantities and they are not unique and not physical state quantities . in presence of dislocations , @xmath5 and @xmath3 are discontinuous ( or multivalued ) fields .
the gauge field @xmath3 may be identified with the negative plastic distortion .
the appearance of incompatibility in form of the translation gauge field gives rise to an incompatibility tensor in the framework of the translation gauge theory of dislocations which is called the dislocation density tensor or torsion tensor .
accordingly , the dislocation density tensor is given in terms of the gauge field @xmath6 and , alternatively , in terms of the elastic distortion @xmath7 in the present framework , the elastic distortion tensor @xmath2 and the dislocation density tensor @xmath8 are the physical state quantities .
due to the form of the dislocation density tensor ( [ t1 ] ) and ( [ t2 ] ) it fulfills the translational bianchi identity @xmath9 eq .
( [ bi ] ) is the well - known conservation law of dislocations and states that a dislocation line can not end inside the body . in the static case of the ( linear ) translation gauge theory of dislocations
, the lagrangian density is of the bilinear form @xmath10 where @xmath11 denotes the stored energy density .
the canonical conjugate quantities ( response quantities ) are defined by @xmath12 where @xmath13 and @xmath14 are the force stress tensor and the pseudomoment stress tensor , respectively . here
, the force stress tensor @xmath13 is asymmetric .
the moment stress tensor @xmath15 can be obtained from the pseudomoment stress tensor : @xmath16k}$ ] ( see @xcite ) .
they have the dimensions : @xmath17=\text{force}/(\text{length})^2 \stackrel{\rm si}{=}\text{pa}$ ] and @xmath18=\text{force}/\text{length}\stackrel{\rm si}{=}\text{n}/\text{m}$ ] .
the euler - lagrange equations with respect to the canonical field variables derived from the total lagrangian density @xmath19 are given by @xmath20 we may add to @xmath21 a so - called null lagrangian , @xmath22 , with the ` background ' stress @xmath23 satisfying : @xmath24 . in terms of the canonical conjugate quantities ( [ can - qua ] ) , eqs . ( [ el-1 ] ) and ( [ el-2 ] )
take the form @xmath25 the force stresses are the sources of the pseudomoment stress .
in fact , one can see in eq .
( [ el-2b ] ) that the source of the pseudomoment stress tensor is an effective force stress tensor @xmath26 ( see , e.g. , @xcite ) .
this effective stress tensor , which is the difference between the gauge - theoretical stress field and the background stress field , drives the dislocation fields . the linear , isotropic constitutive relations for the force stress and the pseudomoment stress are @xmath27 here @xmath28 are the elastic stiffness moduli and @xmath29 denote the resistivity moduli associated with dislocations .
the six material moduli have the dimensions : @xmath30=\text{force}/(\text{length})^2 \stackrel{\rm si}{=}\text{pa}$ ] and @xmath31=\text{force}\stackrel{\rm si}{=}\text{n}$ ] . here
, the material moduli are nonhomogeneous that means they depend on the space coordinates : @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 and @xmath37 .
the constitutive relations for nonhomogeneous media ( [ ce-1 ] ) and ( [ ce-2 ] ) have the same formal form as the relations for homogeneous media .
the only difference is that the constitutive moduli are not constant .
the requirement of non - negativity of the stored energy ( material stability ) @xmath38 leads to the conditions of semi - positiveness of the constitutive moduli .
particularly , the constitutive moduli have to fulfill the following conditions @xcite @xmath39 if we substitute the constitutive equations ( [ ce-1 ] ) and ( [ ce-2 ] ) into the euler - lagrange equations ( [ el-1b ] ) and ( [ el-2b ] ) and use the definition ( [ t2 ] ) , we obtain the explicite form of the euler - lagrange equations for a nonhomogeneous material @xmath40 \nonumber\\ & \
+ c_{1,k}(\beta_{ik , j}-\beta_{ij , k } ) + c_{2,k}(\beta_{ji , k}-\beta_{jk , i}+\beta_{kj , i}-\beta_{ki , j } ) + c_{3,k}\ , \delta_{ij}(\beta_{lk , l}-\beta_{ll , k } ) + c_{3,i } ( \beta_{kk , j}-\beta_{kj , k } ) \nonumber\\ & \ + \lambda\ , \delta_{ij}(\beta_{kk}-\beta^0_{kk } ) + ( \mu+\gamma)(\beta_{ij}-\beta^0_{ij } ) + ( \mu-\gamma)(\beta_{ji}-\beta^0_{ji})=0\ , , \end{aligned}\ ] ] where @xmath41 denotes the ` classical ' elastic distortion tensor .
it can be seen that the space - dependence of the material moduli influences the field equations ( [ eom-1 ] ) and ( [ eom-2 ] ) due to gradients of the material moduli .
( [ eom-1 ] ) is a partial differential equation of first order for the elastic distortion as well as for the constitutive moduli . on the other hand , eq .
( [ eom-2 ] ) contains differential operators of second order acting on the elastic distortion and ` mixed ' gradients of the constitutive moduli and the elastic distortions . rather than looking for the general inhomogeneous behaviour of the constitutive moduli
, we specify here on an inhomogeneous material , which can find application to functionally graded materials .
we assume that the material properties very in a simple , explicite manner . here
, we consider exponential variations of the six constitutive moduli in the following manner @xmath42 where @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 and @xmath51 are the material parameters of the fgm .
the moduli @xmath49 , @xmath50 and @xmath51 are the characteristic parameters for exponentially gradation in @xmath0 , @xmath1 and @xmath52-direction and they have the dimensions : @xmath53=1/\text{length}$ ] . the parameters @xmath49 , @xmath50 and @xmath51 determine the gradation of all six constitutive moduli @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 .
for such an exponentially fgm , the equations ( [ eom-1 ] ) and ( [ eom-2 ] ) reduce to @xmath60 where @xmath61 denotes the three - dimensional nabla operator .
( [ eom-3 ] ) and ( [ eom-4 ] ) are the three - dimensional governing equations of dislocations in exponentially graded materials using the dislocation gauge theory .
we now derive the equations of motion for a screw dislocation in fgms .
we consider an infinitely long screw dislocation parallel to the @xmath52-axis with the burgers vector @xmath62 .
the symmetry of such a straight screw dislocation leaves only the following non - vanishing components of the elastic distortion tensor ( see , e.g. , @xcite ) : @xmath63 , @xmath64 , and for the dislocation density tensor : @xmath65 . for the anti - plane problem of a screw dislocation in nonhomogeneous materials
, the equilibrium condition ( [ eom-1 ] ) reduces to @xmath66 using equation ( [ ec - fgm ] ) , we obtain from ( [ eom-2 ] ) the following equations for the components @xmath63 and @xmath64 of a screw dislocation in nonhomogeneous materials @xmath67 for the two - dimensional anti - plane problem of fgms , we assume that the constitutive moduli are exponentially graded , @xmath68 where @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath49 and @xmath50 are the material constants of the anti - plane problem .
the moduli @xmath49 and @xmath50 are the characteristic parameters of exponentially gradation in @xmath0 and @xmath1-direction , respectively .
if we substitute ( [ fgm - cons ] ) into ( [ eom - bzx1])([eom - bzy2 ] ) , the governing equations are @xmath69 where @xmath70 and @xmath71 denote the two - dimensional laplacian and the nabla operator , respectively .
it can be seen in ( [ eom - bzx - fgm-1])([eom - bzy - fgm-2 ] ) that we have four equations for only two components @xmath63 and @xmath64 .
thus , it is clear that not all four equations ( [ eom - bzx - fgm-1])([eom - bzy - fgm-2 ] ) are independent .
in fact , if we compare ( [ eom - bzx - fgm-1 ] ) with ( [ eom - bzx - fgm-2 ] ) as well as ( [ eom - bzy - fgm-1 ] ) with ( [ eom - bzy - fgm-2 ] ) , we find the relation between the material moduli @xmath72 in addition , we may define the following intrinsic gauge length of the anti - plane problem of a screw dislocation @xmath73 evidently , the length scale @xmath74 is constant for fgms using ( [ fgm - cons ] ) .
this fact is in agreement with gradient elasticity for fgms @xcite where the characteristic gradient length is also constant .
only the constitutive moduli entering the constitutive relations are nonconstant . by means of ( [ rel - c ] ) and ( [ l ] )
the material moduli @xmath57 and @xmath58 can be expressed in terms of the length @xmath74 , and the constitutive moduli @xmath55 and @xmath56 @xmath75 thus , the gradation of @xmath57 and @xmath58 is given by the gradation of the constitutive moduli @xmath55 and @xmath56 .
accordingly , the governing equations for @xmath63 and @xmath64 reduce to @xmath76\beta_{zx}&=\beta^0_{zx}\,,\\ \label{bzy - pde } [ 1-\ell^2(\delta+2 \ba\cdot\!\nabla)]\beta_{zy}&=\beta^0_{zy}\ , .\end{aligned}\ ] ] due to the gradation term @xmath77 , eqs .
( [ bzx - pde ] ) and ( [ bzy - pde ] ) may be called ` perturbed ' helmholtz equations .
differentiating equation ( [ bzx - pde ] ) with respect to @xmath1 and equation ( [ bzy - pde ] ) with respect to @xmath0 and using the definition of the torsion ( [ t2 ] ) , the governing equation for the torsion ( dislocation density ) of a screw dislocation in fgm turns out to be @xmath78\,t_{zxy}&=t^0_{zxy}\,,\end{aligned}\ ] ] where @xmath79 is the dislocation density of a straight screw dislocation in classical elasticity . using the substitution @xmath80 we obtain from eq .
( [ t - pde ] ) the following helmholtz equation @xmath81\,\psi&= -\frac{b}{\ell^2}\ , \delta(x)\delta(y)\,,\end{aligned}\ ] ] with @xmath82 it can be seen that @xmath83 is the inverse intrinsic length characteristic for the anti - plane problem of fgms in the dislocation gauge approach .
therefore , @xmath84 is an effective length scale given in terms of the gauge length scale @xmath74 and the gradation length scale @xmath85 .
the inverse length scale of gradation @xmath86 is also constant .
thus , in addition to @xmath85 , two ( constant ) characteristic length scales , namely @xmath74 and @xmath84 , appear in the gauge - theoretical approach of the anti - plane problem of fgms .
the solution of ( [ psi - pde ] ) is given by @xmath87 where @xmath88 and @xmath89 is the @xmath90th order modified bessel function of the second kind .
finally , the torsion of a screw dislocation in fgm using the dislocation gauge theory is obtained as @xmath91 if we compare the dislocation density ( [ t - screw ] ) with the corresponding one for homogeneous media ( see , e.g. , @xcite ) , we observe the exponential factor and the length scale @xmath83 instead of @xmath92 as difference .
that means the dislocation density ( [ t - screw ] ) has lost the cylindrical symmetry and decays faster in the far - field in fgms . substituting ( [ t - screw ] ) into the constitutive relation ( [ ce-2 ] ) and making use of ( [ rel - c ] ) , we obtain for the components of the pseudomoment stress tensor @xmath93 due to the gradation , the fields ( [ t - screw])([h - yzx ] ) do not possess cylindrical symmetry in contrast to the results for homogeneous media .
it can be seen that the expressions ( [ t - screw])([h - yzx ] ) for @xmath94 and @xmath95 coincide with the result obtained by @xcite for homogeneous materials . to solve the remaining equations ( [ bzx - pde ] ) and ( [ bzy - pde ] )
, we use the constitutive relations ( [ ce-1 ] ) for the components @xmath96 and @xmath97 . in order to fulfill
the equilibrium condition for the forces stresses @xmath98 , we introduce the stress function @xmath99 according to @xmath100 by the help of equations ( [ sigma - zx - f ] ) and ( [ sigma - zy - f ] ) , we are able to express the elastic distortion @xmath63 and @xmath64 in terms of the stress function @xmath99 according @xmath101 of course , eqs .
( [ bzx - f ] ) and ( [ bzy - f ] ) satisfy the condition ( [ ec - fgm ] ) . if we substitute ( [ bzx - f ] ) and ( [ bzy - f ] ) into ( [ bzx - pde ] ) and ( [ bzy - pde ] ) , we find the following partial differential equation for the stress function @xmath99 @xmath102 f&=f^0\,,\end{aligned}\ ] ] where @xmath103 on the right hand side satisfies the equation @xmath104f^0=(\mu+\gamma)^0\ , b\ , \delta(x)\delta(y)\ , .\end{aligned}\ ] ] eq .
( [ f0-pde ] ) is a ` perturbed ' laplace equation .
the solution of ( [ f0-pde ] ) , which is actually the green function , reads ( see @xcite ) @xmath105 in addition , we use the ansatz for @xmath99 @xmath106 and obtain from ( [ f - pde ] ) the helmholtz equation for @xmath107 @xmath108\,\phi&= -(\mu+\gamma)^0\ , b\ , \delta(x)\delta(y)\,.\end{aligned}\ ] ] the solution of ( [ phi - pde ] ) is given by @xmath109 eventually , the stress function @xmath99 is calculated as @xmath110\,.\end{aligned}\ ] ] ( -7.5,2.0)@xmath63 ( -6.8,-0.3)@xmath111 ( 0,-0.3)@xmath112 ( -3.3,-0.3)@xmath113 ( -7.5,3.2)@xmath114 ( -6.8,-0.3)@xmath115 ( 3.6,-0.3)@xmath116 ( 0,-0.3)@xmath117 ( -1.5,1.0)@xmath118 ( -6.5,1.0)@xmath119 ( -6.2,-0.3)@xmath111 ( 0,-0.3)@xmath112 ( 5.5,1.0)@xmath118 ( 0.5,1.0)@xmath119 substituting ( [ f ] ) into ( [ bzx - f ] ) and ( [ bzy - f ] ) , we arrive at the expressions of the elastic distortions as @xmath120 -a_2 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , , \\
\label{bzy } \beta_{zy}&=\frac{b}{2\pi}\ , \e^{-(a_1 x+a_2 y ) } \big(\frac{x}{r}\big[a k_1(a r)-\kappa k_1(\kappa r)\big ] -a_1 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , .\end{aligned}\ ] ] the elastic distortions @xmath63 and @xmath64 are plotted in figs .
( [ fig : strain ] ) and ( [ fig : b3d ] ) .
first of all , it can be seen that the elastic distortions ( [ bzx ] ) and ( [ bzy ] ) are nonsingular unlike the result for fgms using elasticity theory and that they are given in terms of two inverse length scales @xmath86 and @xmath83 .
the component @xmath121 has extremum values near the dislocation line at @xmath122 and it has a finite value at @xmath122 depending on the material moduli . unlike homogeneous media , it does not satisfy : @xmath123 .
also @xmath124 possesses a maximum at @xmath125 and is asymmetric with respect to the plane @xmath125 : @xmath126 .
analogously , @xmath127 has extremum values near the dislocation line at @xmath125 and it has a finite value at @xmath125 .
it does not fulfill , @xmath128 , as for homogeneous materials .
evidently , @xmath129 has a minimum at @xmath122 and is asymmetric with respect to the plane @xmath122 : @xmath130 . for fgms the far - field of the curve of
the elastic distortion @xmath63 is for negative @xmath1 higher than the @xmath131 behavior known from homogeneous materials ( see fig . [
fig : strain]a ) . due to the gradation of the material
the elastic distortion in fgms decays faster than in homogeneous media . using ( [ ce-1 ] )
, the stresses can be derived to be @xmath132 -a_2 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , , \\
\label{txz } \sigma_{xz}&=-\frac{(\mu-\gamma)^0\ , b}{2\pi}\ , \e^{(a_1 x+a_2 y ) } \big(\frac{y}{r}\big[a k_1(a r)-\kappa k_1(\kappa r)\big ] -a_2 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , , \\
\label{tzy } \sigma_{zy}&=\frac{(\mu+\gamma)^0\,b}{2\pi}\ , \e^{(a_1 x+a_2 y ) } \big(\frac{x}{r}\big[a k_1(a r)-\kappa k_1(\kappa r)\big ] -a_1 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , , \\
\label{tyz } \sigma_{yz}&=\frac{(\mu-\gamma)^0\,b}{2\pi}\ , \e^{(a_1 x+a_2 y ) } \big(\frac{x}{r}\big[a k_1(a r)-\kappa k_1(\kappa r)\big ] -a_1 \big[k_0 ( ar)-k_0(\kappa r)\big]\big)\ , .\end{aligned}\ ] ] the stresses @xmath96 and @xmath97 are plotted in figs .
( [ fig : stress ] ) and ( [ fig : t3d ] ) .
( -7.5,2.0)@xmath96 ( -6.8,-0.3)@xmath111 ( 0,-0.3)@xmath112 ( -3.4,-0.3)@xmath133 ( -7.5,3.2)@xmath134 ( -6.8,-0.3)@xmath115 ( 3.6,-0.3)@xmath116 ( 0,-0.3)@xmath117 ( -1.5,1.0)@xmath118 ( -6.5,1.0)@xmath119 ( -6.2,-0.3)@xmath111 ( 0,-0.3)@xmath112 ( 5.5,1.0)@xmath118 ( 0.5,1.0)@xmath119 it is important to note that the stresses ( [ tzx])([tyz ] ) do not possess any singularity . due to the gradation ,
the components ( [ tzx])([tyz ] ) have extremum values near the dislocation line and a finite value at the dislocation line .
for fgms the far - field of the curves of the stresses are less than the @xmath135 behavior known from homogeneous materials ( see fig . [
fig : stress]a and [ fig : stress]d ) .
we note that the expressions ( [ bzx])([tyz ] ) coincide for @xmath94 and @xmath95 with the gauge - theoretical result obtained by @xcite for homogeneous materials and for @xmath94 and @xmath136 with the result given by @xcite using the strain gradient elasticity formulation for fgms .
in this paper , we have investigated a straight screw dislocation in fgms using the gauge theory of dislocations . in this framework , we have derived the field equations of dislocations for general nonhomogeneous media where the constitutive moduli depend on @xmath137 .
because of the gradation , the field equations contain gradients of the constitutive moduli .
subsequent we have specialized the field equations to exponentially graded materials .
we found a new effective inverse length scale @xmath83 characteristic for the gauge - theoretical anti - plane problem of fgms .
we have calculated the elastic distortion , the force stress , the torsion and the pseudomoment stress fields produced by a screw dislocation in fgms in the framework of dislocation gauge theory .
all the analytical solutions are given in terms of the effective length scale @xmath84 as well as the gradation length scale @xmath85 .
we found ` perturbed ' helmholtz equations as governing partial differential equations which we solved .
we found exact analytical gauge - theoretical solutions which can be used by material scientists and design and manufacturing engineers .
one hope is that these gauge solutions will provide new physical insight to possible improvements of designing fgms . in the limit @xmath138
, we recover well - known results of a screw dislocation in the gauge theory for homogeneous media .
chan , y .- s .
, paulino , g.h . , fannjing , a.c . , 2006 .
change of constitutive relations due to interaction between strain - gradient effect and material gradation .
asme journal of applied mechanics * 73 * , 871875
. chan , y .- s . ,
paulino , g.h .
, fannjing , a.c . , 2008 .
gradient elasticity for mode iii fracture in functionally graded materials
part ii : crack parallel to the material gradation .
asme journal of applied mechanics * 75 * , 061015 - 1(11 pages ) .
paulino , g.h . , fannjing , a.c . , chan , y .- s .
gradient elasticity for mode iii fracture in functionally graded materials part i : crack perpendicular to the material gradation .
asme journal of applied mechanics * 70 * , 531542 . | the aim of this paper is to provide new results and insights for a screw dislocation in functionally graded media within the gauge theory of dislocations .
we present the equations of motion for dislocations in inhomogeneous media .
we specify the equations of motion for a screw dislocation in a functionally graded material .
the material properties are assumed to vary exponentially along the @xmath0 and @xmath1-directions . in the present work
we give the analytical gauge field theoretic solution to the problem of a screw dislocation in inhomogeneous media . using the dislocation gauge approach , rigorous analytical expressions for the elastic distortions
, the force stresses , the dislocation density and the pseudomoment stresses are obtained depending on the moduli of gradation and an effective intrinsic length scale characteristic for the functionally graded material under consideration .
+ * keywords : * screw dislocation ; functionally graded material ; dislocation gauge theory ; size - effects . |
it is well established that ( 3 + 1)-dimensional quantum electrodynamics ( qed ) can describe the electromagnetic interaction between charged elementary particles with very high precision .
qed defined on ( 2 + 1)-dimensional space - time , dubbed qed@xmath0 , is safer in the high - energy region than its ( 3 + 1)-dimensional counterpart . in particular
, qed@xmath0 is a superrenormalizable field theory and therefore free of ultraviolet divergence .
extensive investigations have found that qed@xmath0 exhibits a number of interesting physical properties , such as dynamical chiral symmetry breaking ( dcsb ) @xcite , asymptotic freedom @xcite , and permanent confinement @xcite .
apparently , qed@xmath0 is more like four - dimensional qcd than qed@xmath2 .
for this reason , qed@xmath0 is often considered as a toy model of qcd@xmath2 in the context of particle physics .
more interestingly , qed@xmath0 has proven in the past twenty years to be an effective low - energy model for several strongly correlated condensed - matter systems , including high-@xmath3 cuprate superconductors @xcite , spin-@xmath4 kagome spin liquid @xcite , graphene @xcite , quantum critical systems @xcite , and kane - mele model with weak - extended hubbard interaction @xcite .
although qed@xmath0 can be well controlled in the ultraviolet regime , it encounters infrared problems since both dirac fermions and u(1 ) gauge field are massless .
for instance , the theory exhibits infrared divergence if the free gauge boson propagator is utilized in the perturbative calculations .
_ showed that such an infrared divergence can be naturally erased by including dynamical screening effect of massless fermions into the effective gauge boson propagator @xcite .
based on this scheme , appelquist _
@xcite investigated the dyson - schwinger equation ( dse ) of fermion self - energy function , and found that the massless fermions acquire a finite dynamical mass , which induces dcsb , when their flavor is below certain threshold , @xmath5 .
most of the existing analytical and numerical calculations @xcite agree that the critical flavor is @xmath6 at zero temperature . in the application of qed@xmath0 to condensed matter systems ,
dcsb is usually interpreted as the formation of heisenberg quantum antiferromagnetism @xcite .
it was found that dcsb in qed@xmath0 with finite gauge boson mass can still appear if the mass of the gauge boson is not large enough @xcite .
this model can describe the coexistence of antiferromagnetism and superconductivity in high - temperature cuprate superconductors @xcite .
the fate of dynamical mass generation at finite temperature @xcite has also attracted considerable interest .
notice that finite temperature qed@xmath0 is not only interesting from the viewpoint of quantum field theory , but of practical value since qed@xmath0 has wide applications in condensed matter physics .
an important problem is to estimate the critical temperature @xmath3 at which dynamical mass generation is destroyed by thermal fluctuation .
unfortunately , the study of dynamical mass generation at finite temperature seems to be more difficult than the case of zero temperature .
a main obstacle is that the lorentz invariance is explicitly broken at finite temperature .
the integration over 3-momenta @xmath7 that appears in the dse is replaced by an integration over 2-momenta @xmath8 and a summation over imaginary frequency @xmath9 , where @xmath10 is the mastubara frequency with @xmath11 being an integer . as a consequence ,
the dse becomes much more complicated , so it is usually necessary to make certain approximations to perform algebraic calculations .
a number of different approximations have been proposed to study dynamical mass generation in finite temperature qed@xmath0 @xcite .
the most frequently used one is the so - called instantaneous approximation , which assumes that the gauge boson propagator is completely independent of energy , @xmath12 in previous works @xcite , the transverse component of gauge boson propagator , @xmath13 , is usually ignored on top of the instantaneous approximation , i.e. , @xmath14 under these approximations , the dse for fermion mass becomes much simpler and can be easily solved . however , as showed in ref .
@xcite , the transverse component of gauge boson propagator @xmath13 , if incorporated in the dse , induces an infrared divergence at finite temperature . to obtain convergent results ,
lee proposed to simply neglect the contribution from @xmath13 @xcite .
however , neglecting the transverse component of gauge boson propagator is actually problematic .
an important reason for qed@xmath0 to exhibit dynamical fermion mass generation is that the gauge interaction is long - ranged , which is guaranteed by the gauge invariance .
if the gauge boson acquires a finite mass , the dynamical fermion mass generation will be strongly suppressed @xcite . at finite temperature ,
the lorentz invariance is explicitly broken , then the longitudinal and transverse components of gauge boson propagator behave very differently @xcite : the former becomes short - ranged due to thermal screening , whereas the latter remains long - ranged as a consequence of gauge invariance .
therefore , the transverse component of gauge boson propagator should play more significant role than the longitudinal component at finite temperatures .
this judgement is supported by the extensive recent analysis of nontrivial properties induced by gauge interaction @xcite .
overall , the transverse component of gauge boson propagator needs to be included in an appropriate manner in the study of dynamical fermion mass generation at finite temperature , which is the motivation of our work . in this paper
, we revisit the issue of dynamical mass generation in finite temperature qed@xmath0 . in order to obtain physically meaningful results ,
we go beyond the widely used instantaneous approximation , and employ a new approximation that ignores the energy dependence of dynamical fermion mass , i.e. , @xmath15 , but maintains both the longitudinal and transverse components of gauge boson propagator .
different form the instantaneous approximation , the new approximation does not completely neglect the energy dependence of the gauge boson . after performing extensive numerical computations
, we find that the dynamical fermion mass is free of divergence under the new approximation .
we also examine the validity of the new approximation by comparing to the case of zero temperature , and show that it is better than the instantaneous approximation . as aforementioned
, qed@xmath0 can serve as an effective low - energy field theory for several condensed matter systems @xcite . in these systems
, there are always certain amount of disorders , which couple to massless fermions and induce a finite damping rate .
disorders are responsible to the anomalous behaviors of a plenty of observable quantities of interacting systems of dirac fermions @xcite .
therefore , it is also interesting to study dynamical mass generation of qed@xmath0 in the presence of disorders .
to describe the influence of disorders , a finite fermion damping rate @xmath16 is introduced to the fermion propagator ref .
as will be shown below , @xmath16 plays analogous role as temperature @xmath17 , so its effect on dynamical mass generation can be studied in a similar manner to the case of finite temperature .
the rest of paper is organized as follows . in sec .
[ sec : lagrangian ] , we give the lagrangian and the relevant propagators . in sec .
[ sec : gapinstan ] , we show that the dynamical mass is divergent under the instantaneous approximation if temperature or damping rate is finite . in sec . [
sec : gapnewappro ] , we solve the dse for dynamical mass by invoking a new approximation , and find that the results are free of infrared divergence .
the nature of the infrared divergence is also discussed . in sec .
[ sec : gapzero ] , we examine the validity of the new approximation by comparing to the case of zero temperature . in sec .
[ sec : conclusion ] , we summarize the main results . detailed calculations of polarization functions are given in appendix [ appendix : polarization ] .
the lagrangian density for qed@xmath0 with @xmath1 flavors of massless dirac fermions is given by @xmath18 the electromagnetic tensor is related to vector potential @xmath19 as @xmath20 .
the fermion is described by a four - component spinor field @xmath21 , whose conjugate spinor field is @xmath22 .
the @xmath23 gamma matrixes are defined as @xmath24 , which satisfy the standard clifford algebra @xmath25 with the metric being @xmath26 . in ( 2 + 1 ) dimensions , there are two chiral matrices @xmath27 and @xmath28 @xcite , @xmath29 which anticommute with @xmath30 , @xmath31 and @xmath32 . the lagrangian shown in eq .
( [ eq : langrangian ] ) respects a continuous @xmath33 chiral symmetry @xmath34 , where @xmath35 is an arbitrary constant .
once a finite fermion mass is dynamically generated , the global @xmath33 chiral symmetry is spontaneously broken down to its subgroup @xmath36 . in this paper
, we consider a general large @xmath1 and perform perturbative expansion in powers of @xmath37 . for convenience
, we work in units with @xmath38 and set the velocity @xmath39 . in the euclidian space ,
the free propagator of massless fermions is @xmath40 at zero temperature .
the free propagator of gauge boson can be written as @xmath41 in the landau gauge .
after including the dynamical screening effect due to fermions , the effective gauge boson propagator takes the form @xmath42 to the lowest order of @xmath37-expansion , the polarization function can be obtained from the polarization tensor through the relationship @xmath43\nonumber \\ & = & \pi(q)\left(\delta_{\mu\nu}-\frac{q_{\mu}q_{\nu}}{q^{2}}\right),\end{aligned}\ ] ] where @xmath44 with @xmath45 @xcite .
first carefully studied dcsb in qed@xmath46 by means of non - perturbative dseapproach , and showed that dcsb takes place for any finite flavor @xmath1 . however , subsequent analysis of appelquist _ et .
_ @xcite found that dcsb can occur only when the fermion flavor @xmath1 is smaller than a critical value @xmath47 , which is @xmath48 to the lowest order of @xmath37-expansion .
nash @xcite then examined the effect of next - to - leading order correction and obtained a critical flavor @xmath49 .
@xcite included the wave renormalization function and claimed that dcsb can be realized for any flavor @xmath1 , although the corresponding dynamical fermion mass decreases exponentially with increasing @xmath1 .
nevertheless , their analysis ignored the influence of wave renormalization function on the polarization .
later , maris @xcite studied dcsb by solving a set of self - consistent dses for fermion propagator and polarization , which contain vertex corrections and therefore satisfy the ward - takahashi ( wt ) identities .
the calculations of ref .
@xcite arrived at a finite critical flavor @xmath50 , which is close to that of ref .
the key difference between the treatments of pennington _
et al_. @xcite and maris @xcite is that the latter included the interaction correction to the polarization whereas the former did not .
it turns out that an appropriate approximation plays a crucial role in the dse analysis of dcsb .
more refined calculations of fisher _ et al_. revealed a finite critical flavor @xmath51 @xcite .
the gauge invariance of @xmath52 is also discussed @xcite . apart from the dse approach
, this problem can be studied by renormalization group method , which found that @xmath53 @xcite .
the critical flavor @xmath52 obtained in lattice monto carlo simulations @xcite is much smaller than that obtained by means of dses .
however , gusynin _ et al .
_ argued that the difference is attributed to the finite volume effect introduced in lattice simulations @xcite .
it is fairy to say that , although there is still some debate @xcite , most studies have obtained a finite critical flavor for dcsb in qed@xmath46 , which is roughly @xmath6 at @xmath54 .
as we go to finite temperature , the dynamical fermion mass is expected to be rapidly suppressed by thermal fluctuation .
the temperature scale @xmath3 at which the dynamical mass vanishes defines the critical temperature . at @xmath55
, we write the fermion propagator in the standard matsubara formalism as @xmath56 where @xmath57 with @xmath11 being an integer . here , we introduce a constant @xmath16 to represent the fermion damping rate generated by disorder scattering .
this quantity measures the strength of the fermion damping effect . for more explanation of the origin and the physical effect of the constant @xmath16 ,
please see ref .
@xcite . to the leading order of @xmath37-expansion ,
the polarization tensor is given by @xmath58,\end{aligned}\ ] ] where @xmath59 with @xmath60 being an integer and @xmath61 .
the effective propagator of gauge boson now becomes @xmath62 where @xmath63 @xmath64 and @xmath65 are orthogonal and satisfy @xmath66 the functions @xmath67 and @xmath68 are defined as @xmath69 the calculational details of the relevant polarization functions are shown in appendix [ appendix : polarization ] . in the following ,
we will consider two approximations , namely the popular instantaneous approximation and a new approximation to be explained below .
we neglect the energy - dependence of the polarization functions and also the feedback of dynamical mass to the polarization functions .
firstly , in the limit that @xmath70 , @xmath71 , and @xmath72 , the polarization functions are @xmath73 , \label{eq : piat } \nonumber \\ \\
\pi_{b}(\mathbf{q},t ) & = & \frac{8\alpha}{\pi } \int_{0}^{1}dx\sqrt{x(1-x)\mathbf{q}^2 }
\nonumber \\ & & \times \tanh\left(\frac{\sqrt{x(1-x)\mathbf{q}^2}}{2t}\right ) .
\label{eq : pibt}\end{aligned}\ ] ] we have used @xmath74 . secondly , in the limit that @xmath70 , @xmath54 and @xmath72 , the polarization functions are @xmath75 \nonumber \\ & & + \int_{0}^{1}dx\sqrt{x(1-x)\mathbf{q}^{2}}\nonumber \\ & & \left.\times\arctan \left(\frac{\sqrt{x(1 - x)\mathbf{q}^{2}}}{\gamma}\right)\right\},\label{eq : piagamma } \\ \pi_{b}(\mathbf{q},\gamma ) & = & \frac{16\alpha}{\pi^2 } \int_{0}^{1}dx\sqrt{x(1-x)\mathbf{q}^{2 } } \nonumber \\ & & \times \arctan \left(\frac{\sqrt{x(1-x)\mathbf{q}^{2}}}{\gamma}\right ) .
\label{eq : pibgamma}\end{aligned}\ ] ] from these expressions , we can see that the fermion damping rate @xmath16 plays a very similar role to temperature @xmath17 , which allows us to study the effects of temperature and fermion damping using the same scheme . to simplify later calculations , we can further approximate eqs .
( [ eq : piat ] ) and ( [ eq : pibt ] ) by @xcite @xmath76 , \label{eq : piatap } \\
\pi_{b}(\mathbf{q},t ) & \approx&\alpha|\mathbf{q}|\tanh\left ( \frac{c_{2}|\mathbf{q}|}{t}\right ) , \label{eq : pibtap}\end{aligned}\ ] ] where @xmath77 and @xmath78 .
analogously , eqs .
( [ eq : piagamma ] ) and ( [ eq : pibgamma ] ) can be approximated by @xmath79 where @xmath80 .
it can be checked numerically that expressions eqs .
( [ eq : piatap])-([eq : pibgammaap ] ) are very good approximations for both the high- and low - momentum behaviors of eqs .
( [ eq : piat])-([eq : pibgamma ] ) . in the following sections
, we will use eqs .
( [ eq : piatap])-([eq : pibgammaap ] ) to analyze and numerically solve the dses for dynamical fermion mass at finite temperature and/or finite fermion damping rate .
in this section , we present the dse for dynamical mass under the widely used instantaneous approximation . we will show that the solutions of dse are divergent in the infrared region whenever @xmath55 and therefore ill - defined .
such an infrared divergence also exists if @xmath54 but @xmath81 .
the free fermion propagator is @xmath82 due to gauge interaction , the fermion may become massive and the propagator is renormalized to @xmath83 here , to the lowest order of @xmath37-expansion , we neglect the wave renormalization function .
now , @xmath84 is related to @xmath85 through the following dse , @xmath86 where @xmath87 with @xmath88 and @xmath89 .
substituting eqs .
( [ eq : fermionpropagatorfree ] ) and ( [ eq : fermionpropagatorfull ] ) into eq .
( [ eq : selfenergysc ] ) , and then giving an explicit @xmath17- and @xmath16-dependence to the fermion mass , we obtain @xmath90.\label{eq : gapeqoriginal}\end{aligned}\ ] ] in previous studies @xcite , the key assumption beneath the instantaneous approximation is to completely ignore the energy - dependence of gauge boson propagator .
apart from this approximation , the transverse component of gauge boson propagator is widely neglected , i.e. , @xmath91 now we follow the instantaneous approximation and drop the energy dependence of gauge boson propagator .
nevertheless , as illustrated in recent works on the nontrivial properties of qed@xmath0 @xcite , the transverse component of gauge interaction plays more important role than the longitudinal component at finite temperature , since the latter becomes short - ranged after acquiring an effective thermal mass proportional to @xmath17 .
if @xmath92 and @xmath81 , the longitudinal part of gauge interaction also becomes short - ranged due to static screening caused by disorder scattering . in any case
, the gauge invariance ensures that the transverse component of gauge interaction is strictly long - ranged .
it is therefore not appropriate to neglect the transverse component . for completeness , here we keep them both and write the gauge boson propagator as @xmath93 the corresponding dse becomes @xmath94.\end{aligned}\ ] ] once the instantaneous approximation is adopted , the dynamical mass @xmath95 completely loses its dependence on energy .
therefore , the frequency summation in the dse can be written as @xmath96 } = \left(\frac{\beta}{2\pi}\right)^{2}\sum_{n=-\infty}^{\infty } \frac{1}{\left[\left(n+\frac{1}{2 } + x \mathrm{sgn}\left(n+\frac{1}{2}\right)\right)^{2 } + y^2\right]},\end{aligned}\ ] ] where @xmath97 , and @xmath98 . with the help of an identity @xmath99\end{aligned}\ ] ] where @xmath100 is digamma function
, we get @xmath101 } = \frac{\beta^2}{2\pi^2y } \mathrm{im}\left[\psi\left(\frac{1}{2}+x+iy\right)\right].\end{aligned}\ ] ] carrying out frequency summation leads to @xmath102\nonumber \\ & & \times\left[\frac{1}{\mathbf{q}^{2}+\pi_{a}(0,\mathbf{q},t,\gamma ) } + \frac{1}{\mathbf{q}^2+\pi_{b}(0,\mathbf{q},t,\gamma)}\right].\end{aligned}\ ] ] this equation can be further simplified .
first , we assume @xmath71 and @xmath55 , and then rewrite the dse as @xmath103,\label{eq : gapftinstanappro}\end{aligned}\ ] ] where @xmath104 = \frac{\pi}{2}\tanh\left(\frac{\xi}{2t}\right)$ ] is used in the derivation . secondly , at @xmath54 and @xmath81
, we have @xmath105 , \label{eq : gapfgammainstanappro}\end{aligned}\ ] ] where @xmath106 = \arctan\left(\frac{\xi}{\gamma}\right)$ ] . the similarity between the above two equations
can be clearly seen . at @xmath92 ,
the constant fermion damping rate @xmath16 can be considered as certain effective temperature .
this is an important reason for us to discuss these two cases simultaneously in this paper . and
@xmath107 for different temperatures at @xmath71 with @xmath108 in ( a ) and ( b).,title="fig:",width=288 ] and @xmath107 for different temperatures at @xmath71 with @xmath108 in ( a ) and ( b).,title="fig:",width=288 ] and @xmath107 for different @xmath16 at @xmath54 with @xmath108 in ( a ) and ( b ) respectively.,title="fig:",width=288 ] and @xmath107 for different @xmath16 at @xmath54 with @xmath108 in ( a ) and ( b ) respectively.,title="fig:",width=288 ] to solve the dse , it is convenient to introduce an infrared cutoff @xmath109 .
this amounts to assume that the system has a finite volume @xcite . for finite @xmath109 ,
the dynamical fermion mass obtained from dse is always free of infrared divergence .
our aim is to examine whether the dynamical mass is still well defined as @xmath110 .
we employ an infrared cutoff @xmath111 for momentum @xmath112 and replace @xmath113 with @xmath114 .
usually , @xmath111 and @xmath115 should satisfy @xmath116 . for simplicity ,
we assume @xmath117 . now
the dse at @xmath71 and @xmath55 becomes @xmath118,\end{aligned}\ ] ] where @xmath119 is the angle between @xmath120 and @xmath121 . as usual , we choose @xmath122@xcite .
the dse at @xmath54 and @xmath81 can be transformed analogously . for different finite temperatures , the dynamical mass at the lowest momentum @xmath109 with @xmath123 and @xmath124
are shown in figs .
[ fig : gapmuftinstan ] ( a ) and ( b ) respectively . as @xmath125
, @xmath126 does not converge to a finite value , but is divergent . as @xmath17 further decreases , @xmath126 grows more slowly , but still divergent at finite @xmath17 . for different values of @xmath16 ,
the relations between @xmath126 and @xmath109 with @xmath123 and @xmath124 are displayed in fig .
[ fig : gapmufgammainstan](a ) and ( b ) respectively .
numerical calculations show that @xmath126 is ill defined as @xmath127 whenever @xmath81 .
apparently , the instantaneous approximation leads to divergent results .
we have shown in the last section that the widely used instantaneous approximation leads to infrared divergence in the dynamical fermion mass obtained at finite temperature or in the presence of a finite fermion damping rate , when both the longitudinal and transverse components of gauge boson propagator are adopted .
it is therefore necessary to go beyond this approximation and seek a new approximation that could yield convergent results .
we find it very helpful to employ a treatment utilized by ref .
@xcite in the study of excitonic insulating transition in graphene @xcite .
the key assumption of this treatment is to neglect the energy dependence of dynamical fermion mass , which was shown @xcite to generate physically reliable dynamical mass for dirac fermions .
let us start from the dse given by ( 28 ) .
now we assume the dynamical fermion mass does not depend on energy , namely @xmath128 after making this approximation , the dse ( [ eq : gapeqoriginal ] ) can be written in the following form @xmath129 .
\label{eq : gapeqftfgammanewappra}\end{aligned}\ ] ] notice the effective gauge boson propagator still retains an explicit energy dependence , which is thus different from the instantaneous approximation .
however , the summation over @xmath130 can not be performed exactly due to the complicated @xmath130 dependence of polarization functions . to further simplify the dse
, we drop the energy dependence of polarization functions , i.e. , @xmath131 the dses then can be formally written as @xmath132\label{eq : gapeqftnewapprb}\end{aligned}\ ] ] in the case of @xmath55 and @xmath133 , and to @xmath134 \label{eq : gapeqfgammanewapprb}\end{aligned}\ ] ] in the case of @xmath81 and @xmath54 .
the infrared behaviors of these two equations will be further analyzed later .
it is now straightforward to sum over @xmath130 , which leads us to the following expressions : @xmath135 \nonumber \\ & & \left.+\left[\pi_{a}(\mathbf{q},t ) \rightarrow \pi_{b}(\mathbf{q},t)\right]\right\},\\ \label{eq : gapeqftnewappr } m(\mathbf{p},\gamma ) & = & \frac{8\alpha}{n\pi}\int \frac{d^2\mathbf{k}}{(2\pi)^2 } \left\{\frac{m(\mathbf{k},\gamma)}{\left(\gamma^2+\mathbf{k}^{2 } + m^2(\mathbf{k},\gamma)-\mathbf{q}^{2 } - \pi_{a}(\mathbf{q},\gamma)\right)^2 + 4\gamma^2 \left(\mathbf{q}^{2 } + \pi_{a}(\mathbf{q},\gamma)\right)}\right.\nonumber \\ & & \times\left[\frac{\gamma^2-\mathbf{k}^{2 } - m^2(\mathbf{k},\gamma)+\mathbf{q}^{2 } + \pi_{a}(\mathbf{q},\gamma)}{\sqrt{\mathbf{k}^{2 } + m^2(\mathbf{k},\gamma)}}\left(\frac{\pi}{2 } - \arctan\left(\frac{\gamma}{\sqrt{\mathbf{k}^{2 } + m^2(\mathbf{k},\gamma)}}\right)\right)\right.\nonumber \\ & & \left.+\frac{\pi}{2}\frac{\gamma^2+\mathbf{k}^{2 } + m^2(\mathbf{k},\gamma)-\mathbf{q}^{2 } - \pi_{a}(\mathbf{q},\gamma)}{\sqrt{\mathbf{q}^{2 } + \pi_{a}(\mathbf{q},\gamma ) } } - \gamma\ln\left(\frac{\gamma^2+\mathbf{k}^{2 } + m^2(\mathbf{k},\gamma)}{\mathbf{q}^{2 } + \pi_{a}(\mathbf{q},\gamma)}\right)\right ] \nonumber \\ & & \left.+\left[\pi_{a}(\mathbf{q},\gamma)\rightarrow \pi_{b}(\mathbf{q},\gamma)\right]\right\}. \label{eq : gapeqfgammanewappr}\end{aligned}\ ] ] and @xmath107 for different @xmath17 with @xmath108 in ( a ) and ( b ) .
results are obtained under the new approximation.,title="fig:",width=288 ] and @xmath107 for different @xmath17 with @xmath108 in ( a ) and ( b ) .
results are obtained under the new approximation.,title="fig:",width=288 ] and @xmath107 for different @xmath16 with @xmath108 in ( a ) and ( b ) .
results are obtained under the new approximation.,title="fig:",width=288 ] and @xmath107 for different @xmath16 with @xmath108 in ( a ) and ( b ) .
results are obtained under the new approximation.,title="fig:",width=288 ] following the approach used in sec .
[ sec : gapinstan ] , we introduce an infrared cutoff @xmath109 to eqs .
( [ eq : gapeqftnewappr ] ) and ( [ eq : gapeqfgammanewappr ] ) , and then solve these equations numerically . in the case that @xmath55 and @xmath133 , the dependence of dynamical mass @xmath126 on @xmath109 at finite @xmath17 with @xmath123 and @xmath124 is shown in figs .
[ fig : gapftnewappro ] ( a ) and ( b ) respectively . as
the infrared cutoff vanishes , @xmath136 , @xmath126 is saturated to certain finite values , which means the infrared divergence encountered under the instantaneous approximation does not exist under the new approximation .
analogous calculations can be carried out in the case that @xmath92 and @xmath137 .
we display the dependence of @xmath126 on @xmath109 with @xmath123 and @xmath124 in figs .
[ fig : gapfgammanewappro](a ) and ( b ) respectively .
it is easy to see that @xmath126 also approaches finite values as @xmath110 .
apparently , the new approximation adopted in this section leads to convergent results for the dynamical fermion mass , and is therefore more reliable than the instantaneous approximation .
we fix the infrared cutoff at @xmath138 and calculate the dynamical gap @xmath139 under the new approximation .
the calculated @xmath139 under the new approximation at different values of @xmath17 or different values of @xmath16 is shown in figs .
[ fig : gappfnewappro ] ( a ) and ( b ) respectively .
let us now analyze the origin of the infrared divergence . at @xmath140 and @xmath133 ,
the dse has the form @xmath141 .
\label{eq : gaptoriginal}\end{aligned}\ ] ] since @xmath95 is finite after dynamical mass generation and @xmath142 in the low energy limit , the terms appearing in the first and second lines of the integral kernel are safe in the infrared region
. however , the term appearing in the third line contains a potential infrared divergence . to make this transparent
, we divide the summation over @xmath130 as follows , @xmath143 notice that @xmath144 is always finite at finite @xmath17 whenever @xmath145 , so the first term does not yield any infrared divergence . on the contrary ,
the second term is dangerous because @xmath146 with @xmath147 for small momenta .
we now simply focus on the potential divergent term , and find that @xmath148 where @xmath149 .
it is clear that this term is divergent as the infrared cutoff @xmath110 . from the above analysis
, we see that the infrared divergence of dse comes from the zero - energy transfer processes mediated by the singular transverse component of gauge boson propagator .
lee @xcite noticed the existence of infrared divergence , and then simply neglected the transverse component of gauge boson propagator .
the same strategy is widely utilized in other works @xcite .
more recently , lo and swanson @xcite also stressed the existence of infrared divergence and proposed to remove this divergence by choosing an appropriate temperature dependent gauge parameter .
their approach is basically equivalent to considering the following dse , @xmath150,\nonumber\end{aligned}\ ] ] which ignores the zero frequency ( @xmath151 ) contribution of the transverse component of gauge boson propagator .
careful numerical computation of this equation is interesting , but challenging since it is hard to sum over @xmath11 and at the same time integrate over @xmath152 with high precision .
this is subjected to future investigation . under the new approximation ,
the dse of dynamical mass is simply eq .
( [ eq : gapeqftnewapprb ] ) .
the denominator of the kernel of this equation contains a factor of @xmath153 , whose minimum is @xmath154 .
we may consider this term as an effective thermal mass of gauge boson , i.e. , @xmath155 , which serves as an infrared regulator and eliminates the potential infrared divergence of the dynamical fermion mass .
this thermal mass exists only at finite @xmath17 , and vanishes naturally as @xmath156 . to examine to
what extent the new approximation is valid , we will show in the next section that eq .
( [ eq : gapeqftnewapprb ] ) leads to results qualitatively consistent with those obtained at @xmath92 @xcite .
we then turn to the case of finite fermion damping rate . at @xmath92 ,
the dse is given by eq .
( [ eq : gapeqfgammanewapprb ] ) .
the potential infrared divergence can only come from the second term in the bracket .
the most singular part is represented by @xmath157 in order to analyze potential infrared divergence , we replace @xmath158 with @xmath159 , then @xmath160 for small momenta , @xmath159 behaves as @xmath161 with @xmath162 .
it is clear that the potential infrared divergence can be represented by @xmath163 where @xmath164 . as @xmath110
, @xmath165 is definitely not divergent .
this property should be fulfilled no matter what approximation is used to calculate the dse . under the instantaneous approximation ,
the dse is represented by eq .
( [ eq : gapfgammainstanappro ] ) , which contains the following singular contribution , @xmath166 this function is divergent as @xmath110 .
apparently , the instantaneous approximation brings an artificial infrared divergence that should not exist . under the new approximation proposed by us ,
the dse is given by eq.([eq : gapeqfgammanewapprb ] ) .
the only possible singular part can be simply written as @xmath167 which is not divergent and well consistent with eq .
in sec.iv , we have adopted a new approximation to study the dse for dynamical fermion mass in the presence of finite temperature or finite fermion damping . compared to the popular instantaneous approximation
, the main advantage of the new approximation is that , it retains both the longitudinal and transverse components of gauge boson propagator , and at the same time leads to physically meaningful , convergent results . to further see this point ,
it is now interesting to make a more straightforward comparison between these two approximations . in this section
, we consider zero temperature qed@xmath46 in the clean limit , namely @xmath168 .
this model has already been extensively investigated , and the dynamical mass obtained from dse is free of infrared divergence , which then can be considered as a reference to examine the reliability of the results obtained at finite temperature . for an approximation to be reliable
, it should work well at both zero and finite temperatures .
as @xmath169 , @xmath170 oughts to approach a well - defined quantity @xmath171 , which should be free of infrared divergence and as close in quantity as possible to that obtained directly at @xmath92 .
we now examine whether the new approximation is valid according to this criterion .
if one neglects the feedback of fermion mass to the polarization functions , the dse to the lowest order of @xmath37-expansion is known to be @xcite @xmath172 with @xmath44 .
this equation was first solved in ref.@xcite , and the solution is very well known . under the instantaneous approximation ,
the dse is simplified to @xmath173 with @xmath174 . under the new approximation ,
the corresponding dse is @xmath175 where @xmath176 also equals to @xmath177 .
the dynamical fermion mass @xmath126 as a function @xmath109 obtained in three cases are shown in fig .
[ fig : gapzerot ] ( a ) , represented by lines with different colors .
we notice that @xmath126 is saturated to finite values in all these three cases when @xmath110 . however , the fermion mass obtained in the new approximation is closer to that obtained directly from eq .
( [ eq : gapzerotzerogamma ] ) than the instantaneous approximation .
it seems that both of these two approximations lead to convergent results for dynamical mass .
nevertheless , we still need to examine whether these results are robust against higher order corrections .
we then include the feedback of dynamical fermion mass to the polarization . the polarization appearing in eq .
( [ eq : gapzerotzerogamma ] ) becomes @xmath178,\end{aligned}\ ] ] and the polarization appearing in eqs .
( [ eq : gapzerotzerogammainstan ] ) and ( [ eq : gapzerotzerogammanew ] ) can be obtained by replacing @xmath179 of @xmath180 with @xmath113 .
in addition , @xmath181 is substituted by @xmath126 .
the dependence of dynamical mass @xmath126 on @xmath109 is depicted in fig .
[ fig : gapzerot](b ) .
we find that @xmath126 diverges as @xmath110 under the instantaneous approximation .
nevertheless , @xmath126 obtained in the other two cases does not exhibit infrared divergence and remains finite as @xmath110 .
these results further demonstrate that the instantaneous approximation yields unphysical results , and that the new approximation is more reliable .
in this paper , we have studied dynamical fermion mass generation in qed@xmath0 after including the effects of finite temperature or finite fermion damping rate .
many previous dse calculations of dynamical fermion mass adopted an instantaneous approximation , which is often accompanied by simply ignoring the transverse component of gauge interaction @xcite .
as already explained in the context , at finite temperature or at a finite fermion damping rate , the longitudinal component of gauge interaction becomes short - ranged due to static screening , whereas the transverse component of gauge interaction remains long - ranged as required by the local gauge invariance .
it is therefore not appropriate to ignore the more important contribution of gauge interaction .
however , we have showed that , if one adopts the instantaneous approximation and meanwhile includes the complete gauge boson propagator , the dynamical fermion mass exhibits infrared divergence . we have revisited this problem and employed a new approximation to calculate the dse for dynamical fermion mass . under the new approximation ,
both the longitudinal and transverse components of gauge interaction are incorporated , and the results obtained from in the dse are free of infrared divergence . to further examine the validity of the new approximation ,
we have also make a comparison to the results obtained directly at zero temperature . in summary
, our calculations have showed that the new approximation leads to more reliable results for the dynamical fermion mass than the widely used instantaneous approximation .
the existence of infrared divergence is not special to the issue of dynamical mass generation in finite temperature qed@xmath0 . indeed
, similar divergence appears in a number of interacting gauge field theories .
for instance , lee calculated the fermion damping rate caused by gauge interaction within an effective non - relativistic u(1 ) gauge field theory , and found non - fermi liquid behavior at zero temperature @xcite . nevertheless , the fermion damping rate diverges at finite temperature @xcite . recently , analogous divergence is also found in qed@xmath0 defined at finite temperature and finite chemical potential @xcite .
we hope the approach proposed and used in this paper could provide useful insight into this problem .
confinement is an important feature of qed@xmath0 .
it is known from previous studies @xcite that whether this model is confining depends crucially on the behavior of the polarization function in the low energy regime , which is in turn determined by the dynamical fermion mass .
it would be interesting to apply the new approximation proposed here to carefully calculate the polarization function at finite @xmath17 by including the impact of dynamical fermion mass , and then to evaluate the critical temperature for the confinement - deconfinement transition , following the schemes presented in refs.@xcite .
it would also be interesting to examine whether confinement and dynamical fermion mass generation take place simultaneously by analyzing the behavior of wave function renormalization @xcite . to address these issues , one needs to incorporate the wave function renormalization and the vertex functions in the dses , which are subjected to future research .
the authors acknowledge the financial support by the national natural science foundation of china under grants 11274286 , 11174290 , and u1232142 .
in the appendix , we present the detailed calculations for the polarization functions in the presence of finite temperature @xmath17 and finite fermion damping rate @xmath16 .
the calculations are performed within the standard matsubara formalism for finite temperature quantum field theory .
starting from the effective fermion propagator given by eq .
( [ eq : gtgammamass ] ) , we write the polarization functions @xmath183 and @xmath184 in the following forms @xmath185,\label{eq : pi00defappen1 } \\
\pi_{ii}(q_{0},\mathbf{q},t , m_0,\gamma ) & = & \frac{ne^{2}}{\beta } \sum_{n=-\infty}^{+\infty}\int\frac{d^{2}\mathbf{k}}{(2\pi)^{2 } } \mathrm{tr}\left[g(k_{0},\mathbf{k})\gamma_{i}g(k_{0 } + q_{0},\mathbf{k}+\mathbf{q})\gamma_{i}\right],\label{eq : piiidefappen1}\end{aligned}\ ] ] where @xmath186 and @xmath187 with @xmath11 and @xmath60 being integers , to the leading order of @xmath37-expansion .
substituting eq .
( [ eq : gtgammamass ] ) into eqs .
( [ eq : pi00defappen1 ] ) and ( [ eq : piiidefappen1 ] ) , and then using the feynman parametrization formula @xmath188 ^ 2},\ ] ] we can get @xmath189s_2\right.\nonumber \\ & & \left.+ \left[(1 - 2x)q_{0}+\delta\right]s^*\right\ } , \\ \pi_{ii}(q_{0},\mathbf{q},t ,
m_0,\gamma ) & = & -\frac{8ne^{2}}{\beta } \int_{0}^{1}dx\int\frac{d^{2}\mathbf{l}}{(2\pi)^{2}}\left\{s_1- \left[\mathbf{l}^{2}+2x(1-x)\left(q_{0}^{2}+\frac{\mathbf{q}^{2}}{2}\right ) + x\left(2k_{0}+3q_{0}\right)\delta\right]s_2\right . \nonumber \\ & & \left.+ \left[(1 - 2x)q_{0}+\delta\right]s^*\right\},\end{aligned}\ ] ] with @xmath190^{i } } , \\ s^ { * } & = & \sum_{n=-\infty}^{\infty}\frac{l_{0}}{\left[l_{0}^{2 } + \mathbf{l}^{2}+m_{0}^{2}+x(1-x)q^2 + 2x\left(k_{0}+q_{0}\right)\delta\right]^{2}}.\end{aligned}\ ] ] where @xmath191 with @xmath192 , @xmath193 , and @xmath194 $ ] . when @xmath195 , the frequency summation can not be carried out precisely .
there are two ways to make @xmath196 .
first , @xmath197 , corresponding to the static limit .
second , @xmath133 , corresponding to the clean limit of the system ( without any disorder ) .
next we calculate the polarization functions in these two cases respectively . for a general constant @xmath16 , we have @xmath199,\label{eq : pi00defappen2 } \\
\pi_{ii}(\mathbf{q},t , m_0,\gamma ) & = & -\frac{8ne^{2}}{\beta } \int_{0}^{1}dx\int\frac{d^{2}\mathbf{l}}{(2\pi)^{2}}\left[s_1- \left[\mathbf{l}^{2}+x(1-x)\mathbf{q}^2\right]s_2\right],\label{eq : piiidefappen2}\end{aligned}\ ] ] with @xmath200^{i}},\end{aligned}\ ] ] where @xmath201 and @xmath202 .
summing over @xmath11 , it is easy to get @xmath203,\end{aligned}\ ] ] which then leads to @xmath204 - \frac{\beta^4}{16\pi^4y^2 } \frac{\partial\mathrm{im } \left[\psi(\frac{1}{2}+x+iy)\right]}{\partial y}.\end{aligned}\ ] ] substituting the expressions of @xmath205 into eqs .
( [ eq : pi00defappen2 ] ) and ( [ eq : piiidefappen2 ] ) , we have @xmath206 where @xmath207 and @xmath208 $ ] .
now we are interested in the limiting behavior of @xmath183 and @xmath184 at zero temperature . as @xmath156 , we know that @xmath209 = \arctan\left(\frac{t}{\gamma}\right).\ ] ] therefore , at zero temperature the polarization functions can be written as @xmath210 with @xmath211 and @xmath212 .
[ [ calculation - of - polarization - functions - in - the - clean - limit - gamma0 ] ] calculation of polarization functions in the clean limit @xmath71 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ for general @xmath213 , we have @xmath214s_2+(1 - 2x)q_{0}s^*\right\ } , \label{eq : pi00defappen3 } \\
\pi_{ii}(q_{0},\mathbf{q},t , m_0 ) & = & -\frac{8ne^{2}}{\beta } \int_{0}^{1}dx\int\frac{d^{2}\mathbf{l}}{(2\pi)^{2}}\left\{s_1- \left[\mathbf{l}^{2 } + 2x(1-x)\left(q_{0}^{2 } + \frac{\mathbf{q}^{2}}{2}\right)\right]s_2 + ( 1 - 2x)q_{0}s^*\right\}.\label{eq : piiidefappen3}\end{aligned}\ ] ] with @xmath215^{i } } , \\ s^ { * } & = & \left(\frac{\beta}{2\pi}\right)^{3 } \sum_{n=-\infty}^{\infty}\frac{n+\frac{1}{2}+x}{\left[\left(n + \frac{1}{2}+x\right)^{2 } + y^2\right]^{2}}.\end{aligned}\ ] ] where @xmath216 and @xmath217 .
carrying out the frequency summation yields @xmath218 + \frac{\beta^2}{4\pi^2y } \mathrm{im}\left[\psi\left(\frac{1}{2 } - x+iy\right)\right].\label{eq : s1zerogammafiniteq0}\end{aligned}\ ] ] it is then straightforward to obtain @xmath219 - \frac{\partial\mathrm{im}\left[\psi(\frac{1}{2}+x+iy)\right]}{\partial y } + \frac{1}{y } \mathrm{im}\left[\psi\left(\frac{1}{2 } - x+iy\right)\right]\right.\nonumber \\ & & \left.-\frac{\partial\mathrm{im}\left[\psi(\frac{1}{2 } - x+iy)\right]}{\partial y}\right\},\label{eq : s2zerogammafiniteq0 } \\
s^ { * } & = & -\frac{\beta}{4\pi}\frac{\partial s_{1}}{\partial x } = -\frac{\beta^3}{16\pi^3y}\left\{\frac{\partial \mathrm{im}\left[\psi(\frac{1}{2}+x+iy)\right]}{\partial x } + \frac{\partial\mathrm{im}\left[\psi(\frac{1}{2}-x+iy)\right]}{\partial x}\right\}.\label{eq : starzerogammafiniteq0}\end{aligned}\ ] ] substituting eqs .
( [ eq : s1zerogammafiniteq0 ] ) , ( [ eq : s2zerogammafiniteq0 ] ) , and ( [ eq : starzerogammafiniteq0 ] ) into eqs .
( [ eq : pi00defappen3 ] ) and ( [ eq : piiidefappen3 ] ) , the polarization functions can be written as @xmath220 where @xmath221 , @xmath222 , and @xmath223 . here , @xmath224 + \mathrm{im}\left[\psi(\frac{1}{2}-\frac{xq_{0}}{2\pi t } + i\frac{t}{2\pi t})\right]$ ] . since @xmath225
, we have @xmath226.\label{eq : f2appen}\end{aligned}\ ] ] substituting eq .
( [ eq : f2appen ] ) into eqs .
( [ eq : pi00defappen4 ] ) and ( [ eq : piiidefappen4 ] ) , the polarization functions become @xmath227\right ) \right.\nonumber \\ & & \left.-\frac{1}{k_{3}}\frac{\left[m_0 ^ 2 + x(1-x)q_{0}^{2}\right]\sinh(\beta k_{3 } ) } { \cosh^2(\frac{1}{2}\beta k_{3 } ) - \sin^2(\frac{1}{2}\beta xq_{0 } ) } - \frac{1}{2}\frac{(1 - 2x)q_{0}\sin(\beta xq_{0})}{\cosh^2(\frac{1}{2 } \beta k_{3 } ) - \sin^2(\frac{1}{2}\beta xq_{0})}\right\ } , \\ \pi_{ii}(q_{0},\mathbf{q},t ,
m_0 ) & = & \frac{ne^{2}}{2\pi } \int_{0}^{1}dx\left\{\frac{1}{k_{3}}\frac{x(1-x)\left(2q_{0}^{2 } + \mathbf{q}^{2}\right)\sinh(\beta k_{3 } ) } { \cosh^2(\frac{1}{2 } \beta k_{3 } ) - \sin^2(\frac{1}{2}\beta xq_{0 } ) } + \frac{(1 - 2x)q_{0}\sin(\beta xq_{0})}{\cosh^2(\frac{1}{2}\beta k_{3 } ) -\sin^2(\frac{1}{2}\beta xq_{0})}\right\},\end{aligned}\ ] ] with @xmath228 . at zero temperature ,
the polarization functions are simplified to @xmath229 , \\
\pi_{ii}(q_{0},\mathbf{q},m_{0 } ) & = & \frac{ne^{2}\left(2q_{0}^{2 } + \mathbf{q}^2\right)}{\pi } \left[\frac{m_0}{2q^{2 } } + \frac{q^2 - 4m_0 ^ 2}{4q^{3}}\arcsin\left(\frac{q}{\sqrt{q^2 + 4m_0 ^ 2}}\right)\right].\end{aligned}\ ] ] in the limit @xmath70 , the polarization functions are @xmath230 - \frac{m_0 ^ 2}{k_{1}}\tanh\left(\frac{k_{1}}{2t}\right)\right\ } , \nonumber \\
\pi_{ii}(\mathbf{q},t , m_0 ) & = & \frac{ne^{2}}{\pi } \int_{0}^{1}dx\frac{x(1-x)\mathbf{q}^{2}}{k_{1 } } \tanh\left(\frac{k_{1}}{2t}\right).\end{aligned}\ ] ] | extensive investigations show that qed@xmath0 exhibits dynamical fermion mass generation at zero temperature when the fermion flavor @xmath1 is sufficiently small .
however , it seems difficult to extend the theoretical analysis to finite temperature .
we study this problem by means of dyson - schwinger equation approach after considering the effect of finite temperature or disorder - induced fermion damping . under the widely used instantaneous approximation ,
the dynamical mass displays an infrared divergence in both cases .
we then adopt a new approximation that includes an energy - dependent gauge boson propagator and obtain results for dynamical fermion mass that do not contain infrared divergence .
the validity of the new approximation is examined by comparing to the well - established results obtained at zero temperature . |
most massive elliptical galaxies can be divided into two groups with different physical properties .
bright ellipticals are slow anisotropic rotators , have boxy distorted isophotes , are radio - loud , and are surrounded by gaseous x - ray halos ( bender et al . , 1989 ) .
faint elliptical galaxies are preferentially oblate isotropic rotators with disky isophotes . in contrast to boxy ellipticals they are radio - quiet and show no x - ray emission in excess to their discrete source contribution .
gravitational n - body simulations , starting with the work by toomre & toomre ( 1972,1977 ) using a restricted three - body approximation , and continued by others ( e.g. : barnes , 1988 ) using a hierarchical tree algorithm , have shown that mergers of equal mass disk - galaxies can produce slowly - rotating remnants with a de vaucouleurs - like surface - brightness profile and disky or boxy isophotes depending on the viewing angle ( heyl et al . , 1994 ) .
it is generally believed that merger remnants are supposed to have properties resembling observed boxy elliptical galaxies , such as slow figure rotation , kinematically distinct cores , so the question of how disky and fast rotating ellipticals have formed is still unanswered .
here we study a possible formation mechanism for disky elliptical galaxies in an encounter scenario .
an approximated massive galaxy perturbs an equal mass disk galaxy and flies away .
we analyze the triaxial shape , the rotation properties and the isophotal shape of the remnant .
the initial conditions for our simulations were derived following hernquist ( 1993 ) as described briefly below .
the n - body model consists of an exponential disk surrounded by a dark halo .
the thickness of the disk is determined by the velocity dispersion which is a function of radius .
the halo is initially spherical and has an isothermal density profile with a core and a cutoff - radius to reduce the computational costs .
velocities are initialized by taking moments of the collisionless boltzmann equation and approximating the distribution function in phase space by gaussians .
this produces stable models that are nearly in equilibrium .
the toomre q - parameter is normalized to the value of 1.5 at a radius similar to the solar neighborhood .
the mass and size of the disk is scaled to physical values of the milky way , i.e. scale length h=3.5 kpc and disk mass @xmath1 . for the perturbing galaxy
, we used an equal mass particle realization of a profile approximating a dark halo . at the beginning of the simulation
the perturber is on a prograde hyperbolic orbit with an impact paramter of 63 kpc , and a relative velocity of @xmath2 km / s .
the simulation was performed with a direct n - body code using the special hardware device grape ( sugimoto et al . ,
we followed the simulation for 6 gyrs .
the simulation presented here has 50,000 disk particles , 100,000 halo particles , and 20,000 particles representing the perturber .
figure 1 shows a sequence of snapshots for our simulation .
the encounter induces a strong bar in the disk , although the disk is stable against bar formation if simulated in isolation .
this bar becomes unstable to bending oscillations ( raha et al .
1991 , merritt & sellwood , 1994 ) due to an increase of the velocity dispersion in radial direction . as a result of this instability
the initially disklike system becomes spheroidal . to estimate the three - dimensional shape of the remnant we computed the axis ratios of the distribution of disk particles from the eigenvalues of the moment - of - inertia tensor .
for the 60% most tightly bound particles the triaxiality parameter @xmath3 is 0.88 . the effective hubble types for projections along the three principal axes are e5.7 , e7.3 and e4 respectively .
the surface density follows a de vaucouleurs - like profile over a large radial interval and is comparable to the remnants of merger simulations . after the strong rotating bar has formed , its dynamical friction with the live halo component leads to an effective transport of angular momentum to the halo .
the rotation velocity in the inner parts decreases rapidly .
figure 2 shows the rotation properties of the system after 6 gyrs .
we have plotted @xmath4 ( @xmath5 : maximum rotation velocity , @xmath6 : central velocity dispersion ) against the ellipticity .
the points show the values for the same remnant in different projections .
one can see that some projections follow the line for oblate rotators , but we also have anisotropic remnants and those with very high rotation velocities at low ellipticities
. this effect can be influenced by the determination of the ellipticity ; some observers see the same effect in their data ( see nieto et al . ,
1988 ) .
we also investigated the deviations of the isophotes from pure ellipses applying the method described by bender el al .
( 1988 ) after binning the particle distribution and convolving it with a gaussian with a fwhm comparable to the seeing conditions of the observations .
we find that the remnant has disky isophotes ( positive @xmath7 ) for almost all projections .
the value of @xmath8 is higher for remnants with higher ellipticity ( figure 2 ) .
with increasing radius the @xmath8 coefficient shows basically two global features : either a disky inner part changing to boxy in the outer parts , or a continuously rising positive @xmath8 .
those features could be explained as suggested before ( nieto et al . , 1991 ) by a faint disk surrounded by a spheroidal component or , in the case of rising @xmath8 , tidal extensions of a round inner part . in our simulation
the shape of the profile just depends on the projection angle .
barnes , j. , e. 1988 , , 331 , 699 bender , r. , dbereiner , s. & mllenhoff , c. 1988 , , 74 , 385 bender , r. , surma , p. , dbereiner , s. , mllenhoff , c. & madejsky , r. 1989 , , 217 , 35 davies , l. , d. , efstathiou , g. , fall , s. , m. , illingworth , g. & schechter , p. 1983
, , 266 , 41 hernquist , l. 1993 , , 86 , 389 heyl , j. , s. , hernquist , l. & spergel , d. n. 1994 , , 463 , 69 nieto , j .- l .
, capaccioli , m. , held , e. v. 1988 , , 195 , l1 nieto , j .- l . , poulain , p. , davoust , e. & rosenblatt , p. 1991
, , 88 , 559 merritt , d. , sellwod , j. 1994 , , 425 , 551 raha , n. , sellwood , j. , a. , james , r. , a. & kahn , f. d. 1991 , nature , 352 , 411 sugimoto , d. , chikada , y. , makino , j. , ito , t. , ebisuzaki , t. & umemura , m. 1990 , nature , 345 , 33 toomre , a & toomre , j. 1972 , , 178 , 623 toomre , a 1977 , in the evolution of galaxies and stellar populations , ed .
b. tinsley & r.larson ( new haven : yale univ . press),p . | we study a possible formation mechanism for elliptical / s0 galaxies using n - body simulations with grape .
a close galaxy encounter which does not lead to a merger can induce a strong bar in an initially axisymmetric disk .
this bar is unstable to bending oscillations and loses angular momentum to the dark halo .
after 6 gyrs the remnant is spheroidal with a de vaucouleurs - type surface density profile and has ellipticities and rotation properties comparable to observed elliptical / s0 galaxies : the system rotates fast @xmath0 and has disky isophotes . |
perhaps more than any other technological contribution in baseball , the deployment of the pitchf / x system has proven to be an invaluable resource to teams and to fans of the game in their statistical analyses of baseball , both in its original form and as augmented by estimates of pitch spin parameters ( known as `` psuedo - spin '' ) .
end users of the system , particularly brooksbaseball.net and mlb advanced media ( mlb - am ) , have classified these pitches into common pitch types , both by hand and using neural network classification methods .
we hypothesize that the current method for classifying pitches in the database can be refined due to evidence from graphical exploration and previous research [ 1 ] .
+ we improve the current pitch classification system by comparing the current neural network classification results to these from various statistical clustering methods , such as k - means , hierarchical clustering , and model - based clustering with a multivariate gaussian mixture model ( mbc ) .
we ultimately propose an alternative basis for pitch clustering and classification . in section 2
we describe the current methods used and the pitchf / x database . in section 3
, we introduce the model - based clustering method , which has several advantages for examining pitch types including pitches with high variance , pitch evolution across time , and pitches with similar characteristics .
we also address implementation of the mbc method as well as the selection of the clusters and the stability of this clustering method . in section 4 ,
we propose a novel algorithm for classifying pitches based on their characteristics , before concluding with future work in section 5 .
publicly available pitchf / x data is available from several sources . ]
our data subset consists of pitches thrown by roughly 900 pitchers in the 2010 and 2011 seasons ; we exclude data from before the 2010 season due to reported inconsistencies within the pitchf / x system .
one important point to make is that ground truth is not known , which is why we have measured stability of our method in section 3.2 , in terms of cluster memberships .
+ the raw database contains trajectory information on each pitch , including acceleration and velocity , though not the spin of each pitch , which is useful in pitch classification because it can help distinguish between different pitch types .
some spin variables can be estimated using physics and a number of simplifying assumptions [ 3 ] ; the name `` psuedo - spin '' is given to these quantities due to this .
+ there are several important variable definitions that we use throughout our paper and in our figures : * the pitch s * start speed * measured in miles per hour at the release point .
this is measured using radar guns and is commonly acknowledged as the speed of the pitch . * the * back spin * of the baseball measured in radians per second .
a positive number represents back spin .
most fastballs have back spin , while off - speed pitches have a tendency to have more `` top - spin '' , or negative back spin . *
the * side spin * of the baseball measured in radians per second .
a positive number represents left - to - right spin , or a left - handed pitcher s curveball or slider .
a negative number represents right - to - left spin , or the direction of a right - handed pitcher s curveball or slider .
figure 1a plots these three variables for pitches thrown by barry zito ; these pitches have been classified with the system developed via a neural network classification system for each pitcher developed by mlb advanced media .
specifics pertaining to the method , model , and training data used are not publicly available [ 1 ] .
+ we have reason to believe barry zito threw 5 types of pitches in the 2010 and 2011 seasons due to the obvious 5 different clusters observed graphically , and according to http://www.brooksbaseball.net[brooks baseball ] , whose pitchf / x classifications are considered by some the most accurate available to the public , those pitches are a four - seam fastball , sinker , slider , curveball , and changeup ; but they do not make their method or data available , only their results [ 3 ] . moreover , we speculate that the pitches in the two - seam fastball cluster should instead be classified as sinkers since brooks baseball classifies these pitches as such . the neural net classifies brooks baseball s sinker cluster as two - seam fastballs .
we investigate whether or not barry zito s two - seam fastball should be labeled as a sinker in section 4 by implementation of a new classification algorithm .
we ultimately agree with brooks baseball and label the cluster as a sinker .
overall , our model tends to split up four - seam fastballs and two - seam / sinker clusters differently than the neural networks method in a way that more closely resembles brooks baseball classification and empirically makes more sense .
we explore this further in section 4 when we discuss the cliff lee classification example . + in addition
, the neural net classification appears to have obvious misclassifications .
four - seam fastballs are classified as curveballs , and curveballs / changeups are classified as sliders , which are clearly labeled in the incorrect clusters in figure 1a .
mixture models for clustering rely on a straightforward generative premise : there is a series of simple probabilistic models for how an event can be generated , and a weight on which model will be used to generate the observation .
this description lines up directly with pitcher intent : while we as observers may not know what type of pitch is intended , the pitcher himself makes a choice of a specific pitch type ( fastball , slider , curveball , etc ) with a basic profile : a grip and arm motion that gives the ball a desired speed , spin and trajectory . + a multivariate gaussian model for any particular pitch profile makes intuitive sense .
each coordinate has a mean value for example , a typical four - seam fastball might have an initial velocity of 95 miles an hour , a back spin of 100 radians per second and a side spin of 10 radians per second .
the resulting pitch is then affected by many different sources of noise , both in the pitcher s delivery and in other external factors like the wind , and this noise can affect multiple pitch characteristics at once .
the resulting pattern in three dimensions is an ellipsoid ; in figures 1b , 2 , 3 , and 5b , we visualize the mbc results for a variety of pitchers .
one particular advantage to this approach is that we can detect when two clusters overlap , since geometry , and not just proximity , is an important factor .
+ given a set of gaussian clusters and weights , this routine determines the probability that each observed pitch belongs to a given cluster as the relative probability density for the pitch if it were a member of each cluster , factoring in each cluster s relative weight most pitchers , for example , throw more fastballs than other off - speed pitches , and this is taken into account directly . in our classifications
we declare a pitch to belong to the class with the highest probability under the model .
+ we use the ` mclust ` library in the ` r ` statistical programming software to perform our clustering operations , with some modifications that we describe to account for additional information . given a pre - selected number of clusters
, we use an expectation - maximization algorithm to calculate the maximum likelihood estimates ( mles ) for each cluster location , shape , and weight .
once we run this across a range of cluster counts , we use the bayesian information criterion ( bic ) model selection criteria to determine the optimal number of clusters in the model , which attempts to maximize the likelihood of the data while penalizing excessive numbers of parameters .
+ empirically , mbc creates clusters of pitches that are more tightly confined than current pitch classifications , suggesting there are very few curveballs , changeups , or sliders misclassified .
a major concern is the difference between the four - seam fastball and the sinker / two - seam fastball clusters , as seen in the figures section .
we expect two - seam fastballs / sinkers to have a slightly slower start speed than four - seam fastballs because it is known four - seam fastballs to be a pitchers fastest pitch , but this is not the case with the pitchf / x neural networks classification method . however , mbc weights the velocity as an important factor between the two clusters and splits them accordingly .
+ figure 1b shows the mbc as it performs for the pitches of barry zito ; the red cluster is the faster pitch and light blue cluster is the slightly slower pitch with vertical spin closer to zito s other off - speed pitches .
the mbc s potential four - seam fastball cluster ( red ) has the smallest number of pitches , but other sources suggest that zito threw his four - seam fastball most often , leading us to suspect that while the labels ( from section 4 ) may need adjusting , the proper clusters are still being detected .
this also may indicate there is not much difference between zito s four - seam and two - seam / sinker pitches from the batter s perspective . to further support subtle advantages of mbc over the neural networks classification , we evaluate the overall stability of our clustering method in section 3.2 .
+ in its original form , the model tends to choose more clusters than fewer for the pitchers we have tested . on inspecting the clusters produced , it is clear that the method is favoring relatively `` thin '' clusters , which have high internal correlations between variables , which is highly unrealistic for the physical examples we consider . limiting the model to a smaller number of pitches is not generally feasible , and while manual inspection is possible , it would be far more preferable to automate the method to remove this issue .
+ we develop our own criterion for choosing the number of clusters , called the adjusted bayesian information criterion , or bic@xmath0 . since we observe that most pitch clusters are close to spherical , and our prior knowledge suggests that flat ellipsoidal clusters are unlikely , we are motivated to constrain the creation clusters which have high intra - cluster correlations between the three variables of interest
. currently , each cluster @xmath1 has three parameter sets : @xmath2 , the cluster mean ; @xmath3 , the standard deviation of each dimension ; and @xmath4 , the intra - cluster correlation matrix .
it is the terms of @xmath5 that need to be kept small in absolute value ( 1 or -1 indicates perfect correlation , 0 indicates no correlation ) .
+ to account for this , we develop an additional penalty term for the current bic formula that adds a value proportional to each intra - cluster correlation term .
using bic@xmath0 , if the clusters the model finds with @xmath1=6 have high intra - cluster correlations , compared to @xmath1=5 , than the correlation penalty term will be large , and bic@xmath0 will be smaller for @xmath1=5 .
we choose our @xmath1 based off of the minimum bic or bic@xmath0 .
+ figure 1b displays the clustering for barry zito chosen with adjusted bic ( bic@xmath0 ) , which chooses five clusters , which agrees with the pitch number selection of both the neural network classification and brooks baseball . for this application
, bic@xmath0 is a substantial improvement over bic , and is the method we use going forward in choosing the number of pitch clusters . + in order to fully grasp all of the benefits the mbc method can offer , we investigate how it performs when clustering pitchers other than barry zito .
in fact , we have investigated mbc s performance on the entire 2010 and 2011 season , and simply show a few illustrations from this large analysis performed .
+ figure 2 shows how our method is ideal for detecting how a pitch can evolve over time .
the purple and black clusters represent jon lester s curveballs .
the difference between the two clusters is the speed and horizontal break of the pitch . in 2010 , jon lester s curveball averaged 78 mph while in 2011 it averaged 76 mph with slightly more horizontal break a subtle but important difference .
figure 3 visualizes how our method can also cluster pitches with high variances such as tim wakefield s knuckleball ( orange ) ; the pseudo - spins , corresponding to additional break , are considerably wider than for other pitches due to the unpredictable nature of stitch position .
+ we evaluate quality of the mbc method by assessing the _ stability _ of the method in terms of how sensitive our model is to smaller sample sizes .
we go about accomplishing this by taking the data for each pitcher and running the clustering on an 80% subset of the data , the remaining 20% subset , and then on the full data .
next , we calculate the number of pitches in both the 80% and 20% subsets that do not change clusters compared to the full dataset . in order to be confident in our results
, we repeat this process 20 times for each pitcher and find the mean and standard error .
we find that after 20 samples our standard error is sufficiently small and we are confident with our stability sample mean estimates .
figure 4 is the two distributions for the 80% and 20% subsets and the proportion of pitches that are in the same cluster as the full data for all pitchers .
overall , for both the 80% and 20% subsets the majority of pitchers have 80% or more of their pitches clustered in the same cluster .
we found that this stability holds on sample sizes as low as 100 pitches , or the average number of pitches a starting pitcher throws in one start .
it is important to note that we kept the number of clusters chosen ( k ) the same on both subsets as when we ran it on the full data . +
in order to have a fully automated pitch clustering and classification system , we propose a simple classification algorithm , which improves pitch clustering by assigning sensible pitch labels to the respective clusters from our heuristic .
although ground truth is unknown , we make comparisons with the current classification labels in the pitchf / x database , brooks baseball classifications , and graphical visualization to determine how well our method works .
we found that our classification algorithm has many strengths including consistent performance and labeling multiple clusters as the same pitch type when appropriate , such as the jon lester curveball evolution over time scenario referenced in section 3.2 and visualized in figure 2 .
our clustering algorithm appears to have similar classification results as brooks baseball ( which is the closest to the `` truth '' that is , a pitcher s actual choice of pitch to throw against which we can compare ) .
+ our algorithm classifies and names a key subset of pitches : four - seam fastballs , two - seam fastballs , sinkers , change - ups , cut - fastballs , sliders , curveballs , and knuckleballs . for each classification , the algorithm begins by assigning the cluster mean with the highest starting velocity as a four - seam fastball . for each additional cluster ,
the algorithm goes through a series of constraints that are derived from each cluster s mean and variance to determine the label for each cluster .
for example , if a cluster mean has the same side spin direction as the four - seam fastball then it checks if the speed differs from the four - seam fastball by more than 6 mph and if the side spin varies less than 60 rotations per second from the four - seam fastballs .
if it does , it assigns the cluster as a change - up .
if not , it determines if the side or back spin difference from the four - seam fastball is greater .
if the side spin difference is the greater of the two , it assigns the cluster as a two - seam fastball ; if the back spin difference is larger , it assigns the cluster as a sinker .
the rest of the classification algorithm follows similar decision processes .
+ we evaluate how well our classification algorithm performs by taking a sample of various starting and relief pitchers with differing pitch repertoires .
we found after analyzing and comparing the results to brooks baseball and the neural network classification system , in 23 of the 25 scenarios our classification algorithm empirically appear correct and has similar classification results as brooks baseball .
the two pitchers that the algorithm does not classify correctly are barry zito and derek lowe .
it interchanges their sinker and four - seam fastball clusters because in the data sample , the measured speed of the sinkers are bigger than those for the four - seam fastball , a characteristic that is not commonly true for other pitchers .
+ in figure 5 , we visualize cliff lee s mlb - am neural network classification compared to our method .
of particular note , our method separates and labels the two - seam and four - seam fastball by assigning the four - seam fastball to the fastest pitch with the least amount of back and top spin relative to lee s other off - speed pitches .
our method splits the two fastballs similar to brooks baseball , but brooks baseball instead labels the two - seam fastball as a sinker .
we also are able to observe that our method s slider ( brown ) cluster is clearly defined and classified with no other obvious misclassified pitches .
the neural network classification has obvious misclassification in the slider cluster with pitches labeled as changups , curveballs , and cut - fastballs .
our method clearly improves both the clustering and classification of cliff lee compared to the neural network classification .
we have proposed a new clustering method and classification algorithm and tested both approaches using the pitchf / x database for 2010 - 2011 .
our analysis illustrates better clustering than the current neural network method based upon our plots and measuring the stability of our method .
furthermore , based upon our model based clustering method , we recommend a simple algorithm to classify each pitch based on the individual pitcher , which performs extremely well in most situations .
its strongest features are correcting any obvious misclassification in the neural network model , accounting for pitch evolution over time , and performing well for pitches with comparatively high internal variance .
our method also performs well in a highly debated topic in mlb , namely distinguishing two - seam and four - seam fastballs .
+ these improvements suggest others can be made in assigning pitchers themselves to clusters , a problem highly motivated by the need to assess a pitcher s likely performance against unfamiliar players or teams . in these situations ,
most hitters likely have never faced the current pitcher , and thus , there is no or very little data to leading to inference about the hitter s history against the current pitcher .
placing pitchers into similar groups and looking for relationships with groups of batters would be invaluable to mlb .
this information can lead to advanced and more detailed scouting reports of what type of pitches are a hitter s strength or weakness .
moreover , since we have prior information about many pitchers , we can utilize advanced bayesian clustering methods and optimize the choice of @xmath1 for better and more stable results .
* references * + 1 . foster , adam .
_ scouting with pitchf / x _ , june 12 , 2012 .
http://www.baseballprospectus.com/ + article.php?articleid=17327 .
2 . fast , mike .
_ how to build a pitch database _ , august 23 , 2012 .
http://fastballs.wordpress.com/2007/08/23/ + how - to - build - a - pitch - database/ 3 .
nathan , alan m. _ the physics of baseball _ , accessed september 20 , 2012 .
http://webusers.npl.illinois.edu/@xmath6a - nathan / pob/.
.named pitch types with corresponding colors . [ cols="^,^,^,^,^,^",options="header " , ] the pitches thrown by barry zito .
figure 1a ( left ) displays the mlb advanced media classification developed via a neural network system .
there are obvious misclassifications : a small number of four - seam fastballs ( which should be red ) are classified as sliders ( brown ) , as are some curveballs ( black ) and changeups ( green ) .
figure 1b ( right ) displays the mbc model and tightly clusters the pitches , as well as splits up the four - seam and sinker ( light blue ) clusters in an empirically sensible way.,title="fig : " ] the pitches thrown by barry zito .
figure 1a ( left ) displays the mlb advanced media classification developed via a neural network system .
there are obvious misclassifications : a small number of four - seam fastballs ( which should be red ) are classified as sliders ( brown ) , as are some curveballs ( black ) and changeups ( green ) .
figure 1b ( right ) displays the mbc model and tightly clusters the pitches , as well as splits up the four - seam and sinker ( light blue ) clusters in an empirically sensible way.,title="fig : " ] the pitches thrown by jon lester classified using model - based clusters , which shows two distinct clusters for curveballs ( purple and black ) corresponding to a change over time .
the difference between the two clusters is the speed and horizontal break of the pitch .
this is one example of how the mbc model can detect subtle but important pitch evolution differences . ]
the pitches thrown by tim wakefield classified using model - based clusters .
the mbc model can also detect and classify pitches well with relatively higher variances like the pseudo - spins for tim wakefield s ( spinless ) knuckleball ( orange ) , compared to the fastball ( red ) and curveball ( black ) . ]
80% and 20% stability : percent of pitches in the same cluster in subset and full dataset , across 20 replications of the procedure .
for both the 80% and 20% subsets the majority of pitchers have 80% or more of their pitches clustered in the same cluster .
we also found that this stability holds on sample sizes as low as 100 pitches , or the average number of pitches a starting pitchers throws in one start . ]
cliff lee : figure 5a ( left ) displays mlb s classification system developed via a neural network classification system .
as before , there are obvious misclassifications , but our methods clusters are clearly defined and classified with no other obviously misclassified pitches .
the mbc model in figure 5b ( right ) separates and labels the two - seam ( grey ) and four - seam ( red ) fastballs by assigning the four - seam fastball to the fastest pitch with the least amount of back and top spin relative to lees other off - speed pitches .
this method agrees with the manually corrected data from brooks baseball.,title="fig : " ] cliff lee : figure 5a ( left ) displays mlb s classification system developed via a neural network classification system .
as before , there are obvious misclassifications , but our methods clusters are clearly defined and classified with no other obviously misclassified pitches .
the mbc model in figure 5b ( right ) separates and labels the two - seam ( grey ) and four - seam ( red ) fastballs by assigning the four - seam fastball to the fastest pitch with the least amount of back and top spin relative to lees other off - speed pitches .
this method agrees with the manually corrected data from brooks baseball.,title="fig : " ] | the pitchf / x database has allowed the statistical analysis of of major league baseball ( mlb ) to flourish since its introduction in late 2006 .
using pitchf / x , pitches have been classified by hand , requiring considerable effort , or using neural network clustering and classification , which is often difficult to interpret . to address these issues
, we use model - based clustering with a multivariate gaussian mixture model and an appropriate adjustment factor as an alternative to current methods .
furthermore , we describe a new pitch classification algorithm based on our clustering approach to address the problems of pitch misclassification .
we illustrate our methods for various pitchers from the pitchf / x database that covers a wide variety of pitch types . |
the silicon nanorod arrays were fabricated from a single crystalline si wafer by adapting a previously described dry etching method @xcite .
the 380@xmath17 m thick , @xmath25100@xmath26 cut and p - doped wafers exhibit a resistivity of 1 - 10@xmath27 cm .
they were cleaned by sonication , first in acetone then in isopropyl alcohol ( ipa ) , each for 5 min .
they were then thoroughly rinsed with deionized water and dried in a stream of n@xmath28 .
the clean si wafers were spin coated with mma resist ( copolymer resist el9 , microchem ) at 4000rpm for 60s , followed by baking at 180@xmath29c on a hotplate for 2min .
pmma resist ( polymer resist a2 , microchem ) was deposited on the mma layer by spinning at 2000rpm for 60s , followed by baking at 180@xmath29c on a hotplate for 2min .
an array of 250 nm diameter circular dots with 1@xmath17 m spacing was written using a raith 150 ultrahigh - resolution e - beam lithography system ( raith gmbh , dortmund , germany ) .
the patterned wafer was developed in mibk / ipa 1:3 for 1min , followed by rinsing with ipa for 20s and drying in an n@xmath28 stream .
a nickel dot array was prepared by e - beam evaporating a 100 nm thick nickel layer at a base pressure of @xmath30torr with a rate of 1 / s .
finally , the remaining resist was lifted - off in an acetone / ipa 1:1 solution , washed with ipa and dried .
the nickel nanodot arrays served as masks in the following dry etching .
vertical silicon nanowire arrays were fabricated by applying time - multiplexed reactive ion etching in an inductively coupled plasma deep reactive ion etching machine ( icp - drie , plasmatherm slr770 ) . in order to form nanopillars with a well defined breaking point ,
we have implemented a bosch process , i.e. passivation followed by anisotropic etching .
a variation of the ratio between the etching and the passivation times varies the scalloping and leads to different rod diameters .
based on that we set up a three - stage etching process : six passivation - etching sequences with a time ratio of 1:1 allowed us to form 700 nm long nanorod segments . in order to create
the breaking points , the ratio of the time windows was logarithmically changed to 0.7:1 during four further steps .
finally , a wider base was formed by changing the interval ratio linearly to 2:1 in six further steps .
the silicon rods were cleaned by removing the nickel caps chemically . in order to compute the normalized scattered light intensity @xmath31 , the rods are modelled as thin , homogeneous , dielectric needles of length @xmath32 and diameter @xmath33 . adopting the scattering theory for dielectric needles @xcite to a standing - wave situation
, we find that the light intensity in the direction perpendicular to both the cavity axis @xmath34 and the field polarisation axis @xmath35 is proportional to @xmath36 e^{-2(x^2+y^2)/w_0 ^ 2},\end{aligned}\ ] ] with @xmath37 the wave number , @xmath38 the cavity waist , and @xmath39 $ ] . here
, we denote the center - of - mass position of the rod by @xmath40 , and the orientation of the rod is determined by the radial unit vector @xmath41 .
the internal field points in the direction @xmath42 .
figures s[fig : s1 ] , s[fig : s2 ] and s[fig : s3 ] display the simulated dynamics of a @xmath43 nm long and @xmath44 nm thick silicon rod with mass @xmath45 and moment of inertia @xmath46 , which rotates in the plane perpendicular to the cavity @xmath47-axis . in this case
the rod can be treated as a sub - wavelength particle at position @xmath48 , and its orientation with respect to the field polarization @xmath49-axis is described by @xmath50 .
given a constant intra - cavity field amplitude @xmath51 , the rod s motion is governed by the following classical equations of motion : @xmath52 \sin\left(2 k z\right ) e^{-2 \frac{\left(v_x t\right)^2}{\omega^{2}_0 } } \nonumber \\
\ddot{\phi}(t ) & = & - \frac{e^2_0}{4 \theta } \left(\alpha_{\| } - \alpha_{\perp } \right ) \sin\left(2 \phi\right ) \cos^2 \left(k z\right ) e^{-2 \frac{\left(v_x t\right)^2}{\omega^{2}_0 } } \nonumber\end{aligned}\ ] ] with @xmath53 the cavity waist , @xmath54 the vertical velocity of the rod , and @xmath55 the polarizability components , as given in the main text .
particle trajectories that do not pass through the cavity center , but slightly off - axis , can be accounted for by decreasing the field amplitude @xmath51 below the cavity value @xmath56 .
the corresponding scattering intensity is obtained by evaluating the expression given in appendix [ sec : appb ] along the simulated trajectory .
since the scattering signal is here proportional to @xmath57 , the center - of - mass trajectory of a freely rotating rod can be reconstructed from the measured scattering signal by averaging over the fast rotation period .
we ensure that we capture the full transit through the cavity mode by carrying out each simulation over the time interval between @xmath58 . in order to compare this to the measured data ,
the scattering signal must be normalized , and the time offset of the simulation must be adjusted , accordingly .
our work has been supported by the european commission ( 304886 ) as well as by the austrian science fund ( fwf ) : w1210 - 3 and p27297 .
we acknowledge support by s. puchegger and the faculty center for nanostructure research at the university of vienna in imaging the nanorods .
f.p . acknowledges the legacy program ( israel science foundation ) for its support .
krahne , r. ; manna , l. ; morello , g. ; figuerola , a. ; george , c. ; deka , s. _ physical properties of nanorods _ ; springer , 2013 hornberger , k. ; gerlich , s. ; haslinger , p. ; nimmrichter , s. ; arndt , m. colloquium : quantum interference of clusters and molecules .
phys . _ * 2012 * , _ 84 _ , 157173 arndt , m. ; hornberger , k. insight review : testing the limits of quantum mechanical superpositions . _ nature phys . _ * 2014 * , _ 10 _ , 271277 bateman , j. ; nimmrichter , s. ; hornberger , k. ; ulbricht , h. near - field interferometry of a free - falling nanoparticle from a point - like source .
commun . _ * 2014 * , _ 5 _ , 4788 arndt , m. ; nairz , o. ; vos - andreae , j. ; keller , c. ; van der zouw , g. ; zeilinger , a. wave - particle duality of c60 molecules .
_ nature _ * 1999 * , _ 401 _ , 680682 eibenberger , s. ; gerlich , s. ; arndt , m. ; mayor , m. ; txen , j. matter - wave interference with particles selected from a molecular library with masses exceeding 10 000 amu .
phys . _ * 2013 * , haslinger , p. ; drre , n. ; geyer , p. ; rodewald , j. ; nimmrichter , s. ; arndt , m. a universal matter - wave interferometer with optical ionization gratings in the time domain .
_ nature phys . _ * 2013 * , _ 9 _ , 144 148 bassi , a. ; lochan , k. ; satin , s. ; singh , t. p. ; ulbricht , h. models of wave - function collapse , underlying theories , and experimental tests . _ rev .
phys . _ * 2013 * , _ 85 _ , 471527 diosi , l. a universal master equation for the gravitational violation of quantum mechanics . _
physics letters a _ * 1987 * , _ 120 _ , 377381 penrose , r. on gravity s role in quantum state reduction . _ gen .
grav . _ * 1996 * , _ 28 _ , 581600 li , t. ; kheifets , s. ; raizen , m. g. millikelvin cooling of an optically trapped microsphere in vacuum .
_ nature phys . _ * 2011 * , _ 7 _ , 527 530 gieseler , j. ; deutsch , b. ; quidant , r. ; novotny , l. subkelvin parametric feedback cooling of a laser - trapped nanoparticle .
lett . _ * 2012 * , _ 109 _ , 103603 kiesel , n. ; blaser , f. ; delic , u. ; grass , d. ; kaltenbaek , r. ; aspelmeyer , m. cavity cooling of an optically levitated nanoparticle .
usa _ * 2013 * , _ 110 _ , 14180 14185 millen , j. ; fonseca , p. z. g. ; mavrogordatos , t. ; monteiro , t. s. ; barker , p. f. cavity cooling a single charged levitated nanosphere .
lett . _ * 2015 * , _ 123602 _ , 15 asenbaum , p. ; kuhn , s. ; nimmrichter , s. ; sezer , u. ; arndt , m. cavity cooling of free silicon nanoparticles in high - vacuum .
commun . _ * 2013 * , _ 4 _ , 2743 shi , h. ; bhattacharya , m. coupling a small torsional oscillator to large optical angular momentum .
opt . _ * 2013 * , _ 60 _ , 382386 yin , z. ; geraci , a. ; li , t. optomechanics of levitated dielectric particles .
b _ * 2013 * , _ 27 _ , 1330018 mller , t. ; reinhardt , c. ; sankey , j. c. enhanced optomechanical levitation of minimally supported dielectrics . _ physical review a _ * 2015 * , _ 91 _ , 110 bonin , k. d. ; kourmanov , b. light torque nanocontrol , nanomotors and nanorockers . _ opt . expr . _ * 2002 * , _ 10 _ paterson , l. ; macdonald , m. p. ; arlt , j. ; sibbett , w. ; bryant , p. e. ; dholakia , k. controlled rotation of optically trapped microscopic particles
. _ science _ * 2001 * , _ 292 _ , 9124 jones , p. h. ; palmisano , f. ; bonaccorso , f. ; gucciardi , p. g. ; calogero , g. ; ferrari , a. c. ; marago , o. m. rotation detection in light - driven nanorotors .
_ acs nano _ * 2009 * , _ 3 _ , 30773084 tong , l. ; miljkovic , v. d. ; kall , m. alignment , rotation , and spinning of single plasmonic nanoparticles and nanowires using polarization dependent optical forces . _ nano lett . _ * 2010 * , _ 10 _ , 26873 marago , o. m. ; jones , p. h. ; gucciardi , p. g. ; volpe , g. ; ferrari , a. c. optical trapping and manipulation of nanostructures .
_ nature nanotech . _ * 2013 * , _ 8 _ , 80719 arita , y. ; mazilu , m. ; dholakia , k. laser - induced rotation and cooling of a trapped microgyroscope in vacuum .
commun . _ * 2013 * , _ 4 _ hulst , h. ; van de hulst , h. _ light scattering by small particles _ ; dover , 1957 bishop , a. i. ; nieminen , t. a. ; heckenberg , n. r. ; rubinsztein - dunlop , h. optical application and measurement of torque on microparticles of isotropic nonabsorbing material . _
* 2003 * , _ 68 _ , 033802 gieseler , j. ; novotny , l. ; quidant , r. thermal nonlinearities in a nanomechanical oscillator .
_ nature phys . _ * 2013 * , _ 9 _ , 806810 gieseler , j. ; quidant , r. ; dellago , c. ; novotny , l. dynamic relaxation of a levitated nanoparticle from a non - equilibrium steady state .
_ nature nanotech . _ * 2014 * , _ 9 _ , 35864 millen , j. ; deesuwan , t. ; barker , p. ; anders , j. nanoscale temperature measurements using non - equilibrium brownian dynamics of a levitated nanosphere .
_ nature nanotech . _
* 2014 * , _ 9 _ , 425 429 ashkin , a. ; dziedzic , j. m. optical trapping and manipulation of viruses and bacteria .
_ science _ * 1987 * , _ 235 _ , 15171520 romero - isart , o. ; juan , m. ; quidant , r. ; cirac , j. toward quantum superposition of living organisms .
_ new j. phys . _ * 2010 * , _ 12 _ , 033015 pevzner , a. ; engel , y. ; elnathan , r. ; ducobni , t. ; ben - ishai , m. ; reddy , k. ; shpaisman , n. ; tsukernik , a. ; oksman , m. ; patolsky , f. knocking down highly - ordered large - scale nanowire arrays .
_ nano letters _ * 2010 * , _ 10 _ , 12021208 schiffer , r. ; thielheim , k. light scattering by dielectric needles and disks .
_ j. appl .
phys . _ * 1979 * , _ 50 _ , 24762483 | optical control of nanoscale objects has recently developed into a thriving field of research with far - reaching promises for precision measurements , fundamental quantum physics and studies on single - particle thermodynamics . here
, we demonstrate the optical manipulation of silicon nanorods in high vacuum .
initially , we sculpture these particles into a silicon substrate with a tailored geometry to facilitate their launch into high vacuum by laser - induced mechanical cleavage .
we manipulate and trace their center - of - mass and rotational motion through the interaction with an intense intra - cavity field .
our experiments show optical forces on nanorotors three times stronger than on silicon nanospheres of the same mass .
the optical torque experienced by the spinning rods will enable cooling of the rotational motion and torsional opto - mechanics in a dissipation - free environment
. nanoparticles often exhibit unique optical , mechanical , or electro - magnetic properties because of quantum effects in confined geometries and low dimensions @xcite .
complementary to that , our present study is part of a long - term effort to control the quantum properties of the objects motion @xcite .
first experiments demonstrating de broglie wave optics with macromolecules @xcite were triggered by the question whether the superposition principle of quantum mechanics holds on all scales .
they have led to the observation of quantum interference with masses beyond 10,000amu @xcite .
an even higher mass regime , which might give insight to the quantum - classical transition , can be reached with novel coherent manipulation schemes @xcite .
models of a spontaneous localization of the wave function @xcite , and non - standard effects of gravity @xcite , will become relevant for delocalized particles in the mass range of @xmath0amu , and above @xmath1amu , respectively .
tests of such models will require neutral , size- and shape - selected , cold , and slow nanoparticles that are mechanically isolated from their environment .
this has motivated new experiments to launch and cool dielectric nanospheres in optical tweezers @xcite , ion traps @xcite , and in free - flight @xcite . here
we extend this research to rod - shaped dielectrics of tailored geometry and anisotropic polarizability .
once their rotational motion can be controlled sufficiently well , these nanorods may be suitable for realizing torsional optomechanics @xcite .
earlier experiments have shown that optomechanical torques can be exerted on nanorods in solution using the polarization or orbital angular momentum of a light field @xcite .
the manipulation of nanoparticles in a dissipation - free environment , however , has remained challenging @xcite .
the coupling between rotational and motional degrees of freedom was recently demonstrated with optically trapped birefringent microspheres in a low pressure environment @xcite .
mj pulse energy , 6ns duration and @xmath2 nm wavelength ) hitting the rear side of the substrate cause stress - induced breaking of the nanorods at predefined trenches on the front side and launch them into the high vacuum .
individual , freely moving , and rotating rods need to pass a 500@xmath3 m diameter and 4 mm long aperture before entering a strongly pumped high - finesse infrared cavity ( @xmath4 nm , 400w intra - cavity power , 330,000 finesse , @xmath5 m waist ) .
their velocities and rotation rates are deduced from the scattered light , which is collected by a multi - mode fiber placed perpendicular to the cavity axis and the direction of field polarisation.,scaledwidth=45.0% ] we have prepared periodic arrays consisting of more than a million silicon nanorods per mm@xmath6 by dry - etching crystalline silicon wafers ( see appendix[sec : appa ] ) .
a scanning electron microscopy image of such an array is displayed in figure [ fig : sem]a .
the individual nanorods exhibit a length of @xmath7 nm and a diameter of @xmath8 nm corresponding to a mass of @xmath9amu .
the sample was positioned in a chamber evacuated to @xmath10pa underneath an optical cavity ( see figure [ fig : setup ] ) .
the backside of the sample was locally heated by a pulsed laser beam focused to @xmath11 m , which desorbs the rods by laser - induced thermomechanical stress ( lithmos)@xcite .
small kerfs etched into the nanorod base define the desired break - off conditions ( see figure [ fig : sem]c & d ; details in appendix[sec : appa ] ) . in figure
[ fig : sem]a and b we show an electron micrograph of a sample spot before and after the lithmos pulses .
it demonstrates that the rods can be reproducibly broken off at the tailored constrictions .
figure [ fig : sem]e depicts a close - up of the etched conical tip of a rod after launch and recapture .
the standing light wave field of a high - finesse cavity allows us to track the translational and rotational motion of the particles .
it is optically pumped by a linearly polarized , distributed - feedback laser locked close to the cavity resonance . at the laser wavelength of 1560 nm silicon exhibits a high relative permittivity and minimal absorption . in a homogeneous field , the polarisability assumes a maximum value of @xmath12@xmath13 and a minimum value of @xmath14@xmath13 when the rods are oriented parallel and perpendicular to the field , respectively @xcite . even for rotating rods ,
the polarizability averaged over all possible rotation axes , @xmath15@xmath13 , is larger than for a silicon nanosphere of the same mass , @xmath16@xmath13 .
[ ] we can trace each nanorotor using the light it scatters into the direction perpendicular to both the cavity axis and the field polarisation .
we collect this light in a 1 mm multimode fiber placed at a distance of 200@xmath17 m from the cavity center .
the detected intensity depends on the rod s position in the standing wave and also on its orientation ( see appendix[sec : appb ] ) .
when a symmetric rotor enters the cavity and moves freely along the cavity axis , we expect a modulation of the scattering signal at two distinct frequencies : one is the translational frequency @xmath18 of the particle passing the standing - wave nodes with velocity @xmath19 , the second one is twice the rotation frequency , @xmath20
. figure [ fig : rotation]a displays the normalized scattering intensity of a freely rotating nanorod , @xmath21 , where @xmath22 is the measured scattering signal and @xmath23 the simultaneously recorded intra - cavity intensity .
panel c shows that it agrees well with the theoretical expectations for light scattering at dielectric needles , see appendix[sec : appb ] .
we provide a full comparison of the measured signal in panel a and theory in appendix[sec : appc ] figure s[fig : s1 ] .
the corresponding fourier spectrum , depicted in panel b , exhibits the distinct frequency contributions of translation and rotation .
[ ] for slow rods we could observe cavity - induced translational channelling , i.e. one - dimensional trapping along an anti - node of the standing light wave .
one such case is displayed in figure [ fig : channelling ] , where the scattering signal ( panel a ) differs significantly from figure [ fig : rotation ] . when averaged over the rotational period of the rod ,
the scattering signal ( red dotted curve ) does not drop to zero while the particle is close to the center of the gaussian beam . during this time
the frequency related to the transverse motion of the rod is influenced significantly . in panel b
we reconstruct the particle trajectory through the cavity mode from the time evolution of the light scattering curve @xcite .
this is reproduced in a simulation of the rod s motion under the influence of the cavity field ( see appendix[sec : appc ] figure s[fig : s2 ] ) .
the optical channelling effect benefits from the geometrically enhanced induced dipole moment , due to the strong anisotropy of the rods @xcite . for silicon nanorods rotating in the plane
perpendicular to the cavity axis the orientation - averaged polarizability in the light field , @xmath24 , is enhanced by a factor of 2.7 in comparison to silicon spheres of the same mass . in figure
[ fig : channelling ] we observe an enhancement of the trapping potential by a factor of 2.6 . [ ] in addition to the optical force on the particle s center of mass , the cavity field exerts a torque on the rods , which tends to align them along the field polarisation axis @xcite .
figure [ fig : rotforce ] displays the scattering curve of a slowly rotating nanorod manipulated in both its center - of - mass motion and its rotation . in analogy to figure [ fig : channelling ] , we observe transverse channelling , but the rotation rate ( black circles ) is now influenced significantly by the optical torque .
this indicates that the motional and rotational degrees of freedom exchange energy via the optical potential .
the influence of the optical potential on the rotation rate is most pronounced at around 12@xmath17s . in conclusion , we have presented a method to tailor , launch , track , and manipulate high - mass silicon nanorods with well - defined geometry and high aspect ratio .
light scattering inside a high - finesse infrared cavity allows us to follow the translational and rotational motion of the nanorotors in real time . for some of the rods , cavity assisted 1d - trapping and
even rotational forces could be demonstrated .
our results are in good agreement with theoretical expectations and show that the rod - like shape enhances the interaction between the particles and the cavity field significantly , compared to silicon nanospheres of the same mass .
optomechanical trapping and cooling of the center - of - mass motion @xcite will benefit from aligning the rods along the axis of polarisation .
recent studies on single particle thermodynamics @xcite may be extended to rotating systems .
our results represent a first step towards realizing torsional nano - optomechanics @xcite and rotational cooling , which may become applicable to delicate biological nanomaterials such as the similarly shaped tobacco mosaic virus @xcite . |
the standard model suffers several fine tuning problems .
they are all related to the symmetry of the nature and its breaking . the gauge hierarchy problem which is one of the main motivation of the upcoming experiment , lhc has its origin in the gauge symmetry breaking sector .
another problem , strong cp problem is deeply related to a discrete symmetry , cp . the fact that the almost of the suggestions to solve them are more or less related to the invisible missing matter in the universe is interesting .
the strong cp problem can be stated as that a cp - odd operator @xmath4 which can not be forbidden is unmeasurably small @xmath5 @xcite .
the introduction of the axion can solve this problem beautifully @xcite .
the axionic solution introduces an irrelevant operator @xmath6 .
it can be thought of as making @xmath7 a dynamical field .
if then , the vafa - witten theorem tells that cp conserving point @xmath8 becomes a minimum of the free energy by the strong interaction @xcite . however , the introduction of the axion coupling solely can not solve the strong cp problem .
one needs one more ingredient , the relaxation mechanism for axion .
the standard big bang cosmology provides it naturally , so the axionic solution to the strong cp problem is intrinsically cosmological one .
this point was recognized soon after the suggestion of the axion @xcite .
the axion is one of the possible candidate of the cold dark matter .
after the early studies of cosmological implication of the axion , turner @xcite made a quantitative , numerical estimate of the axion energy density with a anharmonic correction and more precise axion mass formula .
all of their work depends on the assumption of the adiabatic process .
however , the assumption can be broken in the two phases of the evolution , at the time of starting to roll down and at the time of qcd phase transition @xcite .
we considered the effects of these two phases on the axion energy density and reanalyzed anharmonic effect .
in addition , many parameters of qcd physics and the cosmology are changed @xcite .
we finally represent the explicit formula in terms of these parameters .
just after the inflation , a very light scalar field like axion tends to be frozen at a certain value @xcite , @xmath9 , which is determined stochastically .
the dynamics of a scalar field in the expanding universe with the hubble parameter @xmath10 is governed by the equation of motion @xmath11 .
the second term in the left hand side of it prevents roll down of the scalar field , while second term forces it to move to the minimum .
therefore , after the inflation , the scalar fields stays at their initial value @xmath12 until the forces @xmath13 to roll down overcome the hubble friction @xmath14 .
if the scalar potential is harmonic one , it starts rolling down and oscillates when @xmath15 becomes small as the mass @xmath16 . in the hydrodynamic description ,
such a coherently oscillating scalar fields behaves as like the pressureless cold(@xmath17 ) matter .
if the interaction with the known particles is sufficiently weak and its mass is light enough , it can be a cold dark matter(cdm ) candidate .
since the invisible axion has the interactions and the mass originating from quantum effect , they can be small enough .
so it s a natural candidate of this kind cdm .
the potential of the axion field can be parameterized by @xmath18 , where @xmath19 and @xmath20 .
if the initial miss alignment angle @xmath21 is small enough , it becomes the harmonic potential with the mass @xmath22 .
at the temperature @xmath23 satisfying @xmath24 , the axion field starts rolling down . in our parametrization , @xmath25 . under the adiabatic assumption , @xmath26 ,
the number of the axion is conserved .
so the axion energy density at present can be easily calculated as @xmath27 .
to give a prediction , we should know @xmath28 and @xmath29 .
although , we can not solve it fully , all we have to know is the mass in the two different regime in which we can estimate it numerically or phenomenologically .
two regimes are zero temperature hadronic phase and the temperature enough higher than @xmath30 @xcite . at the zero temperature axion mass is given by @xmath31 , where @xmath32 .
at the high enough temperature , axion mass can be obtained by the perturbative instanton calculus under the dilute gas approximation @xcite . the instanton density with the size
@xmath33 is @xmath34\right\}},\nonumber\end{aligned}\ ] ] where @xmath35 and @xmath36 for @xmath37 . with the 3-loop running coupling constant ,
the numerical results are well fitted at the @xmath38 by the form @xcite
our rough estimation relies on two assumptions , the harmonic limit(@xmath40 ) and the adiabatic approximation(@xmath41 ) . we considered possible effects of them on the axion energy density .
there is also a possible correction from the qcd phase transition .
however , it turns out to be negligible , if one does not include fine tuning @xcite . in the previous section ,
we use the fact that the axion number is conserved
. however , the axion field is in the coherent state which is not the number eigenstate .
therefore the concept of the number conservation is not obvious except for the case of the non - interacting harmonic limit .
the conservation quantity , total number of axion is actually an adiabatic invariant of the system under the slowly varying potential . in the anharmonic limit , @xmath42 , there is no interpretation like the number conservation .
the adiabatic invariant is given by the area in the phase space swept by one period , @xmath43 @xcite . in our case
, its effect is appeared as a correction factor @xcite @xmath44 . now
the axion energy density with the anharmonic correction is @xmath45 .
there are two adiabatic conditions we assumed .
one is for the expansion rate to be much smaller than the axion oscillation , @xmath46 .
another is for the change rate of the mass to be smaller than the oscillation .
however , at the time of starting to roll down , the previous one can not be satisfied because the expansion rate is comparable with the mass , @xmath47 at that time .
therefore the adiabatic approximation does not work in the first few oscillations , and the numerical analysis must be done in this period . to include its effect in our calculations
, we defined the another temperature @xmath48 and the new misalignment angle @xmath49 at the temperature @xmath48 , such that the adiabatic condition is well satisfied after @xmath48 .
we explicitly checked that the values after the first half oscillation are good enough .
we obtained @xmath50 and @xmath51 numerically , and they are presented in the fig .
as one can sees , the correction factors , @xmath52 and @xmath53 , becomes larger at the large @xmath21 .
it is because of that the anharmonic terms make the field stays longer time at the around of the top of the potential .
so this initial overshoot correction includes other anharmonic correction . to split them , it is useful to factor out overall constant as like @xmath54 and @xmath55 , where the function with a tilde goes to @xmath56 as @xmath21 is going to zero . with the corrections @xmath57 , @xmath52 and @xmath53 ,
the relic density becomes @xmath58 where the function @xmath59 which is shown in fig .
[ fig : combincorr ] .
as one can see , there is a @xmath60 enhancement in the axion relic density .
with all of corrections we considered , the relic density of the axion is estimated as @xmath61 where we used @xmath62 , @xmath63135.5 mev , @xmath6493 mev and @xmath65(present)@xmath663.91 , @xmath67 is the anharmonic correction given in fig .
[ fig : combincorr ] and @xmath68 is the entropy increase ratio .
# 1#2#3phys . rep .
* # 1 * , # 2 ( # 3 ) # 1#2#3rev .
# 1 * , # 2 ( # 3 ) # 1#2#3annu .
nucl . part .
sci . * # 1 * , # 2 ( # 3 ) # 1#2#3nucl .
* b#1 * , # 2 ( # 3 ) # 1#2#3phys .
* b#1 * , # 2 ( # 3 ) # 1#2#3phys .
* d#1 * , # 2 ( # 3 ) # 1#2#3phys .
lett . * # 1 * , # 2 ( # 3 ) # 1#2#3jhep * # 1 * , # 2 ( # 3 ) # 1#2#3jcap * # 1 * , # 2 ( # 3 ) # 1#2#3z .
# 1 * ( # 3 ) # 2 # 1#2#3euro .
j. * c#1 * , # 2 ( # 3 ) # 1#2#3int .
j. mod .
phys . * # 1 * , # 2 ( # 3 ) # 1#2#3mod .
* a#1 * , # 2 ( # 3 ) # 1#2#3astrophys .
j. * # 1 * , # 2 ( # 3 ) # 1#2#3nature * # 1 * , # 2 ( # 3 ) # 1#2#3sov .
j. nucl .
phys . * # 1 * , # 2 ( # 3 ) # 1#2#3astrophys .
j. * # 1 * , # 2 ( # 3 ) # 1#2#3int .
j. mod .
phys . * # 1 * , # 2 ( # 3 ) # 1#2#3mod .
* a#1 * , # 2 ( # 3 ) # 1#2#3nature * # 1 * , # 2 ( # 3 ) # 1#2#3nucl .
* b#1 * , # 2 ( # 3 ) # 1#2#3astropart .
* b#1 * , # 2 ( # 3 ) | we update cosmological bound on axion model . the contribution from the anharmonic effect and the newly introduced initial overshoot correction
are considered .
we present an explicit formula for the axion relic density in terms of the qcd scale @xmath0 , the current quark masses @xmath1 s and the peccei - quinn scale @xmath2 , including firstly introduced @xmath3 factor which is from the initial overshoot .
[email protected] , address = department of physics and astronomy and center for theoretical physics , seoul national university , seoul 151 - 747 , korea |
the galaxies hosting powerful radio - loud active galactic nuclei ( agn ) at high redshift are believed to evolve into the most massive galaxies in the present - day universe , which reside locally in rich cluster environments ( e.g. , best et al . , 1998 ) .
high - redshift radio galaxies ( hzrgs ) are therefore popular targets in studies of the co - evolution of massive galaxies , their central agn , and their environment on larger scales . in recent years
, observations with chandra and xmm - newton have detected extended x - ray emission in an increasing sample of hzrgs ( fabian et al .
2003 , 2003b ; carilli et al . 2002 ; yuan et al . 2003
; scharf et al . 2003 ; belsole et al . 2004 ; overzier et al .
2005 ; erlund et al . 2006 ; blundell et al .
generally associated with the radio structure , the extended x - ray emission is believed to arise in most cases partly or wholly from inverse - compton ( ic ) scattering of cosmic microwave background ( cmb ) photons by relativistic electrons in the jets and/or lobes of the radio galaxy . as noted by schwartz ( 2002 ) ,
the steep @xmath6 dependence of the density of cmb photons on redshift suggests this process will be increasingly significant toward higher redshifts .
based on extrapolation from the radio source number counts , celotti & fabian ( 2004 ) estimate that at redshifts above @xmath7 the majority of luminous ( @xmath8 erg s@xmath4 ) extended x - ray sources may result from ic scattering of the cmb in hzrgs .
this new class of x - ray source may provide crucial insight into the effect of agn activity on the evolution of massive galaxies and their larger scale environment .
the identification of bubble - like features in low - and intermediate - redshift clusters indicates agn jets can inject a significant amount of energy into their environs ( e.g. , mcnamara et al .
2000 ; fabian et al .
similar processes have been suspected to exert significant influence on the evolution of structure at high redshift , e.g. , through the regulation of gas accretion and star formation in massive galaxies ( e.g. , silk & rees , 1998 ) and as a source of non - gravitational heating within forming cluster cores ( e.g. , bower et al . , 2001 ) .
however , the mechanisms and magnitude of agn feedback at high redshift are unknown .
the detection of ic - scattered x - ray emission in hzrgs has begun to clarify the evolution and energetics of high - redshift radio lobes .
as energy losses from the more energetic electron population responsible for the radio emission are more rapid than those of the less energetic population involved in ic scattering , the x - ray lobe emission can outlast the radio emission and provide a measure of past radio activity ( e.g. , fabian et al .
2003 ; blundell et al .
x - ray measurements have allowed estimates of the minimum lorentz factors , @xmath9 , in radio hotspots ( blundell et al . , 2006 ) and
the total energy contained within x - ray luminous radio lobes ( erlund et al .
2006 ) at @xmath10 .
scharf et al . ( 2003 ) , noting the relative positions of x - ray and ly@xmath11 flux in 4c 41.17 at @xmath12 , suggest that the ic - scattered x - ray photons could themselves give rise to a secondary feedback mechanism by photoionising the surrounding medium .
both the direct mechanical feedback from the jets and this second radiative mechanism could influence the structure and energetics of forming structures on galactic and cluster scales and , interestingly , would preferentially affect the most massive structures at high redshift . in this letter
we report the detection of extended x - ray emission associated with the radio galaxy 4c 23.56 ( @xmath13 ; rttgering et al . ,
1997 ) in xmm - newton observations obtained as part of a larger study of the x - ray properties of hzrgs .
we begin with a summary of the multiwavelength properties of 4c 23.56 in 2 , present the x - ray observations and analysis in 3 and briefly discuss our findings in 4 . throughout this letter
we assume a @xmath14cdm cosmology with @xmath15=0.3 , @xmath16 , and @xmath17 km s@xmath4 kpc@xmath18 .
unless specified otherwise , x - ray fluxes are given in the observed - frame 0.58 kev band , x - ray luminosities are corrected for galactic absorption and given in the rest - frame 0.58 kev band , and errors are quoted at the 90% confidence level .
4c 23.56 is an ultra - steep spectrum , fanaroff - riley class ii radio galaxy which is among the largest known at high redshift . the core and radio lobes lie along an axis with a position angle ( p.a . ) of @xmath19 e of n and an angular size of roughly @xmath20 mpc ( chambers et al . , 1996a ) .
multiwavelength observations indicate a heavily obscured agn in a dusty host galaxy ( chambers et al .
, 1996b ; knopp & chambers , 1997 , archibald et al . ,
an enormous ly-@xmath11 emission - line nebula comprised of two opposing ionization cones with opening angles of @xmath21 covers an area of @xmath22 kpc@xmath23 parallel to the radio axis .
the near - symmetry of the emission - line morphology indicates the radio axis lies nearly in the plane of the sky , while the relative brightness of the southwestern lobe implies it is somewhat nearer to us .
moderate excess counts of both ly@xmath11 emitters and faint , red galaxies in the vicinity of 4c 23.56 suggest it may reside in an over - dense cluster or proto - cluster environment ( knopp & chambers , 1997 ) .
we obtained 32 ks xmm - newton observations of the 4c 23.56 field centred at 21@[email protected]@xmath26 , 23@[email protected]@xmath29 on 29
november 2004 ( obsid 0204400201 ) .
the data were taken in full frame mode through the thin filter and reduced using version 6.0.0 of the xmm science analysis system ( sas ) .
the observations were minimally affected by periods of background flaring , which we identified in 7 15 kev lightcurves and removed .
the remaining useful exposures were 27.7 ks in the pn camera and 30.8 ks in the mos cameras .
we applied the standard recommended pattern , flag , and energy filters to the data to produce event lists suitable for imaging and spectral analysis .
we produced images for each of the epic cameras in the full ( 0.58 kev ) , soft ( 0.52 kev ) , and hard ( 28 kev ) bands by binning the filtered events into @xmath30 pixels .
we constructed a combined epic image by summing the individual instrument images .
to improve visibility , we also created smoothed versions of all images by convolving the counts images with a gaussian kernel roughly the size of the xmm psf ( @xmath31 ) .
figure 1a shows a @xmath32 region of the unsmoothed , full - band , epic image centred on the position of 4c 23.56 .
the radio galaxy is clearly detected in the x - ray and appears slightly extended toward the ne and sw .
the extended emission is more easily identified in the smoothed images of the same region . while the hard - band image ( figure 1b ) shows only a point - source at the position of the radio galaxy , the soft - band emission ( figure 1c )
is unambiguously extended along the ne - sw axis .
an unrelated soft x - ray point source , visible in the unsmoothed counts image , is also seen at the southern edge of the figure . to quantify the significance of the extended emission detection , we extracted hard and soft band counts within the pie shaped regions centred on the x - ray point source shown in figure 1c .
two of the sectors are aligned with the axis of the extended emission while the other four serve as background regions . to the extent
that the xmm psf is spherically symmetric , these regions should contain roughly the same amount of contamination from the central point source . across all three detectors ,
we find 144 soft - band counts within the two source regions and 156 counts within the four background regions .
correcting for area , we expect @xmath33 background counts within the source region , implying an excess detection within the source regions of 58 counts , which is significant at the @xmath34 level .
the net source counts are split roughly evenly between the ne and sw regions .
considering the pn and the mos counts separately , we find significant soft - band detections of the extended flux at the @xmath35 and @xmath36 levels , respectively .
we find no statistically significant excess of hard - band counts . , scaledwidth=44.0% ] to maximise the contrast of the faint , soft , extended flux , we subtracted the smoothed soft - band image from the smoothed hard - band image to produce the difference image shown in figure 1d .
as both the central point source and the background have harder x - ray spectra than the extended emission , this method of displaying the data effectively de - emphasizes these sources of contamination , allowing the two lobes of soft emission to stand out distinctly on either side of the x - ray core .
we stress that the morphology of this difference image is not directly indicative of the surface brightness of the extended emission and must be interpreted with caution .
for example , the deficit of soft emission around the point source does not indicate the extended component is fainter near to the radio galaxy , rather that the contribution of hard flux in the wings of the point source is dominant over the soft emission at these radii .
we indicate in figure 1d the 6 cm radio positions of the core of 4c 23.56 and the peaks of its ne and sw radio lobes ( knopp & chambers , 1997 ) .
the peak of the x - ray point source is coincident with the radio position of the core and the bulk of the extended x - ray emission is aligned with the radio lobes at p.a.@xmath37 .
the peak of the excess soft flux coincides with the ne radio lobe at a distance of @xmath38 from the radio core ( @xmath39 kpc at @xmath13 ) . in the sw , the excess soft flux appears to peak interior to the radio lobe , at a radius of @xmath38 , and slightly out of alignment , forming an axis with the core with p.a.@xmath40 .
a further , fainter extension of the excess soft x - ray flux just reaches the position of the sw radio lobe at a distance of @xmath41 ( @xmath42 kpc ) from the centre .
the full extent of the extended emission along p.a.@xmath43 is @xmath44 ( @xmath45 kpc ) .
unfortunately , the large xmm psf and the substantial contamination of the soft emission by the wings of the central point source prohibit more precise characterisation of the morphology of the surface brightness of the extended emission . , scaledwidth=44.0% ] we extracted x - ray spectra for the central point source and for the extended emission from each epic camera separately .
for the core emission , we chose a circular aperture of @xmath46 radius shown in figure 1b .
as the extended emission prohibited extraction of an annular background , suitable nearby regions were selected .
for the extended emission , we extracted source and background counts from the pie - shaped sectors described above and shown in figure 1c .
response files were created for each camera using the standard sas especget tool , which we ran in extended source mode for the lobe emission .
the spectra were grouped such that each bin contained 10 counts and spectra from the three cameras were fit simultaneously to models within the xspec package .
fits were minimized using the @xmath47 gehrels statistic as appropriate for low count data .
90% confidence levels were estimated from projection of the statistic surface .
a galactic absorption column of @xmath48 @xmath49 was included in all the fits .
a single power - law fit indicates that the unresolved radio galaxy core is a hard x - ray source with a photon index @xmath50 .
although we can not distinguish between absorption and an intrinsically hard spectrum with the present data , such a hard spectrum usually indicates the presence of photoelectric absorption in an agn .
fixing the photon index to a typical @xmath51 ( miyaji et al . , 2006 ) , the best - fitting model ( reduced @xmath52=1.2 ) yields a substantial column density of @xmath53 cm@xmath18 and a luminosity for the underlying power - law of @xmath55 erg s@xmath4 , implying an aperture corrected luminosity of @xmath56 erg s@xmath4 .
the relatively poor fit statistic for the absorbed powerlaw results from the inability of this model to fit an excess of soft emission in the lowest - energy bins .
we therefore add a second powerlaw without local obscuration to represent the soft , extended component within the core aperture .
a much improved fit , shown in figure [ fig : core_spec ] , is obtained for a model with @xmath57 .
the complexity of the model strains the statistical limits of the dataset , however , and the four free parameters can not be constrained simultaneously . freezing @xmath58 at the preferred value of 2.6 yields an absorption of @xmath59 cm@xmath18 , a core luminosity of @xmath60 erg s@xmath4 , and a luminosity in the soft power - law component of @xmath61 erg s@xmath4 .
however , the fit can only rigourously provide a joint upper limit to these luminosities of @xmath62 erg s@xmath4 .
the statistical quality of the spectra of the extended source is extremely poor due to the faintness of the source emission .
as can be seen in figure [ fig : ext_spec ] , there are only a few bins in the mos spectra and many of the bins in the pn spectrum are statistically equivalent to zero .
the positive pn bins at 45 kev most likely result from imperfect subtraction of the background .
we fit a single powerlaw model to the data .
the best - fitting solution ( reduced @xmath63 ) yields a soft photon index @xmath64 and a luminosity of @xmath65 erg s@xmath4 .
exclusion of the possibly contaminated bins at 45 kev does not significantly alter the fit .
for completeness , we note that an equally good fit ( reduced @xmath66 ) was obtained for a thermal mekal model with a temperature of 1.7 kev , a luminosity of @xmath67 erg s@xmath4 , and an abundance of 0.3 solar .
the @xmath68 limit on the luminosity of any thermal emission on the scale of the detected lobes outside the radio axis is @xmath69 erg s@xmath4 .
we can also estimate the luminosity of the extended component directly from the counts extracted from the pn camera in [ imaging ] .
following the procedure described in xmm survey science centre memo ssc - lux_tn-0059 , issue 3.0 and adopting a @xmath70 powerlaw spectrum with galactic absorption , we find a 0.52 kev pn energy conversion factor of @xmath71 counts @xmath72 erg@xmath4 .
dividing the net counts in the soft band by this value and the exposure time yields a soft - band flux of @xmath73 erg s@xmath4 cm@xmath18 , implying an unabsorbed full - band luminosity of @xmath74 erg s@xmath4 which is in excellent agreement with the spectroscopically derived value .
the analyses above indicate soft , extended , x - ray emission coincident with the radio lobes of 4c 23.56 over a scale of @xmath75 mpc with a luminosity of @xmath76 erg s@xmath4 . while there are several possible emission mechanisms for extended x - ray flux in hzrgs ( see , e.g. , overzier et al . , 2005 and references therein ) , few can account for the morphology and magnitude of the flux seen in 4c 23.56 . extrapolating the synchrotron spectrum from the radio waveband
, we find the observed x - ray luminosity is two orders of magnitude higher than would be expected from synchrotron emission .
synchrotron self - compton emission , often invoked to explain x - ray hotspots in local radio galaxies ( e.g. , harris et al . 2000 ) , can not produce x - rays over a large area because the synchrotron photon density can only exceed that of the cmb in compact regions .
ic scattering of optical / ir photons from the central agn ( brunetti et al . ,
2001 ) can not account for emission so far from the core .
shock heating of an ambient medium by the expanding radio source a proposed mechanism for the extended x - ray emission in pks1138 - 262 ( carilli et al . ,
2002 ) would require an enormous reservoir of hot , high filling - factor gas which is not detected in the off - axis regions .
finally , thermal emission from hot intracluster gas can not be ruled out , but the surface brightness of the observed emission is much higher than expected and a contrived containment model would be needed to explain an extremely elongated morphology aligned with the radio lobes .
we believe that in 4c 23.56 , as in most previously reported hzrgs with x - ray - bright radio lobes , ic scattering of cmb photons is the most likely emission mechanism for the majority , if not the entirety , of the extended x - ray flux .
fourteen @xmath77 hzrgs currently in the literature ( see references in 1 ) have x - ray lobes on scales ranging from a few tens to a few hundreds of kpc with typical luminosities of a few times @xmath78 erg s@xmath4 . the x - ray lobes in 4c 23.56 are the second largest discovered to date exhibits x - ray emission associated with hotspots and lobes which are separated by @xmath79 kpc ( blundell et al .
2006 ) . ] , and the first @xmath80 mpc - scale lobes discovered at the @xmath2 epoch suspected to be the predominant era of massive galaxy formation .
we now briefly consider the energetics of the x - ray detected lobes , assuming the entirety of the extended x - ray emission arises in ic scattering of cmb photons .
particles with a large lorentz factor , @xmath81 , will upscatter cmb photons of frequency @xmath82 to higher frequencies @xmath83 as @xmath84 scattering to x - ray energies therefore requires electrons with a lorentz factor @xmath85 .
if we assume that all electrons responsible for the x - ray lobes have this lorentz factor , we can use the observed x - ray luminosity and the known cmb photon energy density to estimate the energy stored in the ic - emitting electron population .
the energy flux in scattered x - rays is @xmath86 while the energy of the relativistic electrons is @xmath87 where @xmath88 is the number density of relativistic electrons , @xmath89 is the thompson cross - section , @xmath90 is the speed of light , and @xmath91 is the rest mass of the electron .
@xmath92 is the energy density of the cmb , which increases from the local value with redshift as @xmath6 .
the energy in the scattering electron population is then given by @xmath93 where @xmath94 is the 210 kev x - ray luminosity in units of @xmath78 erg s@xmath4 .
as noted by erlund et al .
( 2006 ) , this calculation provides only a lower limit on the energy in the scattering population . assuming a more realistic power - law distribution of lorentz factors reveals dependencies of the total electron energy on the minimum and maximum lorentz factors and on the x - ray photon index .
these parameters are poorly constrained ; adopting typical values increases the estimated energy by a factor of a few .
in addition , this estimate considers only the energy in relativistic electrons . a substantial population of relativistic protons ( e.g. , celotti & fabian , 1993 ; sikora et al . , 1995 ) could further increase the implied lobe energies by a factor of up to @xmath95 . converting our measured 0.58 kev luminosity to 210 kev with @xmath70
, we find a lower limit to the energy stored in relativistic electrons in the lobes of 4c 23.56 of @xmath96 erg ; adopting a more typical @xmath97 , @xmath98 erg . comparing these estimates with those of erlund et al .
( 2006 ) for three @xmath10 hzrgs we note that while the x - ray lobes in 4c 23.56 are at least twice as large as the lower - redshift sources , the minimum energy stored in ic - emitting electrons appears to be roughly comparable . considering only losses through ic emission ,
the radiative lifetime of the 4c 23.56 x - ray lobes is @xmath99 yr . an enormous reservoir of energy equivalent to @xmath100 supernovae is stored in the ic - emitting electron population within the radio lobes of 4c 23.56 . as the cooling timescale of this population through ic - scattering to x - ray wavelengths
is relatively long , a large fraction of this energy could potentially be injected into the surroundings during a given epoch of agn activity .
the vast extent of the x - ray luminous lobes suggests a mechanism by which significant feedback could take place at radii of up to 250 kpc from the central nucleus .
it is not clear , however , if the energy available in the ic scattering electron population is sufficient to be a major source of feedback influencing the evolution of either the host galaxy or the forming cluster environment . assuming a velocity dispersion of 300 km s@xmath4 , the lobe energy estimates derived above are 1020 times lower than the binding energy of a @xmath101 m@xmath102 elliptical and roughly 500 times lower than the @xmath103 erg m@xmath104 estimated by benson et al .
( 2003 ) to be required to inhibit the excessive growth of such a galaxy at high redshift . in terms of heating of the intracluster medium , wu , fabian , & nelson ( 2000 )
estimate non - gravitational heating of 12 kev per particle is needed to explain the observed deviation of galaxy cluster properties from the theoretical scaling relations . for a cluster with a gas mass of @xmath105 m@xmath102 ,
this is equivalent to @xmath106 erg or roughly 300 times our minimum energy estimate for the ic - scattering population in the lobes of 4c 23.56 . unless the assumption of a single lorentz factor , @xmath107 , is dramatically wrong or there is a significant proton component to the jet , the energy contained in the ic - scattering population in the lobes appears to fall short of the required feedback energies for the host galaxy and cluster environment by at least two orders of magnitude , even assuming perfectly efficient energy transfer acting over the required scales .
in addition , the observation of significant cooling in ic - scattering of cmb photons is evidence that a significant fraction of this energy does not get injected into the local environment .
nevertheless , ic scattering could provide a very efficient mechanism for producing hard , ionising photons on scales far from the central agn . given reasonable assumptions for the halo gas density , x - ray luminosity at the level observed in 4c 23.56 could potentially ionize a large volume of halo gas over scales of several hundred kpc , as outlined in scharf et al .
moreover , several additional mechanisms could increase the feedback potential of the x - ray lobes dramatically .
energy stored within the magnetic field , found to be of order @xmath108 in the erlund et al .
( 2006 ) sources , may increase the total energy of the jet .
more significant may be the @xmath109 work associated with the expansion of lobes into the surrounding medium , which has been estimated to provide energies of up to @xmath110 erg in lower redshift radio galaxies and observed to significantly heat the surrounding intracluster medium ( e.g. , brzan et al .
, 2004 ; mcnamara et al . , 2005 ) . finally , bursts of agn activity may occur several times over the course of a gyr , re - accelerating the radio lobes and driving subsequent periods of environmental feedback .
oj acknowledges pparc for support through a postdoctoral fellowship .
oa and pnb thank the royal society for support through its university research fellowship scheme . | we present an xmm - newton observation of the radio galaxy 4c 23.56 at @xmath0 which reveals extended x - ray emission coincident with the radio lobes spanning @xmath1 mpc .
these are the largest x - ray - bright lobes known at @xmath2 . under the assumption that these features result from inverse - compton scattering of cosmic microwave background photons by relativistic electrons in the radio source lobes ,
the measured luminosity of @xmath3 erg s@xmath4 implies a minimum energy stored within the lobes of @xmath5 erg .
we discuss the potential of the large - scale radio / x - ray lobes to influence evolution of the host galaxy and proto - cluster environment at high redshift .
galaxies : individual : 4c 23.56 galaxies : high - redshift galaxies : evolution x - rays : galaxies : clusters radio continuum ; galaxies |
the basic idea behind the nonequilibrium spectral density approach ( nsda ) is to choose the selfenergy in such a way , that the first four spectral moments @xmath80 of the spectral density are reproduced exactly .
the moments are calculated with the help of the following exact relation : @xmath81_-\dots , h]_-,[h\dots[h , c_{j_r\sigma}^+]_-\dots]_-]_+\right\rangle\ ] ] where the total number of commutators on the right hand side must be equal to @xmath82 .
inserting the hamiltonian ( [ eq : modelh ] ) into this expression yields after some calculation : @xmath83 the moments of the transport selfenergy @xmath84 can be derived in the same way .
one gets : @xmath85 the band correction @xmath37 is given by @xmath86 where @xmath87 is the hopping integral between lattice sites @xmath88 and @xmath89 .
the two higher correlation functions can be reduced to single - particle lesser green s functions@xcite .
we find @xmath90\end{aligned}\ ] ] and @xmath91 the non - diagonal lesser green s function @xmath92 is closely related to the right green s function and the transport selfenergy : @xmath93 putting all these expressions into the band correction leads to the result in eq .
[ eq : bandcorrection ] . +
the dyson equation of the right ferromagnet reads @xmath94 inserting the high - energy expansion for both selfenergies and the green s function @xmath95 yields a system of equations for the unknown moments @xmath96 of the interaction selfenergy .
it can be solved by sorting according to the order of @xmath97 and the use of the moments given earlier in this appendix .
the results are quite simple : @xmath98 these expressions are formally identical to the equilibrium case , therefore the selfenergy will also have the same form eq .
( [ eq : selfenergy])@xcite . for high energies it is acceptable to neglect higher order terms of the expansion in eq .
( [ eq : sigma_exp ] ) .
thus : @xmath99 | we study the behaviour of the magnetization in a half - metallic ferromagnet / nonmagnetic insulator / ferromagnetic metal / paramagnetic metal ( fm1/ni / fm2/pm ) tunnel junction .
it is calculated self - consistently within the nonequilibrium keldysh formalism .
the magnetic regions are treated as band ferromagnets and are described by the single - band hubbard model .
we developed a nonequilibrium spectral density approach to solve the hubbard model approximately in the switching magnet . by applying a voltage to the junction it is possible to switch between antiparallel ( ap ) and parallel ( p ) alignment of the magnetizations of the two ferromagnets .
the transition from ap to p occurs for positive voltages while the inverse transition from p to ap can be induced by negative voltages only .
this behaviour is in agreement with the slonczewski model of current - induced switching and appears self - consistently within the model , i.e. without using half - classical methods like the landau - lifshitz - gilbert equation .
there has been considerable interest in the phenomenon of current - induced switching of magnetization since it was first proposed over 10 years ago@xcite .
the basic idea behind this effect is as follows : the spin direction of electrons moving in a ferromagnet will be mostly aligned parallel to the magnetization axis .
when these spin - polarized electrons are transported to a second ferromagnet , e.g. by applying a voltage , then the spin angular momentum of the itinerant electrons will exert a torque on the local magnetic moment .
this torque is known as the spin - transfer torque .
it will have an influence on the direction of magnetization .
if the parameters of the materials are chosen in the right way and if the current through the junction is high enough it is even able to switch the magnetization of one ferromagnet from parallel to antiparallel or vice versa relative to the other one .
this effect was seen both in all - metallic junctions@xcite such as co / cu / co and in magnetic tunnel junctions ( mtjs ) consisting of two ferromagnets divided by a thin nonmagnetic insulator@xcite . in this paper
we focus on a special case of the latter , where the ferromagnetic lead is half - metallic , i.e. there are only electrons of one spin direction present at the fermi energy .
+ some of the possible technological applications of spin - transfer torques in mtjs have been discussed by diao et al.@xcite .
most of the theoretical work in this area of research has been focused on the landau - lifshitz - gilbert ( llg ) equation@xcite , which is a macroscopic , half - classical equation .
the torques entering this equation were usually calculated in a microscopic picture while treating the interactions on a mean field level . in this paper
we propose a model which takes interactions beyond mean field into account .
we make no use of the llg equation or other macroscopic approaches and thus we stay on the quantum mechanical level throughout this paper .
+ we will start the presentation of the theory by introducing a model hamiltonian which describes the magnetic tunnel junction shown schematically in fig .
[ fig : mtj ] .
there are two ferromagnetic metals ( @xmath0 and @xmath1 ) divided by a non - magnetic insulator ( @xmath2 ) and additionally a paramagnetic metal ( @xmath3 ) which is necessary to have a well - defined chemical potential on the right side of the second ferromagnet .
each region consists of a single s - like band .
the two outer leads @xmath0 and @xmath3 are treated as semi - infinite .
+ the total hamiltonian consists of several parts : @xmath4 @xmath5 describes the left ( right ) ferromagnet , @xmath6 the insulator and @xmath7 the paramagnet . both insulator and paramagnet are assumed to be non - interacting , so their hamiltonians consist of the kinetic energy only : @xmath8 @xmath9 is the annihilation ( creation ) operator of an electron with wave vector @xmath10 and spin @xmath11 .
@xmath12 is the dispersion of the lattice which throughout this paper is chosen as a tight - binding bcc lattice .
the applied voltage @xmath13 will shift the center of gravity of the paramagnet by @xmath14 and half of that amount for the insulator , @xmath15 .
positive voltage , @xmath16 , will shift the bands to lower energies , while negative applied voltages result in a shift towards higher energies .
+ the hamiltonians of the left ( l ) and right ( r ) ferromagnet are formally almost identical .
besides the kinetic energy they also include on - site coulomb interaction .
they are given in a mixed bloch - wannier representation : @xmath17 where @xmath18 stands for either @xmath0 or @xmath1 .
the hubbard-@xmath19 determines the interaction strength .
@xmath20 is the occupation number operator .
the voltage @xmath13 shifts only the band center of the right ferromagnet , @xmath21 , while the left ferromagnet is not directly influenced by @xmath13 , i.e. @xmath22 .
+ the remaining three terms of the hamiltonian are responsible for the coupling between the different regions .
these couplings act as a hybridization@xcite between the bands and therefore the hamiltonians are ( @xmath23 ) : @xmath24 they are characterized by the coupling constants @xmath25 which determine the strength of the hybridization between the different bands .
in general the couplings are wave vector dependent but for the sake of simplicity we neglect this dependence , @xmath26 . furthermore we assume the coupling between the ferromagnets and the insulator to be equal , so that @xmath27 . altogether there remain two couplings @xmath28 and @xmath29 which can not be calculated within this model , so they will be treated as parameters .
+ the main topic of this work will be the calculation of the non - equilibrium magnetization @xmath30 of the right ferromagnet within the keldysh formalism@xcite .
it can be calculated with the help of the fourier transform of the so - called lesser green s function defined as @xmath31 : @xmath32 @xmath33 is the occupation number of particles with spin @xmath11 in the right ferromagnet . in order to derive the lesser green s function
one first has to calculate the retarded one , @xmath34 . by using
the equation of motion method one finds @xmath35 two different selfenergies appear in the green s function .
first there is the interaction selfenergy which can only be calculated approximately for the hubbard model .
we propose a non - equilibrium spectral density approach ( nsda ) .
the basic idea behind this approach is to choose the selfenergy in such a way that the first four spectral moments are reproduced by the theory .
some details of its derivation are given in appendix [ app : nsda ] .
the mean field ( stoner ) solution of the hubbard model on the other hand satisfies only the first two moments .
one finds for the selfenergy @xmath36 this expression is coincidentally formally identical to the equilibrium spectral density approach@xcite .
the difference is in the spin - dependent band correction @xmath37 which is given by @xmath38\label{eq : bandcorrection}\end{aligned}\ ] ] it has to be calculated self - consistently since @xmath37 also appears on the right hand side as part of the lesser green s function .
@xmath39 is the center of gravity of the right ferromagnet .
the second selfenergy is the transport selfenergy which is due to electrons hopping between the different materials .
its retarded ( lesser ) component is given by @xmath40 @xmath41 is the green s function of the insulator when it is only coupled to the left ferromagnet , i.e. @xmath42 since we neglected the wave vector dependence of the couplings , the transport selfenergy is only formally dependent on the wave vector .
@xmath43 and @xmath44 are the equilibrium green s functions of the left ferromagnet and the paramagnet , respectively
. they can be easily calculated by the equation of motion method .
one finds for @xmath45 : @xmath46 the paramagnet does not include interactions so that @xmath47 .
since we are mainly interested in the properties of the right ferromagnet , we assume that the left one is half - metallic so that its minority states play no role for small voltages .
this is done by using the mean field selfenergy @xmath48 with sufficiently large @xmath49 .
thus the retarded green s function is known .
+ the lesser green s function follows immediately from the keldysh equation : @xmath50 where the advanced green s function is simply the complex conjugated of the retarded one , @xmath51 .
furthermore we need the lesser component of the transport selfenergy which was already defined in eq .
( [ eq : delta ] ) .
the lesser part of the insulator green s function can again be calculated with the help of the keldysh equation : @xmath52 since the green s functions in the left ferromagnet and the paramagnet are equilibrium quantities , their lesser parts read @xmath53 where @xmath54 is the fermi function in lead @xmath55 with chemical potential @xmath56 .
they are related by @xmath57 .
thus we have a closed set of equations for calculating the magnetization of the ferromagnet . + between the metals and the insulator .
the arrows indicate in which direction the voltage was changed .
parameters : band occupation @xmath58 ; band widths : @xmath59 ev and @xmath60 ; interaction strengths : @xmath61 ; band center of the insulator @xmath62 ; hybridization strength : @xmath63 ev ; temperature : @xmath64 in fig .
[ fig : switch ] a typical numerical solution for the voltage - dependent magnetization is shown .
we will first discuss the black curve , which was calculated with a hybridization strength of @xmath65 ev .
for the calculation we started with parallel alignment of the two magnetizations ( point @xmath66 in the figure )
. then a negative voltage is applied , i.e. the right ferromagnet is shifted to higher energies compared to the left one . at a critical voltage the parallel alignment becomes unstable and the magnetization reverses its sign ( @xmath67 to @xmath68 ) .
thus the magnetizations are now antiparallel .
when the voltage is further decreased the magnetization stays more or less constant until point @xmath69 is reached .
then the process is reversed and the voltage is reduced to zero again .
the magnetization follows the same line as before until the switching point @xmath68 is reached .
there it does not switch back to parallel alignment but rather stays at about the same level .
when @xmath70 is reached the direction of the voltage is reversed , i.e. the right ferromagnet will now be shifted to lower energies . for small voltages
there is only a slight increase until a critical voltage is reached ( @xmath71 ) .
this voltage has approximately the same value as the first one at point @xmath67 , but of course with an opposite sign .
there the antiparallel alignment is no longer stable and the systems returns to its initial parallel state which is not influenced by higher voltages ( @xmath72 to @xmath73 ) .
then the voltage is turned off and the system will be at its starting point @xmath66 again , so the hysteresis loop is complete .
+ as another test we start again at point @xmath66 , but this time we turn on a positive voltage
. then no switching occurs and the system will move reversibly to point @xmath73 .
a similar reversible behaviour is seen when the alignment is antiparallel ( @xmath70 ) and the voltage is decreased .
this is shown by the arrows in the figure .
so , one has to conclude that switching of magnetization from parallel to antiparallel alignment is only possible for negative voltages and the reverse process will only appear for positive voltage .
the behaviour just described is one of the hallmarks of current - induced switching of magnetization and thus our proposed model is indeed able to simulate this effect without leaving the microscopic picture ! + for @xmath74 and @xmath65 ev .
upper picture for parallel alignment of the magnetizations ( point a ) , lower picture for antiparallel alignment ( point e ) .
spin up is shown along the positive , spin down along the negative axis .
the black line is the qdos of the right ferromagnet , treated within the nsda , and the red broken line shows the qdos of the left ferromagnet in mean field . ]
now we want to give a short explanation on how exactly our model is able to provide these results .
the key to the understanding lies in the effect the hybridization parts of the hamiltonian have on the quasiparticle density of states ( qdos ) of the switching magnet and in the polarization of the current .
a hybridization between two bands generally will lead to a repulsion between them , i.e. the energetic distance between their respective centers of gravity will increase the stronger the effect of the hybridization is .
the magnitude of this shift is mainly influenced by three quantities : the strength of the hybridization itself ( @xmath29 and @xmath28 in our case ) , the energetic distance between the two bands ( the closer they are to each other , the stronger they will be repelled ) and their spectral weight ( higher spectral weight leads to stronger repulsion ) . in the upper part of fig .
[ fig : qdos ] we plotted a typical qdos for the nsda without applied voltage for parallel alignment of the two magnetizations .
the dashed line represents the density of states of the left ferromagnet .
the splitting of the right qdos into lower and upper hubbard band at @xmath75 and @xmath76 is clearly visible .
additionally there are contributions of the insulator at @xmath77 ev which are due to the hybridization .
what happens when a voltage is turned on depends on its sign .
the left ferromagnet fm1 is not influenced by the voltage , so its qdos will be the same .
let us first discuss the case @xmath16 where both spin bands of the right ferromagnet fm2 are shifted to lower energies .
but due to the repulsion between the spin up bands of fm1 and fm2 the shift of the spin up band will actually be stronger than for spin down , where the repulsion is much weaker .
so a positive voltage leads to a stabilization of the magnetization .
this effect is enhanced by the current .
for positive voltages it will flow from left to right . since the left ferromagnet is fully polarized there are only spin up electrons tunneling into the right ferromagnet .
these are the reasons for the slight increase of magnetization in fig .
[ fig : switch ] between points a and h and also an explanation why there can be no switching in this case . on the other hand ,
if we apply a negative voltage , @xmath78 , the right spin bands will be shifted to higher energies .
for the same reasons as discussed above the shift of the spin up band will be enhanced by the hydridization .
thus the difference between the centers of gravity of both spin bands will be decreased .
if the hybridization strength is sufficiently large this additional shift together with the selfconsistency will be enough to push the spin up band above the spin down band and thus the magnetization changes sign .
the selfconsistency is important since it will enhance the shift , because the occupation in one band depends on the occupation in the other one . in this case
current flows from the right to the left lead .
since there are only spin up states available for tunneling in the left ferromagnet , the tunneling current flowing out of the switching magnet will consist of spin up electrons only .
this leads to an additional decrease of the right magnetization .
this explains the behaviour shown by the parallel aligned curve in fig .
[ fig : switch ] .
+ the antiparallel case can be explained in a very similar way . in the lower part of fig .
[ fig : qdos ] the corresponding quasiparticle density of states is shown , again without applied voltage .
the obvious difference to the case discussed above is that the center of gravity of the lower spin down band is below the lower spin up band .
this is the reason for the negative magnetization of course , but it is also responsible for the reversed behaviour with respect to the applied voltage . in this case
a negative voltage can not push the center of gravity of the spin up band below the spin down band .
it rather has the opposite effect because the repulsion pushes the spin up band to even higher energies and thus leads to a more stable magnetization which can also be seen in fig .
[ fig : switch ] between points e and d. for a positive voltage the spin up band of fm2 moves below the spin up band of fm1 so that the hybridization will shift it to lower energies compared to the spin down band . again
, if the hybridization strength is larger than a critical value this additional shift will be enough to reverse the two spin directions such that the magnetization changes sign .
for the same reasons as discussed for the parallel case , a positive voltage will increase the magnetization , while a negative voltage has the opposite effect .
therefore the behaviour of the magnetization in fig .
[ fig : switch ] can be understood in terms of the quasiparticle density of states .
+ in order to prove the explanation based on the hybridization we plotted a second magnetization curve in fig .
[ fig : switch ] with smaller hybridization strength @xmath79 ev . obviously in this case no switching occurs .
there is only a slight change of magnetization . starting from parallel ( antiparallel ) alignment
the magnetization is reduced for negative ( positive ) voltages .
this is in agreement with the explanation given above .
since the hybridization is weaker the repulsion between the bands is also reduced .
it is not strong enough to push the spin down band above the spin up band or vice versa .
thus the direction of magnetization is not changed .
the current density through the junction is closely linked to the coupling strength between the materials@xcite : smaller @xmath28 corresponds to a weaker current . from the results shown in fig .
[ fig : switch ] we can conclude that in order to switch the magnetization the current has to exceed a certain value . + to summarize , we presented a selfconsistent calculation of the voltage - dependent magnetization in a magnetic tunnel junction within a microscopic non - equilibrium framework .
the magnetization shows a hysteresis behaviour similar to that seen in experiments .
the reason for this effect was explained to be the hybridization between left and right ferromagnets which could be seen with the help of the quasiparticle density of states .
it should be noted that the behaviour discussed above does only appear for very special parameter sets ( such as low band occupation , small u ) when one uses the mean field approximation for the right ferromagnet .
this seems reasonable because it is known that mean field strongly overestimates the stability of ferromagnetism .
thus it should be more difficult to switch the direction of magnetization .
we have to conclude that higher correlations seem to be an important factor when describing current - induced switching of magnetization within this model .
one might argue that the kondo peak is missing in the nsda which should have considerable influence on the magnetization .
however , we investigated the strong - coupling regime only , where it is known that the kondo peak does not play a major role . on the other hand
it would be a very interesting expansion of the model to examine its weak - coupling behaviour .
another important extension would be the inclusion of spin - orbit coupling which is widely believed to be the microscopic origin of phenomenological damping effects@xcite which play a crucial role in the macroscopic description of switching of magnetization . |
many complex networks have richer structures than graphs can have . inherently they have hypergraph structures : interconnections often cross multiple nodes . treating these networks as graphs causes a loss of some structures . nonetheless , it is still popular to use graph tools to study these networks ; one of them is the laplacian spectrum .
let @xmath0 be a graph on @xmath1 vertices .
the _ laplacian _
@xmath2 of @xmath0 is the @xmath3-matrix @xmath4 , where @xmath5 is the adjacency matrix and @xmath6 is the diagonal matrix of degrees .
let @xmath7 be the eigenvalues of @xmath2 , indexed in non - decreasing order .
it is known that @xmath8 for @xmath9 .
if @xmath0 is connected , then @xmath10 .
the first nonzero laplacian eigenvalue @xmath11 is related to many graph parameters , such as the mixing rate of random walks , the graph diameter , the neighborhood expansion , the cheeger constant , the isoperimetric inequalities , expander graphs , quasi - random graphs , etc @xcite . in this paper , we define a set of laplacians for hypergraphs .
laplacians for regular hypergraphs was first introduced by chung @xcite in 1993 using homology approach .
the first nonzero laplacian eigenvalue can be used to derive several useful isoperimetric inequalities .
it seems hard to extend chung s definition to general hypergraphs .
other researchers treated a hypergraph as a multi - edge graph and then defined its laplacian to be the laplacian of the corresponding multi - edge graph .
for example , rodrguez @xcite showed that the approach above had some applications to bisections , the average minimal cut , the isoperimetric number , the max - cut , the independence number , the diameter etc .
what are `` right '' laplacians for hypergraphs ? to answer this question ,
let us recall how the laplacian was introduced in the graph theory .
one of the approaches is using geometric / homological analogue , where the laplacian is defined as a self - joint operator on the functions over vertices .
another approach is using random walks , where the laplacian is the symmetrization of the transition matrix of the random walk on a graph .
chung @xcite took the first approach and defined her laplacians for regular hypergraphs . in this paper , we take the second approach and define the laplacians through high - ordered random walks on hypergraphs . a high - ordered walk on a hypergraph @xmath12 can be roughly viewed as a sequence of overlapped oriented edges @xmath13 . for @xmath14 ,
we say @xmath13 is an @xmath15-walk if @xmath16 for each @xmath17 in @xmath18 .
the choice of @xmath15 enables us to define a set of laplacian matrices @xmath19 for @xmath12 . for @xmath20 ,
our definition of laplacian @xmath21 is the same as the definition in @xcite . for @xmath22 , while we restrict to regular hypergraphs , our definition of laplacian @xmath23 is similar to chung s definition @xcite .
we will discuss their relations in the last section . in this paper
, we show several applications of the laplacians of hypergraphs , such as the mixing rate of high - ordered random walks , the generalized diameters , and the edge expansions .
our approach allows users to select a `` right '' laplacian to fit their special need .
the rest of the paper is organized as follows . in section 2 ,
we review and prove some useful results on the laplacians of weighted graphs and eulerian directed graphs .
the definition of laplacians for hypergraphs will be given in section 3 .
we will prove some properties of the laplacians of hypergraphs in section 4 , and consider several applications in section 5 .
in last section , we will comment on future directions .
in this section , we review some results on laplacians of weighted graphs and eulerian directed graphs .
those results will be applied to the laplacians of hypergraphs later on . in this paper , we frequently switch domains from hypergraphs to weighted ( undirected ) graphs , and/or to directed graphs . to reduce confusion , we use the following conventions through this paper .
we denote a weighted graph by @xmath0 , a directed graph by @xmath24 , and a hypergraph by @xmath12 .
the set of vertices is denoted by @xmath25 , @xmath26 , and @xmath27 , respectively .
( whenever it is clear under the context , we will write it as @xmath28 for short . )
the edge set is denoted by @xmath29 , @xmath30 , and @xmath31 , respectively .
the degrees @xmath32 and volumes @xmath33 are defined separately for the weighted graph @xmath0 , for the directed graph @xmath24 , and for the hypergraph @xmath12 .
readers are warned to interpret them carefully under the context . for a positive integer @xmath15 and a vertex set @xmath28 ,
let @xmath34 be the set of all ( ordered ) @xmath15-tuples consisting of @xmath15 distinct elements in @xmath28 .
let @xmath35 be the set of all unordered ( distinct ) @xmath15-subset of @xmath28 .
let @xmath36 be the row ( or the column ) vector with all entries of value @xmath37 and @xmath38 be the identity matrix . for a row ( or column )
vector @xmath39 , the norm @xmath40 is always the @xmath41-norm of @xmath39 .
a _ weighted graph _
@xmath0 on the vertex set @xmath28 is an undirected graph associated with a weight function @xmath42 satisfying @xmath43 for all @xmath44 and @xmath45 in @xmath25 .
here we always assume @xmath46 for every @xmath47 .
a simple graph can be viewed as a special weighted graph with weight @xmath37 on all edges and @xmath48 otherwise .
many concepts of simple graphs are naturally generalized to weighted graphs . if @xmath49 , then @xmath44 and @xmath45 are _ adjacent _ , written as @xmath50 .
the graph distance @xmath51 between two vertices @xmath44 and @xmath45 in @xmath0 is the minimum integer @xmath52 such that there is a path @xmath53 in which @xmath54 for @xmath55 .
if no such @xmath52 exists , then we let @xmath56 .
if the distance @xmath51 is finite for every pair @xmath57 , then @xmath0 is _
connected_. for a connected weighted graph @xmath0 , the _ diameter _ ( denoted by @xmath58 ) is the smallest value of @xmath51 among all pairs of vertices @xmath57 .
the _ adjacency matrix _
@xmath5 of @xmath0 is defined as the matrix of weights , i.e. , @xmath59 for all @xmath60 and @xmath61 in @xmath28 .
the _ degree _ @xmath62 of a vertex @xmath60 is @xmath63 .
let @xmath6 be the diagonal matrix of degrees in @xmath0 .
the _ laplacian _
@xmath2 is the matrix @xmath4 .
let @xmath64 be the eigenvalues of @xmath2 , indexed in the non - decreasing order .
it is known @xcite that @xmath65 for @xmath9 .
if @xmath0 is connected , then @xmath10 . from now on ,
we assume @xmath0 is connected .
the first non - trivial laplacian eigenvalue @xmath11 is the most useful one .
it can be written in terms of the rayleigh quotient as follows ( see @xcite ) @xmath66 here the infimum is taken over all functions @xmath67 which is orthogonal to the degree vector @xmath68 . similarly , the largest laplacian eigenvalue @xmath69 can be defined in terms of the rayleigh quotient as follows @xmath70 note that scaling the weights by a constant factor will not affect the laplacian .
a weighted graph @xmath0 is _ complete _ if @xmath71 for some constant @xmath72 such that @xmath73 , independent of the choice of @xmath57 with @xmath74 .
we say @xmath0 is _ bipartite _ if there is a partition @xmath75 such that @xmath76 for all @xmath77 and all @xmath78 .
we have the following facts ( see @xcite ) .
1 . @xmath79 for each @xmath80 .
2 . the number of @xmath48 eigenvalues equals the number of connected components in @xmath0 .
if @xmath0 is connected , then @xmath10 .
3 . @xmath81 if and only if @xmath0 has a connected component which is a bipartite weighted subgraph .
4 . @xmath82 if and only if @xmath0 is a complete weighted graph .
it turns out that @xmath11 and @xmath69 are related to many graph parameters , such as the mixing rate of random walks , the diameter , the edge expansions , and the isoperimetric inequalities . a random walk on a weighted graph @xmath0 is a sequence of vertices @xmath83 such that the conditional probability @xmath84 for @xmath85 .
vertex probability distribution _ is a map @xmath86 such that @xmath87 for each @xmath45 in @xmath0 and @xmath88 .
it is convenient to write a vertex probability distribution into a row vector .
a random walk maps a vertex probability distribution to a vertex probability distribution through multiplying from right a transition matrix @xmath89 , where @xmath90 for each pair of vertices @xmath44 and @xmath45 .
we can write @xmath91 .
the second largest eigenvalue @xmath92 , denoted by @xmath93 for short , is @xmath94 .
let @xmath95 for each vertex @xmath44 in @xmath0 .
observe @xmath96 is the stationary distribution of the random walk , i.e. , @xmath97 .
a random walk is _ mixing _ if @xmath98 for any initial vertex probability distribution @xmath99 .
it is known that a random walk is always mixing if @xmath0 is connected and not a bipartite graph . to overcome the difficulty resulted from being a bipartite graph ( where @xmath81 ) , for @xmath100
, we consider an @xmath101-lazy random walk , whose transition matrix @xmath102 is given by @xmath103 for each @xmath44 and @xmath104 for each pair of vertices @xmath44 and @xmath45 with @xmath74 .
note that the transition matrix is @xmath105 let @xmath106 and @xmath107 . since @xmath108 is a symmetric matrix , we have @xmath109 it turns out that the mixing rate of an @xmath101-lazy random walk is determined by @xmath110 . [ thm1 ] for @xmath111 , the vertex probability distribution @xmath112 of the @xmath101-lazy random walk at time @xmath52 converges to the stationary distribution @xmath96 in probability . in particular , we have @xmath113 here @xmath99 is the initial vertex probability distribution .
* proof : * notice that @xmath114 and @xmath115 .
we have @xmath116 for each subset @xmath117 of @xmath25 , the _ volume _ @xmath118 is @xmath119 .
if @xmath120 , then we write @xmath121 instead of @xmath122
. we have @xmath123 if @xmath124 is the complement set of @xmath117 , then have @xmath125 .
for any two subsets @xmath117 and @xmath126 of @xmath25 , the _ distance _ @xmath127 between @xmath117 and @xmath126 is @xmath128 .
[ thm2 ] in a weighted graph @xmath0 , for @xmath129 with distance at least 2 , we have @xmath130 a special case of theorem [ thm2 ] is that both @xmath117 and @xmath126 are single vertices , which gives an upper bound on the diameter of @xmath0 .
[ cor1 ] if @xmath0 is not a complete weighted graph , then we have @xmath131 where @xmath132 is the minimum degree of @xmath0 . for @xmath129 ,
let @xmath133 be the set of edges between @xmath117 and @xmath126 .
namely , we have @xmath134 we have the following theorem . [ thm3 ] if @xmath117 and @xmath126 are two subsets of @xmath25 , then we have @xmath135 the laplacian of a general directed graph was introduced by chung @xcite .
the theory is considerably more complicated than the one for undirected graphs , but when we consider a special class of directed graphs eulerian directed graphs , it turns out to be quite neat . let @xmath24 be a directed graph with the vertex set @xmath26 and the edge set @xmath30 . a directed edge from @xmath60 to @xmath61
is denoted by an ordered pair @xmath136 or @xmath137 .
the _ out - neighborhood _ @xmath138 of a vertex @xmath60 in @xmath24 is the set @xmath139 .
the _ out - degree _
@xmath140 is @xmath141 .
similarly , the _ in - neighborhood _ @xmath142 is @xmath143 , and the _ in - degree _ @xmath144 is @xmath145 .
a directed graph @xmath24 is _ eulerian _ if @xmath146 for every vertex @xmath60 . in this case , we simply write @xmath147 for each @xmath60 . for a vertex subset @xmath148 ,
the _ volume _ of @xmath148 , denoted by @xmath149 , is @xmath150 . in particular , we write @xmath151 .
eulerian directed graphs have many good properties . for example , a eulerian directed graph is strongly connected if and only if it is weakly connected .
the _ adjacency matrix _ of @xmath24 is a square matrix @xmath5 satisfying @xmath152 if @xmath153 and 0 otherwise .
let @xmath6 be the diagonal matrix with @xmath154 for each @xmath155 .
let @xmath156 , i.e. , @xmath157 note that @xmath158 is not symmetric .
we define the laplacian @xmath2 of @xmath24 to be the symmetrization of @xmath158 , that is @xmath159 since @xmath2 is symmetric , its eigenvalues are real and can be listed as @xmath160 in the non - decreasing order .
note that @xmath11 can also be written in terms of raleigh quotient ( see @xcite ) as follows @xmath161 chung @xcite proved a general theorem on the relationship between @xmath11 and the diameter . after restricting to eulerian
directed graphs , it can be stated as follows .
[ see @xcite][thm4 ] suppose @xmath24 is a connected eulerian directed graph , then the diameter of @xmath24 ( denoted by @xmath162 ) satisfies @xmath163 where @xmath11 is the first non - trivial eigenvalue of the laplacian , and @xmath132 is the minimum degree @xmath164 .
the main idea in the proof of the theorem above is using @xmath101-lazy random walks on @xmath24 . a random walk on a eulerian directed graph
@xmath24 is a sequence of vertices @xmath83 such that for @xmath85 , the conditional probability @xmath165 equals @xmath166 for each @xmath167 and @xmath48 otherwise . for @xmath168
, the @xmath101-lazy random walk is defined similarly .
the transition matrix @xmath102 of the @xmath101-lazy random walk satisfies @xmath169 chung @xcite considered only @xmath170-lazy random walks .
here we prove some results on @xmath101-lazy random walks for @xmath171 .
let @xmath172 for each @xmath173 .
note that @xmath96 is the stationary distribution , i.e. @xmath174 .
let @xmath175 .
the key observation is that there is a unit - vector @xmath176 such that @xmath176 is both a row eigenvector and a column eigenvector of @xmath108 for the largest eigenvalue @xmath37 . here
let @xmath177 .
we have @xmath178 let @xmath179 be the orthogonal complement of @xmath176 in @xmath180 .
it is easy to check @xmath181 maps @xmath179 to @xmath179 .
let @xmath182 be the spectral norm of @xmath181 when restricting to @xmath179 .
an equivalent definition of @xmath182 is the second largest singular value of @xmath181 , i.e. , @xmath183 [ lemma1 ] we have the following properties for @xmath182 . 1 . for every @xmath184 , we have @xmath185 .
2 . @xmath186 3 .
@xmath187 * proof : * item 1 is from the definition of @xmath182 . since the largest eigenvalue of @xmath108 is 1 , we have @xmath188 .
in particular , @xmath189 .
note that @xmath190 .
let @xmath191 .
it follows that @xmath192 choose @xmath193 such that the raleigh quotient ( [ eq : rl2 ] ) reaches its minimum at @xmath194 , i.e. , @xmath195 we have @xmath196 in the last step , we use the following argument .
@xmath197 since @xmath198 is the maximum over all @xmath199 , we get @xmath200 . for item 3 , we have @xmath201 [ thm5 ] for @xmath202 , the vertex probability distribution @xmath112 of the @xmath101-lazy random walk on a eulerian directed graph @xmath24 at time @xmath52 converges to the stationary distribution @xmath96 in probability .
in particular , we have @xmath203 here @xmath99 is the initial vertex probability distribution .
the proof is omitted since it is very similar to the proof of theorem [ thm1 ] .
notice that when @xmath204 , we have @xmath205 by lemma [ lemma1 ] .
we have the @xmath101-lazy random converges to the stationary distribution exponentially fast . for two vertex subsets @xmath117 and @xmath126 of @xmath26 ,
let @xmath133 be the number of directed edges from @xmath117 to @xmath126 , i.e. , @xmath206 .
we have the following theorem on the edge expansions in eulerian directed graphs .
[ thm6 ] if @xmath117 and @xmath126 are two subsets of the vertex set @xmath28 of a eulerian directed graph @xmath24 , then we have @xmath207 * proof : * let @xmath208 be the indicator variable of @xmath117 , i.e. , @xmath209 if @xmath210 and 0 otherwise .
we define @xmath211 similarly .
assume @xmath212 and @xmath213 , where @xmath214 and are unit vectors .
since @xmath176 is a unit vector , we have @xmath215 and @xmath216 thus @xmath217 similarly , we get @xmath218 it follows that @xmath219 the proof of this theorem is completed .
@xmath220 if we use @xmath221 instead of @xmath198 , then we get a weaker theorem on the edge expansions .
the proof will be omitted since it is very similar to the proof of theorem [ thm6 ] .
[ thm7 ] let @xmath24 be a eulerian directed graph .
if @xmath117 and @xmath126 are two subsets of @xmath26 , then we have @xmath222 for @xmath223 , let @xmath224 .
we have the following upper bound on @xmath127 . [ thm8 ]
suppose @xmath24 is a connected eulerian directed graph .
for @xmath225 and @xmath226 , we have @xmath227 in particular , for @xmath228 , the diameter of @xmath24 satisfies @xmath229 where @xmath230 .
* remark : * from lemma [ lemma1 ] , we have @xmath231 we can choose @xmath101 to minimize @xmath182 . if @xmath232 , then we choose @xmath233 and get @xmath234 ; if @xmath235 , then we choose @xmath236 and get @xmath237 .
combining two cases , we have @xmath238 it is easy to check @xmath239 here the inequality is strict if @xmath240 .
we have @xmath241 theorem [ thm8 ] is stronger than theorem [ thm4 ] in general . *
proof : * similar to the proof of theorem [ thm6 ] , let @xmath242 and @xmath211 be the indicator functions of @xmath117 and @xmath126 , respectively .
we have @xmath243 where @xmath244 and are unit vectors and @xmath245 are given by equations ( [ eq : a0])-([eq : b1 ] ) .
let @xmath246 .
we have @xmath247 thus there is a directed path starting from some vertex in @xmath117 and ending at some vertex in @xmath126 , that is @xmath248 . for the diameter result , we choose @xmath249 and @xmath250 .
note that @xmath251 , @xmath252 , @xmath253 , and @xmath254 .
the result follows .
let @xmath12 be an @xmath255-uniform hypergraph with the vertex set @xmath27 ( or @xmath28 for short ) and the edge set @xmath31 .
we assume @xmath256 and @xmath257 . for a vertex subset @xmath148 such that @xmath258 , the _ neighborhood _ @xmath259 is @xmath260 .
let the _ degree @xmath261 _ of @xmath148 in @xmath12 be the number of edges containing @xmath148 , i.e , @xmath262 . for @xmath14 ,
an @xmath15-walk of length @xmath52 is a sequence of vertices @xmath263 together with a sequence of edges @xmath264 such that @xmath265 for @xmath266 . here
are some examples of @xmath15-walks as shown in figure [ 2walk ] . a 1-walk in a 3-graph a 2-walk in a 3-graph a 2-walk in a 4-graph for each @xmath17 in @xmath267 , the _
@xmath17-th stop _
@xmath268 of the @xmath15-walk is the ordered @xmath15-tuple @xmath269 .
the initial stop is @xmath270 , and the terminal stop is @xmath271 .
an @xmath15-walk is called an @xmath15-path if every stop ( as an ordered @xmath15-tuple ) is different from each other .
if @xmath272 , then an @xmath15-walk is _
closed_. an @xmath15-cycle is a closed @xmath15-path . for @xmath14 and @xmath273 ,
the @xmath15-distance @xmath274 is the minimum integer @xmath52 such that there exists an @xmath15-path of length @xmath52 starting from @xmath60 and ending at @xmath61 .
a hypergraph @xmath12 is _ @xmath15-connected _ if @xmath274 is finite for every pair @xmath136 .
if @xmath12 is @xmath15-connected , then the _ @xmath15-diameter _ of @xmath12 is the maximum value of @xmath274 for @xmath275 .
a random @xmath15-walk with initial stop @xmath270 is an @xmath15-walk generated as follows .
let @xmath270 be the sequence of visited vertices at initial step .
at each step , let @xmath148 be the set of last @xmath15 vertices in the sequence of visited vertices .
a random @xmath276-set @xmath6 is chosen from @xmath259 uniformly ; the vertex in @xmath6 is added into the sequence one by one in an arbitrary order . for @xmath168 ,
an @xmath101-lazy random @xmath15-walk is a modified random @xmath15-walk such that with probability @xmath101 , one can stay at the current stop ; with probability @xmath277 , append @xmath278 vertices to the sequence as selected in a random @xmath15-walk . for @xmath279 ,
let @xmath280 $ ] be the @xmath15-set consisting of the coordinates of @xmath60 . for @xmath281
, we define a weighted undirected graph @xmath282 over the vertex set @xmath34 as follows .
let the weight @xmath283 be @xmath284 \sqcup [ y ] \subseteq f\}|$ ] . here
@xmath280\sqcup [ y]$ ] is the disjoint union of @xmath280 $ ] and @xmath285 $ ] . in particular , if @xmath280\cap [ y]\not=\emptyset$ ] , then @xmath76 . for @xmath279 ,
the degree of @xmath60 in @xmath282 , denoted by @xmath286 , is given by @xmath287}\binom{r - s}{s}s!.\ ] ] here @xmath288}$ ] means the degree of the set @xmath280 $ ] in the hypergraph @xmath12 .
when we restrict an @xmath15-walk on @xmath12 to its stops , we get a walk on @xmath282 .
this restriction keeps the length of the walk .
therefore , the @xmath15-distance @xmath274 in @xmath12 is simply the graph distance between @xmath60 and @xmath61 in @xmath282 ; the @xmath15-diameter of @xmath12 is simply the diameter of the graph @xmath282 .
a random @xmath15-walk on @xmath12 is essentially a random walk on @xmath282 .
it can be constructed from a random walk on @xmath282 by inserting additional random @xmath289 vertices @xmath290 between two consecutive stops @xmath268 and @xmath291 at time @xmath17 , where @xmath290 is chosen uniformly from @xmath292\cup [ x_{i+1}])$ ] and inserted between @xmath268 and @xmath291 in an arbitrary order .
therefore , we define the _ @xmath15-th laplacian _ @xmath19 of @xmath12 to be the laplacian of the weighted undirected graph @xmath282 .
the eigenvalues of @xmath19 are listed as @xmath293 in the non - decreasing order .
let @xmath294 and @xmath295 .
for some hypergraphs , the numerical values of @xmath296 and @xmath297 are shown in table [ tab:1 ] at the end of this section .
for @xmath298 , we define a directed graph @xmath299 over the vertex set @xmath34 as follows .
for @xmath275 such that @xmath300 and @xmath301 , let @xmath136 be a directed edge if @xmath302 for @xmath303 and @xmath280\cup [ y]$ ] is an edge of @xmath12 . for @xmath304 , the out - degree @xmath140 in @xmath299 and the in - degree @xmath144 in @xmath299 satisfy @xmath305}(r - s)!=d^-_x.\ ] ]
thus @xmath299 is a eulerian directed graph .
we write @xmath306 for both @xmath140 and @xmath144 .
now @xmath299 is strongly connected if and only if it is weakly connected .
note that an @xmath15-walk on @xmath12 can be naturally viewed as a walk on @xmath299 and vice versa .
thus the @xmath15-distance @xmath274 in @xmath12 is exactly the directed distance from @xmath60 to @xmath61 in @xmath282 ; the @xmath15-diameter of @xmath12 is the diameter of @xmath299 .
a random @xmath15-walk on @xmath12 is one - to - one corresponding to a random walk on @xmath299 . for @xmath307
, we define the @xmath15-th laplacian @xmath19 as the laplacian of the eulerian directed graph @xmath299 ( see section 2 ) .
the eigenvalues of @xmath19 are listed as @xmath293 in the non - decreasing order .
let @xmath294 and @xmath295 .
for some hypergraphs , the numerical values of @xmath296 and @xmath297 are shown in table [ tab:1 ] at the end of this section .
let @xmath308 be the complete @xmath255-uniform hypergraph on @xmath1 vertices .
here we compute the values of @xmath296 and @xmath297 for some @xmath308 ( see table [ tab:1 ] ) .
.the values of @xmath309 and @xmath297 of some complete hypergraphs @xmath310 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] * remark : * from the table above , we observe @xmath311 for some complete hypergraphs .
in fact , this is true for any complete hypergraph @xmath310 .
we point out the following fact without proofs . for an @xmath255-uniform hypergraph @xmath12 and an integer @xmath15 such that @xmath312 , @xmath313 holds if and only if @xmath20 and @xmath12 is a @xmath314-design .
in this section , we prove some properties of the laplacians for hypergraphs .
[ lemma2 ] for @xmath281 , we have the following properties . 1 .
the @xmath15-th laplacian has @xmath315 eigenvalues and all of them are in @xmath316 $ ] .
2 . the number of @xmath48 eigenvalues is the number of connected components in @xmath282 .
the laplacian @xmath317 has an eigenvalue @xmath314 if and only if @xmath318 and @xmath282 has a bipartite component .
* proof : * items 1 and 2 follow from the facts of the laplacian of @xmath282 .
if @xmath19 has an eigenvalue 2 , then @xmath282 has a bipartite component @xmath6 .
we want to show @xmath318 .
suppose @xmath319 .
let @xmath320 be an edge in @xmath6 .
for @xmath321 and @xmath322 , let @xmath323 and @xmath324 .
observe @xmath325 form an odd cycle in @xmath282 .
@xmath220 the following lemma compares @xmath326 and @xmath327 for different @xmath15 .
[ lemma3 ] suppose that @xmath12 is an @xmath255-uniform hypergraph .
we have @xmath328 * remark : * we do not know whether similar inequalities hold for @xmath329 . + * proof : * let @xmath330 be the diagonal matrix of degrees in @xmath282 and @xmath331 be the rayleigh quotient of @xmath19 .
it suffices to show @xmath332 for @xmath333 .
recall that @xmath296 can be defined via the rayleigh quotient , see equation ( [ eq:3 ] ) .
pick a function @xmath334 such that @xmath335 and @xmath336 .
we define @xmath337 as follows @xmath338 where @xmath339 is a @xmath340-tuple consisting of the first @xmath340 coordinates of @xmath60 with the same order in @xmath60 .
applying equation [ eq:4 ] , we get @xmath341 } \binom{r - s}{s}s!.\ ] ] we have @xmath342 } & = & \sum_{x } \sum_{f : [ x ] \subseteq f } g(x)\\ & = & \sum_{x ' } \sum_{f : [ x ' ] \subseteq f } ( r - s+1 ) f(x')\\ & = & \sum_{x ' } d_{[x']}(r - s+1 ) f(x')\\ & = & \frac{r - s+1}{\binom{r - s+1}{s-1}(s-1 ) ! } \sum_{x ' } f(x')d_{x'}^{(s-1)}=0 .
\end{aligned}\ ] ] here the second last equality follows from equation [ eq:4 ] and the last one follows from the choice of @xmath39 .
therefore , @xmath343 thus @xmath344 .
similarly , we have @xmath345 putting them together , we obtain @xmath346 by the similar counting method , we have @xmath347 \sqcup [ y ] \subseteq f}(g(x)-g(y))^2\\ & = & \sum_{x ' \sim y ' } \sum_{f : [ x ' ] \sqcup [ y ' ] \subseteq f}(r - s+1)(r - s+2 ) ( f(x')-f(y'))^2\\ & = & ( r - s+1)(r - s+2)\sum_{x ' \sim y'}(f(x')-f(y'))^2w(x',y').\end{aligned}\ ] ] thus , @xmath348 by the choice of @xmath39 . as @xmath296 is the infimum over all @xmath194 , we get @xmath332 .
the inequality ( [ lambdamax ] ) can be proved in a similarly way . since @xmath327 is the supremum of the raleigh quotient ,
the direction of inequalities are reversed .
@xmath220 [ lemma4 ] for @xmath298 , we have the following facts . 1 .
the @xmath15-th laplacian has @xmath315 eigenvalues and all of them are in @xmath316 $ ] .
the number of @xmath48 eigenvalues is the number of strongly connected components in @xmath299 .
if @xmath314 is an eigenvalue of @xmath317 , then one of the @xmath15-connected components of @xmath12 is bipartite .
the proof is trivial and will be omitted .
we show some applications of laplacians @xmath19 of hypergraphs in this section . for @xmath349 and @xmath281 , after restricting an @xmath101-lazy random @xmath15-walk on a hypergraph @xmath12 to its stops ( see section [ section3 ] ) , we get an @xmath101-lazy random walk on the corresponding weighted graph @xmath282 . let @xmath350 for any @xmath279 , where @xmath62 is the degree of @xmath60 in @xmath282 and @xmath351 is the volume of @xmath282 . applying theorem [ thm1 ] , we have the following theorem . for @xmath281 ,
suppose that @xmath12 is an @xmath15-connected @xmath255-uniform hypergraph @xmath12 and @xmath296 ( and @xmath297 ) is the first non - trivial ( and the last ) eigenvalue of the @xmath15-th laplacian of @xmath12
. for @xmath352 , the joint distribution @xmath112 at the @xmath52-th stop of the @xmath101-lazy random walk at time @xmath52 converges to the stationary distribution @xmath96 in probability .
in particular , we have @xmath353 where @xmath354 and @xmath298 , when restricting an @xmath101-lazy random @xmath15-walk on a hypergraph @xmath12 to its stops ( see section 2 ) , we get an @xmath101-lazy random walk on the corresponding directed graph @xmath299 .
let @xmath350 for any @xmath279 , where @xmath62 is the degree of @xmath60 in @xmath299 and @xmath351 is the volume of @xmath299 .
applying theorem [ thm5 ] , we have the following theorem . for @xmath298 ,
suppose that @xmath12 is an @xmath15-connected @xmath255-uniform hypergraph and @xmath296 is the first non - trivial eigenvalue of the @xmath15-th laplacian of @xmath12 .
for @xmath204 , the joint distribution @xmath112 at the @xmath52-th stop of the @xmath101-lazy random walk at time @xmath52 converges to the stationary distribution @xmath96 in probability .
in particular , we have @xmath355 where @xmath356 , and @xmath99 is the probability distribution at the initial stop .
* remark : * the reason why we require @xmath204 in the case @xmath357 is @xmath358 for @xmath359 .
let @xmath12 be an @xmath255-uniform hypergraph .
for @xmath14 and @xmath360 , the @xmath15-distance @xmath274 is the minimum integer @xmath52 such that there is an @xmath15-path of length @xmath52 starting at @xmath60 and ending at @xmath61 . for @xmath361 ,
let @xmath362 .
if @xmath12 is @xmath15-connected , then the @xmath15-diameter @xmath363 satisfies @xmath364 for @xmath312 , the @xmath15-distances in @xmath12 ( and the @xmath15-diameter of @xmath12 ) are simply the graph distances in @xmath282 ( and the diameter of @xmath282 ) , respectively . applying theorem
[ thm2 ] and corollary [ cor1 ] , we have the following theorems .
suppose @xmath12 is an @xmath255-uniform hypergraph .
for integer @xmath15 such that @xmath365 , let @xmath296 ( and @xmath297 ) be the first non - trivial ( and the last ) eigenvalue of the @xmath15-th laplacian of @xmath12 .
suppose @xmath366 .
for @xmath367 , if @xmath368 , then we have @xmath369 here @xmath370 are volumes in @xmath282 .
* remark : * we know @xmath371 if and only if @xmath12 is @xmath15-connected .
the condition @xmath372 holds unless @xmath20 and every pair of vertices is covered by edges evenly ( i.e. , @xmath12 is a @xmath314-design ) .
suppose @xmath12 is an @xmath255-uniform hypergraph .
for integer @xmath15 such that @xmath365 , let @xmath296 ( and @xmath297 ) be the first non - trivial ( and the last ) eigenvalue of the @xmath15-th laplacian of @xmath12 . if @xmath366 , then the @xmath15-diameter of an @xmath255-uniform hypergraph @xmath12 satisfies @xmath373 here @xmath374 and @xmath375 is the minimum degree in @xmath282 . when @xmath376 , the @xmath15-distances in @xmath12 ( and the @xmath15-diameter of @xmath12 ) is the directed distance in @xmath299 ( and the diameter of @xmath299 ) , respectively
. applying theorem [ thm8 ] and its remark , we have the following theorems .
let @xmath12 be an @xmath255-uniform hypergraph .
for @xmath377 and @xmath378 , if @xmath12 is @xmath15-connected , then we have @xmath379 here @xmath296 is the first non - trivial eigenvalue of the laplacian of @xmath299 , and @xmath370 are volumes in @xmath299 . for @xmath377 ,
suppose that an @xmath255-uniform hypergraph @xmath12 is @xmath15-connected .
let @xmath296 be the smallest nonzero eigenvalue of the laplacian of @xmath299 .
the @xmath15-diameter of @xmath12 satisfies @xmath380 here @xmath381 and @xmath375 is the minimum degree in @xmath299 . in this subsection
, we prove some results on the edge expansions in hypergraphs .
let @xmath12 be an @xmath255-uniform hypergraph . for @xmath382 , we recall that the volume of @xmath148 satisfies @xmath383 here @xmath62 is the degree of the set @xmath60 in @xmath12 . in particular , we have @xmath384 the _ density _ @xmath385 of @xmath148 is @xmath386 .
let @xmath387 be the complement set of @xmath148 in @xmath388 .
we have @xmath389 for @xmath390 , @xmath382 , and @xmath391 , let @xmath392 note that @xmath393 counts the number of edges contains @xmath394 for some @xmath395 and @xmath396 . particularly , we have @xmath397 [ thm11 ] for @xmath398 , @xmath382 , and @xmath399 , let @xmath400 .
we have @xmath401 * proof : * let @xmath282 be the weighed undirected graph defined in section [ section3 ] .
define @xmath402 and @xmath403 ( sets of ordered @xmath15-tuples ) as follows @xmath404\in s\};\ ] ] @xmath405 \in t \}.\ ] ] let @xmath406 ( or @xmath407 ) be the complement set of @xmath402 ( or @xmath403 ) in @xmath34 , respectively .
we make a convention that @xmath408 denotes volumes in @xmath282 while @xmath33 denotes volumes @xmath12 .
we have @xmath409 let @xmath410 be the number of edges between @xmath402 and @xmath403 in @xmath282 .
we get @xmath411 applying theorem [ thm3 ] to the sets @xmath402 and @xmath403 in @xmath282 , we obtain @xmath412 combining equations ( [ eq : start]-[eq : end ] ) and the inequality above , we obtain inequality [ eq : exp1 ] .
@xmath220 now we consider the case that @xmath329 .
due to the fact that @xmath413 , we have to use the weaker expansion theorem [ thm7 ] .
note that @xmath414 we get the following theorem .
[ t : exp2 ] for @xmath415 , @xmath382 , and @xmath399 , let @xmath400 . if @xmath416 for any @xmath395 and @xmath396 , then we have @xmath417 * proof : * recall that @xmath299 is the directed graph defined in section [ section3 ]
. let @xmath404\in s\};\ ] ] @xmath418\in t\}.\ ] ] we also denote @xmath406 ( or @xmath407 ) be the complement set of @xmath402 ( or @xmath403 ) in @xmath34 , respectively .
we use the convention that @xmath419 denotes the volumes in @xmath299 while @xmath33 denotes the volumes in the hypergraph @xmath12 .
we have @xmath420 let @xmath421 ( or @xmath422 ) be the number of directed edges from @xmath402 to @xmath403 ( or from @xmath403 to @xmath402 ) in @xmath299 , respectively .
we get @xmath423 from the condition @xmath424 for each @xmath395 and each @xmath396 , we observe @xmath425 applying theorem [ thm7 ] to the sets @xmath402 and @xmath403 in @xmath299 , we obtain @xmath426 combining equations ( [ eq : start2]-[eq : end2 ] ) and the inequality above , we get inequality [ eq : exp2 ] . @xmath220 + nevertheless
, we have the following strong edge expansion theorem for @xmath427 . for @xmath428 ,
let @xmath429 be the set of edges of the form @xmath430 for some @xmath431 and @xmath396 .
namely , @xmath432 observe that @xmath433 [ t : exp3 ] for @xmath307 and @xmath428 , let @xmath434 .
we have @xmath435 * proof : * let @xmath404\in s\};\ ] ] @xmath436 let @xmath406 ( or @xmath407 ) be the complement set of @xmath402 ( or @xmath403 respectively ) in @xmath34 .
we use the convention that @xmath419 denotes the volumes in @xmath299 while @xmath33 denotes the volumes in the hypergraph @xmath12 .
we have @xmath437 let @xmath421 ( or @xmath422 ) be the number of directed edges from @xmath402 to @xmath403 ( or from @xmath403 to @xmath402 ) in @xmath299 , respectively .
we get @xmath438 applying theorem [ thm7 ] to the sets @xmath402 and @xmath403 on @xmath299 , we obtain @xmath426 combining equations ( [ eq : start3]-[eq : end3 ] ) and the inequality above , we get inequality [ eq : exp3 ] . @xmath220
in this paper , we introduced a set of laplacians for @xmath255-uniform hypergraphs . for @xmath14 ,
the @xmath15-laplacian @xmath19 is derived from the random @xmath15-walks on hypergraphs . for @xmath365 ,
the @xmath15-th laplacian @xmath19 is defined to be the laplacian of the corresponding weighted graph @xmath282 .
the first laplacian @xmath21 is exactly the laplacian introduced by rodrguez @xcite . for @xmath439 ,
the @xmath19 is defined to be the laplacian of the corresponding eulerian directed graph @xmath299 . at first glimpse
, @xmath440 might be a good parameter .
however , it is not hard to show that @xmath358 always holds , which makes theorem [ thm6 ] useless for hypergraphs .
we can use weaker theorem [ thm7 ] for hypergraphs .
our work is based on ( with some improvements ) chung s recent work @xcite on directed graphs .
let us recall chung s definition of laplacians @xcite for regular hypergraphs .
an @xmath255-uniform hypergraph @xmath12 is @xmath441-regular if @xmath442 for every @xmath443 .
let @xmath0 be a graph on the vertex set @xmath444 .
for @xmath445 , let @xmath446 be an edge if @xmath447 and @xmath448 such that @xmath449 is an edge of @xmath12 .
let @xmath5 be the adjacency matrix of @xmath0 , @xmath6 be the diagonal matrix of degrees in @xmath0 , and @xmath450 be the adjacency matrix of the complete graph on the edge set @xmath444 .
chung @xcite defined the laplacian @xmath2 such that @xmath451 this definition comes from the homology theory of hypergraphs .
firstly , @xmath2 is not normalized in chung s definition , i.e. , the eigenvalues are not in the interval @xmath316 $ ] . secondly , the add - on term @xmath452
is not related to the structures of @xmath12 .
if we ignore the add - on term and normalize the matrix , we essentially get the laplacian of the graph @xmath0 .
note @xmath0 is disconnected , then @xmath453 and it is not interesting .
thus chung added the additional term .
the graph @xmath0 is actually very closed to our eulerian directed graph @xmath454 .
let @xmath455 be the adjacency matrix of @xmath454 .
in fact we have @xmath456 , where @xmath457 is a rotation which maps @xmath458 to @xmath459 .
since @xmath460 , @xmath457 and @xmath6 commute , we have @xmath461 here we use the fact @xmath462 .
this identity means that the singular values of @xmath463 is precisely equal to @xmath37 minus the laplacian eigenvalues of the graph @xmath0 .
our definitions of laplacians @xmath19 are clearly related to the quasi - randomness of hypergraphs .
we are very interested in this direction .
many concepts such as the @xmath15-walk , the @xmath15-path , the @xmath15-distance , and the @xmath15-diameter , have their independent interest .
d. aldous and j. fill , _
reversible markov chains and random walks on graphs _ , in preparation .
n. alon , eigenvalues and expanders , _ combinatorica _ * 6*(1986 ) , 86 - 96 . f. chung , diameters and eigenvalues , _ j. of the amer .
* 2 * 1989 , 187 - 196 .
f. chung , the laplacian of a hypergraph , in j. friedman ( ed . ) , _ expanding graphs ( dimacs series ) _ * 1993 * , 21 - 36 .
f. chung , v. faber , and t.a .
manteuffel , an upper bound on the diameter of a graph from eigenvalues associated with its laplacian , _ siam .
_ 1994 , 443 - 457 .
f. chung , _ spectral graph theory _ , ams publications , 1997 .
f. chung , laplacians and the cheeger inequality for directed graphs , _ annals of comb .
_ , * 9 * 2005 , 1 - 19 .
f. chung , the diameter and laplacian eigenvalues of directed graphs , _ electronic journal of combinatorics _
, * 13 * 2006 , r#4 .
rodrguez , laplacian eigenvalues and partition problems in hypergraphs , _ applied mathematics letters _ * 2009 * -22 , 916 - 921 . | despite of the extreme success of the spectral graph theory , there are relatively few papers applying spectral analysis to hypergraphs .
chung first introduced laplacians for regular hypergraphs and showed some useful applications .
other researchers treated hypergraphs as weighted graphs and then studied the laplacians of the corresponding weighted graphs . in this paper , we aim to unify these very different versions of laplacians for hypergraphs .
we introduce a set of laplacians for hypergraphs through studying high - ordered random walks on hypergraphs .
we prove the eigenvalues of these laplacians can effectively control the mixing rate of high - ordered random walks , the generalized distances / diameters , and the edge expansions . |
let @xmath0 be a simple , simply connected algebraic group defined over an algebraically closed field @xmath1 of characteristic zero .
fix a maximal torus @xmath2 .
let @xmath3 be the weyl group and @xmath4 be the group of characters .
the main results of this article are as follows : * ( theorem 1 ) .
* _ let @xmath0 be a simple simply connected algebraic group over an algebraically closed field @xmath1 of characteristic zero .
let @xmath5 let @xmath6 be a dominant weight such that for each @xmath7 either @xmath8 is dominant or @xmath9 is singular .
then @xmath10 for all @xmath11 _ let @xmath12 .
for given @xmath13 , in proposition 2 we give conditions on @xmath14 s so that @xmath6 satisfies the hypothesis of theorem 1 .
for each simple root @xmath15 , we define the _ coxeter number of @xmath15 _ , denoted by @xmath16 , as the number @xmath17 , where @xmath18 .
we prove in theorem 9 that @xmath19 , and for a shorter simple root @xmath20 , where @xmath21 is the coxeter numberof @xmath0 . as a corollary of theorem 1 and proposition 2 ,
we get the following vanishing theorem : @xmath22 let rank@xmath23 and @xmath24 in theorem 12 we give strictly dominant weight @xmath6 such that @xmath25 we also derive the following corollaries : 1 .
we show that the bott vanishing property fails to hold for generalised flag varieties
@xmath26 is not a toric variety .
@xmath26 can not be degenerated to a smooth toric variety in such a manner that ample cone degenerates to ample cone .
we recall some standard facts about algebraic groups .
a standard reference for this material is @xcite . for a finite dimensional @xmath2-module @xmath27 and @xmath28 , set @xmath29 the finitely many @xmath30 such that @xmath31 are called weights of the @xmath2-module @xmath27 and @xmath32 the weight spaces .
then @xmath27 decomposes into a direct sum of weight spaces , i.e. @xmath33 , where @xmath34 are weights .
let @xmath35 and @xmath36 be the lie algebras of @xmath0 and @xmath2 respectively .
let @xmath2 act on @xmath35 via the adjoint representation .
the non - zero weights of this representation are called roots and we denote the set of all roots by @xmath37 .
the weight space corresponding to the zero weight is @xmath36 .
we identify @xmath38 canonically as a subset of @xmath39 .
let @xmath40 denote the set of all simple roots , that is a subset of @xmath37 such that ( i ) @xmath40 is a basis of @xmath41 , ( ii ) for each root @xmath42 there exist integers @xmath43 of like sign such that @xmath44 .
the number of elements in @xmath40 is called the rank of the group @xmath0 .
the set of all roots @xmath42 for which the @xmath14 are not negative is denoted by @xmath45 and an element of @xmath45 is called a positive root .
@xmath46 is the set of negative roots .
let @xmath47 and @xmath48 .
let @xmath49 be the borel subgroup of @xmath0 such that the lie algebra @xmath50 of @xmath49is @xmath51 .
there is a natural faithful action of @xmath52 on @xmath53 and there exists a @xmath52-invariant bilinear form @xmath54 on @xmath55 . for @xmath56 , we denote @xmath57 by @xmath58 for each @xmath59 , we have reflection @xmath60 defined by @xmath61 , @xmath62 .
the set @xmath63 generates @xmath52 . for each @xmath64 there exists a weight @xmath65 such that @xmath66 ( here @xmath67 is the krnecker delta . ) these @xmath68 are called fundamental weights .
the action of @xmath52 permutes the roots .
for @xmath69 we define the length of @xmath70 , denoted by @xmath71 to be the number of positive roots moved by @xmath70 to negativeroots .
following jantzen @xcite , we denote the sheaf associated to a @xmath49-module @xmath27 by @xmath72 ( i.e. the sheaf of sections of the associated vector bundle on @xmath26 ) . when @xmath73 , we denote @xmath74 by @xmath75 . the tangent sheaf on @xmath26
( @xcite , p. 229 ) is @xmath76 .
the @xmath2 weights on @xmath77 are @xmath78 .
we say a weight @xmath79 is dominant ( strictly dominant ) if @xmath80 we denote the set of dominant weight by @xmath81 . for @xmath82
we have the ` dot ' action given by @xmath83 where @xmath84 is half sum of positive roots .
now we recall the borel weil bott theorem .
let @xmath85 . if @xmath86 then @xmath87 only when @xmath88 , and @xmath89 is the set of all simple @xmath0-modules . if there is a @xmath69 such that @xmath90 ( such a @xmath70 will be unique ) then @xmath91 only when @xmath92 and @xmath93 and moreover if there is no such @xmath70 ( equivalently if @xmath94 is singular , i.e. @xmath95 for some root @xmath59 ) then @xmath96 .
recall that in @xcite , bott proved the following vanishing theorem for the complex projective space : @xmath97 we say that a smooth projective variety @xmath98 satisfies the bott vanishing property if the twisted holomorphic forms or dolbeault cohomology @xmath99 @xmath100 for all ample line bundle @xmath101 where @xmath102 stands for sheaf associated to the vector bundle @xmath103 and @xmath104 we use the partial order ` @xmath105 ' in @xmath38 , for @xmath106 , we say @xmath107 if @xmath108 sum of some simple roots . for a simple algebraic group ,
the coxeter number is defined as @xmath109 . for the ordering of the simple roots we follow the following diagrams @xcite . in the following diagram
the arrows are pointing towards shorter simple root .
if there is no arrow it means that the adjacent roots are of the same length .
now we introduce some new notations . for any subset @xmath110 of @xmath38 and @xmath111 , @xmath112 and @xmath113
s are @xmath114 . for any finite set @xmath110 of @xmath38
, we define an @xmath115-matrix associated to @xmath110 as follows : we represent @xmath116 as tuples @xmath117 , and write the elements of @xmath110 one below the other to get a @xmath118 matrix .
we call the matrix corresponding to @xmath119 the positive roots matrix .
for @xmath120 we define the coxeter number of @xmath121
in this section , we study conditions on dominant weights @xmath6 so that @xmath122 for all @xmath123 and @xmath13 .
let @xmath5 let @xmath6 be a dominant weight such that for each @xmath124 either @xmath8 is dominant or @xmath125 is singular .
then @xmath126 for all @xmath11 the proof follows from the following observations : 1 . by the hypothesis and by the borel
bott theorem , for any weight @xmath127 of @xmath128 , we have @xmath129 for all @xmath130 .
any lowest weight @xmath131 of @xmath132 will give a one - dimensional @xmath49-submodule @xmath133 .
let @xmath134 be @xmath135 .
then we will have a short exact sequence of @xmath49-modules ; @xmath136 3 .
the above short exact sequence induces the following long exact sequence : @xmath137 \hskip -.55 cm & \rightarrow h^1(g / b , \l({\lambda_1 } ) ) \rightarrow h^{p,1}(g / b , \l({\lambda } ) ) \rightarrow h^1(g / b,\l(v_1))\\[.2pc ] \hskip -.55 cm & \rightarrow h^2(g / b , \l({\lambda_1 } ) ) \rightarrow h^{p,2}(g / b , \l({\lambda } ) ) \rightarrow \cdots \cdots \cdots\end{aligned}\ ] ] 4 . by @xmath138 and @xmath139
, we have @xmath140 for all @xmath141 .
now we repeat steps @xmath142 and @xmath139 with the @xmath49-module @xmath134 .
let @xmath143 be the lowest weight of @xmath134 and @xmath144 . as before we get @xmath145 for all @xmath130 .
after a finite number of steps we end up with a two - dimensional @xmath49-module @xmath146 such that @xmath147 for all @xmath141 .
this proves the theorem.@xmath148 we now make explicit computations to give more delicate bound on the @xmath6 for vanishing results .
for this we use the classification of simple algebraic groups .
@xmath149 1 .
let @xmath0 be a simply connected algebraic group of type @xmath150 .
let @xmath151 then @xmath152 if @xmath153 n_k \geq n & \hbox{when } \hskip .3 cm n \leq p \leq \displaystyle\frac{n^2-n+2}{2}\\ & \\[-.7pc ] n_k \geq \displaystyle\frac{n^2+n+2 - 2p}{2 } & \hbox{when } \hskip .3cm\displaystyle\frac{n^2-n+4}{2 } \leq p \leq \frac{n(n+1)}{2 } \end{cases}.\ ] ] 2 .
let @xmath0 be a simply connected algebraic group of type @xmath154 .
let @xmath155 then @xmath152 if @xmath156 n_k \geq p , k \not = n;n_n \geq 2n-1 & \hbox{when } \hskip .3 cm n\leq p\leq 2n-3 \\ & \\[-.7pc ] n_k \geq 2n-2 , k \not = n;n_n \geq 2n-1 & \hbox{when } \hskip .3 cm 2n-2\leq p \leq n^2 - 2n+3\\ & \\[-.7pc ] n_k \geq n^2 + 1-p , & \hbox{when } \hskip .3 cm n^2 - 2n+4 \leq
p\\ & \\[-.7pc ] ~~k \not= n;n_n \geq 2n-1 & \hskip 1 cm \leq n^2-n+1\\ & \\[-.7pc ] n_k \geq n^2 + 1-p , & \hbox{when } \hskip .3 cm n^2-n+2 \leq p \leq n^2\\ & \\[-.7pc ] ~~ k \not = n;n_n \geq 2n^2 + 1 - 2p \end{cases}.\ ] ] 3 .
let @xmath0 be a simply connected algebraic group of type @xmath157 .
let @xmath158 .
then @xmath152 if @xmath159
n_k \geq p+1 , k \not = n ; n_n \geq p & \hbox{when } \hskip .3 cm 3\leq p\leq n-1 \\[.2pc ] n_k \geq p+1 , k \not = n ; n_n \geq n & \hbox{when } \hskip .3 cm n\leq p\leq 2n-3 \\[.2pc ] n_k \geq 2n-1 , k \not = n ; n_n \geq n & \hbox{when } \hskip
.3 cm 2n-2\leq p \leq n^2 - 2n+3\\[.2pc ] n_k \geq 2n-3 , k \not= n;n_n \geq n & \hbox{when } \hskip .3 cm p = n^2 - 2n+4\\[.2pc ] n_k \geq n^2 + 1-p , & \hbox{when } \hskip .3 cm n^2 - 2n+5 \\[.2pc ] ~~k \not = n;n_n \geq n & \hskip 1 cm \leq p \leq n^2-n+1\\[.2pc ] n_k \geq n^2 + 1-p , & \hbox{when } \hskip .3 cm n^2-n+2 \leq p \leq n^2\\[.2pc ] ~~ k \not = n;n_n \geq n^2 + 1-p \end{cases}.\ ] ] 4 .
let @xmath0 be a simply connected algebraic group of type @xmath160 .
let @xmath151 then @xmath152 if @xmath161 n_k
\geq 2n-3 & \hbox{when } \hskip .3 cm 2n-3 \leq p \leq n^2 - 3n+4\\[.2pc ] n_k \geq n^2-n+1-p & \hbox{when } \hskip .3 cm n^2 - 3n+5 \leq p \leq n^2-n \end{cases}.\ ] ] 1 .
let @xmath0 be a simply connected algebraic group of type @xmath162 .
let @xmath151 then @xmath163 if @xmath164 n_k \geq 11 & \hbox{when } \hskip .3 cm 11 \leq p \leq 26\\[.2pc ] n_k \geq 37-p & \hbox{when } \hskip .3 cm 27 \leq p \leq 36 \end{cases}.\ ] ] 1 .
let @xmath0 be a simply connected algebraic group of type @xmath165 .
let @xmath151 then @xmath166 if @xmath167 n_k \geq 17 & \hbox{when }
\hskip .3 cm 17 \leq p \leq 47\\[.2pc ] n_k \geq 64-p & \hbox{when } \hskip .3 cm 48 \leq p \leq 63 \end{cases}.\ ] ] 1 .
let @xmath0 be a simply connected algebraic group of type @xmath168 .
let @xmath151 then @xmath166 if @xmath169 n_k \geq 29 & \hbox{when } \hskip .3 cm 29 \leq p \leq 92\\[.3pc ] n_k \geq 121-p & \hbox{when } \hskip .3 cm 93 \leq p \leq 120 \end{cases}.\ ] ] 1 .
let @xmath0 be a simply connected algebraic group of type @xmath170 .
let @xmath171 then @xmath166 if @xmath172 n_1,n_2 \geq p ; n_3,n_4 \geq p+3 & \hbox{when } \hskip .3 cm 5 \leq p \leq7\\[.3pc ] n_1,n_2 \geq 8 ; n_3,n_2\geq 11 = h_k-1 & \hbox{when } \hskip .3 cm 8 \leq p \leq 17\\[.3pc ] n_1,n_2 \geq 25-p ; & \hbox{when } \hskip .3 cm 18\leq p \leq 20\\[.3pc ] n_3,n_4 \geq 45 - 2p \\[.3pc ] n_k \geq 25-p & \hbox{when } \hskip .3 cm 21 \leq p \leq 24 \end{cases}.\ ] ] 1 .
let @xmath0 be a simply connected algebraic group of type @xmath173 .
let @xmath174 .
then @xmath175 if @xmath176 n\geq 2 , m\geq 4 & \hbox{when } \hskip .3 cm p=2,5 \\[.3pc ] n\geq 3 , m\geq 5 & \hbox{when } \hskip .3 cm p=3,4 \end{cases}.\ ] ] we consider the set of weights of @xmath177 . by our notation
this set is @xmath178 .
our aim is to make all @xmath179 either dominant or @xmath180 singular for suitable choice of @xmath6 . for
that we give condition on @xmath181 so that @xmath182 for all @xmath183 .
to avoid complications we include only simple singularities , i.e. we give conditions on @xmath184 so that either @xmath185 , or there exists a simple root @xmath15 such that @xmath186 if we know in each column of the ` positive roots matrix ' ,
what the positive entries , the negative entries , and the zeros are , and how many times they appear , then we can compute easily what the maximum positive value is which can occur in @xmath187 for all @xmath188 , where @xmath189 runs over @xmath190 .
suppose that this number is @xmath14 , then we demand that @xmath6 satisfy the inequalities @xmath191
. the following observation will be useful in writing down the conditions stated in the proposition . in the positive root matrix , suppose there are @xmath192 positive entries and @xmath193 negative entries and @xmath194 zeros in the @xmath195th column .
then the maximum among the positive entries in the @xmath195th column of the matrix corresponding to @xmath196 increases when @xmath197 increases from zero to @xmath192 ( how quickly it increases depends on what the positive entries are ( with repetition ) ) ; it will be non - decreasing when @xmath197 increases from @xmath192 to @xmath198 , and then it starts decreasing ( again , how quickly it decreases depends on how many negative entries there are ( with repetition ) ) .
we give the needed information about the positive roots matrix in the following table . [ cols="<,^,^,^,^,^,^,^,^ " , ] @xmath199
the author would like to thank prof . c s seshadri for his constant encouragement and v balaji for the discussions and comments throughout the preparation of this paper .
the referee s comments and suggestions ( in particular the proof of lemma 8) are gratefully acknowledged .
the author also would like to thank s senthamaraikannan and k v subramanian for carefully reading the preliminary draft of this paper . | in this paper we prove some vanishing theorems for the twisted dolbeault cohomology of the complete flag varieties associated to a simple , simply connected algebraic group .
[ theore]*theorem * [ theorr]*theorem * [ caa]case o l p
= msam10 at 10pt # 1artpath#1 |
in recent times a significant amount of research effort has been directed towards nonsmooth bifurcations because a large number of physical and engineering systems have been found to exhibit this class of bifurcation phenomena @xcite .
such physical systems are generally given by two or more sets of differential equations , and the system switches from one set to another when some switching conditions are satisfied .
it has been shown that such systems give rise to piecewise smooth maps , where the discrete - time state space is divided into two or more compartments with different functional forms of the map , separated by borderlines @xcite . a new type of bifurcation , called _
border collision bifurcation _ occurs when a fixed point collides with a borderline , resulting in an abrupt change in the jacobian matrix .
most past studies on border collision bifurcations have concentrated on piecewise smooth maps that are continuous across the borderlines @xcite .
however , it has also been found that some switching dynamical systems , for example , the colpitts oscillator @xcite , dc - dc converters @xcite , thyristor controlled reactor circuits @xcite , sigma - delta modulators @xcite and many other electronic circuits @xcite give rise to discontinuous maps .
discontinuity in the map arises if , in the @xmath0 dimensional poincar section there exists an @xmath1 dimensional manifold such that infinitesimally close points at the two sides of the manifold map to points that are far apart . in order to explain the bifurcation phenomena in such systems ,
it is necessary to have a bifurcation theory for discontinuous maps .
the bifurcation theory for 1d and 2d continuous piecewise smooth maps has been developed in @xcite . in another important line of work ,
feigin obtained some results on the existence of periodic orbits in @xmath0-dimensional piecewise smooth continuous maps , which were brought to the notice of the english speaking audience in @xcite .
the theory for discontinuous maps is in the preliminary stage of development , with some progress reported for one - dimensional maps in @xcite . in the present paper ,
we follow feigin s approach in case of discontinuous maps , and present new results that are applicable to the general @xmath0-dimensional case .
in particular , we present the conditions of existence of the period-1 and period-2 fixed points which form the basis of the observable bifurcation phenomena in such maps . for this purpose
, we use a piecewise linear normal form where the discontinuity is defined by two parameters .
finally , the application of these results is illustrated by deducing the bifurcation behavior of a two - dimensional discontinuous map , and validating the results by simulation .
we consider discontinuous maps of the form @xmath2 where the discrete - time phase space is divided into two regions @xmath3 and @xmath4 , and @xmath5 and @xmath6 represent the functional forms of the map in the two regions respectively .
since our points of interest are the occurrence of bifurcations with the traversal of a fixed point across the border , we can study the same by applying local linear approximation in neighborhood of the discontinuity : @xmath7 where @xmath8 and @xmath9 here @xmath10 is @xmath11th iteration of @xmath12 matrix @xmath13 , @xmath14 and @xmath15 are @xmath16 matrices , @xmath17 is @xmath12 matrix and @xmath18 are parameters . we can apply a suitable coordinate transformation to obtain the matrices in the normal form , where the matrices @xmath19 $ ] and @xmath20 $ ] are such that @xmath21 if @xmath22 , i.e. , except the first columns of the matrices all other columns are same . in the earlier work on piecewise smooth continuous maps @xcite this condition was required to ensure continuity of the map . in the present work , though we are considering discontinuous maps , we retain this property of the matrices @xmath23 and @xmath15 and move the discontinuity to the parameters @xmath24 and @xmath25 without loss of generality . in @xcite
the length of the discontinuity was considered to be constant . in this paper , in order to allow a variable length of the discontinuity , we have used two parameters @xmath24 and @xmath25 .
we have assumed @xmath26 to be same on both sides , which physically means that the two parameters have proportional effect on the two sides of the border .
let us consider two fixed points @xmath27 and @xmath28 of the sub - mappings @xmath29 and @xmath30 respectively .
that is , @xmath31 @xmath32 where @xmath33 and @xmath34 are the first components of @xmath35 and @xmath36 respectively .
we define @xmath37 and @xmath38 , where @xmath39 is the @xmath16 identity matrix . and @xmath40 . ]
assuming @xmath41 and @xmath42 to be invertible , we find @xmath43 , @xmath44 . now
the @xmath45th element of these matrices can be written as @xmath46 where @xmath47_k$ ] is the @xmath45th element of the column matrix @xmath48 and @xmath49_k$ ] is the @xmath45th element of the column matrix @xmath50 .
now , since in the matrices @xmath14 and @xmath15 , @xmath51 if @xmath22 , the cofactor of first column of @xmath52 and @xmath42 are same . hence @xmath53 where @xmath54 and @xmath55 , are the elements of the first rows of the matrices @xmath56 and @xmath57 respectively and @xmath58 are the elements of the column matrix * c*. using the above in ( [ gen1m1m2 ] ) we obtain @xmath59 we require @xmath60 for @xmath61 to exist and @xmath62 for @xmath63 to exist . with the help of above derivations we can now attempt a classification of the possible behavior at the border collision bifurcation .
equation ( [ m1m2 ] ) points to two possibilities : * case 1 : * @xmath64
. this will be true if both @xmath65 and @xmath66 are positive or if both are negative .
there are two sub - cases .
case 1a : if @xmath67 then only one fixed point will exist .
case 1b : if @xmath68 then either two fixed points will exist or no fixed point will exist .
* case 2 : * @xmath69
. this will be true if @xmath70 and @xmath71 , or if @xmath72 and @xmath73 . here
also there are two sub - cases .
case 2a : if @xmath67 then either two fixed points will exist or there will not be any fixed point .
case 2b : if @xmath68 then only one fixed point will exist .
table [ table1 ] and table [ table2 ] summarize these results for @xmath74 and @xmath75 respectively . from these tables
the bifurcation sequences can be inferred for a given set of parameters , as @xmath24 and @xmath25 are varied .
.[table1]the conditions of existence of fixed points when @xmath74 . [ cols="^,^,^,^,^,^,^ " , ]
suppose that a new two period solution originates at a border collision point .
let the two points of the period-2 orbit be @xmath76 and @xmath77 lying on the half planes @xmath78 and @xmath79 respectively .
now we derive the condition for the existence of the period-2 orbit . using the local map ( [ normal - nd ] ) , we can write @xmath80 i.e. , @xmath81 now subtracting ( [ m * ] ) from ( [ m * * ] ) we get @xmath82 where @xmath83 .
now as @xmath84 if @xmath22 , writing ( [ deltam ] ) in scalar form we obtain : @xmath85 for @xmath86 .
equation ( [ deltamk ] ) can be rewritten by adding and subtracting @xmath87 as @xmath88 similarly by adding and subtracting @xmath89 in equation ( [ deltamk ] ) we get @xmath90 the above two equations can be written in vector form as @xmath91_{1}\ : m_1^ { * * } + { \bf c } ( \mu_1-\mu_2)\label{delma1}\ ] ] @xmath92_{1}\ : m_1^ { * } + { \bf c}(\mu_1-\mu_2)\label{delma2}\ ] ] where @xmath93_{1}$ ] indicate the first column of @xmath94 $ ] . assuming @xmath95 to be invertible , from ( [ delma2 ] ) we obtain : @xmath96 where @xmath97_1 $ ] and @xmath98 . similarly assuming @xmath99 to be invertible , from ( [ delma1 ] ) we obtain @xmath100 where @xmath101_1 $ ] and @xmath102 . now writing ( 12 ) and ( 13 ) in scalar form @xmath103
@xmath104 since @xmath105 for @xmath22 and @xmath106 , we have @xmath107 and @xmath108 .
hence from ( 14 ) and ( 15 ) @xmath109 where @xmath110 . from ( 14 ) @xmath111 substituting ( 17 ) into ( 7 ) , we obtain ( in scalar form ) @xmath112 or equivalently using matrices : @xmath113 where @xmath114 assuming @xmath115 to be invertible , we can now solve @xmath116 . from ( 18 ) @xmath117 now @xmath118 where @xmath119 and @xmath120 is some constant .
hence from ( [ finalm * ] ) and ( [ detahat ] ) we obtain @xmath121 where @xmath122 writing ( [ finalm*2 ] ) in scalar form @xmath123 and substituting this into ( 16 ) we obtain @xmath124 hence in two parameter @xmath0-dimensional discontinuous map ( [ normal - nd ] ) , a period-2 orbit will occur only when the following two conditions are simultaneously satisfied : @xmath125 where @xmath126 and @xmath127 are respectively the first components of @xmath128 and @xmath129 , and @xmath130 the derivation of the conditions for the existence of orbits of periodicity greater than 2 in the @xmath0-dimensional discontinuous map is very cumbersome .
this is still an open problem .
note that ( [ 2fcon1 ] ) and ( [ 2fcon2 ] ) give the conditions of existence of period-2 orbit .
in addition , one has to consider the stability of the orbit in order to determine whether the period-2 orbit will actually be observable in the asymptotically stable behavior of the system .
the normal form for the two dimensional discontinuous map is @xmath131 where @xmath132 @xmath133 is the trace and @xmath134 is the determinant of the jacobian matrix @xmath14 of the system at a fixed point in @xmath135 and @xmath136 is the trace and @xmath137 is the determinant of the jacobian matrix @xmath15 of the system at a fixed point in @xmath138
. the normal form ( [ 2dmap ] ) follows from that derived for piecewise smooth continuous maps in @xcite , with discontinuity added through the parameters @xmath24 and @xmath25 . in case of the map ( [ 2dmap ] ) ,
the terms @xmath139 , @xmath66 , @xmath140 and @xmath141 are given by @xmath142 now if we choose @xmath143 and @xmath73 , i.e. , @xmath144 , application of the results of @xmath0-dimensional discontinuous map ( for @xmath145 ) will yield the following inferences .
if we keep @xmath25 constant and vary @xmath24 , then * if @xmath67 then only one period-1 fixed point will exist , * if @xmath68 then either two period-1 fixed points will exist or no period-1 fixed point will exist .
we now show , using the specific 2d map ( [ 2dmap ] ) that the above results are true .
if a fixed point @xmath146 exists , it is given by the solution of the system of equations @xmath147 now the fixed point will exist if @xmath148 , i.e. , @xmath149 this is true if @xmath150 the above conditions ( [ 2d1 ] ) and ( [ 2d2 ] ) are depicted graphically in fig . [
fig2label](a ) .
( a ) regions of existence of the @xmath151 fixed point as @xmath24 is varied through 0 .
the fixed point exists in the shaded region if @xmath152 , and in the unshaded region if @xmath153 .
( b ) regions of existence of the @xmath63 fixed point as @xmath25 is varied through 0 .
the fixed point exists in the shaded region if @xmath154 , and in the unshaded region if @xmath155 . ] in a similar manner it can be shown that the fixed point @xmath63 will exist if @xmath156 this is true if @xmath157 the above conditions ( [ 2d3 ] ) and ( [ 2d4 ] ) are depicted graphically in fig .
[ fig2label](b ) .
the conditions ( [ 2d1 ] ) , ( [ 2d2 ] ) , ( [ 2d3 ] ) , and ( [ 2d4 ] ) divide the @xmath158 parameter space into four regions @xmath159 and @xmath160 as shown in fig .
[ fig4label ] .
these regions correspond to the four rows in table [ table1 ] . in the region @xmath161 , since @xmath162 and @xmath163 , the period one fixed points @xmath61 and @xmath63 are not stable .
therefore , even though they exist , the trajectory diverges to infinity . as the bifurcation parameter and @xmath164 .
the other parameters @xmath165 and @xmath166 are in @xmath167 . ] as the bifurcation parameter and @xmath168 .
the other parameters @xmath165 and @xmath166 are in @xmath167 . ] in @xmath169 , suppose we keep @xmath25 fixed at a negative value , and vary @xmath24 .
table [ table1 ] shows that the period-1 fixed point @xmath151 will exist so long as @xmath152 , and the fixed point @xmath170 will exist for the whole range of @xmath24 .
now , if we choose the parameters such that @xmath171 and @xmath163 , then @xmath61 is stable and @xmath63 is unstable .
the representative bifurcation diagram for such a situation is shown in fig .
[ 2regmune ] , which shows the period-1 orbit for @xmath152 which disappears for @xmath155 .
if we take @xmath155 and vary @xmath24 , then we infer from table [ table1 ] that @xmath61 will exist ( and is stable ) for @xmath152 , but no fixed point will exist for @xmath153 .
this is what we see in the bifurcation diagram of fig .
[ 2regmu2pe ] . as the bifurcation parameter and @xmath164 .
the other parameter @xmath172 and @xmath166 are in @xmath173 . ] as the bifurcation parameter and @xmath168 .
the other parameters @xmath172 and @xmath166 are in @xmath173 . ] in the region @xmath174 , suppose we fix @xmath154 and vary @xmath24 .
table [ table1 ] shows that for @xmath152 , @xmath61 will exist but @xmath63 will not . for @xmath153
none of the fixed points will exist .
now , if the parameters are chosen such that @xmath171 , then a stable period-1 behavior will be observed so long as @xmath152 .
this is shown in fig .
[ 3regmune ] . on the other hand ,
if @xmath25 is chosen to be positive , for @xmath152 both the fixed points will exist , and for @xmath153 @xmath63 will exist but @xmath61 will not . if the parameters are such that @xmath175 , then @xmath63 will occur for the whole parameter range , while @xmath151 will occur only for @xmath152 .
this is shown in the simulated bifurcation diagram of fig .
[ 3regmu2pe ] . as the bifurcation parameter and @xmath164 .
the other parameters @xmath176 and @xmath166 are in @xmath177 . ] as the bifurcation parameter and @xmath168 .
the other parameters @xmath176 and @xmath166 are in @xmath178 . ]
similarly in @xmath160 , suppose we keep @xmath25 fixed at a negative value , and vary @xmath24 .
table [ table1 ] shows that the none of the fixed points will exist so long as @xmath152 , and for @xmath153 the fixed point @xmath61 will exist ( and will be unstable ) but @xmath170 will not exist .
the representative bifurcation diagram for such a situation is shown in fig .
[ 4regmune ] . when @xmath154 , the period-1 fixed point is not observed over the whole parameter range , but the existence of the unstable fixed point @xmath61 causes high periodic orbits to occur for @xmath153 . on the other hand , if @xmath25 is chosen to be positive , for @xmath152 the fixed point @xmath170 will exist but @xmath61 will not exist and for @xmath153 both the fixed points will exist . if the parameters are chosen such that @xmath175 , then @xmath63 is stable and will occur for the whole parameter range and @xmath61 is unstable .
this is shown in the simulated bifurcation diagram of fig .
[ 4regmu2pe ] . the conditions for existence of period-2 orbits can be obtained from ( [ 2fcon1 ] ) and ( [ 2fcon2 ] ) .
first we have to find the values of @xmath126 , @xmath127 and @xmath179 . from section [ seciii ] , we have @xmath180 @xmath181_1 \\ & = & \mbox{adj}\left ( \begin{array}{llcl } 1+\tau_r & 1 \\ -\delta_r & 1 \end{array}\right ) \left(\begin{array}{c } \tau_l -\tau_r \\ -\delta_l + \delta_r \end{array}\right)\\ & = & \left(\begin{array}{c } \tau_l -\tau_r + \delta_l - \delta_r\\ \delta_r ( 1+\tau_l)-\delta_l ( 1+\tau_r)\end{array}\right ) , \end{aligned}\ ] ] @xmath182 and @xmath183 now with the help of above values , we find @xmath184 substituting the parameters , we obtain @xmath185 and from ( [ k3 ] ) @xmath186 using these expressions , ( [ 2fcon1 ] ) and ( [ 2fcon2 ] ) give @xmath187 @xmath188 as the conditions of existence of the period-2 orbit .
we now show that these are indeed the conditions of the existence of the period-2 fixed points in the 2d map .
since the system is linear in each side of the border , period two ( or a higher period ) fixed points can not exist with all points in @xmath78 or all points in @xmath79 .
however in some regions of the parameter space , a period two fixed point may exist with one point in @xmath78 and one point in @xmath79 .
the conditions for existence of period-2 orbit are obtained by solving @xmath189 @xmath190 where @xmath191 and @xmath192 . now solving the above two equations we find @xmath193 @xmath194 these are the same as those obtained from the theory for @xmath0-dimensional systems .
this implies that if @xmath195 then the period-2 orbit will exist when @xmath196 and if @xmath197 then the period-2 orbit will exist when @xmath198 these two conditions are shown in fig .
[ fig1label3](a ) as the shaded and unshaded regions respectively in the @xmath199 space .
now , for particular values of the parameters @xmath200 , @xmath201 , @xmath202 and @xmath166 the region of existence of period-2 are depicted in the @xmath203 space in fig .
[ fig1label3](b ) . and @xmath136 .
the period-2 points exist in the shaded region if @xmath204 and in the unshaded region if @xmath205 .
( b ) regions of existence of period-2 point in @xmath206 parameter space .
the period-2 points exist in the shaded regions for a set of particularly chosen @xmath207 . ]
these equations allow one to calculate the parameter range of existence of the period-2 orbit , when one of the parameters is varied . to illustrate this
, we show the bifurcation diagram in fig .
[ fig1label4 ] , where @xmath24 varied and all the other parameters are kept fixed .
the stability of the period-2 fixed point is given by the eigenvalues of the matrix @xmath208 .
for the chosen parameter values , these are @xmath209 and @xmath210 .
this implies that if the period-2 fixed point exists , it will be stable .
now from ( [ p2cond - a ] ) and ( [ p2cond - b ] ) we find that the period-2 orbit should exist between @xmath211 and @xmath212 .
again , since @xmath154 , table [ table1 ] predicts that the period-1 orbit should exist so long as @xmath152 .
the bifurcation diagram clearly shows that the above predictions are true . as the bifurcation parameter .
the other parameters are : @xmath200 , @xmath201 , @xmath202 , @xmath166 , @xmath164 . ]
in view of the fact that a large number of systems of practical interest give rise to discontinuous maps , it has become necessary to develop a comprehensive theory of bifurcations in such systems . in this paper
we have taken the first step in that direction by extending feigin s method that was earlier developed in the context of continuous maps .
we have obtained the conditions of existence of the period-1 and period-2 fixed points in general @xmath0-dimensional discontinuous maps , and have demonstrated the application of the derived results in a piecewise linear two dimensional discontinuous map .
it may be noted that the results obtained in this paper speak only of the existence of period-1 and period-2 fixed points . in order to obtain information about the occurrence of high - periodic orbits and chaos
, one has to consider specific dimension of the system , which can not be obtained in the general @xmath0-dimensional context .
one of the author ( p.s.dutta ) thanks csir india for financial support in the form of a csir - jrf .
this work was supported in part by the brns , department of atomic energy , government of india under project
no . 2003/37/11/brns .
1 banerjee s , karthik m s , yuan g h and yorke j a 2000 bifurcations in one - dimensional piecewise smooth maps theory and applications in switching circuits _ ieee transactions on circuits and systems
i _ * 47 * 38994 | in this paper we report some important results that help in analizing the border collision bifurcations that occur in @xmath0-dimensional discontinuous maps . for this purpose , we use the piecewise linear approximation in the neighborhood of the plane of discontinuity .
earlier , feigin had made a similar analysis for general @xmath0-dimensional piecewise smooth continuous maps .
proceeding along similar lines , we obtain the general conditions of existence of period-1 and period-2 fixed points before and after a border collision bifurcation .
the application of the method is then illustrated using a specific example of a two - dimensional discontinuous map . |
shear melting is a non - equilibrium process where a crystalline lattice is ruptured and eventually melted by an external shearing force .
shear melting and related phenomena were experimentally studied in soft condensed matter systems such as colloids @xcite and complex ( dusty ) plasmas @xcite . a complex , or dusty plasma is a suspension of small solid particles in an ionized gas @xcite .
particles get charged by collecting electrons and ions from the ambient plasma and interact with each other via a screened coulomb pair potential .
when these interparticle forces are balanced by other forces ( gravity , neutral and ion drag , ambipolar electric fields ) , the particles can form regular structures called plasma crystals . in the presence of gravity , a single - layer ( two - dimensional , 2d ) plasma crystal can form @xcite . in a 2d plasma crystal
, the in - plane interparticle pair potential can be well approximated by the yukawa potential .
the shear modulus of such crystals is on the scale of @xmath0 n / mm @xcite , and therefore they can be easily manipulated , e.g. by the radiation pressure force of a focused laser beam @xcite .
using plasma crystals as model systems , shear melting was studied experimentally in refs .
proliferation of supersonic ( for shear sound ) dislocations was suggested as a possible mechanism of shear melting @xcite .
shear flows in melted complex plasmas and their transport properties , e.g shear viscosity were studied in refs .
one important aspect of shear melting that has not been studied so far is the role of relative angular orientation of the crystal and applied shearing force . in this work
, we studied the response of a 2d complex plasma crystal to a suddenly applied shear stress .
the crystal was pre - oriented to align the closely packed rows of particles at @xmath1 or @xmath2 ( two principal orientations ) relative to the shearing force direction .
the force was applied by the radiation pressure of a focused laser beam .
( color online ) shear stress is applied to a two - dimensional plasma crystal by the radiation pressure force of a rastered laser beam .
the laser force is applied in the stripe @xmath3 in the positive @xmath4 direction .
particle trajectories in the camera field of view are shown during @xmath5 s of steady - state shear flow ( corresponds to @xmath6 w in figs .
[ vy_t],[y_t ] ) .
analysis was performed in the region indicated by the dashed - line rectangle , @xmath7 was calculated in the smaller area indicated by the dotted - line square . before the laser was switched on , the plasma crystal was rotated so that the closely packed rows of particles formed an angle @xmath8 ( see inset ) or @xmath1 with the laser beam . ]
our experimental setup was a modified gec ( gaseous electronics conference ) rf reference cell @xcite .
plasma was produced using a capacitively coupled rf discharge in argon at @xmath9 pa . a single layer of dust particles
was suspended in the plasma sheath of the lower rf electrode .
the microspheres made of melamine formaldehyde had a diameter of @xmath10 @xmath11 m , a mass @xmath12 kg , and acquired an electric charge of @xmath13 @xcite .
the suspension included around @xmath14 particles and had a diameter of @xmath15 mm .
the mean interparticle distance in the center was @xmath16 mm [ measured from the first peak of the pair correlation function @xmath17 .
the wigner - seitz radius was @xmath18 mm , where @xmath19 is the areal number density of particles .
this corresponds to the screening parameter @xmath20 , where @xmath21 is the screening length @xcite .
the neutral gas damping rate was @xmath22 . in our experimental conditions , the particle suspension
self - organized in a highly ordered triangular lattice , as indicated by the high value of the ( local ) coupling parameter @xmath23 , where @xmath24 is the particle kinetic temperature . when this plasma crystal settled , we rotated it to align one of the closely packed rows at @xmath8 or @xmath1 with respect to the laser beam , see inset in fig .
[ exp_layout ] . to rotate the crystal
, we temporarily placed a small permanent magnet on the upper glass window @xcite .
after the crystal was properly aligned , the magnet was removed .
shear stress was applied to the plasma crystal by pushing a stripe of particles by the radiation pressure force of a rastered laser beam , as shown in fig .
[ exp_layout ] .
a millennia pro 15sj solid - state @xmath25 nm laser with a maximum cw output power of @xmath26 w was used .
the laser beam coming at a grazing angle of @xmath27 was focused on the particle suspension to a spot smaller than the interparticle distance and then rapidly scanned to draw a rectangular stripe on the suspension .
the scanning was performed by galvanometer - driven mirrors oscillating at a frequency of around @xmath28 hz .
the particles within the illuminated stripe reacted to the average radiation pressure force and were pushed in the direction of the laser beam .
shear stress was thus created in the particle suspension , its magnitude was controlled by varying the laser output power .
the lattice layer was illuminated by a @xmath29 nm , @xmath30 mw diode laser , with its output expanded into a horizontal sheet by a cylindrical lens .
a video camera ( photron fastcam 1024 pci ) was mounted above the chamber , capturing a top view of the particle suspension with a size of @xmath31 and resolution @xmath32 .
the camera lens was equipped with a narrow - band interference filter to admit only the illumination laser light scattered off the particles .
the recording rate was set at @xmath33 frames per second .
we applied increasing levels of shear stress to the crystal . at low stress ,
the crystal deformed elastically . at higher stress ,
the following stages were observed : defect generation while in a solid state , onset of plastic flow , and fully developed shear flow . in this paper , we will study in detail the last two stages . averaged particle velocity @xmath34 in the laser - illuminated stripe as a function of time , for two principal orientations of the plasma crystal .
different curves are for the manipulation laser power @xmath35 ranging from @xmath36 w to @xmath37 w , in increments of @xmath36 w. the laser was switched on at @xmath38 and off at @xmath39 s. ] ( color online ) ( a),(c ) particle trajectories during @xmath40 s and ( b),(d ) the lattice vorticity @xmath41 , where @xmath42 is the velocity of individual particles , at the onset of plastic deformation .
the particles in the shaded area were pushed by the laser ( @xmath6 w ) , as indicated by the arrows .
the lattice orientation before the laser was applied was ( a),(b ) @xmath43 and ( c),(d ) @xmath8 .
the lattice vorticity reveals shear motion and is used here to visualize moving dislocations @xcite . ] displacement of the laser - illuminated particle stripe as a function of time , for two principal orientations of the triangular lattice .
different curves are for the manipulation laser power @xmath35 ranging from @xmath44 w to @xmath37 w , in increments of @xmath36 w. the laser was switched on at @xmath38 and off at @xmath39 s. ] for an overview of the succession of stages leading to shear melting , we measured the average velocity @xmath34 @xcite of all particles in the laser - illuminated stripe as a function of time , see fig .
[ vy_t ] . for @xmath43 ( upper panel ) , there is a rather gradual change in the @xmath45 curves for increasing laser power . the particle suspension becomes liquid and flowing at higher laser powers , but there is no evident threshold for shear melting .
in contrast , for @xmath8 ( lower panel ) the shear melting has a pronounced threshold character : by increasing the laser output power from @xmath46 w to @xmath47 w , the @xmath45 curve changes dramatically . for laser powers
@xmath48 w , the applied stress apparently exceeds the yield stress and the lattice eventually shear - melts . for @xmath49 w , the lattice defects ( mostly dislocation pairs )
are continuously generated which however do not lead to shear melting on the time scale of the present experiment .
the finite values of @xmath34 in this case reveal a kind of a creep flow in the lattice @xcite .
shear melting occurs more easily when closely packed rows of the crystal are aligned with the laser beam ( @xmath43 ) .
this configuration favors dislocation nucleation and subsequent motion , since they glide in the direction of the laser beam @xcite , see figs .
[ disl](a),(b ) . for @xmath8 ,
the plastic deformation of the lattice occurs via dislocations gliding at a certain angle with respect to the laser beam , see figs .
[ disl](c),(d ) .
therefore , the melting happens at higher @xmath35 and has a more pronounced threshold character .
the first peak at @xmath50 s in fig .
[ vy_t ] corresponds to in - cage oscillations of particles .
its presence is a manifestation of a gradual crossover from elastic to plastic deformation .
it is naturally more pronounced for @xmath8 .
we measured the displacement of the laser - illuminated particle stripe by integrating the average velocity @xmath34 @xcite of all particles in the stripe , see fig .
[ y_t ] . after the laser
is switched on at @xmath38 the displacement scales as @xmath51 , which indicates that the particles accelerate due to the action of the applied force . after @xmath52 s , the particle motion slows down .
as long as @xmath35 is below a threshold the curves tend to saturate , whereas above the threshold the lattice shear - melts .
then @xmath53 , where @xmath54 , depending on the laser power .
notice the final push - back , even of a liquid complex plasma after the laser was switched off at @xmath39 s. this is caused by the elastic response of unmelted lattice and by the pressure gradient in the liquid state .
shear - induced reordering in a 2d complex plasma .
the left column shows snapshots of particle positions , the right column shows corresponding static structure factor @xmath7 ( all from the area indicated by the dotted - line square in fig .
[ exp_layout ] ) .
undisturbed triangular lattice was initially oriented at @xmath8 , first row .
the lattice was distorted elastically @xmath55 s after the laser was applied ( @xmath6 w ) , second row . during the subsequent steady - state shear flow ,
the complex plasma was in a liquid state which is , however , not completely isotropic and has a nascent triangular order with @xmath43 , third row [ @xmath7 averaged over @xmath56 s ] . after the laser was switched off , the lattice re - crystallized with @xmath43 , last row [ @xmath7 averaged over @xmath57 s ] . ]
we observed shear - induced reordering in a flowing liquid complex plasma .
we made two related observations .
first , a flowing liquid complex plasma is not completely isotropic . rather , on top of isotropic particle distribution appears a weak triangular order with @xmath43 even in the cases when the original triangular lattice was oriented at @xmath2 ( see fig .
[ s_k ] , third line ) .
similar anisotropy in a liquid complex plasma was observed in ref .
@xcite , where the initial triangular order with @xmath43 was not completely destroyed in a shear melted state .
second , after the laser was switched off , the liquid complex plasma re - crystallized in a triangular lattice with @xmath43 ( see fig .
[ s_k ] , last line ) , even though the lattice was oriented at @xmath8 before melting . a flowing liquid complex plasmas , therefore , favors particle ordering where `` closely packed lines '' are oriented along the flow .
it was further observed that even outside of the shear flow the adjacent crystalline regions gradually change their orientation from @xmath8 to @xmath1 and stay in that orientation after cessation of the flow .
the observed particle ordering is similar to the formation of particle strings aligned in the flow direction in sheared 2d colloidal dispersions @xcite and simulated liquid of soft disks @xcite . here , we observed it for the first time in a sheared complex plasma . the underlying mechanism might be similar to the phenomenon of lane formation observed in binary systems where particles of one sort were driven through a cloud of particles of another sort , e.g. in colloidal dispersions @xcite and complex plasmas @xcite . in this section ,
we discuss the rheology of a steady - state shear flow in liquid complex plasma .
the time @xmath58 for the flow to reach steady state varied depending on the initial crystal orientation and the laser power applied , see fig .
for example , @xmath59 s for @xmath8 and @xmath6 w. for this case , the particle trajectories in a steady - state shear flow are shown in fig .
[ exp_layout ] .
the laser force is applied in the stripe @xmath3 .
particles far from the flow are only slightly disturbed and the original crystalline structure here is intact .
there is a boundary layer between the flow and unmelted crystal .
profiles of time - averaged components of the particle velocity @xmath60 are shown in fig .
[ viscosity](a ) , and the longitudinal shear strain rate @xmath61 is plotted in fig .
[ viscosity](c ) ( solid curve ) .
the profiles of the coupling parameter are shown in fig .
[ viscosity](b ) , where @xmath62 corresponds to the particle kinetic temperature @xmath63 @xcite .
the values of @xmath64 far from the flow indicate a highly ordered crystal .
there is a small anisotropy in the coupling parameter , in particular at the edges of the flow , where @xmath65 , indicating noticeable deviation of the system ( driven in the @xmath4 direction ) from equilibrium .
we note that the values of coupling parameter @xmath66 in the middle of the flow and @xmath67 at its edges are still above the _ equilibrium melting line _ in the @xmath68 coordinates as reported in simulations @xcite .
this is because melting in our experiment is essentially a non - equilibrium phenomenon driven by external forces . indeed , by comparing the relative flow velocity @xmath69 with the thermal velocity at the flow edges , @xmath70
, we conclude that these are about the same .
therefore , the effective `` non - equilibrium '' coupling parameter ( in the @xmath4 direction ) , which is inversely proportional to @xmath71 , is substantially reduced with respect to the `` equilibrium '' value of @xmath72 .
( a ) averaged particle velocity @xmath73 , ( b ) coupling parameter @xmath62 , ( c ) shear stress @xmath74 and shear rate @xmath61 , and ( d ) kinematic viscosity @xmath75 ( where @xmath76 is the areal mass density ) as functions of the transverse coordinate in the steady - state shear flow .
the dimensionless viscosity @xmath77 is also shown , where @xmath78^{1/2}$ ] is the relevant plasma frequency and @xmath79 is the wigner - seitz radius .
the averaging was performed in narrow rectangular bins with the width of @xmath80 mm during @xmath81 s. the dotted line in ( d ) is a fit of experimental data with @xmath82 , where @xmath83 and @xmath84 .
the lattice was oriented at @xmath8 before the laser was switched on , the laser power was @xmath6 w. ] to characterize the flow , we calculate the shear stress @xmath74 in the particle suspension using the positions of individual particles , as explained below .
we used a method where the stress at the position of a particle is built up from individual contributions of its neighbors . in this approach ,
the stress tensor is given by @xcite : @xmath85 where @xmath86 is pair correlation function and @xmath87 is the pair interaction potential .
then for our 2d system we obtain in polar coordinates : @xmath88 to calculate @xmath74 from eq .
( [ stress2 ] ) , @xmath87 was approximated by the yukawa potential and @xmath89 shells of neighbors were included in @xmath90 .
the result is shown in fig .
[ viscosity](c ) ( dotted curve ) .
note that we calculate only the potential part of shear stress , since the remaining kinetic part is negligible in our experimental conditions ( @xmath91 ) @xcite .
it is clear from eq .
( [ stress2 ] ) that the shear stress is determined by the @xmath92 component of @xmath90 . as it was shown in ref .
@xcite , the interparticle spacing @xmath93 [ representative of the first peak of @xmath90 ] in a sheared liquid complex plasma indeed has a strong @xmath92 component . in fig .
[ g_r_phi ] , we present the complete @xmath90 , where the @xmath92 component shows up as an elliptical distortion particularly clearly seen in the first peak , see panels ( c ) and ( d ) .
pair correlation function @xmath90 of complex plasma ( a ) in a crystalline area far from the shear flow , @xmath95 mm , ( b ) in the middle of the shear flow , ( c),(d ) at its edges , @xmath96 mm and @xmath97 mm , respectively ( see fig . [ viscosity ] ) .
the peak amplitudes are up to @xmath98 in ( a ) and therefore appear saturated in this color coding . elliptically distorted @xmath90 in ( c),(d ) indicates strong shear stress .
only @xmath99 shells of neighbors are shown here , shear stress was calculated using @xmath89 shells . ]
given the shear stress and strain rate , the kinematic viscosity of liquid complex plasma was calculated as @xmath75 , where @xmath76 is the areal mass density . the result for @xmath100
is shown in fig .
[ viscosity](d ) ( solid curve ) .
shear viscosity has two local minima at the edges of the flow , @xmath96 mm and @xmath101 mm , where the shear rate is maximal .
this indicates shear thinning , i.e. , a kind of non - newtonian behavior of a liquid whose viscosity diminishes for higher strain rates .
shear thinning in complex plasmas was demonstrated in experiments @xcite and simulations @xcite . in the region @xmath102 , where @xmath103 the shear viscosity can not be calculated as @xmath104 and therefore is not shown in fig .
[ viscosity](d ) .
the found values of kinematic viscosity compare well with the results of ref .
@xcite . in the present work ,
@xmath105 at the edges of shear flow where @xmath106 . in ref .
@xcite , @xmath107 , for similar values of @xmath108 and @xmath109 . in both experiments ,
shear thinning was apparently well pronounced . in ref .
@xcite , viscosity was calculated from the fits of experimental particle velocity profiles to the solution of a modified navier - stokes equation taking neutral gas drag into account .
it is remarkable that the two very different methods of calculating shear viscosity give similar results .
kinematic viscosity @xmath100 is shown as a function of @xmath110 , revealing shear thinning .
the line is a fit of experimental data with @xmath82 , where @xmath83 and @xmath84 .
the dimensionless viscosity @xmath77 is also shown . ] to characterize shear thinning , we plot kinematic viscosity @xmath100 as a function of @xmath110 , see fig .
[ nu_gammagot2 ] .
viscosity declines roughly three times when @xmath110 increases from @xmath111 to @xmath112 .
note that simultaneously the coupling parameter @xmath108 varies by @xmath113 , see fig .
[ viscosity](b ) , due to the viscous heat release @xcite and higher temperature at the edges of the flow , where the shear rate is higher .
the variations in @xmath114 , @xmath108 , and @xmath100 are self - consistent due to the dependence of shear viscosity on temperature @xcite .
the value of viscosity in the limit of low shear , @xmath83 can be compared to the results of equilibrium simulations @xcite .
it corresponds to dimensionless viscosity @xmath115 , where @xmath78^{1/2}$ ] and @xmath79 . in the equilibrium 2d md simulation of ref .
@xcite , where viscosity was calculated using the green - kubo formula , @xmath116 for the highest reported values of @xmath117 , for a comparable value of @xmath109 .
shear induced melting of a 2d plasma crystal was studied experimentally .
shear stress was applied by a focused laser beam rastered over a rectangular area in the crystal . for the first time , the study was angle - resolved , i.e. , the closely packed rows of the crystal were oriented at @xmath43 or @xmath2 ( two principal orientations ) with respect to the shearing force .
shear melting was found to be strongly anisotropic .
the transition of the crystal to flowing liquid with increasing shear stress was rather gradual for @xmath43 and had a distinctly threshold character for @xmath2 .
this is explained by the role of dislocations in the crystal plastic deformation .
their nucleation and motion was facilitated for @xmath1 and retarded for @xmath2 .
the rheology of the steady - state shear flow in liquid complex plasma was studied .
shear stress and strain rate were measured with spatial resolution and from these , the shear viscosity was calculated .
shear viscosity had minima in the regions with high velocity shear , demonstrating shear thinning .
shear - induced reordering was observed in flowing liquid complex plasma , where particles tended to form strings aligned in the direction of the flow .
different from laning in driven binary complex plasmas , formation of strings in sheared complex plasma was observed for the first time .
a. ivlev , h. lwen , g. morfill , c. p. royall , _ complex plasmas and colloidal dispersions : particle - resolved studies of classical liquids and solids _ , series in soft condensed matter vol .
5 ( world scientific , singapore , 2012 ) . to take into account the return flow of particles at the suspesion perifery , @xmath34 was calculated in every frame as @xmath119 , where the averaging was performed respectively for all particles in the laser - illuminated stripe and in the complete field of view .
the average @xmath120 component was on the scale of @xmath121% of @xmath34 .
p. hartmann , a. douglass , j. c. reyes , l. matthews , t. w. hyde , a. kovcs , z. donk , slow deformation of 2d plasma crystals , poster presented at : 13th workshop on the physics of dusty plasmas , 2012 may 20 - 23 , waco , tx .
k. r. stterlin , a. wysocki , a. v. ivlev , c. rth , h. m. thomas , m. rubin - zuzic , w. j. goedheer , v. e. fortov , a. m. lipaev , v. i. molotkov , o. f. petrov , g. e. morfill , and h. lwen , phys .
lett . * 102 * , 085003 ( 2009 ) .
du , k. r. stterlin , k. jiang , c. rth , a. v. ivlev , s. khrapak , m. schwabe , h. m. thomas , v. e. fortov , a. m. lipaev , v. i. molotkov , o. f. petrov , y. malentschenko , f. yurtschichin , y. lonchakov , and g. e. morfill , new j. phys . * 14 * , 073058 ( 2012 ) . | a two - dimensional plasma crystal was melted by suddenly applying localized shear stress .
a stripe of particles in the crystal was pushed by the radiation pressure force of a laser beam .
we found that the response of the plasma crystal to stress and the eventual shear melting depended strongly on the crystal s angular orientation relative to the laser beam .
shear stress and strain rate were measured , from which the spatially resolved shear viscosity was calculated .
the latter was shown to have minima in the regions with high velocity shear , thus demonstrating shear thinning .
shear - induced reordering was observed in the steady - state flow , where particles formed strings aligned in the flow direction . |
traditional machine learning methods usually assume that there are sufficient training samples to train the classifier .
however , in many real - world applications , the number of labeled samples are always limited , making the learned classifier not robust enough .
recently , cross - domain learning has been proposed to solve this problem @xcite , by borrowing labeled samples from a so called source domain " for the learning problem of the target domain " in hand .
the samples from these two domains have different distributions but are related , and share the same class label and feature space .
two types of domain transfer learning methods have been studied : * classifier transfer * method which learns a classifier for the target domain by the target domain samples with help of the source domain samples , while * cross domain data representation * tries to map all the samples from both source and target domains to a data representation space with a common distribution across domains , which could be used to train a single domain classifier for the target domain @xcite . in this paper , we focus on the cross domain representation problem .
some works have been done in this field by various data representation methods .
for example , blitzer et al .
@xcite proposed the structural correspondence learning ( scl ) algorithm to induce correspondence among features from the source and target domains , daume iii @xcite proposed the feature replication ( fr ) method to augment features for cross - domain learning .
pan et al .
@xcite proposed transfer component analysis ( tca ) which learns transfer components across domains via maximum mean discrepancy ( mmd ) @xcite , and extended it to semi - supervised tca ( sstca ) @xcite .
recently , sparse coding has attracted many attention as an effective data representation method , which represent a data sample as the sparse linear combination of some codewords in a codebook @xcite .
most of the sparse coding algorithms are unsupervised , due to the small number of labeled samples .
some semi - supervised sparse coding methods are proposed to utilize the labeled samples and significant performance improvement has been reported @xcite . in this case
, it would be very interesting to investigate the use of cross - domain representation to provide more available labeled samples from the source domain .
to our knowledge , no work has been done using the sparse coding method to solve the cross - domain problem to fill in this gap , in this paper , we propose a novel cross - domain sparse coding method to combine the advantages of both sparse coding and cross - domain learning . to this end , we will try to learn a common codebook for the sparse coding of the samples from both the source and target domains . to utilize the class labels ,
a semi - supervised regularization will also be introduced to the sparse codes .
moreover , to reduce the mismatch between the distributions of the sparse codes of the source and target samples , we adapt the mmd rule to sparse codes .
the remaining of this paper is organized as follows : in section [ sec : crodomsc ] , we will introduce the formulations of the proposed cross - domain sparse coding ( crodomsc ) , and its implementations .
section [ sec : exp ] reports experimental results , and section [ sec : concl ] concludes the paper .
in this section , we will introduce the proposed crodomsc method .
we denote the training dataset with @xmath0 samples as @xmath1 , where @xmath0 is the number of data samples , @xmath2 is the feature vector of the @xmath3-th sample , and @xmath4 is the feature dimensionality . it is also organized as a matrix @xmath5\in \mathbb{r}^{d\times n}$ ] .
the training set is composed of the source domain set @xmath6 and target domain set @xmath7 , i.e. , @xmath8 . we also denote @xmath9 and @xmath10 as the number of samples in source and target domain set separatively . all the samples from the source domain set @xmath6
are labeled , while only a few samples from the target domain @xmath7 are labeled . for each labeled sample @xmath2
, we denote its class label as @xmath11 , where @xmath12 is the class label space . to construct the objective function , we consider the following three problems : sparse coding problem : : given a sample @xmath13 and a codebook matrix @xmath14\in \mathbb{r}^{d\times k}$ ] , where the @xmath15-th column is the @xmath15-th codeword and @xmath16 is the number of codewords in the codebook , sparse coding tries to reconstruct @xmath17 by the linear reconstruction of the codewords in the codebook as @xmath18 , where @xmath19^\top \in \mathbb{r}^{k}$ ] is the reconstruction coefficient vector for @xmath2 , which should be as sparse as possible , thus @xmath20 is also called sparse code
. the problem of sparse coding can be formulated as follows : @xmath21 where @xmath22\in \mathbb{r}^{k \times n}$ ] is the sparse code matrix , with its @xmath3-th collum the sparse code of @xmath3-th sample .
semi - supervised sparse coding regularization : : in the sparse code space , the intra - class variance should be minimized while the inter - class variance should be maximized for all the samples labeled , from both target and source domains .
we first define the semi - supervised regularization matrix as @xmath23\in \{+1,-1,0\ } ^{n\times n}$ ] , where @xmath24 we define the degree of @xmath2 as @xmath25 , @xmath26 , and @xmath27 as the is the laplacian matrix
. then we formulate the semi - supervised regularization problem as the following problem : @xmath28=tr(v l v^\top ) \end{aligned}\ ] ] in this way , the @xmath29 norm distance between sparse codes of intra - class pair ( @xmath30 ) will be minimized , while inter - class pair ( @xmath31 ) maximized . reducing mismatch of sparse code distribution
: : to reduce the mismatch of the distributions of the source domain and target domain in the sparse code space , we adopt the mmd @xcite as a criterion , which is based on the minimization of the distance between the means of codes from two domains .
the problem of reducing the mismatch of the sparse code distribution between source and target domains could be formatted as follows , @xmath32=tr(v \pi v^\top ) \end{aligned}\ ] ] where @xmath33^\top \in \mathbb{r}^{n}$ ] with @xmath34 the domain indicator of @xmath3-th sample defined as @xmath35 and @xmath36 . by summarizing the formulations in ( [ equ : sc ] ) , ( [ equ : semisupervised ] ) and ( [ equ : mismatch ] ) , the crodomsc problem is modeled as the following optimization problem : @xmath37 + \gamma tr[v \pi v^\top ] + \alpha \sum_{i:{{\textbf{x}}}_i\in \mathcal{d } } \|{{\textbf{v}}}_i\|_1 \\ & = \|x - uv\|^2_2 + tr[v e v^\top ] + \alpha \sum_{i:{{\textbf{x}}}_i\in \mathcal{d } }
\|{{\textbf{v}}}_i\|_1 \\ s.t.&\|{{\textbf{u}}}_k\|\leq c,~k=1,\cdots , k \end{aligned}\ ] ] where @xmath38 . since direct optimization of ( [ equ
: objective ] ) is difficult , an iterative , two - step strategy is used to optimize the codebook @xmath39 and sparse codes @xmath40 alternately while fixing the other one . by fixing the codebook @xmath39
, the optimization problem ( [ equ : objective ] ) is reduced to @xmath41 + \alpha \sum_{i:{{\textbf{x}}}_i\in \mathcal{d } } \|{{\textbf{v}}}_i\|_1 \end{aligned}\ ] ] since the reconstruction error term can be rewritten as @xmath42 , and the sparse code regularization term could be rewritten as @xmath43 $ ] @xmath44 , ( [ equ : object_v ] ) could be rewritten as : @xmath45 when updating @xmath20 for any @xmath46 , the other codes @xmath47 for @xmath48 are fixed .
thus , we get the following optimization problem : @xmath49 with @xmath50 .
the objective function in ( [ equ : scv ] ) could be optimized efficiently by the modified feature - sign search algorithm proposed in @xcite . by fixing the sparse codes @xmath40 and removing irrelevant terms , the optimization problem ( [ equ : objective ] ) is reduced to @xmath51 the problem is a least square problem with quadratic constraints , and it can be solved in the same way as @xcite .
the proposed * cross domain sparse coding * algorithm , named as * crodomsc * , is summarized in algorithm [ alg : crodomss ] .
we have applied the original sparse coding methods to the samples from both the source and target domains for initialization .
* input * : training sample set from both source and target sets @xmath8 ; initialize the codebooks @xmath52 and sparse codes @xmath53 for samples in @xmath54 by using single domain sparse coding . update the sparse code @xmath55 for @xmath2 by fixing @xmath56 and other sparse codes @xmath57 for @xmath48 by solving ( [ equ : scv ] ) . update the codebook @xmath58 by fixing the sparse code matrix @xmath59 by solving ( [ eqe : object_u ] ) .
* output * : @xmath60 and @xmath61 .
when a test sample from target domain comes , we simply solve problem ( [ equ : scv ] ) to obtain its sparse code .
in the experiments , we experimentally evaluate the proposed cross domain data representation method , crodomsc . in the first experiment , we considered the problem of cross domain image classification of the photographs and the oil paintings , which are treated as two different domains .
we collected an image database of both photographs and oil paintings .
the database contains totally 2,000 images of 20 semantical classes .
there are 100 images in each class , and 50 of them are photographs , and the remaining 50 ones are oil paintings .
we extracted and concatenated the color , texture , shape and bag - of - words histogram features as visual feature vector from each image . to conduct the experiment , we use photograph domain and oil painting domain as source domain and target domain in turns .
for each target domain , we randomly split it into training subset ( 600 images ) and test subset ( 400 images ) , while 200 images from the training subset are randomly selected as label samples and all the source domain samples are labeled .
the random splits are repeated for 10 times .
we first perform the crodomsc to the training set and use the sparse codes learned to train a semi - supervised svm classifier @xcite .
then the test samples will also be represented as sparse code and classified using the learned svm .
we compare our crodomsc against several cross - domain data representation methods : sstca @xcite , tca @xcite , fr @xcite and scl @xcite .
the boxplots of the classification accuracies of the 10 splits using photograph and oil painting as target domains are reported in figure [ fig : figpho ] . from figure
[ fig : figpho ] we can see that the proposed crodomsc outperforms the other four competing methods for both photograph and oil painting domains .
it s also interesting to notice that the classification of the fr and sci methods are poor , at around 0.7 .
sstca and tca seems better than fr and sc but are still not competitive to crodomsc . in the second experiment , we will evaluate the proposed cross - domain data representation method for the multiple user based spam email detection .
a email dataset with 15 inboxes from 15 different users is used in this experiment @xcite .
there are 400 email samples in each inbox , and half of them are spam and the other half non - spam . due to the significant differences of the email source among different users , the email set of different users could be treated as different domains . to conduct the experiment , we randomly select two users inboxes as source and target domains . the target domain will further be split into test set ( 100 emails ) and training set ( 300 emails , 100 of which labeled , and 200 unlabeled ) .
the source domain emails are all labeled .
the word occurrence frequency histogram is extracted from each email as original feature vector .
the crodomsc algorithm was performed to learn the sparse code of both source and target domain samples , which were used to train the semi - supervised classifier .
the target domain test samples were also represented as sparse codes , which were classified using the learned classifier .
this selection will be repeated for 40 times to reduce the bias of each selection .
figure [ fig : figspam ] shows the boxplots of classification accuracies on the spam detection task . as we can observed from the figure , the proposed crodomsc always outperforms its competitors .
this is another solid evidence of the effectiveness of the sparse coding method for the cross - domain representation problem .
moreover , sstca , which is also a semi - supervised cross - domain representation method , seems to outperform other methods in some cases .
however , the differences of its performances and other ones are not significant .
in this paper , we introduce the first sparse coding algorithm for cross - domain data representation problem .
the sparse code distribution differences between source and target domains are reduced by regularizing sparse codes with mmd criterion . moreover ,
the class labels of both source and target domain samples are utilized to encourage the discriminative ability .
the developed cross - domain sparse coding algorithm is tested on two cross - domain learning tasks and the effectiveness was shown .
this work was supported by the national key laboratory for novel software technology , nanjing university ( grant no .
kfkt2012b17 ) .
j. blitzer , r. mcdonald , and f. pereira .
domain adaptation with structural correspondence learning . in _
2006 conference on empirical methods in natural language processing , proceedings of the conference _ , pages 120 128 , 2006 . | sparse coding has shown its power as an effective data representation method .
however , up to now , all the sparse coding approaches are limited within the single domain learning problem . in this paper , we extend the sparse coding to cross domain learning problem , which tries to learn from a source domain to a target domain with significant different distribution .
we impose the maximum mean discrepancy ( mmd ) criterion to reduce the cross - domain distribution difference of sparse codes , and also regularize the sparse codes by the class labels of the samples from both domains to increase the discriminative ability .
the encouraging experiment results of the proposed cross - domain sparse coding algorithm on two challenging tasks image classification of photograph and oil painting domains , and multiple user spam detection show the advantage of the proposed method over other cross - domain data representation methods . |
nano - technology requires an accurate control of the interacting components , both in terms of detection and preparation .
this is a major motivation for the considerable attention that was devoted recently to sub - wavelength localization of single particles .
several remarkable schemes were proposed to achieve this goal @xcite . of related interest
is the problem of localizing and distinguishing two nearby particles
. optical resolution of two molecules at the nanometer scale and manipulation of their degree of entanglement was experimentally demonstrated in @xcite , and the collective interaction between pairs of oriented nanostructures was considered in @xcite .
also the measurement of the relative position of two atoms assisted by spontaneous emission @xcite or the measurements of interparticle separations on a scale smaller than the emission wavelength @xcite were investigated in detail .
somewhat related , considerable effort is devoted to quantum lithography @xcite .
the above localization schemes , however , have in common that they apply to the localization of a single particle or to the measurement of relative position of two individual particles . in many cases of interest , however , an ensemble of particles is concentrated to a small region in space such that the ensemble properties become relevant , while the properties of the individual ensemble constituents can not be resolved or are rapidly fluctuating in time . therefore , here we describe a scheme capable of localizing an ensemble of two - level atoms which are bunched together in a volume much smaller than an emission wavelength .
possible realizations include small clusters , few - atom impurities , or atoms trapped , e.g. , in optical lattices .
the localization relies on the coherent interaction with a standing - wave electromagnetic field .
since the interatomic distances are small , the atoms interact collectively via the environmental vacuum modes .
one consequence of this is the appearance of superfluorescence , i.e. , the scattered light intensity scales with the number of atoms @xmath0 squared ( @xmath1 ) .
we find that the fluorescence light emitted collectively by the ensemble is a function of the ensemble position in the standing wave .
in particular , for suitable standing wave parameters and for ensemble positions around the nodes of the standing wave field , the emitted fluorescence intensity sharply drops to a minimum over a narrow spatial region .
the narrow width of the dip in the spatial intensity profile is a direct consequence of the collectivity .
since this collective fluorescence intensity profile is our main observable , we discuss the profile in detail in terms of the available free parameters , and show that the profile can be tailored to suit a given localization setup . based on these results ,
we then propose two schemes which exploit this spatial fluorescence intensity profile to localize an ensemble of atoms .
first , we assume the sample to be fixed within the standing wave field . in this case
, the spatial fluorescence intensity profile can be scanned along the standing wave axis by changing the relative phase of the laser fields forming the standing wave .
a continuous measurement of the intensity of the scattered light throughout this scan reveals the position of the sample on a sub - wavelength scale .
we further show that this setup also enables one to measure the distance between two samples , the number of atoms in a sample , or the linear dimension of the sample .
second , we consider an atom cluster flying through the standing wave field . here
a scanning is not possible due to the short interaction time of ensemble and standing wave field .
rather , the absolute intensity of the scattered light can be used to recover the crossing position of the ensemble .
the article is organized as follows . in sec .
[ sec - theory ] , we present our theoretical model for the ensemble interaction with the standing wave field and derive the intensity profile as our main observable . sec .
[ sec - results ] consists of three parts . in the first part sec .
[ sec - int ] , we in detail study the collective fluorescence profile numerically . in sec .
[ sec - sweep ] , the results are applied to the localization of an ensemble fixed in the standing wave field . the last part sec .
[ sec - single ] discusses the localization of an ensemble flying through the standing wave field .
finally , sec .
[ sec - summary ] discusses and summarizes the results .
in the usual mean - field , dipole , and rotating - wave approximations the interaction of such an atomic sample with an external laser field and the surrounding vacuum modes , in a frame rotating with the laser frequency @xmath2 , is described by the hamiltonian @xmath3 where @xmath4 \\ % h_{l}&=\sum^{n}_{j=1}\hbar\bigl ( \omega(\vec r_{j})s^{+}_{j } + \omega^{\ast}(\vec r_{j})s^{-}_{j } \bigr ) , \nonumber \allowdisplaybreaks[2]\\ % h_{i}&= i\sum_{k}\sum^{n}_{j=1}(\vec g_{k}\cdot \vec d_{j } ) a^{\dagger}_{k}s^{-}_{j}e^{-i\vec k\cdot \vec r_{j } } + \textrm { h.c.}\ , .
\label{hm}\end{aligned}\ ] ] here @xmath5 are the raising and lowering operators for the @xmath6th atom , positioned at @xmath7 , and obeying the commutation relations @xmath8=2s_{zj}\delta_{jl}$ ] , and @xmath9=\pm s^{\pm}_{j}\delta_{jl}$ ] with @xmath10 being the inversion operator .
@xmath11 and @xmath12 are the radiation creation and annihilation operators satisfying the commutation relations @xmath13=\delta_{kk^{'}}$ ] , and @xmath14=[a^{\dagger}_{k},a^{\dagger}_{k^{'}}]=0 $ ] . in eq .
( [ hm ] ) , @xmath15 represents the free electromagnetic field ( emf ) and free atomic hamiltonians , respectively . the second term
, i.e. @xmath16 , describes the interaction of the atomic system with an external standing - wave coherent field .
in general , the rabi frequencies of the atoms in a standing wave are position - dependent since @xmath17 where @xmath18 while @xmath19 is the amplitude of the electromagnetic field intensity with a wave vector @xmath20 and @xmath21 is the dipole moments of the atoms .
note that the scheme described here can also be generalized to multiphoton transitions .
then , the rabi frequency can be written as @xmath22 where @xmath23 is a multiphoton rabi - frequency arising from an adiabatic elimination of intermediate states , and @xmath24 denotes the number of photons involved in the multiphoton process . in this way
, the wavelength @xmath25 can be reduced to the effective wavelength @xmath26 , thus increasing the spatial resolution @xcite . the last term in eq .
( [ hm ] ) , @xmath27 , takes into account the interaction of all atoms with the environmental vacuum modes . in the born - markov approximations the quantum dynamics of the driven multi - atom sample
( each atom having identical transition frequency @xmath28 ) is governed by the master equation @xcite : @xmath29 = \nonumber \\ & \quad \sum^{n}_{i , j=1}\{\gamma_{ij}(\omega_{0})[s^{+}_{j},s^{-}_{l}\rho ] + \textrm{h.c . } \}\
, , \label{me}\\ % \tilde h_{0}&=\hbar\sum_{j}[\delta s_{zj}/2 + \omega(\vec r_{j})s^{+}_{j } + \textrm{h.c.}]\,.\end{aligned}\ ] ] here , @xmath30 is the detuning of atomic levels from the frequency of the driving field .
further , @xmath31 where the collective parameters describing the mutual interactions among any atomic pair in the sample are given , respectively , by @xcite @xmath32\frac{\sin(\omega r_{jl}/c)}{\omega r_{jl}/c } + [ 1 - 3\cos^{2}{\xi_{jl } } ] \nonumber \\ & \times&\bigl [ \frac{\cos(\omega r_{jl}/c)}{(\omega r_{jl}/c)^{2}}- \frac{\sin(\omega r_{jl}/c)}{(\omega r_{jl}/c)^{3}}\bigr ] \bigr \ } , \nonumber \\
\omega_{jl}(\omega ) & = & \frac{3\gamma}{4}\bigl \{[\cos^{2}{\xi_{jl}}-1]\frac{\cos(\omega r_{jl}/c)}{\omega r_{jl}/c } + [ 1 - 3\cos^{2}{\xi_{jl } } ] \nonumber \\ & \times & \bigl [ \frac{\sin(\omega r_{jl}/c)}{(\omega r_{jl}/c)^{2 } } + \frac{\cos(\omega r_{jl}/c)}{(\omega r_{jl}/c)^{3}}\bigr ] \bigr \ } , \label{colp}\end{aligned}\ ] ] with @xmath33 being the single - atom spontaneous decay rate . here
, we have assumed that all the dipole moments are identical and parallel , i.e. @xmath34 , and then @xmath35 is the angle between the dipole moments @xmath36 and @xmath37 . inspecting the master equation ( [ me ] )
one can easily distinguish the part of it describing the coherent evolution of atoms under the influence of the laser field , i.e. the term containing the hamiltonian @xmath38 , from that characterizing the collective spontaneous emission due to the vacuum modes , that is the terms proportional to @xmath39 .
the dipole - dipole interactions between the two - level atoms are described by the terms proportional to @xmath40 .
if the interparticle separations are small enough , that is @xmath41 ( @xmath42 ) , then to second order in this parameter , eqs . ( [ colp ] )
reduce to @xmath43\ } , \nonumber \\
\omega_{jl}(k ) & = 3\gamma\{[\cos^{2}\xi_{jl } -1](2/kr_{jl}-kr_{jl } ) + [ 1- \nonumber \\ 3&\cos^{2}\xi_{jl } ] [ ( kr_{jl})^{-1 } - kr_{jl}/4 + 2(kr_{jl})^{-3}]\}/8 .
\label{apcolp } \end{aligned}\ ] ] it is easy to realize that in this case @xmath44 while @xmath45 reduces to the static dipole - dipole interaction potential , i.e. @xmath46 for lower atomic densities the collective parameters @xmath47 and @xmath45 vanish because the atoms react independently from each other in this particular case .
finally , the master equation ( [ me ] ) describes adequately driven atomic samples of any shapes providing that retardation effects are negligible . when dealing with smaller atomic systems of an arbitrary irregular shape it is hard to specify the orientation of dipole moments relative to the interparticle separations .
that is to say , we do not have a privileged angular distribution of photons as they are emitted equally in all directions . since there is no information on the dipole orientations
, we average over all directions the dipole - dipole interaction potential .
interestingly , the static dipole - dipole interaction given by eq .
( [ sdd ] ) vanishes in this case .
then , according to the second term in eq .
( [ apcolp ] ) , the averaged dipole - dipole interactions among the two - level emitters are given by the expression : @xmath48 thus , the dipole - dipole potential reduces from a short - range to a long - range interaction , although the radiators are close to each others .
moreover , the influence of the dipole - dipole interactions , in a two - atom system , was shown to be negligible in practice for interparticle separations such that @xmath49 @xcite . in what follows
we shall apply eqs .
( [ me]-[av_dd ] ) to the localization of a small atomic system within an emission wavelength .
suppose that the linear dimension of the atomic sample is much less than the emission wavelength ( say , for instance , smaller than @xmath50 ) . under this assumption
the two - level emitters are almost in an equivalent position relative to the driving standing - wave laser and we can omit the atomic indices from the expression characterizing the rabi frequency , i.e @xmath51 . the master equations ( [ me ] ) transforms then into : @xmath52 \nonumber \\ & = & \gamma \ { [ s^{+},s^{-}\rho ] + [ \rho s^{+},s^{- } ] \}.
\label{mer}\end{aligned}\ ] ] here @xmath53 and @xmath54 are collective atomic operators satisfying the standard commutation relations of su(2 ) algebra @xcite while @xmath55 is the dipole - dipole interaction potential considered identical for all radiators .
note that eq .
( [ mer ] ) describes a small driven atomic system that involves symmetrized multiparticle states only .
the antisymmetric states are decoupled from the dynamics within our current framework , and in the following we assume that they are not populated initially . the dipole - dipole interaction @xmath55 considerably shifts the symmetric states from the field resonance if the interparticle separations are very small @xcite . even though the hypothesis of identical dipole - dipole interactions is , in general ,
not fulfilled , it admits to solve analytically the above master equation in the long - time limit .
thus , in order to get some insight on how the dipole - dipole interactions affect the localization processes of a small system as a whole we shall accept the hypothesis of identical dipole - dipole interactions between any pair in the sample . the solving procedure of eq .
( [ mer ] ) was described in @xcite for running wave lasers . adopting it to the case of a standing wave field ,
one arrives at the steady - state solution @xmath56 where @xmath57 with @xmath58 and @xmath59 .
the normalization constant @xmath60 is chosen such that @xmath61 , i.e. @xmath62 the trace can be performed using the following relations @xmath63 where the collective dicke states @xmath64 , with @xmath65 and @xmath66 , are the eigenstates for the operator @xmath67 and the operator of the total `` spin '' @xmath68 @xcite : @xmath69 thus , @xmath70 and , then @xmath71 the intensity of the collective resonance fluorescence emitted by driving the multiparticle system is calculated taking into account that this quantity is proportional to the first order atomic correlator , i.e. @xmath72 .
then , using eq .
( [ sss ] - [ z ] ) , one obtains : @xmath73
we now turn to the discussion of eq .
( [ it ] ) .
first , we study the dependence of the collective fluorescence intensity on the various external parameters .
second , we show how the collective fluorescence intensity can be used to precisely locate a sample of particles which is fixed in space inside the standing wave field .
third , we discuss the localization of a collection of particles flying through the cavity field . the dependence of the collective steady - state resonance fluorescence intensity @xmath74 as function of @xmath75 .
the solid , dashed and dotted curves are for @xmath76 and @xmath77 , respectively .
other parameters are : @xmath78 , @xmath79 and @xmath80.,height=170 ] fig .
[ fig-1 ] shows the collective fluorescence intensity versus the position of the multiparticle collection in the standing wave field .
note that @xmath81 corresponds to a node of the standing wave field , and thus the fluorescence intensity vanishes for particles located at this point in space .
the parameters in fig .
[ fig-1 ] are number of atoms @xmath80 , dipole - dipole coupling constant @xmath78 , and detuning @xmath79 .
the solid line shows a standing wave rabi frequency @xmath82 , the dashed line is for @xmath83 , and the dotted one is for @xmath84 .
it can be seen that with decreasing rabi frequency , the width of the dip in the fluorescence intensity around the node at @xmath81 increases .
thus a strong driving field allows for a narrow region in space that leads to vanishing fluorescence intensity , while a weaker field gives rise to fluorescence intensity over a wider range of positions .
( color online ) the steady - state collective resonance fluorescence intensity @xmath74 as function of @xmath75 and @xmath85 .
here @xmath80 , @xmath83 and @xmath86 .
the black line on top of the surface plot indicates the position @xmath87.,width=302 ] ( color online ) the steady - state collective resonance fluorescence intensity @xmath74 as function of @xmath75 and @xmath85 . here
@xmath88 , @xmath83 and @xmath89 .
the black line on top of the surface plot indicates the position @xmath90 around which the two - atom symmetric collective state is located.,width=302 ] the second free parameter is the detuning @xmath91 between the driving field frequency and the bare transition frequency @xmath92 of the individual atoms in the sample .
[ fig-2 ] shows the dependence of the collective fluorescence intensity versus the position of the sample on this detuning .
note that the y - axis of this figure is a scaled detuning @xmath93 .
thus , @xmath94 corresponds to the resonance case @xmath95 , whereas @xmath96 indicates @xmath97 and thus @xmath98 .
a qualitative understanding of this figure can be gained from the case of a two atom sample , as shown in fig .
[ fig-3 ] . in a collective state basis
, the two - atom sample corresponds to a collective ground state @xmath99 at energy 0 , where each of the two atoms @xmath100 is in its respective ground state , an excited collective state @xmath101 at energy @xmath102 , and a symmetric [ antisymmetric ] collective state at energy @xmath103 [ @xmath104 : @xmath105 , the symmetric state is located at @xmath90 , since the dipole - dipole coupling is chosen as @xmath106 .
collective fluorescence versus atom ensemble position in the standing wave field for different sizes of the ensemble .
the parameters are @xmath89 , @xmath107 , and @xmath108 .
the solid line is for number of atoms @xmath109 , the dashed shows the case @xmath110 , and the dotted is for @xmath111.,width=302 ] it can be seen that the width of the intensity dip is minimal if the driving field frequency is in resonance with the symmetric state , and becomes wider in moving away from the resonance .
no additional structure is visible around @xmath112 , where the asymmetric state is located , since it is decoupled .
note that for vanishing dipole - dipole interaction , @xmath113 , figs .
[ fig-2 ] and [ fig-3 ] would exhibit features symmetric with respect to both the planes given by @xmath114 and @xmath81 . these results from the two - atom case
directly carry over to the many - particle sample .
the minimum width of the intensity dip is close to the position where symmetric state combinations can be expected , whereas no structure can be found towards asymmetric collective states .
a direct identification of the position of the detuning with minimum dip width is difficult , however , since the collective - state basis of a multi - particle sample includes many symmetric collective states . in the example of fig .
[ fig-2 ] , in a very small range around the node @xmath81 the dip at @xmath115 is narrowest , but its width increases faster in moving away from the node than it does for slightly lower values of @xmath116 .
the many - particle case also shows an additional structure at @xmath117 , see fig .
[ fig-2 ] , which , however , is not of interest for our current localization scheme . finally , in fig .
[ fig-4 ] , we show the dependence of the collective intensity on the number of particles in the sample .
this is different from the previous results , since by changing the number of atoms , both the scaled rabi frequency @xmath118 and the scaled detuning @xmath119 are changed at the same time .
it can be seen from fig .
[ fig-4 ] that for a given standing wave intensity and a given detuning , varying the number of atoms in the ensemble changes the width of the intensity dip at the nodes .
this can be understood by noting that a change of the number of atoms effectively shifts the position of the symmetric state resonance . since the laser field frequencies are kept fixed in fig .
[ fig-4 ] , this corresponds to moving along the @xmath116 axis in figs .
[ fig-2 ] and [ fig-3 ] . therefore
, different widths of the intensity dip can be observed . the maximum intensity changes with @xmath0 in fig .
[ fig-4 ] , since in this figure the unscaled rabi frequency @xmath120 is kept fixed .
it should be noted that the parameters in fig .
[ fig-4 ] are such that for any shown number of atoms , the scaled rabi frequency @xmath118 dominates the dynamics .
a further increase of the number of atoms which leads to @xmath121 shifts the relevant collective states out of the laser field resonance such that the total fluorescence intensity vanishes .
( color online ) scanning - dip scheme .
the figure depicts a possible experimental implementation of our scheme .
an atom ensemble ( green dot ) is assumed fixed inside the standing wave field .
a detector measures the scattered fluorescence light , while the phase of the standing wave field is varied .
the black curve symbolizes the fluorescence intensity profile as , e.g. , shown in fig .
[ fig-1 ] . if the intensity dip does not coincide with the ensemble position , then a high intensity of fluorescence is detected .
but if the dip sweeps across the ensemble position , then the measured intensity drops to a minimum over a narrow spatial range , thus providing a sub - wavelength localization.,width=226 ] after the discussion of the collective fluorescence intensity as our main observable , we now turn to the application of this observable to the localization of a collection of atoms . as the first setup
, we consider a collection of atoms which is fixed inside the standing wave field at an unknown position . in order to detect the position of the sample ,
the total collective fluorescence intensity is continuously monitored , which may already provide a coarse position measurement .
then the relative phase of the two counter - propagating fields forming the standing wave is changed , such that the nodal structure of the field shifts along the standing wave propagation axis . throughout this shift ,
the detected intensity is modulated in time with the collective fluorescence intensity profile , as depicted in fig .
[ fig - sweep ] .
if the intensity profile is located such that its dip does not coincide with the actual position of the sample , then the intensity is near its maximum value , see fig .
[ fig - sweep](a ) .
but if the two positions coincide , then the intensity vanishes , see fig .
[ fig - sweep](b ) . obviously , for this scheme it is desirable to have the intensity dip as narrow as possible in order to achieve a localization well below the usual diffraction limit .
according to our results of sec .
[ sec - int ] , this can be achieved by using a strong standing wave field and by tuning it close to the symmetric collective state resonance .
for instance , the solid curve in fig . [ fig-1 ] has a width of about @xmath122 , corresponding to @xmath123 .
note , however , that the obtained accuracy is also limited by the spatial size of the ensemble , if the dip width is smaller than the linear dimension of the sample .
( color online ) single - pass localization .
( a ) shows the schematic setup . an ensemble indicated by the green dot passes through a standing wave field .
the intensity of the scattered fluorescence light is measured .
( b ) via the fluorescence intensity profile , the measured intensity can be related to the sub - wavelength particle position . the higher the slope of the intensity profile , the better is the accuracy of the localization .
in this figure , the horizontal bars indicate two measurement outcomes with a certain uncertainty .
the vertical bars show the corresponding potential positions .
the black curve is a fluorescence intensity profile as , e.g. , shown in fig .
[ fig-1].,width=226 ] the present scheme can also be used to measure the distance between the center of masses of two ensembles by relating the required phase shifts between two positions with vanishing intensity to the measured intensity profile . further , since the width of the intensity dip depends on the number of atoms in the ensemble if all other parameters remain fixed [ see fig .
[ fig-4 ] ] , this scheme can also be used to obtain the number of atoms in the ensemble located in the standing wave field by measuring the width of the intensity dip via the sweep of the standing wave phase .
finally , the spatial dimension of the sample can be measured from the width of the intensity dip if the other parameters are known .
it should be noted that , in contrast to the precision position detection , these measurements relying on the determination of the width of the intensity dip are relative measurements in the sense that they do not require a reference to obtain the absolute standing wave phase
. only the change of the phase is relevant , and thus also knowledge of the actual position of the sample within the standing wave field is not required for these relative measurements .
the maximum attenuation of the fluorescence intensity depends on the width of the collection relative to the width of the dip in the intensity profile .
if the profile dip is narrower than the sample , then the dip only affects part of the sample .
still , this will result in a sudden reduction of the fluorescence intensity , which is sufficient to determine the onset of the overlap of intensity dip and atom sample .
also , a continuous scan of the standing wave phase results in repeated intensity dips over time which can be used to suppress statistical errors in the measurement . in this section
, we discuss a different experimental setup .
we now assume that a collection of atoms ( atomic cluster ) flies through a standing wave field at an unknown position on the standing wave field axis .
the aim is to gain as much information on the position as possible by measuring the collective resonance fluorescence . obviously , the scanning - dip scheme described in sec .
[ sec - sweep ] is not suitable for this kind of setup , since the change of the standing wave phase is too slow as compared to the interaction time of field and atom ensemble . in the present scheme
, we only have to require that the time of flight @xmath124 through the standing wave field is much larger than the time @xmath125 needed to evolve into the steady state .
for example , for a thermal beam with velocity @xmath126 m / s and a standing wave field width of @xmath127 mm , the flight time is about @xmath128 s. the steady - state time @xmath125 is of the order @xmath129 . for @xmath130
mhz and @xmath131 , one obtains @xmath132 s and thus @xmath133 . note that the preparation of atom clusters has been discussed in @xcite .
we now make use of the fact that position information can be gained from the absolute value of the scattered light intensity during the flight of the ensemble through the field . the schematic setup is shown in fig .
[ fig - single](a ) . the measured intensity allows to fix a horizontal section in a collective fluorescence intensity plot versus ensemble position as shown in fig .
[ fig - single](b ) .
ideally , this section provides a set of few discrete points where the intensity profile crosses the measured intensity .
these points correspond to the potential positions of the ensemble . from the two examples in fig .
[ fig - single](b ) it is clear that the localization for a given intensity measurement with a measurement uncertainty becomes better with increasing slope of the intensity profile . a high slope , however , leads to a pronounced plateau with almost constant intensity in between the dips .
this means that , in case of a narrow dip , for large parts of the single wavelength width only a rather inaccurate localization is possible .
therefore , in contrast to the sweep scheme in sec .
[ sec - sweep ] , in this setup a wide intensity dip is desirable in order to achieve sub - wavelength localization for all possible positions .
the reason is that for a wide intensity dip , wide plateaus in the intensity profile are avoided .
a wide intensity dip can be achieved , for example , by working with weaker standing wave fields or far away from the symmetric collective state resonance .
we have described a scheme to localize small atomic samples with sub - wavelength accuracy .
the scheme relies on measuring the super - fluorescence radiation scattered in a standing wave field .
we have demonstrated that external parameters such as the strength of the applied lasers or the detuning from the atomic resonance are convenient tools to tailor the localization region for a given experimental setup . based on these results ,
two possible experimental situations have been considered .
first , for fixed samples , a scanning - dip spectroscopy was proposed . here , the standing wave field phase is changed in order to scan the fluorescence intensity profile along the cavity axis in order to reveal the actual position of the sample .
this setup also allows for a number of relative measurements , for example , of distance between two collections , of the number of atoms in a sample , or of the linear dimension of the sample .
second , for samples passing through the standing wave field only once , a single - pass scheme was discussed , which relates the maximum intensity measured to the passing position of the sample .
our scheme can be generalized to the multi - photon case .
s. qamar , s .- y .
zhu , and m. s. zubairy , phys .
a * 61 * , 063806 ( 2000 ) ; k. t. kapale , s. qamar , and m. s. zubairy , phys . rev .
a * 67 * , 023805 ( 2003 ) ; m. sahrai , h. tajalli , k. t. kapale , and m. s. zubairy , phys .
a * 72 * , 013820 ( 2005 ) .
a. n. boto , p. kok , d. s. abrams , s. l. braunstein , c. p. williams , and j. p. dowling , phys .
* 85 * , 2733 ( 2000 ) ; m. dangelo , m. v. chekhova and y. shih , phys .
* 87 * , 013602 ( 2001 ) ; j. xiong , d. z. cao , f. huang , h. g. li , x. j. sun , and k. wang phys .
lett . * 94 * , 173601 ( 2005 ) ; c. h. keitel and s. x. hu , appl .
lett . * 80 * , 541 ( 2002 ) . | the sub - wavelength localization of an ensemble of atoms concentrated to a small volume in space is investigated .
the localization relies on the interaction of the ensemble with a standing wave laser field .
the light scattered in the interaction of standing wave field and atom ensemble depends on the position of the ensemble relative to the standing wave nodes .
this relation can be described by a fluorescence intensity profile , which depends on the standing wave field parameters , the ensemble properties , and which is modified due to collective effects in the ensemble of nearby particles .
we demonstrate that the intensity profile can be tailored to suit different localization setups .
finally , we apply these results to two localization schemes .
first , we show how to localize an ensemble fixed at a certain position in the standing wave field .
second , we discuss localization of an ensemble passing through the standing wave field . |
after more than 2 years of operation , it is time to review the contribution that bepposax is providing to our knowledge of seyfert galaxies .
as the best characteristic of bepposax is the wide energy band ( 0.1200 kev ) , the results will mainly concern broad band spectra .
moreover , as shown in the last section , the mecs instrument , which has the best spatial resolution in the 210 kev band so far , is also giving valuable information on faint agn and on the x ray background ( xrb ) .
bepposax ( see boella et al .
1997 for a comprehensive description of the mission ) is composed by several instruments : 4 co aligned narrow field instruments ( nfi ) ; two wide field cameras pointing opposite each other and perpendicularly to the nfi ; an all sky @xmath0-ray burst monitor .
we are interested here in three of the nfi : lecs , mecs and pds .
the mecs is presently composed of two units ( after the failure of a third one in march 1997 ) ; the working band is 1.510.5 kev , the energy resolution is @xmath18% and the angular resolution is @xmath10.7 arcmin ( fwhm ) at 6 kev
. the characteristics of the lecs are similar to those of the mecs in the overlapping band , but its energy band extends down to 0.1 kev .
the pds is a passively collimated detector ( about [email protected] degrees f.o.v . )
working in the 13200 kev energy range .
a significant fraction of the bepposax observing time ( 80% the first year , 60% the second year , 50% afterwards ) is devoted to a core program " ( reserved to collaborations led by italian or dutch pis , with a fraction also to mpe and esa / ssd ) which consists of major programs aiming to observe classes of sources in a sistematic way .
in this paper i will assume that the unification model for seyfert galaxies ( antonucci 1993 ) is valid , i.e. seyfert 1 and 2 galaxies are intrinsically identical ( at least as far as the nuclear properties are concerned ) , and surrounded by a @xmath1pc - scale molecular torus . if the line of sight does not intercept the torus , the nucleus can be directly observed and the source is classified as type 1 ; if the line of sight is blocked by the torus , the source is classified as type 2 .
i will come back to unification models in sec . 2 .
this review is necessarily incomplete and inevitably biased towards my own interests .
i will discuss selected results on : bright , classical seyfert 1 galaxies ( see brandt , this volume , for results on narrow line seyfert 1 galaxies ) ; seyfert 2 galaxies ( expecially compton thick ones , and a sample of optically selected sources ) ; the seyfert 1 galaxy ngc 4051 , which switched off in may 1998 and on again a month later ; and the hellas program , i.e. a sample of hard x ray selected sources , whose identification program is in progress but which is already giving valuable results on the x ray background .
before describing bepposax results on seyfert 1 galaxies , let me briefly recall the picture which has emerged after ginga , rosat , cgro and asca observations ( see e.g. mushotzky , done & pounds 1993 ; fabian 1996 ) .
the main component is a power law , cut offing at high energies , and very likely originating from inverse compton emission from relativistic electrons to uv / soft x
ray photons , the latter possibly emitted by the accretion disc ( see e.g. svensson 1996 for a review ) .
a significant fraction of the primary radiation is intercepted and reprocessed by optically thick matter , either the accretion disc or the torus , or both . if the matter is neutral , the shape of the reflected component is determined basically by the competition between photoelectric absorption ( whose cross section depends on the energy , after each photoabsorption edge , as @xmath3 ) and compton scattering ( whose cross section is constant , at least up to a few tens of kev ) . for cosmic abundances ,
the two cross sections are equal at about 10 kev .
this so called compton reflection component has been studied in detail in several papers ( see e.g. lightman & white 1988 ; george & fabian 1991 ; matt , perola & piro 1991 ; magdziarz & zdziarski 1995 ) ; its spectrum is a broad hump , peaking around 30 kev .
when added to the primary component , it hardens the total spectrum above a few kev and steepens it above a few tens of kev . besides this compton reflection continuum , the illumination of neutral matter by the primary radiation results also in a strong iron 6.4 kev fluorescent line , emitted by iron atoms after removal of a k electron by a x ray photon .
kinematic and gravitational effects in the inner accretion disc modify the line profile ( see e.g. reynolds , this volume , and references therein ) , as firstly observed by asca in mcg6 - 30 - 15 ( tanaka et al .
1995 ) and later on realized to be common in seyfert 1s ( nandra et al .
1997 ) . at low energies ( below @xmath11 kev )
a further component ( soft excess " ) , of rather unclear nature , may arise . among the possible explanations , the tail of thermal emission from the accretion disc , and reflection from ionized matter ( ross & fabian 1993 ) are the most popular .
it is worth noticing that most of the observations on soft excesses are based on relatively narrow band instruments ( like rosat ) or on non simultaneous broad band observations .
all these components , which most likely originate close to the black hole , ( with the possible exception of the reflection from the torus ) may pass throughout ionized matter ( the warm absorber " ) .
the main signatures of this matter are absorption edges of high ionization ions , mainly oxygen ( halpern 1984 ; nandra & pounds 1992 ; fabian et al .
1994 ) , which has been unambiguosly observed by asca in a large fraction of seyfert 1s ( reynolds 1997 ; george et al .
resonant absorption lines may also be important ( matt 1994 ; krolik & kriss 1995 ; nicastro , fiore & matt 1998 ) and detectable by future missions featuring high resolution spectrometers .
two major bepposax core programs are devoted to classical seyfert 1s : one ( pi : g.c .
perola ) , aiming to study broad band spectra of bright sources ; the second ( pi : l. piro ) , looking for spectral variability .
about a dozen sources have been observed on aggregate so far ( see table 1 ) .
i am not going to discuss in detail any of these sources , but i will rather try to outline the main results of seyfert 1s as a class that are emerging by these observations .
let me discuss separately the iron line , the compton reflection continuum , the high energy cut off and the soft excess .
lllll & & & & source & s.e . &
@xmath4 & r & e@xmath5 ( kev ) & & & & & & & & ngc 4151@xmath6 & ?
& 1.21.5 & 0.10.5 & 70@xmath715 ngc 5548@xmath8 & no & 1.551.65 & 0.7 & 200@xmath7140 ngc 5506@xmath9 & ?
& 2.05 & 1.1 & @xmath10 400 ( @xmath10200 ) ngc 7469@xmath8 & yes ? & 2.00 & 1.0 & @xmath10 200 ic 4329a@xmath9 & no & 1.86 & 0.6 & 350@xmath11 mcg-6 - 30 - 15@xmath12 & no & 2.05 & 1.3 & 170@xmath13 ngc 4593@xmath14 & no & 1.87 & 1.1 & @xmath10 150 fairall 9@xmath9 & ? & 2.01 & 0.9 & @xmath10 260 ngc 3516 & no & 1.77 & 1.3 & @xmath10 100 mrk 509 & no & 1.94 & 1.7 & 105@xmath15 mrk 766 & no & 2.15 & ? & ? & & & & @xmath6 piro
1998 @xmath8 preliminary results : piro , private communication @xmath9 perola et al .
1998 ; matt et al .
1998a @xmath12 guainazzi et al .
1998a @xmath14 guainazzi et al .
1998b due to the poorer energy resolution of bepposax with respect to asca , no major advances about the iron line , especially when narrow , are to be expected , even if the broad band may help better estimating the underlying continuum .
when the line is very broad , and the s / n good enough , like in the case of mcg6 - 30 - 15 , the mecs energy resolution is however sufficient to study the profile ; this is demonstrated in fig . 1 , where the best fit model and residuals ( with the line normalization set to zero for the sake of illustration ) are shown . the best fit parameters for the line are consistent with those of tanaka et al . (
1995 ) . in less favourable cases
the line is usually still significantly broad , but the fit can not discriminate between gaussian and relativistic disc profiles ; the relativistic line hypotesis is nevertheless to be preferred on plausibility arguments ( fabian et al .
1995 ) . where bepposax is expected to improve previous observations is in the higher energy part of the spectrum , and in particular in the measurement of the reflection component .
the importance of the broad band in this respect is best illustrated by fig .
2 ( left panel ) where a model with a reflection component included is fitted to the mecs data alone of ngc 5506 ( first observation ) , and then extrapolated to higher energies : the upper panel shows still an excess of counts above 15 kev ( due to the fact that the reflection component is actually not required by the mecs data ) , which is well cured when the mecs and pds data are fitted together ( lower panel ) .
the right panel of fig . 2 gives the probability contours for the power law index , @xmath4 , and the relative normalization of the reflection component , r = @xmath16 , where @xmath17 is the solid angle subtended by the reflecting matter at the illuminating source .
the two parameters are very strongly correlated , and only the very broad energy band permits a good estimate of @xmath18 . from table 1 , it is clear that the reflection component is ubiquitous in seyfert 1s , so confirming the ginga results ( nandra & pounds 1994 ) .
the values of @xmath18 in the table have been estimated assuming a face on slab , and therefore the actual values of @xmath16 may be somwaht higher . assuming a typical error of 3050 percent , the measuread values are consistent with accretion disc reprocessing .
high energy cut
offs , expected in presently popular thermal comptonization models and required to explain the hard x ray background in terms of agn ( e.g. comastri et al . 1995 ) , can also be searched for by bepposax if they are at low enough energies .
for some of the sources in table 1 the results are rather inconclusive , due to the limited s / n . for other sources more stringent lower limits are obtained to the @xmath19folding energy of an exponential cut
off.200 kev are required by the data . ] for a few sources , finally , a positive detection is obtained .
what it is clear from the table is that there is no a universal " cut off energy ( i.e. temperature in thermal models ) , but that the spread in this quantity is significant .
the low energy sensitivity of bepposax down to 0.1 kev , coupled with the wide band which permits a reliable estimate of the intrinsic power law , allows to search for soft excesses .
surprisingly enough , no soft excesses have been detected so far in classical seyfert 1s ( apart maybe from ngc 7469 , this result being however preliminary : piro , private communication ) .
this sharply constrasts with bepposax results on narrow line seyfert 1s , where soft excesses are clearly observed , fully in agreement with both rosat and asca findings ( comastri et al .
1998a , b ; brandt , this volume ) .
also in low redshift quasars soft excesses are usually observed by bepposax ( fiore et al .
1998a ) .
while a quantitative comparison with previous results is still in progress , this promises to be one of the most intriguing bepposax result on seyfert galaxies .
according to unification models , seyfert 2 galaxies are intrinsically identical to seyfert 1s , but are observed through absorbing matter .
indeed , all seyfert 2s detected so far in x
rays show evidence for absorption in excess of the galactic one .
the appearance of a seyfert 2 in x
rays depends strongly on the column density of the absorbing matter .
two regimes may be identified : compton thin and compton thick . in the former case
the absorbing matter has a column density lower than 10@xmath20 @xmath21 , and is therefore optically thin to compton scattering ; the nuclear radiation is then directly visible above a few kev . in the latter case , i.e. when @xmath22 @xmath21 , the absorber is thick to compton scattering : radiation is intercepted and , when @xmath23 @xmath21 , the nucleus is virtually unobservable at all energies ( up to the klein nishina decline , at least ) , because photons are downscattered to energies where photoelectric absorption dominates , and eventually get absorbed . ( for intermediate column densities , a fraction of the nuclear radiation can still escape above @xmath110 kev ; the circinus galaxy , described below , in one example ) .
the obscuration of the nucleus , which in compton
thick sources is complete below 10 kev , permits to study in detail components which would have otherwise been diluted by the nuclear emission . in particular ,
reflection from both cold ( the torus ? ) and ionized ( the optical reflector ? ) matter may become visible .
details of the continuum and line emission expected in compton thick seyfert 2s may be found in e.g. ghisellini , haardt & matt ( 1994 ) , krolik , madau & zicky ( 1994 ) , matt , brandt & fabian ( 1996 ) , netzer ( 1996 ) .
apart from a few scattered observations in the guest observer program , three major core programs on seyfert 2s are ongoing .
the first one , led by l. bassani , is a survey of bright compton thin seyfert 2s .
a summary of the first results can be found in bassani et al .
( 1998a ) . among the five sources observed , two of them ( ngc 7674 , malaguti et al .
1998 ; mrk 3 , cappi et al .
1998 ) turned out to be compton
for the other , genuinely compton
thin sources , broad band bepposax observations confirm the lack of any significant difference in x ray properties ( apart of course from absorption ) with seyfert 1s , in agreement with unification models .
the second and third programs , which i will discuss in more detail , are a survey of bright compton
thick seyfert 2s , and of an optically selected sample of faint seyfert 2s .
two objects have been observed so far in this program : circinus galaxy and ngc 1068 .
the circinus galaxy is probably the nearest active nucleus , its distance being only @xmath14 mpc .
it was observed by bepposax on march 1998 ( matt et al .
the spectrum below 10 kev is consistent with the asca one ( matt et al .
1996 ) , i.e. a reflection component from cold matter ( including a huge iron k@xmath24 line ) and a soft excess of unclear origin .
other lines are also clearly present in the spectrum ( guainazzi et al . , in preparation ) , as evident from fig . 3 , again confirming asca findings . the surprise comes when the best fit lecs+mecs model is extrapolated to the pds band : it dramatically falls short of the data ( fig . 4 , left panel ) .
no known bright sources are present in the pds field of view , so the best interpretation of the spectrum in fig .
4 is that the nuclear emission is piercing through the absorber .
this implies that the column density of the absorber is greater than @xmath110@xmath20 @xmath21 ( to completely block the radiation below 10 kev ) but lower than @xmath110@xmath25 @xmath21 ( to still permit significant transmission ; see ghisellini , haardt & matt 1994 ) .
a similar situation has already been observed in a couple of other sources , namely ngc 4945 ( iwasawa et al .
1993 ; done , madejski & smith 1996 ) and , with bepposax , mrk 3 ( cappi et al . 1998 ) .
in this situation , pure absorber models , like those usually available in xspec , are no longer applicable , unless the covering factor of the absorbing matter is very small . a model which includes
compton scattering is necessary ; we have then constructed one using montecarlo simulations and fitted the data with it .
the resulting column density is about 4@xmath210@xmath20 @xmath21 and the 210 kev unabsorbed nuclear luminosity is @xmath26 erg s@xmath27 .
it was observed around new year 1997 for about 100 ksec ( effective mecs observing time ) , and then again one year later for about 37 ksec
. results from the first observation on the high energy part of the spectrum can be found in matt et al .
the full band analysis of both observations is reported in guainazzi et al .
, in preparation .
the broad band spectrum is rather complex : above about 4 kev the spectrum is dominated by reflection of the nuclear radiation from both cold matter , which dominates above @xmath110 kev , and from warm matter .
the presence of two reflection components was already noted by marshall et al .
( 1993 ) in the bbxrt data , which in fact revealed a complex iron line with both neutral and highly ionized lines , a results later confirmed and improved by asca observations ( ueno et al .
matt , brandt & fabian ( 1996 ) and iwasawa , fabian & matt ( 1997 ) attributed the cold reflection to the inner surface of the torus , and the highly ionized ( he like and h like ) lines to fluorescence and resonant scattering from optically thin ( but possibly thick to resonant absorption ) , photoionized material .
bepposax , thanks to the hard x ray sensitivity provided by the pds , has permitted to confirm this scenario and estimate the relative contributions of the two components ( matt et al .
1997 ; see fig . 5 ) .
[ hard1068 ] below 4 kev , the spectrum is well described by a thermal like component , probably related with the starburst regions ( wilson et al .
1992 ) , plus the warm reflection component . as already known from asca ( ueno et al . 1994
; turner & netzer 1997 ) , the spectrum is very rich in emission lines ( fig .
6 ) . a detailed discussion on the line spectrum can be found in guainazzi et al .
, in preparation
. [ 1068_lines ] long term variability has also been searched for .
the 0.1 - 1 kev lecs light curves are shown in fig . 7 , together with the background
. the left panel refers to the first observation , the right panel to the second one .
a variation between the two observations is apparent , which clearly can not be ascribed to variations in the background .
the gain is also very stable .
unfortunately , however , it is possible that such variations arise from a larger shadowing , in the second observation , from the grid on top of the detector : even if the nominal position of the source in the two observations does not allow for such a change , uncertanties in the aspect reconstruction can not permit to rule out definitely this hypothesis . the variation , if true , would be very important , as it would permit to estimate the size of the emitting region to be at most 1 light year .
the starburst component is very extended ( wilson et al .
1992 ) , and therefore the variation should be ascribed to the warm reflection component .
further long term monitoring of the source will hopefully settle this issue .
the results described above can be easily understood in the framework of unification models .
however , in recent years many works showed that some intrinsic differences bewteen the average properties of seyfert 1s and 2s do exist : enhanced star formation in seyfert 2 galaxies ( maiolino et al .
1997 ) ; different morphologies between galaxies hosting type 1 and 2 nuclei , those hosting type 2 being on average more irregular ( maiolino et al .
1997 , malkan et al .
1998 ) ; a greater dust content in seyfert 2s ( malkan et al .
the aspect angle is clearly not the only relevant parameter .
the next question is , of course , whether the differences involve the nuclear properties or only the environment . in this respect , it is important to observe in hard x rays ( where nuclear activity dominates the emission ) a sample of optically selected sources .
salvati et al .
( 1997 ) and maiolino et al .
( 1998 ) studied with bepposax an oiii flux
limited sample of seyfert 2s , in the assumption that the oiii flux is a good isotropic indicator of seyfert activity ( this is probably not completely true , but it is in any case the best one ) .
the first result of this program is that all sources observed so far have been detected ( 9 out of 9 , risaliti et al .
1998 ) , with typical x ray luminosities exceeding those of normal galaxies .
this is a rather strong indication that _ all seyfert 2 have a type 1 nucleus _ , and then that any difference between the two classes should be searched for in the nuclear environment .
let me suggest a modification to the zero order unification model ( see fig .
8) : all seyferts have a type 1 nucleus plus circumnuclear , @xmath1100 pc scale dust lanes ( malkan et al .
1998 ) , which i presume to be optically thin to compton scattering ( otherwise , too much circumnuclear matter would probably be present ) . only a fraction of seyferts , however , have also the ( compton thick ) molecular torus , which possibly forms preferentially in irregular , disturbed galaxies ( which have also , probably for the same reason , an enhanced star formation activity as well as an overall greater dust content ) . as illustrated in fig . 8 ,
if the nucleus is directly observed , the source is a seyfert 1 . if the line of sight intercepts matter other than the torus ( i.e. a dust lane , or even the galactic disc for highly inclined galaxies ) the source falls in the mixed bag including compton thin seyfert 2 s , intermediate seyfert and nelgs . if , finally , the line of sight intercepts the torus , the source is a compton thick seyfert 2 galaxy . in this scenario ,
seyfert 1s would be preferentially , but not exclusively ( see next section for a seyfert 1 _ with _ the torus ) observed among torus free sources .
it is worth noticing that another results of the program under discussion is that compton
thick sources contributes to a large fraction of the total ( see fig .
moreover , other compton thick sources have been discovered by bepposax ( malaguti et al . , 1998 ; ueno et al . 1998 ; cappi et al .
1998 ) . therefore ,
even if in the proposed scenario the presence of a molecular torus is no longer ubiquitous , it should nevertheless still be rather common ( see also below ) . .
[ matt_um ] one of the aim of the program is to determine the distribution of column densities . as shown in fig .
9 , a large fraction of sources turned out to be compton
however , the sample is biased in favour of high column densities . in fact
, sources have been selected on the basis of the oiii flux , but excluding sources for which good spectra ( from ginga and/or asca ) were already available .
the ginga and asca samples were largely x ray selected , and therefore biased towards low column density ( and therefore large flux ) .
the real n@xmath28 distribution should then be a mixture of that from bepposax and the ones from ginga and asca . collecting all data available in the literature , bassani et al .
( 1998b ) have derived the true " column density distribution : the average value of the column is @xmath13@xmath29 @xmath21 , and the fraction of compton - thick sources is as large as 30% .
[ matt_maio1 ] most of the sources in the maiolino et al .
( 1998 ) sample show evidence for soft x ray emission in excess of the ( absorbed ) hard x
ray emission .
the question is whether this emission is related to nuclear activity ( i.e. scattering of the nuclear radiation from ionized material ) or is rather related to starburst regions . in fig .
10 the soft x
ray luminosities vs. the ir luminosities of the sources in the sample , for which a good estimate of the soft excess is possible , are shown .
the thick solid line indicates the relationship for starburst and normal galaxies ( david et al .
1992 ) ; the shaded area is the 90% confidence limit on this relation . clearly , more data are needed , but at a first glance one would say that in general some nuclear related emission contributes , but not necessarily dominates , the soft x rays .
the spectrum in the left panel of fig . 11 ( from guainazzi et al . 1998c ) closely resembles those of compton thick seyfert 2 galaxies ( see previous section ) , i.e. a very flat continuum and a prominent iron line . and , actually , the interpretation is the same : a reflection dominated spectrum .
the surprising fact is that this is the spectrum of a well known seyfert 1 galaxy , ngc 4051 .
the source was observed by bepposax on 9 - 11 may 1998 , and was caught in an unprecedented low state ( see right panel of fig .
11 ) , which show the bepposax best fit model together with the 1994 asca one ) , 20 times fainter than the average flux .
the source is usually very variable on fairly short time scales , and in the past it has probably been observed at such a low flux level , but usually only for a few thousands seconds , never for a so long time .
the upper limit on the nuclear component implies a dimming of at least a factor 35 with respect to the average value .
we have therefore observed , for the first time , what can be practically considered a switching off of an active nucleus , which has left the reflection component as the only echo of the past activity .
this may also be considered one of the strongest evidence for the presence of substantial circumnuclear matter in seyfert 1s . comparing the bepposax and asca best fit models in the right panel of fig .
11 , it seems that also the reflection component was lower during the bepposax observation , even if this evidence must be taken with great caution .
this could imply that part of the reflection , when the nucleus is active , originates in the accretion disc , which would respond almost simultaneously ( say , within @xmath11 day ) to any variation of the nuclear radiation .
alternatively , if all the reflection comes from a pc scale torus ( or sub pc ; remember that ngc 4051 is one of the less luminous agn , and it is well possible that all distances are scaled down there ) , a fading of the reflection component on time scales of months is expected .
this would have provided a tool to measure " the size of the torus .
we therefore observed the source again about one and half month later .
unfortunately , this time the source did not collaborate : it resumed the normal level of activity , and the spectrum was no longer reflection dominated ( orr et al . 1998 ) .
the good spatial resolution of the mecs instrument onboard bepposax has been exploited to search for serendipitous , hard x ray sources in the field of view of pointed targets ( giommi et al .
one of the aims of this program , called hellas ( high energy llarge area survey ) , is to assess , in a band where its energy density is much greater than in the so far most studied band below @xmath12 kev , the nature of the sources making the x ray background ( xrb ; see barcons 1996 and matt 1995 for recent reviews ) .
the survey presently consists of about 150 sources in the 510 kev energy band ( a fraction of them being detected _ only _ in this band ) over an area of about 50 degrees .
taking properly into account the sky coverage , these numbers translate to a density of @xmath1 20 sources per square degree down to a 510 kev flux of 5@xmath30 erg @xmath21 s@xmath27 , which in turn corresponds to a 35 - 40 percent of the cosmic x - ray background in this band being resolved ( fiore et al . 1998b ) . the next question , of course , is the nature of these sources .
an extensive identification program is in progress , with several night allocated to both northern and southern telescopes .
first runs have already been highly rewarding , as reported in fiore et al .
( 1998b ) .
out of ten sources observed ( for all of them specta were taken of all source , within the mecs error box , down to a r magnitude of 20 ) , 9 of them have been identified .
note that the sources were selected as being the only ones visible from the telescope at that time of the year , and therefore there is no astrophysical bias whatsoever .
8 sources were identified with agns .
( the eight source is identified with a liner , therefore a borderline object ) .
the corresponding chance probability is only 0.15% .
it is worth noticing that the averege spectrum of these sources is consistent with that of the xrb .
therefore , from all the pieces of evidence just summarized , one is authorize to infer that at least a large fraction of the xrb in this band is made by agn . of the 8 agn identified ,
3 are normal , blue " qso , 2 are red " qso , and 3 are intermediate ( 1.81.9 ) seyferts , with redshifts ranging from 0.17 to 1.3 . four sources ( the three intermediate seyferts and one of the red quasars ) show evidence for x ray absorption with column densities in excess of 10@xmath31 @xmath21 . these findings therefore support models in which the hard xrb is made by a mixture of obscured and unobscured agn ( comastri et al .
1995 , and references therein ) .
i thank all my collaborators in the various bepposax observing programs described in this paper . in particular , i thank f. fiore , m. guainazzi , r. maiolino , f. nicastro , and g.c .
perola for many helpful discussions , and l. piro for allowing me to quote results before publication .
i am indebted to the bepposax science data center for frequent help in the data reduction .
i acknowledge financial support from asi and murst .
antonucci r. , 1993 , ara&a 31 , 473 barcons x. , 1996 , memorie della s.a.it .
67 , n. 1/2 , 481 bassani l. , et al . , 1998a , in from nano to tera
all the colours of agn " , in press bassani l. , et al . , 1998b , apj , submitted boella g. , et al .
, 1997a , a&ass , 122 , 299 cappi m. , et al . , 1998 ,
a&a , submitted comastri a. , setti g. , zamorani g. , hasinger g. , 1995 , a&a , 296 , 1 comastri a. , et al .
, 1998a , a&a 333 , 31 comastri a. , et al .
, 1998b , in the active x ray sky `` , p. 523
david l.p . ,
jones c. , forman w. , 1992 , apj , 388 , 82 done c. , madejski g.m . ,
smith d.a . , 1996 ,
apj , 463 , 63 fabian a.c . ,
et al . , 1994 , pasj , 46 , 137 fabian a.c . , et al . , 1995 , mnras , 277 , l11 fabian a.c . , 1996 , in rntgenstrahlung in the universe '' , mpe report 263 , p. 403
fiore f. , et al .
, 1998a , in the active x ray sky " , p. 529 fiore f. , et al .
, 1998b , in the extreme universe " , in press george i.m . , fabian a.c . 1991 , mnras , 249 , 352 george i.m . , et al . , 1998 , apjs , 114 , 73 ghisellini g. , haardt f. , matt g. , 1994 , mnras , 267 , 743 giommi et al . , 1998 , in the active x ray sky " , p. 591
guainazzi m. et al .
, 1998a , a&a , submitted guainazzi m. et al .
, 1998b , a&a , submitted guainazzi m. et al . , 1998c , mnras , in press halpern j.p . , 1984 , apj , 281 , 90 iwasawa k. , et al . , 1993 , apj , 409 , 155 iwasawa k. , fabian a.c . ,
matt g. , 1997 , mnras , 289 , 443 krolik j.h .
, madau p. , ycki p.t .
, 1994 , apj , 420 , l57 krolik j.h , kriss g.a .
1995 , apj , 447 , 512 lightman a.p .
, white t.r . , 1988 ,
apj , 335 , 57 magdziarz p. , zdziarski a.a . , 1995 , mnras , 273 , 837 maiolino r. , ruiz m. , rieke g.h . , papadopoulos p. , 1997 ,
apj , 485 , 552 maiolino r. , et al . , 1998 , a&a , 338 , 781 malaguti g. , et al . , 1998 , a&a , 1998 , 331 , 519 malkan m.a .
, gorjian v. , tam r. , 1998 , apjs , 117 , 25 marshall f.e .
, et al . , 1993 , apj , 405 , 168 matt g. , perola g.c . , piro l. , 1991 , a&a , 247 , 25 matt g. , 1994 , mnras , 267 , l17 matt g. , 1995 , in proc . of the vulcano workshop 1994 , f. giovannelli and g. mannocchi eds .
, bologna : editrice compositori , 163 matt g. , brandt w.n .
, fabian a.c . , 1996 , mnras 280 , 823 matt g. , et al . , 1996 , mnras , 208 , 253 matt g. , et al . , 1997 ,
a&a , 325 , l13 matt g. , et al .
, 1998a , in the extreme universe " , in press matt g. , et al .
, 1998b , a&a , submitted mushotzky r.f . ,
done c. , pounds k. a. , 1993 , ara&a , 31 , 717 nandra k. , pounds k.a . , 1992 ,
nat , 359 , 215 nandra k. , pounds k.a . , 1994 , mnras 268 , 405 nandra k. , george i. m. , mushotzky r. f. , turner t. j. , yaqoob t. , 1997 , apj , 477 , 602 netzer h. , 1996 , apj , 781 , 796 nicastro f. , fiore f. , matt g. , 1998 , apj , submitted orr a. , et al , in proc .
of the 32nd cospar scientific assembly , in press perola g.c . , et al . , 1998 , in
from nano to tera
all the colours of agn " , in press piro l. , et al . , 1998 , in the active x ray sky " , p. 481 reynolds c.s . , 1997 , mnras , 287 , 513 risaliti g. , et al . , 1998 , in
from nano to tera
all the colours of agn " , in press ross r.r .
, fabian a.c . , 1993 , mnras , 261 , 74 salvati m. , et al . , 1997 , a&a , 323 , l1 svensson r. , 1996 , a&as , 120 , 475 tanaka y. , et al . , 1995 , nat , 375 , 659 turner t.j . , netzer h. , 1997 , apj , 488 , 694 ueno s. , et al . , 1994 ,
pasj 46 , l71 ueno s. , et al .
, 1998 , in the active x ray sky " , p. 554
wilson a.s . , elvis m. , lawrence a. , bland
hawthorn j. , 1992 , apj , 391 , l75 | bepposax results on seyfert galaxies are reviewed and discussed . |
cooperative diversity is a highly promising technique for coverage extension and reliability improvement of wireless networks .
it exploits the additional degrees of freedom of the fading environment , which is introduced by the spatially distributed multiple relays utilizing either amplify - and - forward ( af ) or decode - and - forward ( df ) relaying . proposed recently in @xcite , the opportunistic
relaying is a simple yet efficient cooperative diversity protocol , whose diversity - multiplexing tradeoff is identical to that of the more complex distributed space - time coding cooperative schemes @xcite . by selecting a single best " relay among the all available relays ,
the opportunistic relaying achieves full spatial diversity while maintaining the spectral efficiency of a two - hop communication link .
the outage and error probabilities of the opportunistic relaying systems have been studied in @xcite-@xcite , which clearly demonstrate its excellent performances . however , there are some design issues for which the outage and error probabilities criteria are not sufficient , such as , packet or slot lengths and latencies , switching rates @xcite , power and bandwidth allocation or , decision criterion for changing adaptive modulation levels
. these issues can be addressed by investigating the system s second - order outage statistics . to the best of authors knowledge ,
such statistics that describe the outage events of cooperative systems , such as , the average outage rate ( aor ) and average outage duration ( aod ) , have not been studied previously .
we propose the aor and the aod be defined with respect to the capacity outage events derived from information - theoretic capacity of the opportunistic system .
similar definition of the outage statistics has been applied over mimo systems in @xcite . in this letter , we derive exact expressions for the aor and the aod of opportunistic systems , employing either af or df relaying in rayleigh fading environment .
similarly to @xcite , we consider a typical half - duplex dual - hop communication scenario , where the communication between the source @xmath0 and the destination @xmath1 is possible only via @xmath2 relays ( denoted by @xmath3 , @xmath4 ) , as the direct path is assumed blocked by an intermediate wall . in the beginning of each slot ( divided into two equal sub - slots ) , a single best " opportunistic relay is selected out of the @xmath2 possible dual - hop paths for relaying the communication between @xmath0 and @xmath1 . during the first sub - slot , @xmath0 transmits its signal over the first hop , while the selected relay forwards that signal toward @xmath1 over the second hop during the second sub - slot .
the channel is exposed to rayleigh fading and is assumed to remain constant during the entire slot duration . without loss in generality , we assume that @xmath0 and the selected
best " relay transmit with equal powers @xmath5 , rendering the total available transmission power to @xmath6 . denoting the rayleigh - faded channel gains of hops @xmath7 and @xmath8 during a given slot @xmath9 by @xmath10 and @xmath11 , the received signal - to - noise ratios ( snrs ) at the relay @xmath3 and at the destination @xmath1
are expressed as @xmath12 and @xmath13 , with @xmath14 as the noise power .
specifying the average squared channel gains as @xmath15 = \omega_{sk}$ ] and @xmath16 = \omega_{kd}$ ] , the average received snrs at @xmath3 and @xmath1 are respectively given by @xmath17 and @xmath18 .
the relay selection is based on the estimation of the end - to - end performance over the dual - hop path @xmath19 using the _ selection variable _ @xmath20 , which is estimated separately by each relay @xmath3 in the beginning of each slot from its channel state information ( csi ) @xcite . to facilitate channel state estimation by the relays , @xmath0 and @xmath1 previously exchange short control packets . in the beginning of slot @xmath9 , the best
" relay @xmath21 is selected in a distributed manner by using the _ selection policy _ : @xmath22 in the beginning of slot @xmath9 , each df relay estimates the selection variable @xmath23 which actually evaluates the minimal instantaneous received snr among the two hops , @xmath24 .
we assume fixed gain af relays that amplify the received signal from the first hop by @xmath25 and forward it to @xmath1 over the second hop @xcite . in the beginning of slot @xmath9 , each af relay estimates the selection variable @xmath26 where @xmath27 .
actually , each af relay evaluates the dual - hop snr that is relayed over @xmath3 and received at @xmath1 .
note that fixed - gain ( e.g. , semi - blind ) af relays have considerably simpler but yet comparably close performance to that of the variable gain af relays @xcite .
we consider 2-dimensional isotropic scattering around source @xmath0 , relays @xmath3 and destination @xmath1 , all of which are assumed to be mobile and have no line - of - sight with other stations .
thus , each @xmath7 ( @xmath8 ) hop behaves as a mobile - to - mobile rayleigh channel .
its channel gain , @xmath28 ( @xmath29 , is a time - correlated rayleigh random process with known statistical properties ( e.g. , the doppler spectrum ) @xcite .
if a station at one end of a hop is fixed , the mobile - to - mobile rayleigh - fading hop is transformed into the classic " fixed - to - mobile rayleigh - fading hop @xcite .
the time derivative @xmath30 ( @xmath31 ) is independent from the channel gain @xmath32 ( @xmath33 ) , and follows the gaussian probability distribution function ( pdf ) with zero mean and variance [ [ ref7 ] , eq .
( a5 ) ] [ [ ref10 ] , eq . ( 39 ) ] @xmath34 in ( [ 3a])-([3b ] ) , @xmath35 , @xmath36 and @xmath37 denote the maximum doppler rates of @xmath0 , @xmath1 and relay @xmath3 @xmath38 , respectively .
the doppler rate ( and consequently the aor ) is expressed in the unit @xmath39 . expressing the slot duration in seconds , both the doppler rate ( now becoming the doppler frequency ) and the aor
are expressed in hz . in a given slot @xmath9 ,
the opportunistic relaying system experiences a capacity outage event when the mutual information of the dual - hop path over the best " relay @xmath21 drops below some predefined spectral efficiency @xmath40 @xcite , @xmath41 where @xmath42 thus , the resulting time - varying capacity @xmath43 suffers from the random occurrence of capacity outage events , during which the channel is unable to support the specified @xmath40 . transforming ( [ 4 ] ) , the capacity outage event at slot @xmath9 occurs if @xmath44 where the outage threshold is @xmath45 .
note , @xmath46 can be varied by varying @xmath40 ( as in @xcite ) or @xmath47 ( as in this work ) , but the functional dependencies of the increasing @xmath40 or the decreasing @xmath47 ( in db ) have almost same shapes .
using a similar approach to that presented in @xcite for deriving level crossing rates ( lcr ) of classic " selection diversity systems , the joint pdf of @xmath48 and @xmath49 is expressed as @xmath50 where @xmath51 denotes the joint pdf of selection variable @xmath20 and its time derivative @xmath52 . assuming @xmath3
is selected best " relay , @xmath53 denotes the conditional probability that the selection variable @xmath54 of @xmath3 drops below @xmath55 , @xmath56 where @xmath57 denotes the cumulative distribution function ( cdf ) of @xmath58 .
the aor is evaluated based on the standard lcr definition [ [ ref6 ] , chapter 1 ] , yielding @xmath59 where @xmath60 denotes the aor of the dual - hop path over the relay @xmath3 .
the aod is then given by @xmath61 we now focus on the random process @xmath62 , defined by ( [ 1 ] ) .
communication through the relay @xmath3 falls in outage if either one of the two hops fail .
thus , the pdf of @xmath62 is given by @xmath63 since both hops follow the rayleigh pdf , @xmath64 is also determined to follow the rayleigh pdf , @xmath65 with @xmath66 , and the respective cdf @xmath67 .
using @xmath68 and the independence of the channel gains and their respective time derivatives , the required joint pdf is found as @xmath69 in rayleigh fading , ( [ 14 ] ) specializes to @xmath70 thus rendering @xmath54 and @xmath71 as independent rvs , where @xmath72 is given by ( [ 12 ] ) and @xmath73 with @xmath74 and @xmath75 denoting zero mean gaussian pdfs with variances @xmath76 ( [ 3a ] ) and @xmath77 ( [ 3b ] ) , respectively .
thus , the aor of dual - hop path over @xmath3 is obtained as @xmath78 with @xmath79 given by ( [ 12 ] ) . inserting ( [ 17 ] ) into ( [ 9 ] ) and ( [ 10 ] ) , we obtain the aor and the aod of df relaying system . the exact expressions for the cdf and the aor of the random process @xmath80 , defined by ( [ 2 ] ) , are respectively given by [ [ ref9 ] , eq .
( 9 ) ] and [ [ ref10 ] , eq . ( 19 ) ] , as @xmath81 where @xmath82 is the first - order modified bessel function of the second kind .
note that ( [ 19 ] ) can be efficiently and accurately evaluated by applying the gauss - hermite quadrature rule [ [ ref11 ] , eq .
( 25.4.46 ) ] . combining ( [ 18 ] ) and ( [ 19 ] ) into ( [ 9 ] ) and ( [ 10 ] ) , we obtain the aor and the aod of af relaying system .
note , if source @xmath0 and destination @xmath1 are fixed , the approach presented in this section can be applied to derive analogous analytic expressions for aors and aods for more general fading channels ( such as , rice and nakagami-@xmath83 models ) , because such fixed - to - mobile hops have known second - order statistical properties .
in this section , we present illustrative examples for the normalized aor ( fig .
1 ) an the normalized aod ( fig .
2 ) in function of @xmath47 of opportunistic relaying system employing either df or af relays . the source @xmath0 and the destination @xmath1 are fixed ( @xmath84 ) , whereas all the relays are mobile and introduce same maximum doppler rates @xmath85 , @xmath86 .
the aor and the aod are normalized with respect to the doppler rate @xmath87 as @xmath88 and @xmath89 .
the average squared channel gains of all hops are equal to @xmath90 ( i.e. , @xmath91 , @xmath92 ) , thus rendering total available transmission power equal to @xmath5 .
the spectral efficiency is set to @xmath93 bps / hz . the monte carlo simulations clearly validate our derived analytical results .
1 a. bletsas , a. khisti , d. p. reed and a. lippman , a simple cooperative diversity method based on network path selection , " _ ieee j. select . areas .
commun . _ , vol .
3 , pp . 659 - 672 , mar .
2006 j. n. laneman and g. w. wornell , distributed space - time coded protocols for exploiting cooperative diversity in wireless networks , " _ ieee trans .
inform . theory _
10 , pp . 2415 - 2525 , oct .
2003 [ ref3 ] a. bletsas , h. shin , and m. z. win , cooperative communications with outage - optimal opportunistic relaying , " _ ieee trans .
wireless commun .
_ , vol . 6 , no .
9 , pp . 3450 - 3460 , sept .
2007 d. s. michalopoulos and g. k. karagiannidis , performance analysis of single relay selection in rayleigh fading " , _ ieee trans .
wireless commun .
_ , vol . 7 , no . 10 , pp
. 3718 - 3724 , oct .
2008 k .- s .
hwang , y .- c .
ko , and m .- s .
alouini , outage probability of cooperative diversity systems with opportunistic relaying based on decode - and - forward " , _ ieee trans .
wireless commun .
_ , vol . 7 , no . 12 , dec .
2008 d. michalopoulos , a. lioumpas , g. k. karagiannidis and r. schober , selective cooperative relaying over time - varying channels " , _ submitted to ieee trans .
_ , http://arxiv.org/abs/0905.0564v1 b. o. hogstad , m. patzold , n. youssef , v. kontorovitch , exact closed - form expressions for the distribution , the level - crossing rate , and the average duration of fades on the capacity of ostbc - mimo channels " , _ ieee trans . veh .
2 , pp . 1011 - 1016 , feb .
2009 [ ref6]w .
c. jakes , microwave mobile communications , 2nd ed .
piscataway , nj : _ ieee press _ , 1994 .
s. akki and f. haber , a statistical properties of mobile - to - mobile land communication channel , " _ ieee trans . veh .
826 - 831 , nov .
1994 x. dong and n. c. beaulieu , average level crossing rate and average fade duration of selection diversity " , _ ieee commun .
letters _ , vol . 5 , no . 10 , oct .
2001 [ ref9]m . o. hasna and m
alouini , a performance study of dual - hop transmissions with fixed gain relays " , _ ieee trans .
wireless commun .
3 , no . 6 , pp .
1963 - 1968 , nov .
2004 [ ref10]c.s .
patel , g.l .
stuber and t.g .
pratt , statistical properties of amplify and forward relay fading channels , " _ ieee trans . veh .
1 , pp . 1 - 9 ,
[ ref11]m .
abramowitz and i. a. stegun , _ handbook of mathematical functions with formulas , graphs , and mathematical tables _
new york : dover , 1970 | opportunistic relaying is a simple yet efficient cooperation scheme that achieves full diversity and preserves the spectral efficiency among the spatially distributed stations .
however , the stations mobility causes temporal correlation of the system s capacity outage events , which gives rise to its important second - order outage statistical parameters , such as the average outage rate ( aor ) and the average outage duration ( aod ) .
this letter presents exact analytical expressions for the aor and the aod of an opportunistic relaying system , which employs a mobile source and a mobile destination ( without a direct path ) , and an arbitrary number of ( fixed - gain amplify - and - forward or decode - and - forward ) mobile relays in rayleigh fading environment .
average outage rate , average outage duration , opportunistic relaying , doppler effect , rayleigh fading |
for almost as long as they have been known to be located outside the milky way , galaxies have been used to probe the geometry of the universe ( @xcite ) . fairly early in the history of relativistic cosmology
, it was realized that if the number density of galaxies is known , their apparent numbers per solid angle and redshift interval depend directly on basic cosmological parameters ( @xcite ) . applying this `` d@xmath0d@xmath1 test ''
requires a substantial sample of distant galaxies with measured redshifts ; it was first applied by loh and spillar ( 1986 ) using a set of 406 galaxies with estimated photometric redshifts @xmath6 between 0.15 and 0.85 . assuming that the comoving number density of galaxies is constant and that their luminosity function retains similar shape over that redshift range
, they measured @xmath7 .
however , it remains unclear whether the assumptions of loh and spillar hold ; no complete theory or simulation of galaxy formation and evolution as yet exists ( @xcite ) .
it would thus be preferable to apply such a test to objects whose abundance can be calculated semi - analytically or via computer models , eliminating such ambiguities . in this paper , we suggest that the observed numbers of galaxies ( and thus indirectly dark halos ) as a function of linewidth and redshift provide a candidate for improving on the analysis of loh & spillar , particularly because such quantities will be measured by surveys occurring in the next decade . in the process of studying this possibility , we have developed a semi - analytic method for determining the abundance of dark halos as a function of their circular velocity ( @xmath8 ) and redshift that can provide results similar to those of n - body simulations in considerably less computer time and without the effects of limited resolution .
in essence , we begin with the abundances of galaxies as a function of mass predicted by the modified press - schechter formalism of sheth ( 1999 ) , which provides an excellent fit to dark - matter simulations .
we then determine the probability that a halo of given mass has formation time such that it will virialize with the desired circular velocity using the semi - analytic method of lacey and cole ( 1994 ) , which again is very well reproduced by simulations . by integrating over the distribution in mass
we may then determine the total abundance of objects with a given circular velocity at the epoch of interest .
we describe this method in more detail in @xmath92 , and the prospects for our proposed test in @xmath93 .
a number of semi - analytic methods for calculating the veloctity distribution of dark halos based upon the press - schechter ( @xcite ) formalism have been proposed in the past .
narayan & white ( 1988 ) used the assumption that halos have just virialized at the epoch of observation to map the press - schechter distribution of halos in mass into a distribution in velocity for a @xmath10 cold dark matter ( cdm ) universe ; kochanek ( 1995 ) extended their analysis to other common cosmological models .
however , the assumption that objects observed are only just virializing is clearly inappropriate in many cases , particularly for lower - mass halos ; once objects virialize and break away from the cosmological expansion , their circular velocities evolve little if at all save in major mergers ( ghigna 1999 ) . to account for this ,
kitayama and suto ( 1996a ) proposed that the distribution of x - ray temperatures of clusters ( @xmath11 ) be calculated by totalling the rate at which objects of the desired temperature that survive without merging into a much more massive object through the epoch of interest form , using the results of lacey & cole ( 1993 ) .
as the authors of that paper noted , there remain some ambiguities in their formulation , however .
we have attempted to improve upon these prior methods for calculating the velocity distribution of dark matter halos , taking advantage of the progress made over the past few years in extending the press - schechter formalism .
it has been repeatedly demonstrated that standard press - schechter calculations overpredict the abundance of low - mass halos compared to equivalent simulations ; a number of remedies have been suggested recently ( e.g. bond & myers 1996 ) .
one of the simplest to implement is that of sheth ( 1999 ) , which accounts for the average ellipticities of dark matter halos by making the critical overdensity for their collapse mass - dependent .
the resulting analytic expression for the number density of halos of given mass ( given by their equations 1 and 6 ) provides an excellent fit to the results of a variety of cosmological simulations .
the results of sheth allow a prediction of halo abundances much improved over the standard press - schechter formula with only minimal complications to the calculation .
we have therefore used their equations to determine the distribution of halos in mass at a given redshift in our work . because the density of the universe was greater at early times , objects that virialized at high redshift are more compact than those that virialized more recently , and therefore have a higher circular velocity for their mass .
furthermore , once it has collapsed , the circular velocity of a halo changes only minimally ( ghigna 1999 ) .
therefore , in determining the abundance of galaxy - scale halos with a given circular velocity today , it is preferable to take into account the time of formation of those halos in addition to their mass .
fortunately , this may be calculated ; lacey & cole ( 1993 ) derived a semi - analytic formula for the distribution of formation times of halos with a given mass that exist as virialized objects at a given redshift ( defining the formation time as the epoch when the object first had at least half its final mass ) .
this formula provides an excellent fit to the results of simulations ( lacey & cole 1994 ) , even though it assumes spherical collapse while the simulated halos could be ellipsoidal .
we therefore have used their semi - analytic formula ( equation 2.19 of lacey & cole 1994 ) to determine @xmath12 , the probability distribution of the epoch of collapse @xmath13 of halos observed to have mass @xmath14 at time @xmath15 .
note that the integral defining @xmath16 must be evaluated with particular care where the integration variable @xmath17 .
however , an approximate analytic solution exists in that region , and the integral may be safely performed over the remainder of its range numerically . in practice
, we are interested in calculating the abundance of halos at redshift @xmath1 that have circular velocity @xmath18 ; for each halo mass , we therefore need only evaluate @xmath16 at that formation epoch for which a halo of the given mass would virialize with the desired circular velocity .
we may obtain that time from the relationship between the circular velocity and the mass @xmath14 , which is most conveniently parameterized by using the length scale @xmath19 ( @xmath20mpc ) @xmath21 , where @xmath22 is the present - day mass density of the universe . by combining the definitions of circular velocity , of @xmath19 , of @xmath23 ( the ratio of the mean density of a virialized halo to the critical density of the universe when it virialized ) , and of the dimensionless evolution of the hubble parameter @xmath24 we find ( cf .
narayan & white 1988 , kochanek 1995 ) : @xmath25 where we have determined the circular velocity for a halo of mass @xmath26 , more appropriate given our definition of the formation time .
the objection might be raised that we are using a formula for circular velocity that implicitly assumes spherical geometry but are applying it to potentially ellipsoidal dark matter halos ; however , lacey & cole ( 1996 ) found that the velocity dispersions of dark matter halos calculated using a singular isothermal sphere model ( for which @xmath27 ) matched the properties of n - body simulated halos quite well . in our calculations
we have used the fitting formulae for @xmath23 given by bryan and norman ( 1997 ) .
thus , to calculate the comoving abundance of halos with a given circular velocity at an epoch of interest , we simply calculate the distribution of halos in mass on a 500-element grid , determine the probability that those halos have the desired circular velocity , and then integrate numerically over that grid .
this procedure takes @xmath28 sec on a modern workstation to calculate the abundance at a single circular velocity and epoch , far faster than simulations could be run , and has none of the limitations of resolution or sample size that affect even the largest n - body simulations today .
our technique seems to be similar to one employed by kitayama and suto ( 1996b ) to check their methods , save for the use of the ellipsoidal - collapse mass function rather than a traditional press - schechter calculation ; however , details on the latter are minimal .
our procedures could certainly be improved upon .
for instance , the connection between circular velocity , formation time , and mass could be made using a more realistic model than a singular isothermal sphere ( @xcite , bullock 1999 ) .
calculation of formation times using an ellipsoidal - halo abundance paradigm rather than the press - schechter method of cole and lacey might also yield modest improvements .
finally , we must note that the mass function of sheth was based upon a fit to mass functions of halos identified in n - body simulations using the `` friends - of - friends '' algorithm with selection parameter @xmath29 ; in such a method , as in standard press - schechter calculations , smaller subhaloes within a more massive halo ( e.g. galaxies within a cluster ) are not counted separately in the mass function . unless an improved method taking this into account is used
, we can only calculate the abundance of relatively * isolated * dark matter halos ( i.e. , those that are not part of a larger , virialized group or cluster ) . such objects are readily identified observationally in well - sampled redshift surveys , but this is still a significant limitation . despite these caveats , we believe that our technique for semi - analytic calculations provides a substantial advance over methods previously used ; all of them suffer from similar flaws and have fewer advantages .
our methods may also be useful for a variety of cosmological calculations ; for instance , the circular velocity distribution of dark halos is required for calculations of the abundance of strong gravitational lenses , and hence for the limits placed on @xmath30 thereby .
in order to study the evolution of the volume element of the universe , we require some identifiable tracer whose number density evolution is well known . the dark halos which contain galaxies are an obvious candidate ; the results of semi - analytic calculations ( e.g. , sheth 1999 ) and n - body simulations ( e.g. , @xcite ) for the behavior of dark matter are by now fairly well understood .
of course , directly observing the properties of dark halos at high redshift is impossible ; however , some of their parameters may be determined by studying galaxies lying within them . in particular , it is well - known that the motion of material in the outer parts of typical galaxies is dominated by the gravitational influence of dark matter ( fischer 1999 ) ; the depth of the halo potential well may thus be directly derived from observed linewidths or velocity dispersions of galaxies . unlike mass measurements , such observations are relatively insensitive to distance from a galaxy s center ( for instance , the rotation curves of many spiral galaxies are observed to be fairly flat over a great range of radii , as would be predicted for halos which are roughly isothermal spheres ) . under modest assumptions ( e.g. about their radial profiles ) , the expected abundance of dark matter halos as a function of observable linewidth may be determined from simulations or from the semi - analytic method described in @xmath92 .
results of our semi - analytic calculations for the evolution of the differential comoving abundance of isolated halos with circular velocity 200 @xmath31 in some commonly - used cosmological models are depicted in fig .
a cdm - like power spectrum with shape parameter @xmath32 and normalized to match the value of @xmath33 derived by borgani ( 1999 ) from x - ray observations of galaxy clusters was employed in all cases . as it is likely that selection effects will be somewhat similar for galaxies at @xmath34 and @xmath35 , we plot the ratio of the abundance of galaxies at a given redshift to that at @xmath34 .
the abundance of halos at low circular velocities is smaller now than in the recent past because of merging of smaller halos into new , larger objects and because smaller halos lose their individual identities in a press - schechter framework when they are included in a larger virialized object such as a group or cluster .
qualitatively , the predictions of n - body simulations are similar ( see , e.g. , the cluster simulations of ghigna 1999 ) .
our calculations demonstrate that the comoving abundance of halos with fixed linewidth at @xmath35 relative to today is very insensitive to the background cosmological model under reasonable assumptions about the matter power spectrum ; e.g. for models with a low matter density @xmath36 , the relative abundance is only @xmath37 different in a flat model from an open one .
thus , observations of the apparent numbers of such objects per unit redshift per steradian in comparison to their abundance today directly yields a measurement of the cosmological volume element , and hence of fundamental cosmological parameters
. a measurement of @xmath38 will be quite feasible once the deep redshift survey ( @xcite ) is completed .
this survey , to be begun in early 2001 and completed within the five years following , will obtain spectra of @xmath39 galaxies at redshift @xmath40 with a resolution of @xmath41 @xmath31 fwhm , allowing measurements of linewidths for perhaps a quarter to half of the galaxies in the sample .
the number of spectra obtained will be great enough to test for selection effects in redshift or linewidth by comparison to calculations or simulations at the few percent level .
for instance , smaller halos are predicted to be distributed in circular velocity according to a power - law with a slope well - determined by both n - body simulations and our techniques ; deviation from such a distribution would be a strong indicator of observational selection effects , allowing us to correct for or avoid regions of parameter space suffering from incompleteness .
complementary information on the linewidth distribution of galaxies at low redshift should be available from the sloan digital sky survey ( @xcite ) , 2df ( @xcite ) , or other local surveys , allowing accurate normalization to the abundance of halos at @xmath34 .
the determination of the volume element afforded by such measurements can place constraints not only on the simplest models of the universe which include only matter and a cosmological constant , but also on so - called `` quintessence '' models ( @xcite ; @xcite ) in which the equation of state of a non - material component , @xmath42 , can take arbitrary value between -1 and 1 ( @xmath43 for matter , while @xmath44 corresponds to a cosmological constant ) .
the amount of comoving volume per unit redshift per steradian at @xmath35 depends strongly on @xmath3 , as illustrated in fig .
2 . the test we propose in combination with a determination of @xmath5 ( e.g. from velocity statistics of galaxies ) should allow direct determination of the equation of state of any dark energy present , as illustrated in figure 3 . in comparison ,
cmb measurements planned for the near future and ground - based observations of the apparent magnitudes of type ia supernovae can determine @xmath3 to @xmath45% at best ( hu 1999 ) ; similar accuracy might be achieved using the statistics of lensed sources in the sloan digital sky survey ( @xcite ) .
our estimated constraints are similar in strength to those predicted for the proposed snap satellite ( perlmutter et al . 1999 ) , though less sensitive to the present - day curvature and more sensitive in the orthogonal direction ( i.e. , more complementary to cmb measurements ) .
the technique we propose is at least as effective as these even without including complementary information from other tests . as fig .
3 illustrates , since @xmath5 will be increasingly constrained from other experiments in the near future ( including other analyses of the deep dataset ) , the test we propose has the potential to determine @xmath3 to @xmath46 , even without other complementary measurements .
if the deep survey suffers from incompleteness , there must be at least as many halos of a given linewidth as we observe ; thus , a lower limit on the volume element , and hence on @xmath47 , would be provided .
the test we propose could also detect evolution of @xmath3 , if any exists and other cosmological parameters are well - determined .
it has recently been proposed that in reasonable quintessence models @xmath3 may vary with the current matter density as the universe evolves ( so - called `` tracker solutions '' ; cf .
steinhardt 1998 ) or even oscillate over time ; such changes would be quite difficult to measure via methods that study integrated quantities ( e.g. , the luminosity distances of sne ia ) .
however , the test proposed here allows measurement of a more localized quantity , the volume element ; hence , it may be able to measure changes in that quantity even on relatively small scales . by studying the apparent abundances of the dark halos of galaxies , we may test the most recent cosmological models or even those not yet envisioned with techniques having their origin in the earliest days of extragalactic astronomy .
we would like to thank sandy faber for her helpful suggestions and james bullock , eric gawiser , dragan huterer , james robinson , rob thacker , mike turner , martin white , and the anonymous referee for useful suggestions and discussions .
we would also like to thank zoltan haiman and jay mohr for their assistance in including their results in this paper .
this work was supported by nasa grant nag 5 - 7833 .
benson , a.j . , pearce , f.r .
, frenk , c.s . ,
baugh , c.m . , & jenkins , a. 1999 , preprint ( astro - ph/9912220 ) bond , j.r . & myers , s.t .
1996 , apjs , 103 , 1 borgani , s. , plionis , m. , & kolokotronis , v. 1999 , mnras , 305 , 866 bryan , g.l . & norman , m.l .
1996 , apj , 495 , 80 bullock , j. s. 1999 , preprint ( astro - ph/9910301 ) caldwell , r. , dave , r. , & steinhardt , p. 1997
, ap&ss , 261 , 303 colless , m. 1998 , in _ wide field surveys in cosmology _ , ed . s. colombi , y. mellier , & b. raban ( paris : editions frontieres ) , 77 cooray , a. r. & huterer , d. 1999 , apj , 513 , l95 davis , m. & faber , s. 1998 , in _ wide field surveys in cosmology _ , ed .
s. colombi , y. mellier , & b. raban ( paris : editions frontieres ) , 333 fischer , p. 1999 , preprint ( astro - ph/9912119 ) ghigna , s. , moore , b. , governato , f. , lake , g. , quinn , t. , & stadel , j. 1999 , preprint ( astro - ph/991066 ) haiman , z. , mohr , j. j. , & holder , g. p. 2000 , preprint ( astro - ph/0002336 ) hu , w. , eisenstein , d. , tegmark , m. , & white , m. 1998 , preprint ( astro - ph/9806362 ) hubble , e. 1926 , apj , 64 , 321 jenkins , a. 1997 , in asp conf .
117 , dark and visible matter in galaxies , ed .
m. persic & p. salucci ( san francisco : asp ) , 348 kochanek , c.s . , 1995 , apj , 453 , 545 kitayama , t. & suto , y. 1996a , mnras , 280 , 638 kitayama , t. & suto , y. 1996b , apj , 469 , 480 lacey , c. & cole , s. 1993 , mnras , 262 , 627 lacey , c. & cole , s. 1994 , mnras , 271 , 676 loh , e.d .
& spillar , e.j .
1986 , apj , 307 , l1 loveday , j. , 1998 , in _ wide field surveys in cosmology _ , ed . s. colombi , y. mellier , & b. raban ( paris : editions frontieres ) , 317 melchiorri , a. 1999 , preprint ( astro - ph/9911445 ) narayan , r. & white , s.d.m .
1988 , mnras , 231 , 97 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1995 , mnras , 275 , 720 peebles , p.j.e . 1999 , _ principles of physical cosmology _
( princeton : princeton university press ) perlmutter , s. 1999 , apj , 517 , 565 press , w.h .
& schechter , p. 1974 , apj , 187 , 425 sheth , r.k . , mo , h.j . , & torment , g. 1999 , preprint ( astro - ph/9907024 ) steinardt , p.j . ,
wang , l. , & zlatev , i. 1999 , preprint ( astro - ph/9812313 ) tolman , r. c. 1934 , _ relativity thermodynamics and cosmology _ ( oxford : clarendon press ) turner , m. s. & white , m. 1997 , phys .
d , 56 , 4439 | the classical d@xmath0d@xmath1 test allows the determination of fundamental cosmological parameters from the evolution of the cosmic volume element .
this test is applied by measuring the redshift distribution of a tracer whose evolution in number density is known . in the past , ordinary galaxies have been used as such a tracer ; however , in the absence of a complete theory of galaxy formation , that method is fraught with difficulties . in this paper , we propose studying instead the evolution of the apparent numbers of dark matter halos as a function of their circular velocity , observable via the linewidths or rotation speeds of visible galaxies .
upcoming redshift surveys will allow the linewidth distribution of galaxies to be determined at both @xmath2 and the present day .
in the course of studying this test , we have devised a rapid , improved semi - analytic method for calculating the circular velocity distribution of dark halos based upon the analytic mass function of sheth ( 1999 ) and the formation time distribution of lacey & cole ( 1993 ) .
we find that if selection effects are well - controlled and minimal external constraints are applied , the planned deep redshift survey could allow the measurement of the cosmic equation - of - state parameter @xmath3 to @xmath4 ( as little as 3% if @xmath5 has been well - determined from other observations ) .
this type of test has the potential also to provide a constraint on any evolution of @xmath3 such as that predicted by `` tracker '' models .
department of astronomy , university of california , berkeley , ca 94720 - 3411 ; -12pt |
magnetoimpedance ( mi ) consists of a change of total impedance of a magnetic conductor ( usually ferromagnetic ) under application of a static magnetic field @xmath0 .
when an ac current , @xmath1 of magnitude @xmath2 and angular frequency @xmath3 (= @xmath4 , with @xmath5 the ordinary frequency ) flows through the material , it generates , by ampre s law , a transverse magnetic field inducing some magnetization . at low frequency , the change in the transverse magnetization generates an additional inductive voltage @xmath6 across the conductor : @xmath7 where @xmath8 is the resistance .
hence mi can be written as @xmath9 , where the imaginary part is given by the ratio of magnetic flux to ac current and mi field dependence is related to the transverse permeability .
when frequency increases the current gets distributed near the surface of the conductor , changing both resistive and inductive components of the total voltage @xmath10 .
the field dependence of mi is dictated by skin depth @xmath11 where @xmath12 is the velocity of light , @xmath13 the conductivity and @xmath14 the permeability ( for non magnetic ordinary metals , @xmath15 ) .
the current distribution is governed not only by the shape of the conductor and frequency but also the transverse magnetization depending on @xmath0 .
+ typically mi increases with frequency , attains a maximum at frequencies for which the skin effect is strong ( @xmath16 where @xmath17 is a characteristic length scale such as the wire radius or ribbon / film thickness ) and then decreases since permeability becomes insensitive to the field at high enough frequency .
+ mi effect is ordinarily weak and did not attract much attention in the past .
interest in mi was triggered in the early 90 s when panina et al .
@xcite and beach et al .
@xcite reported a very large ( giant ) mi effect in amorphous ferromagnetic fecosib wires with small magnetic fields and at relatively low frequencies ( see fig .
[ fig1 ] ) .
this shows that a large variation of the mi is observable in a finite frequency range for reasons that will be explained in later sections .
later on , machado et al .
@xcite observed a smaller effect in fe@xmath18co@xmath19si@xmath20b@xmath20 thin films and beach et al .
@xcite in ribbons .
+ the dramatic variation of the mi ( that can reach in some cases values larger than 800% ) with small magnetic fields ( a few oersteds ) and at relatively low frequencies ( tens of mhz ) in widely available materials is the origin of the interest in the giant magnetoimpedance ( gmi ) effect .
+ gmi materials share the property of being magnetically soft ( easy to magnetize ) and are now available as wires ( typical diameter of the order of a mm ) , microwires ( typical diameter of the order of a micron ) , ribbons , magnetically coated metallic ( usually nonmagnetic ) wires ( tubes ) , thin films and multilayers , albeit , the effect itself occurs with widely differing magnitude depending on the geometry , the constituent materials and their layering . +
in order to clearly identify gmi , several observations should be made : 1 .
a very large change ( of the order of at least 100% variation ) of the impedance should occur with an external dc magnetic field @xmath0 .
the change expressed in % is defined by the largest value of the ratio : @xmath21 where @xmath22 is the impedance measured in the presence of the dc magnetic field @xmath0 and @xmath23 is the impedance measured at the saturation limit when the magnetization does not change any longer with the applied field .
2 . the external dc magnetic field @xmath0 should be on the order of a few oersteds only ( see table i on magnetic units and quantities ) .
the frequency range is on the order of mhz or tens of mhz ( excluding any effect based on ferromagnetic resonance ( fmr ) where the frequencies are typically in the ghz range @xcite ) . in many materials
this means that the skin depth @xmath24 ( typically microns at these frequencies ) is larger than the thickness of the material ( typically a fraction of a micron ) .
when the frequency is in the ghz , @xmath24 is generally very small with respect to the thickness .
it should be stressed that in ordinary metals , the skin depth does not depend on permeability , whereas in magnetic materials , the behavior of the permeability on geometry ( see for instance section iv ) , temperature , stress , composition and so on , is reflected in the skin depth . in addition , permeability might be changed by post - processing the material after growth with annealing under presence or absence of magnetic field or mechanical stress ...
gmi is a _ classical phenomenon _ that can be explained thoroughly on the basis of usual electromagnetic concepts @xcite in sharp contrast with giant magnetoresistance ( gmr ) effect where resistance is changed by a magnetic field .
gmr requires quantum mechanical concepts based on the spin of the carriers and their interactions with the magnetization of the magnetic material .
several general conditions must be satisfied by any material in order to show a gmi . 1
. the material should be magnetically soft .
that is , it should be easily magnetised or in other words must have a relatively narrow hysteresis curve implying , in general , small losses in the course of the magnetization cycle .
2 . the material should have a well defined anisotropy axis .
that means there must be a direction along which the magnetization of the material lies on the average ( easy axis ) .
however the value of the anisotropy field @xmath25 ( see fig .
[ fig2 ] ) should be relatively small ( on the order of a few oersteds )
. the typical ratio of @xmath25 to @xmath26 must be about 20 .
that insures observation of large magnetoimpedance effects , typically .
3 . the coercive field @xmath26 must be small ( fraction of an oersted ) and the hysteresis loop thin and narrow . since @xmath26 and the shape of the hysteresis loop change ( see fig .
[ fig2 ] ) with the angle the magnetic field makes with the easy axis ( or anisotropy axis ) of the material , these are taken at the reference point when the field is along the easy axis .
the ac current @xmath1 , injected in the material , should be perpendicular to the easy axis ( or anisotropy direction ) and the magnetic field it creates @xmath27 should be small with respect to @xmath25 ( see table i on magnetic units and quantities ) .
the material must have a small resistivity ( @xmath28.cm ) since it carries the ac current .
this is important , since many magnetic materials have large resistivities .
amorphous metals are interesting in that respect since , typically , their resistivities at room temperature are in the 100 @xmath29.cm range .
the material should have a large saturation magnetization @xmath30 in order to boost the interaction with the external magnetic field .
above arguments are equivalent to a very large permeability at zero frequency ( the ratio @xmath31 gives a rough indication for this value ) .
this means several 1000 s ( see table ii ) .
the material should have a small magnetostriction ( ms ) .
this means , mechanical effects caused by application of a magnetic field should be small .
mechanical stress due to ms alters the soft properties of the material by acting as an effective anisotropy .
this alters the direction of the anisotropy , displacing it from the transverse case and thereby reducing the value of the mi .
typical case materials are displayed along with their ms coefficient in table ii .
the general theory of the mi effect is widely available in classic textbooks @xcite ( for a long cylinder ) and it has been shown experimentally that a large mi often occurs at frequencies of a few mhz . changing the dc biasing field @xmath0 , the maximum @xmath32 can be as large as a few times the value of @xmath33 the dc resistance . at low frequency @xmath32
has a peak around @xmath34 and as the frequency increases , the peak moves toward @xmath35 where @xmath25 is the anisotropy field .
therefore , @xmath32 as a function of @xmath0 possesses a single or a double peak as the frequency increases ( fig . [ fig3 ] ) . when the direction of the anisotropy field is well defined the peaks are sharp
. + the behavior of @xmath32 versus @xmath0 follows very closely the behavior of the real part of the transverse permeability versus @xmath0 as we will show in later sections on wires and ribbons .
therfore it is very important to develop an understanding for the processes controlling the permeability . + material permeability depends on sample geometry , nature of exciting field , temperature , frequency , stress distribution in the material as well as internal configuration of the magnetization that might be altered by processing or frequency . for instance
, some materials should be annealed under the presence of a magnetic field or a mechanical stress in order to favor some direction for the magnetization or to release the stress contained in them .
+ regarding frequency , when it is large enough ( @xmath36 1mhz is sufficient in many materials ) domain wall displacements ( dwd ) are considerably reduced by eddy - currents and therefore magnetization varies by rotation or switching as if in a single domain ( see fig . [ fig2 ] ) . as a consequence , the rotational motion of the magnetization controls the behaviour of the permeability , through the skin depth . + considering @xmath17 as a typical thickness ( in the case of films / ribbons ) or radius ( in the case of wires or microwires ) , frequencies in the tens of mhz , lead to @xmath37 for the observation of gmi .
this condition depends strongly on geometry .
for instance , @xmath37 in 2d structures like films is satisfied at much higher frequencies ( ghz ) than in 1d structures like wires .
this is simply due to the optimal circular shape of wires that contains in an optimal fashion the flux while , simultaneously , carrying the ac current .
+ in terms of gmi performance , multilayered films ( such as f / m / f where f is a ferromagnet and m is metallic non magnetic material ) are preferred with respect to single layered films since they allow to inject the ac electric current in the metallic layer and sense the magnetic flux in neighbouring or sandwiching magnetic layers .
these can be in direct contact with the metallic layer or separated from it by an insulator or a semiconductor .
flux closure , that increases gmi , occurs when the width of the film ( transverse with respect to the ac current ) is large or that the metallic layer is entirely buried in the magnetic structure to trap the flux .
+ the progress of gmi is thrusted towards the increase of the largest value of the ratio @xmath38 and the sensitivity given by the derivative of the ratio with respect to the field ( see table iii for some illustrative values ) . + this sensitivity
is simply estimated by looking at the behaviour of the permeabilty @xmath39 versus @xmath0 as the frequency is changed .
we ought to have a large variation of @xmath39 about @xmath34 .
this happens , in general , at low frequencies that is when @xmath37 . at high frequencies
, we will show in later sections that this sensitivity is either lost or one has to go to higher @xmath0 to observe it hampering the use of the effect in small magnetic field detection .
+ applications of gmi range from tiny magnetic field detection and sensing ( such as magnetic recording heads ) to magnetic field shielding ( to degauss cathode ray tubes ( crt ) monitors ) .
the reason is that gmi materials , being soft , possess large permeabilities that are required in magnetic shielding . +
this work is organised as follows : in section ii , a general discussion on soft materials is presented ; in section iii , gmi in wires is described . in section iv , gmi in films and ribbons
is dicussed .
section v highlights potential applications of gmi effect in electronic ans sensing devices while section vi bears conclusions and perspectives of gmi .
soft ferromagnetic materials ( sfm ) used in gmi applications play a key role in power distribution , make possible the conversion between electrical and mechanical energy , underlie microwave communication , and provide both the transducers and the active storage material for data storage in information systems .
the fingerprint of any magnetic material is its hysteresis loop ( hl ) whose characteristic shape stems from two essential properties of magnetic materials : non - linearity and delay between input and output signals .
non - linearity is given by the shape of the loop whereas the delay is given by the width of the hl .
the quantities associated with the hl are displayed in fig .
[ fig2 ] . in general , magnetic materials being anisotropic
their hl width varies with the angle the external magnetic field makes with some given direction .
the easy axis direction is defined as the direction for which the hl width ( or coercive field @xmath40 ) is largest .
+ the coercive field @xmath40 represents the effort to demagnetise any magnetic material . for instance , hard magnetic materials exhibit large resistance to demagnetization and are therefore used in materials requiring permanent magnets whereas sfm are used in devices demanding little effort to demagnetise or remagnetise .
hard magnets are used as permanent magnets for many electrical applications .
some rare - earth alloys based on smco@xmath41 and sm(co , cu)@xmath42 are used for small motors and other applications requiring an extremely high energy - product magnetic materials .
fe - cr - co alloys are used in telephone receivers and nd - fe - b is used for automotive starting motor .
+ in contrast , sfm possess a narrower loop than hard materials and the area within the hysteresis curve is small .
this keeps the energy losses small , during each magnetic cycle , in devices based on these materials .
+ in order to qualify the different requirements for read heads and storage media in magnetic recording , an area belonging to the application portfolio of sfm ( besides sensing and shielding ) , table iv displays the desired characteristics .
+ sfm have high permeability ( i.e. easily magnetized ) with a low coercivity ( i.e. easily demagnetized ) . examples of soft ferromagnetic materials include fe alloys ( with 3 to 4% si ) used in motors and power transformers and generators .
ni alloys ( with 20 to 50% fe ) used primarily for high - sensitivity communications equipment .
this illustrates the versatility of sfm applications that range from mechanical to electrical and from power to communications systems .
+ sfm exhibit magnetic properties only when
they are subject to a magnetizing force such as the magnetic field created when current is passed through the wire surrounding a soft magnetic core .
they are generally associated with electrical circuits where they are used to amplify the flux generated by the electric currents .
these materials can be used in ac as well as dc electrical circuits .
several types of sfm exist : 1 .
soft mono and polycrystalline ferrites 2 .
powder composite magnetic materials 3 .
rapidly quenched ferromagnetic materials 4 .
amorphous magnetic materials 5 .
nanocrystalline magnetic materials the soft ferromagnetic behavior in these materials arises from a spatial averaging of the magnetic anisotropy of clusters of randomly oriented small particles ( typically @xmath43 fraction of a micron ) . in some cases , the ms of these materials is also reduced to near zero .
metallic glasses obtained by rapid quench ( e.g. splat - cooling ) are magnetically very soft .
this property is used in multilayered metallic glass power transformer cores .
amorphous magnetic alloys such as cop first reported in 1965 and splat - cooled materials with attractive soft ferromagnetic properties are based either on 3d transition metals ( t ) or on 4f rare - earth metals ( r ) . in the first case , the alloy can be stabilized in the amorphous state with the use of glass forming elements such as boron , phosphorus and silicon : examples include fe@xmath44 b@xmath45 , fe@xmath46 ni@xmath46p@xmath47 b@xmath48 , and co@xmath49 fe@xmath50 b@xmath51 si@xmath52 ( t@xmath53 m@xmath54 , with @xmath55 at % , approximately ) . the transition metals of late order ( tl where tl = fe , co , ni ) can be stabilized in the amorphous state by alloying with early order transition metals ( te ) of 4d or 5d type ( te= zr , nb , hf ) : some examples are co@xmath56 zr@xmath57 , fe@xmath58 nb@xmath59 b@xmath60 , and co@xmath61 nb@xmath47 b@xmath60 ( te@xmath54 tl@xmath53 , where @xmath62 is roughly @xmath63 at % ) . +
nanocrystalline soft magnetic materials are derived from crystallizing amorphous ribbons of a specific composition such as the ( fe , b ) based alloy family .
this class of materials is characterized by 10 - 25 nm sized grains of a body - centered - cubic ( fe , x ) phase consuming 70 - 80% of the total volume , homogeneously dispersed in an amorphous matrix . +
two families of alloys show the best performance characteristics and have emerged as the best candidates to major sfm applications : fe - cu - nb - b - si ( the `` finemet '' family , see table iii ) and fe - zr-(cu)-b-(si ) ( the `` nanoperm '' family ) .
the finemet family is characterized by an optimum grain size of about 15 nm , exhibits good properties at high frequencies and is comparable to some of the best ( and relatively costly ) co based amorphous materials . on the other hand , the grain sizes consistent with optimum performance are larger , around about 25 nm , in the nanoperm family .
the distinguishing feature of the nanoperm family of alloys is the very low energy loss exhibited at low frequencies ( 60 hz ) , offering the potential for application in electrical power distribution transformers . +
a typical amorphous alloy with a small ms coefficient @xmath64 , is co@xmath19 fe@xmath18 si@xmath20 b @xmath57 .
it is obtained by alloying fesib that has a positive @xmath65 with cosib that has a negative @xmath66 coefficient . some materials that result from this alloying
can reach a very small value of @xmath64 .
for instance , ( co@xmath67 fe@xmath68)@xmath69 si@xmath70 b @xmath20 has @xmath71 that can be considered as zero since typical values of @xmath64 are units or tens of @xmath72 .
this was obtained by varying systematically the concentration @xmath62 in the compound ( co@xmath54 fe@xmath53)@xmath69 si@xmath70 b@xmath20 .
starting with @xmath73 the ms coefficient @xmath74 decreases steadily to a near zero value @xmath75 for @xmath76 .
let us indicate , at this point , that a commercial product exists called vitrovac 6025 with the composition co@xmath77
fe@xmath60 mo@xmath78 si@xmath79 b@xmath59 that exhibits a value @xcite of @xmath80 .
+ a small negative value of @xmath64 is exploited in wires , as discussed in the next section , produces a magnetization profile that is circular in a plane perpendicular to the wire axis whereas a positive @xmath64 produces a radial magnetization profile .
the search for gmi in soft magnetic wires and microwires is a topic of interest related to possible applications as tiny field sensors .
+ at low frequencies ( weak skin effect or large skin depth ) , the first order expansion term of the impedance @xmath81 versus frequency is responsible for the voltage field dependence .
this term is represented by an inductance , which is proportional to the transverse permeability . when the skin effect is strong , the total impedance including resistance @xmath8 and reactance @xmath82 is field dependent through the penetration depth .
+ the general theory of the mi effect in a long cylinder is widely available in classic textbooks @xcite and it has been shown experimentally that a large mi often occurs at frequencies of a few mhz . changing the dc biasing field @xmath0 , the maximum @xmath32 can be as large as a few times the value of @xmath33 for amorphous wires . following beach et al .
@xcite the behaviour of the permeability can be understood on the basis of a simple phenomelogical model based on a single relaxation time @xmath83 for the magnetization that yields for the permeability : @xmath84 the behaviour of @xmath39 depicted in fig . [ fig4 ] indicates that the strongest variation of @xmath39 with @xmath0 happens at low frequencies as required for a sensitive sensor whereas the variation gets totally smeared out at large frequencies .
it also illustates the existence of a finite frequency range for the gmi effect as already observed in fig .
the range is determined by the inverse time 1/@xmath83 that controls relaxation processes of the transverse magnetization . +
while gmr is usually attributed to the differential scattering of conducting electrons whose spins make particular angles with the local magnetization of different scattering centers , it was believed , like gmr , this effect also resulted from the electron scattering by ac current - induced domain wall oscillations .
many different models exist for the description of permeability of metallic samples .
such models start with a uniform dc permeability throughout the sample , and then consider a class of domain structures with a magnetization process of dwd or domain magnetization rotations ( dmr ) . in the case of ac impedance ,
standard quantitative models are used without considering domain structures .
therefore , some current explanations of mi have to be made based on usual considerations with some additional modifications to include concepts of dwd , dmr , ferromagnetic resonance , and magnetic relaxation . for a straight wire of radius @xmath17 , conductivity @xmath13 and permeability @xmath14 ( see fig . [ fig5 ] ) the expression for impedance is given by @xcite : @xmath85 where @xmath86 is the i - th order bessel function and @xmath87 where the skin depth is given by @xmath11 .
@xmath12 is the speed of light and @xmath13 the conductivity .
+ since the ac current is applied along the wire axis perpendicularly to the anisotropy ( fig .
[ fig5 ] ) , it is again the transverse permeability @xmath14 that is considered here . in order to show that the behavior of @xmath88 is dictated by the permeability , let us consider firstly the low frequency case , i.e. @xmath89 . expanding the bessel functions , yields : @xmath90 in the opposite high frequency case ( @xmath16 or @xmath91 ) , taking @xmath92 1 mm whereas @xmath93 m , we can expand the bessel function to obtain : @xmath94
this indicates that @xmath95 could reach a value of several 1000 s at high frequencies .
the same result is obtained in the case of microwires by rescaling all lengths ( since we have , in this case , @xmath96 m ) .
it should be pointed out that since the ac current is along the wire axis , the magnetic field produced is circular along or against the rotational magnetization profile ( see fig .
[ fig5 ] ) .
note that at the frequencies of interest ( @xmath36 1mhz ) , where we have the largest sensitivity to @xmath0 , eddy - currents heavily damp dwd . + in wires made of materials having a small negative @xmath64 ,
the magnetization and anisotropy field @xmath25 run in a circular fashion in a plane transverse to the ac current as illustrated in fig .
the dc field @xmath0 acts to realign the magnetization along its direction , decreasing therefore the transverse permeability .
therefore , a positive @xmath64 can not be used for gmi since the magnetization profile is rather radial .
the skin depth dependent classical formula does not account for the presence of domain walls that can alter the scattering of the carriers .
panina et al .
@xcite and chen et al .
@xcite accounted for domain wall scattering and showed that the latter is efficient when the inter - domain distance @xmath97 ( see fig .
[ fig5 ] ) to the wire radius @xmath17 ratio @xmath98 for frequencies @xmath99 1mhz or greater . introducing a parameter @xmath100
, they argue that a change of mi by a factor of 10 ( assuming a change of @xmath101 by a factor between 7 and 700 ) yields a change of @xmath102 by a factor of 100 to 10 , 000 . using a resistivity
@xmath103125 @xmath104 and 2@xmath17=0.124 mm , the frequency @xmath105 mhz yields a factor of 4 to 10 change in @xmath32 .
chen et al .
@xcite proved also that the classical limit is recovered when @xmath106 .
the interpretation of a single or a double peak in the mi versus @xmath0 is interpreted as resulting from scattering by dmr .
the scattering increases as @xmath0 gets closer to the anisotropy field @xmath25 and then decreases after reaching its maximum for @xmath107 . wires with very small
anisotropy possess a single peak since the scattering maximum occurs at @xmath34 .
frequency effects produce the same result as seen in fig .
[ fig3 ] . in the case of gmi effect in coated tubes such as the electroplated feni , fenico , and cop microwires , kurlyandskaya et al .
@xcite found a huge enhancement factor approaching 800% in fe@xmath45co@xmath48ni@xmath49 layers ( 1 @xmath14 m - thick ) electroplated onto cube nonmagnetic microwire ( 100 @xmath14 m diameter ) at a frequency of 1.5 mhz .
that shows the way for producing the largest gmi in these systems .
sputtered or otherwise produced films or ribbons possess several advantages with respect to wires due to the possibilities of size reduction and increase of efficiency .
nevertheless , oblong or elongated ribbon or thin film geometries are preferred in order to emulate the wire case where the largest gmi ratios were obtained . +
since the main ingredients of gmi are a metal that carries an ac current and a nearby magnetic material that must sense strongly the flux , 2d materials or multilayers offer more flexibility than wires and experimentally , several excitation configuration become possible when we are dealing with a ribbon or a film . + the general mi measurement configuration ( longitudinal , perpendicular , transverse ) in ribbons and films
is shown in fig .
[ fig6 ] . in films and ribbons ,
the domain wall configuration is similar to what goes on in wires in the sense domain walls are transverse with respect to the ac current . in fig .
[ fig7 ] neighbouring domains have magnetizations pointing in different directions and the magneto - impedance effect result from averaging over transverse domains like in the wire case .
+ when a probe current @xmath108 is applied to a film of thickness @xmath109 , the impedance is written as : @xmath110 where @xmath33 is the dc resistance of the film and @xmath87 with @xmath24 the skin depth given by , as before , @xmath11 . since the current is applied ( in the ribbon case ) along the ribbon axis , again the transverse permeability is considered .
as in the wire case , it is possible to expand @xmath95 at low ( or @xmath111 ) and high frequencies ( or @xmath112 ) and show that @xmath113 controls its behavior with @xmath0 . at low frequencies ( @xmath111 ) ,
we get : @xmath114 whereas , at high frequencies ( @xmath112 ) : @xmath115 the domain wall configuration ( see fig .
[ fig7 ] ) permits to estimate the permeability in the case of weak anisotropy field ( @xmath25 small but finite ) .
it should be stressed that domains , @xmath25 and permeability are all transverse .
when @xmath0 is longitudinal , it plays the role of a hard axis field since it is acting on the magnetization , to rotate it , hence decreasing the transverse permeability , similarly like in wires . + in order to increase the mi response , multilayers are preferred to single layered films / ribbons . if a sandwiched metallic layer carries the ac current that creates the flux in the neighbouring magnetic layers , the mi effect is considerably enhanced ( table iii ) . moreover , when the width @xmath116 is large ( fig .
[ fig7 ] ) the ac flux loop is closed and stray magnetic fields small @xcite .
+ multilayers of permalloy ( ni@xmath117fe@xmath118 denoted as py ) and ag of the type @xmath119py(x )/ag ( y )@xmath120_{n}$ ] with thickness x = 107 and 88 , y= 7 and 24 and the number of layers n=15 and 100 were grown by sommer et al .
@xcite with magnetron sputtering techniques .
the case n=100 displayed a gmi effect of the order of 5 to 25 % .
the thickness of the py layer has a strong effect on the coercive field and consequently on the softness of the material .
the coercive field @xmath26 varies from 0.85 oersted for x=88 to reach 13.6 oersted for x=5000 . as the ag thickness varied from 7 to 24 the mi effect changed from single peak to double peak structure in a fashion akin to the effect of frequency .
sommer and chien @xcite showed also that in the ribbon case , a double or a single peak structure in the mi results from the direction of the applied magnetic field @xmath0 . when it is longitudinal ( along the ribbon axis , see fig . [ fig7 ] ) a double peak at @xmath107is observed , but when it is transverse the demagnetization factor perpendicular to the ribbon axis becomes important .
that is responsible of an apparent _ longitudinal _ permeability that gives birth to a single peak mi .
the ribbon mi is similar to the wire case , however mi in single layer films remains small and it is still an open problem to find ways to increase it dramatically like in coated microwires .
nevertheless , when multilayers are used ( see table iii ) the gmi ratio may attain 700% .
this enhancement is due to the electrical insulation of the metallic layer carrying the ac current by sandwiching it with sio@xmath121 layers , the outer magnetic layers embodying the ac inductance created by the current .
since gmi uses soft magnetic materials possessing a large permeability , the first immediate application is in devices related to magnetic shielding since it requires the soft properties of the material .
a host of other applications involves the gmi per se .
+ the first of these applications are the detection of very small magnetic fields . in order to have a general idea of
the scale of current magnetic fields table v displays some orders of magnitude of natural and artificial fields .
+ detection of magnetic field is also important and magnetic field sensors ( see hauser et al .
@xcite ) are broadly classified in three categories ( see table vi ) : 1 .
medium to large field detection by hall and magnetoresistive devices .
2 . small to medium magnetic field detection by magnetoimpedance and flux gate sensors .
3 . very small to small magnetic field detection by squid s ( superconducting quantum interference devices ) .
the possible devices are recording read heads , magnetic guidance devices in vehicles , boats and planes ( with or without gps , i.e. global positioning system ) , brain imaging ( magneto - encephalogram or meg devices ) and heart mapping ( magneto - cardiogram or mcg devices ) etc ... + the detection of the earth magnetic field has a host of applications for instance in petroleum or minerals exploration or in shielding used for degaussing of high performance crt monitors ... + the general features required for sensors are not only high sensitivity , flexibility , large bandwidth and lost cost but linearity is also a desired feature . in order to increase the sensitivity of gmi devices with respect to the dc magnetic field
, devices were developed @xcite possessing non - symmetric variation of the mi with respect to @xmath0 .
the materials used are field annealed co - based amorphous ribbons ( co@xmath77 fe@xmath60 nib@xmath47 si@xmath20 ) ( see table iii ) .
the asymmetry of the mi profile allows a very sensitive detection of the magnetic field specially around @xmath122 and when the asymmetry of the profile is so sharp that it becomes step - like , we have a so - called `` gmi valve '' based device . in these devices
, the sensitivity might be enhanced to reach 1000%/oe ( compare with values in table iii ) .
asymmetry may also be induced by torsional stress such as in wires of fe@xmath123si@xmath42 b @xmath20 @xcite .
altering the gmi response with mechanical stress paves the way toward the development of strain sensors that can be used in several areas of engineering and science .
good candidates for gmi are amorphous cobalt rich ribbons , wires , glass covered microwires and multilayers made of a metallic layer sandwiched between ferromagnetic materials ( with or without intermediate insulating layers ) . + zero - field annealing or annealing under a magnetic field or with the application of some mechanical stress favors orienting the magnetization along a desired direction . in order to get a fast reduction of the transverse permeability ( see fig .
[ fig4 ] ) under the application of an external field @xmath0 , there must be an optimal distribution of the magnetization around the desired direction @xcite and that can be obtained with suitable growth or annealing conditions .
reduction of the stress contained into the as - grown materials by sputtering or quenching can also be obtained with annealing .
+ the current activity in gmi studies is oriented toward the development of devices using a built - in magnetic field rather than an applied external field .
this is reminescent of the p - n , schottky , heterojunctions etc ... where the built - in field is electrical .
+ the discovery of magnetic bias produced a revolution in magnetism because of the many potential applications in storage , sensing , spintronics ... nevertheless , there are other ways to produce built - in magnetic fields . for instance ueno et al .
@xcite produced susch an internal field by superposing two sputtered films of co@xmath124 fe@xmath125 b @xmath45 with crossed anisotropy axes , i.e. with the anisotropy axis of the bottom layer making angles of opposite sign with some common direction . +
while gmi in wires and ribbons is steadily progressing , the case of single layered films is still lagging behind in favor of multilayered films .
ways must be developed in order to increase the sensitivity of single or multilayered structures in order to reach the level observed in wires and microwires .
the range of applications will substantially explode once gmi in thin films will be competitive with wires and ribbons .
+ * acknowledgement * + the authors wish to acknowledge correspondance with p. ripka and his kind providing of re(pre)prints of his work as well as friendly discussions with a. fessant regarding characterisation and impedance measurements .
+ 99 l. v. panina and k. mohri , appl .
65 , 1189 ( 1994 ) .
r. s. beach and a. e. berkowitz , appl .
64 , 3652 ( 1994 ) .
f. l. machado , b. l. da silva , s. m. rezende and c. s. martins , j. appl .
75 , 6563 ( 1994 ) .
r. s. beach and a. e. berkowitz , j. appl .
76 , 6209 ( 1994 ) .
m. r. britel , d. menard , l. g. melo , p. ciureanu , a. yelon , r. w. cochrane , m. rouahbi and b. cornut , appl .
77 , 2737 ( 2000 ) .
d. atkinson and p. t. squire , j. appl .
83 , 6569 ( 1998 ) .
m. knobel and k. r. pirota , jmmm 242 - 245 , 33 ( 2002 ) .
r. valenzuela , m. knobel , m. vazquez and a. hernando , j. appl .
78 , 5189 ( 1995 ) .
l. d. landau and e. m. lifshitz , electrodynamics of continuous media , pergamon , oxford , p.195 ( 1975 ) .
n. murillo , j. gonzalez , j. m. blanco , r. valenzuela , j. m. gonzalez and j. echebarria , j. appl .
81 , 5683 ( 1997 ) .
d - x chen , j.l .
munoz , a. hernando and m. vazquez , phys .
57 , 10699 ( 1998 ) .
g. v. kurlyandskaya , j. m. barandiaran , j. l. munoz , j. gutierrez , d. garcia , m. vazquez , and v. o. vaskovskiy , j. appl .
87 , 4822 ( 2000 ) .
a. paton , j. appl .
42 , 5868 ( 1971 ) .
r. l. sommer , a. gndel and c. l. chien , j. appl .
86 , 1057 ( 1999 ) .
r. l. sommer and c. l. chien , appl .
67 , 3346 ( 1995 ) .
h. hauser , l. kraus and p. ripka , ieee inst . and meas .
kim , k.j .
jang , d.y .
kim and s.s .
yoon : appl .
75 , 2114 ( 1999 ) .
ryu , s.c .
yu , c. g. kim , i.h .
nahm and s.s .
yoon , j. appl .
87 , 4828 ( 2000 ) . k. ueno , h. hiramoto , k. mohri , t. uchiyama and l. v. panina , ieee trans . mag .
36 , 3448 ( 2000 ) .
.correspondance between magnetic units in the si and cgs unit systems .
note that magnetic field units are a / m and oe and induction s are tesla and gauss . in vacuum or non - magnetic materials in cgs values in oe and gauss are same . [ cols="<,^,>",options="header " , ] | the status of giant magneto - impedance effect is reviewed in wires , ribbons and multilayered soft ferromagnetic thin films .
after establishing the theoretical framework for the description of the effect , and the constraints any material should have in order to show the effect , experimental work in wires , ribbons and multilayered thin films is described .
existing and potential applications of the effect in electronics and sensing are highlighted . |
vehicular ad - hoc networks ( vanets ) are formed by a collection of vehicles , wirelessly connected to each other to form a communication network .
these networks are typically one dimensional as they dynamically and rapidly self - organise on roads and highways , although communication with road side access point infrastructure is possible under different scenarios . primarily ,
vehicle to vehicle ( v2v ) and vehicle to infrastructure ( v2i ) communications involve safety related issues , such as collision warnings aimed at preventing imminent car accidents through broadcasting and relaying messages , thereby increasing local situation awareness @xcite .
v2v communications can also be exploited for applications such as intelligent cruise control or platooning , traffic information and management , as well as internet access and advertising .
the ieee 802.11p standard defines a wireless area network ( wlan ) for dedicated short range communication ( dsrc ) among vehicles .
the standard defines protocols for the physical and mac layers , has a 75 mhz bandwidth allocated at 5.9 ghz , and is the prime candidate currently being deployed in order to get ieee 802.11p equipped cars on the roads @xcite . under the standard , it is possible to bundle together information on position , speed , direction , brake information , steering wheel angle , threat - events , etc . , and append them to a _
basic safety message _ ( bsm ) which is then broadcasted @xcite .
vehicles within range can then actuate on this information , edit it , or append to the content message , and re - broadcast , thus locally flooding the network .
flooding algorithms are commonplace in ad hoc networks , however here the algorithm is also spatially constrained to run along a one - dimensional road network .
such networks are typically modelled as random geometric graphs @xcite formed by a 1d poisson point process ( ppp ) and a communication range @xmath0 ( see fig.[fig : graph ] ) directly related to snr @xcite thus lending themselves to mathematical analysis and engineering .
a major challenge in vanets is the timeliness and latency in which information must arrive to be useful to a fast approaching vehicle .
hop - count statistics find application in a variety of other settings , e.g. , in gas pipe sensor networks @xcite , nanowires @xcite , and map navigation problems in general .
therefore , hop - count statistics have been extensively studied in 1d @xcite and 2d networks @xcite .
they were first studied by chandler @xcite , who looked at the probability that two wireless network nodes can communicate in @xmath1 hops .
such information can further assist the calculation of network centrality measures @xcite , or achieve range - free localisation @xcite . )
used to model a vanet .
the geodesic length between the two extreme nodes is @xmath2 . ] in this paper we are concerned with the statistical properties of the _ shortest _ multihop paths , also referred to as _
geodesics _ , between nodes in 1d vanets .
to this end , we calculate for the first time the first few moments of the number of geodesics @xmath3 between nodes in a 1d vanet , as a function of the euclidean distance @xmath4 between them and the vehicle density @xmath5 .
clearly for @xmath6 , the shortest possible path is of length @xmath1 hops , employing just @xmath7 relay nodes , thus defining a fundamental upper limit on the latency involved with such transmissions . on
the other hand , due to the broadcast nature of wireless transmissions , multiple bsms containing similar information may arrive via different @xmath1-hop paths almost simultaneously .
it is therefore of interest to understand the statistical properties of the number of @xmath1-hop paths @xmath8 , as a function of @xmath0 , @xmath4 , and @xmath5 .
such statistics can be used to enhance throughput @xcite , validate threat events , protect against collusion attacks , infer location information , and also limit redundant broadcasts thus reducing interference .
must involve at least 1 relay node located in each of the shaded lenses " . ]
consider a source node @xmath9 located at the origin , and a destination node @xmath10 a distance @xmath11 to the right of @xmath9 along the positive real line .
further , consider a 1d ppp of density @xmath5 vehicles per unit length forming on the real line , with each point ( node ) representing a vehicle along an infinite stretch of road .
nodes are then connected via communication links whenever their euclidean distance is less than a predefined communication range @xmath0 ( see fig .
[ fig1 ] ) , thus forming a 1d network .
the source and destination nodes are unable to communicate directly and must employ multihop communications in order to share information .
depending on the density of vehicles @xmath5 , there may be none , one , or several multihop paths connecting @xmath9 and @xmath10 .
the _ length _ of these paths is the number of hops required for a message to pass between the two vehicles .
it follows that the length of the _ shortest _ multihop paths is @xmath12 .
therefore , paths of length @xmath1 are geodesic . running a breadth - first search ( bfs ) algorithm can find all geodesic paths in linear - time since the underlying graph is neither directed , nor weighted .
let the set of all geodesics be described by @xmath13 .
then the number of geodesic paths is @xmath14 .\end{split}\ ] ] monte carlo simulations of the pmf of @xmath3 are shown in fig .
we will first demonstrate the difficulties with obtaining the distribution of @xmath8 for the case of @xmath15 , and then calculate its first few moments for the general case of @xmath16 .
the cases of @xmath17 and @xmath18 are trivial and therefore omitted . for @xmath19 and @xmath20 such that geodesics are of length @xmath21 , and @xmath22 respectively .
let @xmath23 and @xmath15 as in fig .
[ fig1 ] such that there are two sub - domains @xmath24 and @xmath25 within which relay nodes must be situated in order for a 3-hop path to exist .
we call these sub - domains lenses , since in two dimensions they are formed by the intersection of two equal disks .
this is because the first relay node located at a maximum distance of @xmath0 can form a 3-hop path by connecting with any node in @xmath26 $ ] . by symmetry
@xmath27 $ ] such that the two lenses are of equal widths @xmath28 .
the number @xmath29 of relay nodes in @xmath24 is therefore a poisson random variable with mean @xmath30 , where we have defined @xmath31 .
moreover , for each relay node in @xmath24 there corresponds a subset of @xmath25 within which a second relay node must be located as to form a 3-hop path from @xmath9 to @xmath10 . labelling the @xmath29 relays in descending distances @xmath32 from the source ( i.e. , @xmath33 ) we can identify subsets @xmath34\subseteq l_2 $ ] within which if located a second relay can successfully form a 3-hop path . defining the sub - domains @xmath35 $ ] , for @xmath36 with @xmath37 and @xmath38
it can be seen that a relay node in @xmath39 connects to @xmath40 relays in @xmath24 .
we therefore arrive at a simple expression for the number of shortest 3-hop paths @xmath41 where the @xmath42 is the number of relays in @xmath39 and are thus poisson random variables with mean @xmath43 ( see fig .
[ fig : windows ] ) .
the widths @xmath39 are also random variables however must satisfy the constraint that @xmath44 , i.e. , the @xmath42 are correlated .
nodes in the left lens @xmath24 , and the corresponding sub - domains @xmath39 in the right lens @xmath25 .
note that @xmath45 is within range from all three nodes and therefore a fourth node located in @xmath45 will connect to all in @xmath24 , to form three @xmath21-hop paths from @xmath9 to @xmath10 .
in contrast , @xmath46 is in range of nodes 1 and 2 ( not 3 ) , @xmath47 is only in range of node 1 ( not 2 or 3 ) , and @xmath48 is not in range from any of the nodes in @xmath24 . ]
the pmf of @xmath3 can be expressed as follows : @xmath49 \!=\ ! \mathbb{e}_{n_1,\mathbf{w}}\big [ \mathbb{p } [ \sigma_{3 } \!=\ ! x \big| n_1 , \mathbf{w } ] \big ] \end{split}\ ] ] where @xmath50 and any configuration of widths @xmath51 is equally likely .
we can attempt to obtain the pmf of @xmath52 through the use of probability generating functions ( pgfs ) .
namely , we have that the pgf of the random variable @xmath53 is given by @xmath54\!=\!g_{n_i}(z^i)\!=\!e^{\lambda w_i ( z^i -1)}$ ] since @xmath42 is poisson distributed with mean @xmath43 .
it follows that the sum of @xmath29 such random variables has a pgf given by @xmath55 and the corresponding pmf given by @xmath56 = \ ! \\
\sum_{k=0}^\infty \mathbb{p}[n_i\!=\!k ] \
! \int_{[0,3r_0-l]^{n_1 } } \!\!\!\!\!\
! \mathbf{1}(\mathbf{w})\frac{c}{k ! } \frac { { \textrm{d}}^k g_{\sigma_{3}}}{{\textrm{d}}z^k}\big|_{z=0 } { \textrm{d}}w_1 \ldots { \textrm{d}}w_{n_1}\end{gathered}\ ] ] where @xmath57= \frac{\lambda_3^k}{k!}e^{-\lambda_3}$ ] and @xmath58 is the indicator function equal to @xmath59 whenever @xmath60 and zero otherwise such that @xmath61 , and @xmath62 is some normalisation constant .
geometrically , the indicator function defines a simplex polytope with @xmath63 vertices at @xmath64 .
the integral is therefore over the surface of the @xmath29-simplex .
recall that the @xmath29-simplex is a triangle , a tetrahedron , a 5-cell , for @xmath65 and @xmath22 respectively , and therefore is an evermore complex polytope embedded in the positive hyperoctant of @xmath66 for which the integration of becomes intractable .
for this reason we next restrict our study to the mean and variance of @xmath52 .
we now describe a method which allows us to analytically derive the moments of @xmath67 .
this involves dividing up the lenses into many small parts and making a simplifying approximation about the interactions .
this allows us to treat the problem as one involving many independent random variables rather than trying to account for dependence .
the final step is to take the limit of the number of divisions of the lenses to infinity , in which our approximation becomes exact .
we firstly split the lenses @xmath68 into a large number @xmath69 of equally sized , disjoint domains @xmath70 where @xmath71 and @xmath72 .
the number of relay nodes in each @xmath70 is then a poisson distributed random variable @xmath73 with mean @xmath74 . for finite @xmath75
we make the approximation that all relay nodes in @xmath76 connect with all those in @xmath77 , all those in @xmath78 connect with all in @xmath77 and @xmath79 etc .
the number of shortest 3-hop paths is then given by @xmath80 using the independence of the @xmath73 we calculate the mean @xmath81&=\lim_{l \rightarrow \infty}\sum_{q=1}^{l } \sum_{r=1}^{q}{\mathbb{e}}[y_{1q}]{\mathbb{e}}[y_{2r}]\\ & = \lim_{l \rightarrow \infty}\left(\frac{\lambda_3 ^ 2}{l^2}\right)\left(\frac{l^2+l}{2}\right)=\frac{\lambda_3 ^ 2}{2 } \label{expectation2 } \end{split}\ ] ] to extract the variance we first define the random variable @xmath82 given that the variance of a sum is equal to the sum of the variances plus the covariances we have @xmath83 we first evaluate the variance of @xmath84 .
we use the independence of @xmath85 and @xmath86 and note that @xmath87 is a poisson random variable with mean @xmath88 .
in addition we use the mean of the square of a poisson random variable with mean @xmath89 is equal to @xmath90 and derive @xmath91 using ( [ vartq ] ) we evaluate the limit of the first sum in ( [ var ] ) as @xmath92 for the covariance terms in eq.([var ] ) we let @xmath93 and use the relation @xmath94-{\mathbb{e}}[t_s]{\mathbb{e}}[t_t]$ ] .
the expectation of @xmath95 is given by @xmath96=\frac{s\lambda_3 ^ 2}{l^2 } \label{expts } \end{split}\ ] ] for the expectation of the product we have via ( [ tq ] ) @xmath97={\mathbb{e}}\big[\big(y_{1s}\sum_{r=1}^{s}y_{2r}\big ) y_{1t}\big ( \sum_{r=1}^{s}y_{2r}+\sum_{r = s+1}^{t}y_{2r } \big)\big ] \label{eq : exprod } \end{split}\ ] ] where by splitting the sum in @xmath98 we can factorise using the mutual independence of the terms as @xmath99 \!=\ ! { \mathbb{e } } [ y_{1s } ] { \mathbb{e}}[y_{1t}]\big ( { \mathbb{e}}\big [ \big ( \sum_{r=1}^{s}\!y_{2r } \!\big)^2\big ] \!+\ ! { \mathbb{e}}\big [ \sum_{r=1}^{s}\!y_{2r}\big]\ ! { \mathbb{e}}\big[\!\!\sum_{r = s+1}^{t } \!\!\ ! y_{2r}\big ]
\big ) \label{eq : exprod2 } \end{split}\ ] ] evaluating the individual expectations and combining we have @xmath99=\frac{\lambda_3 ^ 4st}{l^4}+\frac{\lambda_3 ^ 3 s}{l^3 } \label{eq : exprod3 } \end{split}\ ] ] combining ( [ expts ] ) and ( [ eq : exprod3 ] ) we have that @xmath100
. we can now evaluate the sum of covariances in ( [ var ] ) @xmath101 taking the limit @xmath102 and combining it with ( [ vartqsumlimit ] ) we may extract the variance @xmath103 similarly we can extract higher order moments of the distribution using this technique . for example the third moment @xmath104)^3]\!=\ ! - \frac{5\lambda_3 ^ 5}{6}-\frac{\lambda_3 ^ 4}{5 } $ ] , which can be used to analyse the skewness of the distribution . as a function of @xmath105 calculated numerically from ensembles of @xmath106 realisations for a range of values of @xmath5 and @xmath107
. also illustrated is the analytical results of eq.([expectation2 ] ) and eq([variancesigma ] ) in ( a ) and ( b ) respectfully ( grey line ) . ]
more generally for @xmath108 with integer @xmath1 there will be @xmath7 lenses of equal width @xmath109 .
the method of ( [ sec : meanvariance ] ) can still be used . for general @xmath1
we have @xmath110 where the @xmath111 are poisson with mean @xmath112 . using that @xmath113 we can derive the mean @xmath114=\frac{\lambda_k^{k-1}}{(k-1 ) ! } \label{esigmakhop } .\end{split}\ ] ] now , letting @xmath115 , we recursively define @xmath116 by further defining @xmath117 , such that @xmath118 we can recursively define the expectation of @xmath119 @xmath120=\frac{\lambda}{l}\sum_{r=1}^{l}{\mathbb{e}}(\tau_l^{(n ) } ) , \label{exptrec } \end{split}\ ] ] where @xmath121 is the mean of the poisson variables @xmath122 .
similarly , for the variance of @xmath123 we calculate @xmath124 using the recurrence relation we have @xmath125 for the covariance we have @xmath126-{\mathbb{e}}[t_{s}^{(n)}]{\mathbb{e}}[t_{t}^{(n ) } ] \\ & \!=\!\frac{\lambda^2}{l^2 } \big({\mathbb{e}}\big [ \sum_{p=1}^s t_{p}^{(n ) } \big(\sum_{p=1}^s t_{p}^{(n ) } \!+\!\!\ ! \sum_{p = s+1}^t \!\ !
t_{t}^{(n)}\big ) \big]\!-\ ! { \mathbb{e}}[\tau_s^{(n ) } ] { \mathbb{e } } [ \tau_t^{(n ) } ] \big)\\ & \!=\!\frac{\lambda^2}{l^2 } \big ( \sum_{p=1}^s \sum_{r = s+1}^t \ ! { \mathbb{e } } [ t_{p}^{(n)}t_{r}^{(n ) } ] \!+\ ! { \mathbb{e}}[(\tau_s^{(n)})^2]-{\mathbb{e}}[\tau_s^{(n ) } ] { \mathbb{e } } [ \tau_t^{(n ) } ] \big ) \label{eq : covrecur } \end{split}\ ] ] where we have used that @xmath127 and split the sum into parts .
we now have @xmath128 { \mathbb{e } } [ \tau_t^{(n)}]\\ & + \ ! { \mathbb{e}}[\tau_s^{(n)}]^2 \!+\ ! \sum_{p=1}^s \ !
\sum_{r = s+1}^t \ ! { \mathrm{cov}}\left(t_{p}^{(n)},t_{r}^{(n)}\right )
\frac{\lambda^2}{l^2}{\mathbb{e}}[\tau_p^{(n)}]{\mathbb{e}}[\tau_r^{(n ) } ] \big ] \label{eq : covrecurr } \end{split}\ ] ] letting @xmath129 and @xmath130 we can combine ( [ exptrec ] ) , ( [ varsumtrec ] ) , ( [ vartrec ] ) and ( [ eq : covrecurr ] ) to obtain the variance for @xmath8 .
for example , for @xmath131 we have @xmath132 .
this recursion relation allows us to derive the variance of @xmath133 , which involves evaluating a @xmath134-fold sum of products of random variables ( see ( [ sigma4hop ] ) ) in terms of a simpler @xmath135-fold sum .
motivated by the multihop diffusion of information in vanets , realised through the periodic broadcasts of bsms as mandated by the dsrc standard @xcite , we have studied the statistics of the number @xmath8 of shortest @xmath1-hop paths in 1d random networks .
namely , we have derived simple closed form expressions for the mean and variance of @xmath8 for @xmath136 , provided a recursive formula for general @xmath1 , and have confirmed them numerically using monte carlo simulations ( see fig .
[ fig3 ] ) .
we argue that knowledge of such statistics can be used to enhance throughput @xcite , validate threat events , protect against collusion attacks , infer location information , and also limit redundant broadcasts thus reducing interference . as an example , consider the realistic scenario where there are about @xmath137 vehicles per km , transmission range is @xmath138 km , and a vehicle detects an event and broadcasts a bsm containing relevant safety information which should reach at least a range of @xmath139 km from the epicentre of the detected event .
it follows that the length of the shortest multihop path is @xmath140 , and that the expected number of shortest paths is @xmath141 \!=\ !
1333.33 $ ] ( in either forward or backward direction ) .
this is clearly unnecessary , and only a fraction of @xmath142 $ ] vehicles should re - broadcast the original bsm .
thus inverting , we can calculate the re - broadcast probability @xmath143 , where @xmath144 is the target number of shortest paths , e.g. , setting @xmath145 we estimate that just @xmath146 of vehicles should re - broadcast the original bsm .
the authors would like to thank the directors of the toshiba telecommunications research laboratory for their support .
this work was supported by the epsrc grant number ep / n002458/1 for the project spatially embedded networks .
k. a. hafeez , l. zhao , b. ma , and j. mark , `` performance analysis and enhancement of the dsrc for vanet s safety applications , '' _ vehicular technology , ieee transactions on _ , vol .
62 , no . 7 , pp . 30693083 , 2013 .
s. najafzadeh , n. ithnin , s. a. razak , and r. karimi , `` bsm : broadcasting of safety messages in vehicular ad hoc networks , '' _ arabian journal for science and engineering _ , vol .
39 , no . 2 , pp .
777782 , 2014 .
o. georgiou , c. p. dettmann , and j. p. coon , `` connectivity of confined 3 networks with anisotropically radiating nodes , '' _ wireless communications , ieee transactions on _ , vol . 13 , pp . 45344546 , aug 2014 .
i. stoianov , l. nachman , s. madden , and t. tokmouline , `` pipenet : a wireless sensor network for pipeline monitoring , '' in _
2007 6th international symposium on information processing in sensor networks _ , pp . 264273 ,
april 2007 .
t. w. larsen , k. d. petersson , f. kuemmeth , t. s. jespersen , p. krogstrup , j. nygrd , and c. m. marcus , `` semiconductor - nanowire - based superconducting qubit , '' _
phys.rev.lett._ , vol .
115 , p. 127001
, 2015 . c. nguyen , o. georgiou , and y. doi , `` maximum likelihood based multihop localization in wireless sensor networks , '' in _ 2015 ieee international conference on communications ( icc ) , london , uk _ , pp .
66636668 , 2015 .
g. v. rossi , k. k. leung , and a. gkelias , `` density - based optimal transmission for throughput enhancement in vehicular ad - hoc networks , '' in _ communications ( icc ) , 2015 ieee international conference on _ , pp . 65716576 , ieee , 2015 . | in the ieee 802.11p standard addressing vehicular communications , basic safety messages ( bsms ) can be bundled together and relayed as to increase the effective communication range of transmitting vehicles .
this process forms a vehicular ad hoc network ( vanet ) for the dissemination of safety information .
the number of shortest multihop paths " ( or geodesics ) connecting two network nodes is an important statistic which can be used to enhance throughput , validate threat events , protect against collusion attacks , infer location information , and also limit redundant broadcasts thus reducing interference . to this end , we analytically calculate for the first time the mean and variance of the number of geodesics in 1d vanets . |
the discovery of the accelerated expansion of the universe ( see @xcite and references therein ) has opened a new window in cosmological studies .
indeed , the underlying physical process responsible for this phenomenon is considered as one of the fundamental problems in cosmology . within the framework of general relativity ,
scalar fields provide possible dark energy models which can describe , but not so far explain , this acceleration .
scalar field models require the choice of a self - interaction potential @xmath0 for the scalar field @xmath1 .
the considerations of a specific potential @xmath0 is done by an adhoc requirement ( ansatz ) . in this manner
various candidates have been proposed in the literature , such as an inverse power law , exponential , hyperbolic and the list goes on ( for review see @xcite and references therein ) .
one such ansatz , which we shall consider in the following , has been done by rubano and barrow @xcite ( see also @xcite ) who found that if the scalar field behaves as a perfect fluid then the potential @xmath2 , where the constants @xmath3 are given in terms of observable cosmological parameters , namely the dark - energy equation of state ( eos ) parameter and @xmath4 .
obviously it is important to proposed potentials which are realistic and at the same time lead to exactly soluble models in the flrw spacetime .
the reason for employing the noether symmetries is that noether symmetries provide us with the noether integrals which facilitate the analytic solution of the field equations . as an example , in extended theories of gravity , where the birkhoff theorem is not guaranteed , the noether approach provides a means of describing the global dynamics using the first integrals of motion @xcite .
furthermore , besides the technical possibility of reducing the dynamical system , the first integrals of motion always give rise to conserved currents that are not only present in physical space - time but also in configuration spaces ( see discussion in @xcite ) . in space - time such currents are linear ( momentum , angular momentum etc . )
but in configuration space the conserved quantities appear as relations between dynamical variables @xcite .
the latter implies that in the configuration space ( minisuperspace ) , the first integrals are considered as `` selection rules '' for scalar field potentials and coupling functions in the case of modified gravity theories .
the above features have inspired many authors to propose the admittance of a noether ( point or dynamical ) symmetry by the field equations as a selection rule for dark - energy models , including those of modified gravity @xcite .
* * in particular it has been shown that the dynamical noether symmetries are associated with the killing tensors of the space @xcite and that the noether point symmetries are related to the homothetic algebra of the space ( see @xcite and references therein ) .
therefore the requirement of a noether symmetry is indeed a geometric demand hence independent of the particular dynamics .
in scalar field cosmology the system develops in minisuperspace whose geometry is defined by the field equations . therefore the requirement that the field equations admit a noether symmetry becomes , according to the above results , a geometric requirement on a geometry which is inherent to the system . in this work by using a more general geometric criterion , i.e. by employing the lie point symmetries of the wheeler - dewitt ( wdw ) equation , we extend the work of rubano and barrow @xcite to a general family of hyperbolic scalar - field potentials @xmath0 . in section [ sfc ] ,
we give the basic theory of the scalar - field cosmology in a flrw spacetime .
the basic definitions and results from the lie and noether point symmetries of partial differential equations and the application in the wdw equation are presented in section [ preliminaries ] . in section [ wdwsym ]
, we consider our cosmological model , which includes a hyperbolic family of scalar field potentials with a perfect fluid with constant equation of state parameter @xmath5 .
we study the existence of lie point symmetries of the wdw equation , where we find that for our model the wdw equation admits lie point symmetries if the free parameters of the potential and the parameter @xmath5 are related . in section [ exactsol ]
we apply the lie point symmetries of the wdw equation in order to construct invariant solutions of the wdw equation and exact solutions of the field equations .
also , in section [ dynanalysis ] we perform a dynamical analysis by studying the fixed points of the field equations in the dimensionless variables for the general model and we show that when the cosmological model is liouville integrable ( the model admits conservation laws ) , there is a unique stable point which describes the de sitter universe . finally , in section [ conclusion ] we draw some conclusions .
we start with the friedmann - lematre - robertson - walker ( flrw ) spacetime with line element ( @xmath6 ) @xmath7 the total action of the field equations is written as @xmath8 where @xmath9is the einstein - hilbert action , @xmath10 is the ricci scalar of the underlying space , @xmath11 is the action of the scalar field @xmath12 , \label{sf.02}\ ] ] and @xmath13 is the matter term .
we assume that @xmath1 inherits the symmetries of the metric ( [ sf.1 ] ) therefore @xmath14 and consequently @xmath15 where @xmath16 . from the action ( [ sf.01 ] )
, we have the einstein field equations @xcite@xmath17 where @xmath18 , @xmath19 is the ricci tensor and @xmath20 is the total energy momentum tensor given by @xmath21 .
@xmath22 is the energy - momentum tensor of baryonic matter and radiation and @xmath23 is the energy - momentum tensor associated with the scalar field @xmath1 .
modeling the expanding universe as a fluid ( which includes radiation , matter and de ) with @xmath24velocity @xmath25 , proper isotropic density @xmath26 and proper isotropic pressure @xmath27 gives @xmath28 , where @xmath29 and @xmath30 .
the variable @xmath31 denotes the energy density of the scalar field and @xmath32 is the corresponding isotropic pressure .
moreover the parameters @xmath33 of the scalar field are given by @xmath34 for the flrw spacetime , for comoving observers ( @xmath35 ) , the einstein field equations ( [ sf.03 ] ) are @xmath36 and @xmath37 where @xmath38 is the hubble function . furthermore , assuming that the scalar field and matter do not interact , we have the two conservation laws@xmath39@xmath40 while the corresponding equation of state ( eos ) parameters are given by @xmath41 and @xmath42 .
in what follows we assume a constant @xmath43 so that @xmath44 ( @xmath45 for cold matter and @xmath46 for radiation ) , where @xmath47 is the matter density at the present time .
generically , some high - energy field theories suggest that the dark energy eos parameter may be a function of cosmic time ( see , for instance , @xcite ) . replacing ( [ sf.04 ] ) in ( [ sf.09 ] ) we have the klein - gordon equation@xmath48 where @xmath49 .
furthermore the corresponding dark energy eos parameter is @xmath50 which means that when @xmath51 then @xmath52 . on the other hand ,
if the kinetic term of the scalar field is negligible with respect to the potential energy , i.e. @xmath53 , then the equation of state parameter is @xmath54 . from the above analysis it becomes clear that the unknown quantities of the problem are @xmath55 @xmath14 and @xmath0 whereas we have only two independent differential equations available namely eqs .
( [ frie2 ] ) and ( [ sf.10 ] ) .
therefore , in order to solve the system of differential equations we need to make an additional assumption ( ansatz ) .
this usually concerns the functional form of the scalar field potential @xmath0 . in the literature ,
due to the unknown nature of de , there are many forms of this potential which describe differently the physical features of the scalar field ( for instance see @xcite ) .
as far as the exact solution of the field equations ( [ sf.07 ] ) , ( [ frie2 ] ) and ( [ sf.10 ] ) is concerned there are few solutions with spatial curvature @xcite and even fewer solutions are known for a perfect fluid and a scalar field @xcite . a special solution for a spatially flat flrw spacetime ( @xmath56 ) which contains a perfect fluid with a constant equation of state parameter @xmath57 and a scalar field with a constant equation of state parameter @xmath58 , has been found in @xcite . specifically in @xcite
it has been shown that under these assumptions one solves the field equations and finds the potential @xmath59 : @xmath60 ^{-\frac{2\gamma_{\phi}}{\gamma-\gamma_{\phi}}}. \label{bgp1}\ ] ] evidently , this solution is a special solution , in the sense that it exists for specific initial conditions , for example with @xmath61 constant . in the following we consider a spatially - flat flrw spacetime with a perfect fluid @xmath62 and a scalar field and assume that the potential has the generic form @xmath63
^{q } , \label{bgp}\ ] ] where @xmath64and @xmath65 are constants .
this potential ( [ bgp ] ) is a generalization of ( [ bgp1 ] ) and our aim is to determine exact solutions of the field equations for a particular relation between the constants @xmath66 of the potential ( [ bgp ] ) and the barotropic parameter @xmath67 of the perfect fluid .
in this section , we show that if the wdw equation admits lie symmetries which form an abelian lie algebra , then the wdw equation is liouville integrable ; that is , the field equations can be solved by quadratures .
a lie symmetry of a differential equation @xmath68 is the generator of the one parameter point transformation which leaves invariant the differential equation @xmath69 .
that means that if @xmath70 is a lie symmetry for @xmath71 then there exists a function @xmath72 such that the following condition holds @xcite @xmath73}(h)=\lambda h~,~\mathbf{mod}h=0\ ] ] where @xmath74}=\mathbf{x}+\eta_{i}^{a}\partial_{u_{,i}^{a}}+\eta _ { ij}^{a}\partial_{u_{,ij}^{a}}\mathbf{,~}$]is the second prolongation vector and @xmath75 is the total derivative . ] of @xmath76 .
the importance of lie symmetries is that each symmetry can be used to reduce the number of dependent variables .
solutions which follow from the application of lie symmetries are called invariant solutions
. for differential equations which arise from a variational principle there exists a special class of lie symmetries , the noether symmetries .
noether symmetries are lie symmetries which leave the action integral invariant .
according to noether s theorem to each noether symmetry there corresponds a conserved noether current .
the condition for a noether symmetry is that there exists a vector field @xmath77 such that the following condition is satisfied : @xmath78}l+ld_{i}\xi^{i}=d_{i}a^{i}~. \label{pr.08}\ ] ] the corresponding noether current @xmath79 is defined by the expression @xmath80 and it is conserved , that is , it satisfies the relation @xmath81 @xcite .
as we discussed above , the method of using the noether symmetries of the cosmological field equations has been applied by many authors in scalar - field cosmology , in @xmath82 gravity and other modified gravity theories .
recently @xcite , it has been proposed that the cosmological model will be determined by the existence of lie symmetries of the wdw equation of quantum cosmology .
this selection rule is more general than that imposed by the noether symmetries of the field equations , because , as it has been shown in @xcite , the wdw equation is possible to admit lie point symmetries while the lagrangian of the field equations does not admit noether point symmetries . in the following
we discuss the application of lie symmetries in the wdw equation .
specifically , we discuss the reduction process and we show how to construct noetherian conservation laws for a conformally related lagrangian of the field equations from the lie point symmetries of the wdw equation .
the lagrangian of the field equations in minimally coupled scalar field cosmology in a spatially flat flrw spacetime with a perfect fluid with a constant equation of state parameter @xmath62 is @xmath83 the field equations are the euler - lagrange equations of ( [ lang.01 ] ) with respect to the variables @xmath84 and are equations ( [ frie2 ] ) and ( [ sf.10 ] ) .
as the lagrangian is independent of time , we also have the hamiltonian constraint ( [ sf.07 ] ) , which , in terms of the momenta @xmath85 becomes @xmath86 finally , the field equations are equivalent to the following hamiltonian system:@xmath87@xmath88 the wdw equation is the klein gordon equation which is defined by the conformal laplacian operator .
the general conformal klein gordon equation is:@xmath89 where @xmath90 is the laplacian operator , @xmath91 is the metric of the space and @xmath92 .
we brake the lagrangian ( [ lang.01 ] ) in two parts .
the kinematic part which we consider as a riemannian space with dimension @xmath93 and line element@xmath94 which we call the minisuperspace , and the dynamic part which is defined by the potential @xmath95 $ ] .
specifically , the minisuperspace is a 2-dimensional lorentz manifold whose coordinates are the scale factor and the scalar field .
the symmetries of this space are related to the lie and the noether symmetries of the dynamical field equations .
the metric of the minisuperspace is defined by the kinematic part of the lagrangian ( [ lang.01 ] ) for the dynamics of the field . the rest of the lagrangian is considered to be the `` effective potential '' of the dynamical system that is defined by the gravitational and scalar fields .
therefore , using the minisuperspace ( [ lang.04 ] ) the wdw equations becomes @xmath96 \psi=0,\label{lang.05}\ ] ] where the laplacian operator @xmath97 is given by@xmath98 in @xcite , it was proved that the lie point symmetries of equation ( [ lang.03 ] ) are related to the conformal algebra of the minisuperspace @xmath91 .
more specifically , it has been shown that : a. the general form of the lie point symmetry vector is @xmath99 \partial_{\psi}\mathbf { , } \label{lang.07}\ ] ] where@xmath100 is a conformal killing vector of the minisuperspace , with conformal factor @xmath101 b. the lie point symmetry condition which constrains the potential is @xmath102 .
we require now that equation ( [ lang.03 ] ) admits as lie point symmetry the vector ( [ lang.07 ] ) . then under the coordinate transformation
@xmath103 so that @xmath104 , the lie symmetry vector ( [ lang.07 ] ) becomes@xmath105 \partial_{\psi}. \label{lang.08}\ ] ] there exist two equivalent methods to reduce the wdw equation by means of the symmetry vector ( [ lang.08 ] ) .
\a ) in the first method we calculate the zero - order invariants from the lagrange system ( @xmath106 @xmath107 which turn out to be@xmath108 . \label{lang.10}\ ] ] \b ) the second method is to write the lie point symmetry as a lie bcklund symmetry .
the lie point symmetry ( [ lang.08 ] ) is equivalent to the contact symmetry@xmath109 from which we obtain the differential equation@xmath110 we set @xmath111 and find that the solution of the reduced equation is@xmath112 \label{lang.13}\ ] ] from which it follows again that the coordinate @xmath113 is factored out from the solution of the wavefunction @xmath114 in the wkb approximation , @xmath115 the wdw equation reduces to a ( null ) hamilton - jacobi equation .
the latter can be seen as the hamilton - jacobi equation of a hamiltonian system moving in the same geometry under the conformal laplace operator of the wdw equation and with the same potential .
specifically , the wdw equation ( [ lang.05 ] ) in scalar field cosmology provides the null hamilton - jacobi equation : @xmath116 furthermore , in @xcite it was also shown that the symmetries of the wdw equation can be used in order to find noether point symmetries for classical particles .
however , the null hamilton - jacobi equation is separable if the @xmath117- dimensional hamiltonian system admits @xmath117 conservation laws @xmath118 ( symmetries ) i.e. @xmath117 corresponding noether symmetries which are independent and in involution , i.e. @xmath119 where @xmath120 denotes the poisson bracket .
if this is the case , then the hamiltonian system is liouville integrable @xcite .
that means that it is possible for the wdw equation to admit an invariant solution and at the same time the classical hamiltonian system to be not integrable .
therefore , in order for the wdw equation to admit an invariant solution and the hamiltonian system to be liouville integrable , the @xmath117- dimensional wdw _
_ equation must admit at least @xmath121 independent lie point symmetries , @xmath122 , which form an abelian lie algebra .
if this is the case , the zero - order invariants of these @xmath121 lie point symmetries will give the solution of the wdw equation in the form@xmath123 \mathbf{,}\ ] ] where @xmath124 are constants , @xmath125 and the function @xmath126 satisfies a linear second - order ode .
that is , when the field equations are liouville integrable by noether point symmetries then there exists a coordinate system where the wdw equation admits @xmath117 oscillatory terms in the solution and _ vice versa_. it is important to note that this result is more general and includes the one given in @xcite when @xmath127 that is , if one considers the killing algebra of the minisuperspace only .
we conclude that for the reduction / solution of the wdw we may consider directly the lie point symmetries of the wdw equation which are given in terms of the ckvs of the space instead of restricting ourselves to the noether point symmetries only , as has been done in @xcite .
below , we study the lie point symmetries and the wdw equation for the potentials of the form ( [ bgp ] ) which generalize the work done in @xcite .
we are considering a scalar field cosmological model which contains a quintessence scalar field with the potential of eq.([bgp ] ) and a perfect fluid with equation of state parameter @xmath128 . under these assumptions
the lagrangian of the field equations ( [ lang.01 ] ) becomes@xmath129 ^{q}-\rho_{m0}a^{-3\left ( \gamma-1\right ) } .
\label{ssf.01}\ ] ] from previous work on noether point symmetries in scalar - field cosmology @xcite , and on dynamical symmetries , @xcite we know that this lagrangian admits conservation laws when : a. the potential reduces to the exponential potential i.e. @xmath130 and b. we have the so called _ unified dark matter _ ( udm ) potential ( see paliathanasis et al . @xcite and references therein ) , i.e. @xmath131 and @xmath132 when the extra fluid is dust , namely @xmath133 . in the following we consider @xmath134 which implies that the current analysis generalizes the previous works of @xcite .
the hamiltonian ( [ lang.02 ] ) of the field equations for the lagrangian ( [ ssf.01 ] ) in terms of the momenta @xmath135 and @xmath136 is @xmath137 ^{q}+\rho_{m0}a^{-3\left ( \gamma-1\right ) } \right ) = 0 , \label{ssf.02}\ ] ] and the wdw equation ( [ lang.05 ] ) is @xmath138 ^{q}+\rho_{m0}a^{-3\left ( \gamma-1\right ) } \right ) ~\psi=0 . \label{ssf.03}\ ] ] applying the results of @xcite , we find that the second - order partial differential equation ( [ ssf.03 ] ) admits the generic lie point symmetry vector @xmath139 where @xmath140 is a constant are ckvs of the minisuperspace ( [ lang.04 ] ) .
] , @xmath141 \label{ssf.05}\\ x_{2 } & = a^{\frac{3\mu}{2}}\left [ \frac{\sqrt{6}}{6}a\cosh\left ( \frac{\sqrt{6}}{4}\mu\phi\right ) \partial_{a}+\sinh\left ( \frac{\sqrt{6}}{4}\mu\phi\right ) \partial_{\phi}\right ] .
\label{ssf.06}\ ] ] and the constants @xmath142 are related as follows:@xmath143 that is , the effective potential of the field equations is @xmath144 ^{-\frac{4}{\mu}-2}+\rho_{m0}a^{-3\left ( \mu+1\right ) } \right ) .
\label{ssf.08}\ ] ] therefore , for @xmath145 , we have that@xmath146 , @xmath147 i.e. we have the udm potential with dust ( for the exact solution and the observation constraints of that model see @xcite ) .
if the perfect fluid is a barotropic fluid , that is the barotropic index @xmath148 $ ] then , from ( [ ssf.07 ] ) , @xmath149 since @xmath150 .
however , if we require the perfect fluid to have a negative equation of state parameter , like a cosmological constant , then @xmath151 which means that @xmath152 . furthermore , when @xmath153 , i.e. there is no extra fluid , we have that @xmath154 . in the following ,
we apply the lie symmetry vector ( [ ssf.04 ] ) in order to construct the invariant solution of the wdw equation ( [ ssf.03 ] ) and to solve the null hamilton - jacobi equation of the hamiltonian ( [ ssf.02 ] ) in order to reduce the order of the field equations . in the following section we study the case @xmath155 and in appendix [ general ] we present the general solution for @xmath156 .
in this section we determine the exact solution of the field equations and of the wdw equation for the quintessence scalar field .
we consider the case @xmath158 ( the case @xmath159 is equivalent to that case ) . under the coordinate transformation .
however when @xmath153 it is possible to have @xmath160 in that case all calculations remain valid provided we replace @xmath161 in ( [ ssf.08 ] ) and in the subsequent coordinate transformations .
] : @xmath162 the effective potential ( [ ssf.08 ] ) becomes@xmath163 and the wdw equation is @xmath164 = 0 \label{ssf.11}\ ] ] where @xmath165 . in these coordinates the lie point symmetry vector ( [ ssf.04 ] ) is @xmath166 therefore , the solution of equation ( [ ssf.11 ] ) admits an oscillatory term , i.e.@xmath167 where@xmath168 furthermore , in the wkb approximation , @xmath169 and equation@xmath170([ssf.11 ] ) becomes@xmath171 = 0 , \label{ssf.13}\ ] ] which is the null hamilton - jacobi equation which describes the field equations .
the solution of ( [ ssf.13 ] ) is @xmath172 therefore , the field equation is reduced to the following two - dimensional system , @xmath173 in order to simplify the system ( [ ssf.15 ] ) further we apply the transformation @xmath174 @xmath175 and the dynamical system becomes @xmath176 the exact solution of the system ( [ ssf.16 ] ) is expressed in terms of elliptic functions then @xmath177 , which is the udm solution for @xmath178 of @xcite . ] .
we perform a numerical integration of the non - linear system ( [ ssf.15 ] ) and in fig .
[ fig1dr ] we give the evolution of the equation of state parameter for the scalar field @xmath179 and for the total fluid @xmath180 for various values of the constant @xmath181 in the case @xmath145 . for @xmath145 the extra perfect
fluid is dust , i.e. @xmath133 . concerning the values of the cosmological parameters , we use the planck priors @xcite , namely @xmath182 and @xmath183km / s / mpc which imply @xmath184 , in units of @xmath18 .
from fig.[fig1dr ] , we observe that the scalar field mimics the cosmological constant for small values of the constant @xmath181 , however for large values of @xmath181 the scalar field has an eos parameter @xmath185 .
we find that within a physical range of the above cosmological parameters the corresponding dark energy eos parameter deviates by @xmath186 .
furthermore , from the evolution of @xmath187 we see that there is a matter - dominated epoch .
however , as the parameter @xmath181 increases , this epoch has shorter duration . in what follows we study the case @xmath188 and express analytically the scalar field and the hubble function in terms of the scale factor . *
* for the scalar field and @xmath180 for the total fluid for the lagrangian ( [ ssf.01 ] ) with @xmath189 in ( [ ssf.15 ] ) we have taken the minus sign and @xmath145 ( dust fluid : @xmath45 ) . for the numerical solution we use @xmath190 @xmath191 with @xmath182 and @xmath192km / s / mpc ( or @xmath193 in units of @xmath194 ) .
the solid line is for @xmath195 ; the dashed line is for @xmath196 the dotted line is for @xmath197,height=264 ] when @xmath188 , from ( [ ssf.15 ] ) we have @xmath198 , hence from ( [ ssf.09 ] ) it follows that @xmath199 .
furthermore , from the transformation ( [ ssf.09 ] ) and from ( [ ssf.15 ] ) we find that @xmath200 -a^{-3\mu}\left ( y_{0}^{2}+a^{-3\mu}\right ) ^{-\frac{2}{\mu}-1}}{y_{0}^{2}\left [ \frac{\omega_{m0}}{\omega_{\lambda0}}+\left ( y_{0}^{2}+a^{-3\mu}\right ) ^{-\frac{2}{\mu}-1}\right ] + a^{-3\mu}\left ( y_{0}^{2}+a^{-3\mu}\right ) ^{-\frac{2}{\mu}-1 } } \label{ssf.19}\ ] ] where we have set @xmath201 here , we would like to note that we call @xmath202 the density parameter of the cosmological constant - like term and @xmath4 the parameter of the perfect fluid , hold @xmath203 ; however the equality holds only when the constant @xmath204 [ @xmath205 which means that the scalar field act as a cosmological constant , i.e. @xmath206 , see eq .
( [ sf.220 ] ) . ] . therefore ,
for the scalar field density @xmath207 we have : @xmath208 .
\label{ssf.20}\ ] ] then ( [ sf.07 ] ) implies that@xmath209 + a^{-3\mu}\left ( y_{0}^{2}+a^{-3\mu}\right ) ^{-\frac{2}{\mu}-1}\right ) .
\label{ssf.21}\ ] ] we note that if @xmath210 then@xmath211 \right ) .
\label{ssf.22}\ ] ] if @xmath204 , then ( [ ssf.22 ] ) becomes @xmath212 which is obvious because when @xmath204 , we have @xmath213 and @xmath214 which means that the scalar field acts as a cosmological constant . furthermore , from ( [ ssf.21 ] ) and @xmath215 , we have the constraint@xmath216 -1=0 . \label{sf.220}\ ] ] in the following section we consider special values of the barotropic constant @xmath217 .
when the perfect fluid is dust then @xmath218 ( @xmath45 for other cases see appendix a ) eq.([ssf.21 ] ) takes the following form @xmath219 = \omega_{m0}a^{-3}+\delta h(a)\;. \label{sf.22a}\ ] ] it should be mentioned that the last term @xmath220 of the normalized hubble function ( [ sf.22a ] ) introduces a cosmological constant - like fluid , dust and stiff matter .
furthermore , from ( [ sf.220 ] ) we have the following algebraic equation@xmath221 hence , the discriminant of the polynomial ( [ sf.22b ] ) ( for @xmath222 ) is@xmath223 and @xmath224 when @xmath225 . recall that @xmath226 , \omega_{\lambda0}\in(0,1],$ ] and because @xmath227 we have the solution @xmath228 let us now compute the effective dark energy eos @xmath229 for the scalar field model introduced above .
it is well known that one can express the effective dark energy eos parameter in terms of the normalized hubble parameter @xcite @xmath230 where @xmath231 . inserting the second equality of eq.([sf.22a ] ) into eq.([eos22 ] ) , the effective dark energy eos parameter takes the following form ( see @xcite ) : @xmath232 which implies that any modifications to the effective eos parameter are included in the second term of eq .
( [ eos222 ] ) . inserting eq .
( [ sf.22a ] ) into eq .
( [ eos222 ] ) it is straightforward to obtain a simple analytical expression for the effective dark energy eos parameter : @xmath233 our dynamical system is integrable in the case of a single perfect fluid . here ,
we introduce two perfect fluids ( for example dust and radiation ) .
consider that we have dust ( @xmath45 ) and another perfect fluid with equation of state parameter @xmath234 ( for radiation @xmath235 ) , and @xmath236 the latter implies that the equation of state parameter which is associated with the two perfect fluids is @xmath237 that is , @xmath238 and @xmath239 .
when @xmath240 we have that @xmath241 and when @xmath242 holds we have @xmath243 .
we replace ( [ ssf.21a ] ) in ( [ ssf.21 ] ) and perform a taylor expansion near @xmath244 ( @xmath147 or @xmath145 ) .
we find@xmath245 where the normalized hubble parameter @xmath246 is given by eq .
( [ sf.22a ] ) and @xmath247 where @xmath248 and when @xmath204 , @xmath249 . in fig .
( [ fig22 ] ) we give the numerical solutions of the total eos parameter @xmath180 for @xmath250 and @xmath188 . for the total fluid for the lagrangian ( [ ssf.01 ] ) with @xmath251 where in ( [ ssf.15 ] ) we take the minus sign and @xmath188 . for the numerical solution we use @xmath252 and
@xmath193.@xmath170the solid line is for @xmath145 [ dust : @xmath133 ] , the dash - dotted line is for @xmath254 and the dotted line is for @xmath255,height=264 ] in appendix [ appena ] , we give the exact solutions for the hubble function , @xmath256 , for the cases where @xmath257 ( radiation fluid ) and @xmath258 ( curvature - like fluid ) .
in order to complete our analysis of the model with lagrangian ( [ ssf.01 ] ) , we perform a dynamical analysis of the field equations by studying the fixed points of the field equations .
we introduce the new dimensionless variables @xcite@xmath259 and the lapse time @xmath260 . in the new variables
the field equations reduce to the following first - order odes @xmath261 \label{ssf.48}\]]@xmath262 \label{ssf.49}\]]@xmath263 where @xmath264 and the friedmann equation ( [ sf.07 ] ) gives the constraint @xmath265 , where @xmath266 . in this case the second friedmann equation ( [ frie2 ] ) becomes@xmath267 which gives that the total eos parameter @xmath268 as a function of @xmath43 @xmath269and @xmath270 : @xmath271 furthermore , the eos parameter @xmath272for the scalar field is @xmath273 .
note that at any point @xmath274 from ( [ ssf.51 ] ) the solution of the scalar factor is a power law as long as @xmath275 that is , @xmath276 for @xmath277 and @xmath278 for @xmath279 @xmath280 in the following we consider in ( [ ssf.01 ] ) @xmath281 so that the potential of the scalar field is @xmath282 @xcite .
for this potential we write @xmath283 as a function of @xmath72 , i.e. @xmath284 , and equation ( [ ssf.50 ] ) becomes @xmath285 equations ( [ ssf.48 ] ) , ( [ ssf.49 ] ) and ( [ ssf.53 ] ) describe an autonomous dynamical system in the @xmath286 space .
furthermore from the constraints @xmath287 the variables @xmath288 are bounded in the ranges @xmath289 $ ] , @xmath290 ~$]from which follows that that the points @xmath288 belong to a half disk ; however for the parameter @xmath72 there is no constraint that implies that @xmath291 @xcite .
furthermore , we consider @xmath292 .
the fixed points of the dynamical system ( [ ssf.48 ] ) , ( [ ssf.49 ] ) and ( [ ssf.53 ] ) and the corresponding cosmological parameters are given in table [ fixedpoints ] .
the eigenvalues of the linearized dynamical system near the fixed points are given in table [ eigenvalues ] @xmath293 .
[ c]cccccc**point * * & @xmath294 & @xmath295 & @xmath296 & @xmath297 & acceleration + @xmath298 & @xmath299 & @xmath300 & @xmath5 & @xmath301 & @xmath302 + @xmath303 & @xmath304 & @xmath305 & @xmath300 & @xmath300 & no + @xmath306 & @xmath307 & @xmath305 & @xmath300 & @xmath300 & no + @xmath308 & @xmath309 & @xmath305 & @xmath310 & @xmath310 & yes + @xmath311 & @xmath312 & @xmath305 & @xmath313 & @xmath313 & @xmath314 + @xmath315 & @xmath316 & @xmath317 & @xmath5 & @xmath5 & @xmath302 + @xmath318 & @xmath319 & @xmath317 & @xmath5 & @xmath5 & @xmath302 + [ fixedpoints ] [ c]cccc**point * * & @xmath320 & @xmath321 & @xmath322 + @xmath298 & @xmath305 & @xmath323 & @xmath324 + @xmath303 & @xmath325 & @xmath326 & @xmath327 + @xmath306 & @xmath325 & @xmath328 & @xmath329 + @xmath308 & @xmath330 & @xmath331 & @xmath332 + @xmath311 & @xmath333 & @xmath334 & @xmath335 + @xmath336 & @xmath337 & @xmath338 $ ] & @xmath339 $ ] + [ eigenvalues ] point@xmath340 exists for all values of the parameter @xmath72 and corresponds to the matter epoch @xmath341 the total eos parameter is @xmath342 .
since there exists at least one positive eigenvalue , @xmath343 the point @xmath298 is always unstable . at this point
the universe accelerates if and only if @xmath302 . at the points @xmath344 and
@xmath306 the universe is dominated by the kinetic energy of the scalar field @xmath345 which means that the scalar field acts as a stiff fluid , i.e. @xmath346 which provides a decelerating universe .
these points exist when @xmath347 for arbitrary @xmath3 . for these points there exist positive eigenvalues of the linearized system , @xmath348 , for @xmath292 , hence these critical points are always unstable .
point @xmath308 is the de sitter solution , @xmath349 where the scalar field acts as a cosmological constant and the matter component vanished .
this point exists for all values of the constants @xmath3 and could be the future attractor of the universe . from the eigenvalues of table [ eigenvalues ] for that point
we have that it is stable when @xmath350 ( for a similar solution see @xcite ) .
the points @xmath311 correspond to a scalar field dominated universe @xmath351 and exist only when @xmath352 the total eos parameter is that of the scalar field @xmath353 which gives an accelerated universe when @xmath354 .
the points @xmath311 are stable for @xmath355 and@xmath356 .
hence , we see that @xmath311 are stable points and describe an accelerated universe when @xmath51 and @xmath354 .
furthermore , in the limit @xmath357 these points correspond to the de sitter universe ; when @xmath358 then @xmath359 . finally , the points @xmath336 are the so called scaling solutions where @xmath360 and the scalar field mimics the matter component of the universe , i.e. @xmath361 the points @xmath362exist when @xmath363 and they are stable when @xmath364 whereas the points @xmath365 exist when @xmath366 and are stable when @xmath355 , @xmath367 .
the total eos parameter is @xmath342 so they lead to an accelerated universe when @xmath302 . in fig .
[ dyne ] , we give the two - dimensional phase portrait in the @xmath368 plane and the three - dimensional phase portrait of the model with values @xmath369 and @xmath370 . we observe that for @xmath371 the two stable points are the points @xmath311 whereas for @xmath372 the stable points are @xmath373 . plane and in the @xmath286 space for the potential @xmath374 left - hand figures are for the variables @xmath375 and the right - hand figures are for the variables @xmath376 for @xmath377the stable points are the points @xmath378 ( scaling solutions ) while for @xmath371 the two stable points are the points @xmath379 .
the solid lines are for the initial condition @xmath380 , the dashed lines for @xmath381 and the dotted lines for @xmath382 . , height=377 ] it is important to study the case when the constants @xmath383 @xmath384are related to the constant @xmath385 so as to render the integrable field equations ( [ ssf.07 ] ) .
this is the case we studied in the previous section .
since we considered@xmath386 we have that @xmath387 .
hence , for the integrable case we have that the points @xmath388 @xmath389 @xmath390exist and they are always unstable .
the point @xmath298 has @xmath391 when @xmath392 and the points @xmath393 describe an accelerated universe so long as @xmath394 .
the point @xmath308 exists and it is the unique stable point . finally ,
the tracker solutions , i.e. points @xmath395 do not exist for @xmath396 .
the existence and the stability of the fixed points for general values @xmath3 and for the integrable case are given in table [ stability ] .
[ c]ccccc**point * * & * existence * & * stability * & * stability for * @xmath397 & * acceleration * + @xmath298 & @xmath398 & unstable & unstable & @xmath392 + @xmath303 & @xmath398 & unstable & unstable & no + @xmath306 & @xmath398 & unstable & unstable & no + @xmath308 & @xmath398 & stable for @xmath399 & stable & yes + @xmath311 & @xmath400 & stable for @xmath401 & unstable & @xmath402 + @xmath403 & @xmath404 & stable for @xmath405 , @xmath406 & @xmath301 & @xmath301 + @xmath365 & @xmath407 & stable for @xmath408 & @xmath301 & @xmath301 + [ stability ] in fig .
[ dyndust ] we give the two - dimensional phase portrait in the @xmath368 plane and the three - dimensional phase portrait of the the model with values @xmath409 which correspond to the integrable case for @xmath145 .
the point @xmath308 is the unique stable point .
we observe that the points @xmath311 act as attractors in the plane @xmath410 for @xmath411 and the solutions reach the boundary where @xmath412 , and move to the de sitter points @xmath413 .
it is important to note that the existence of conservation laws in the field equations which follow from the lie point symmetries of the wdw equation , i.e. the dynamical system is liouville integrable , gives us constraints on the free parameters of the model so that there exists a unique stable point which describes the de sitter universe . plane and in the @xmath286 space for the potential @xmath414 with variables @xmath415 which corresponds to the integrable case for @xmath145 .
the point @xmath308 is the unique stable point .
the points @xmath311 act as attractors in the plane @xmath410 for @xmath411 . from the right - hand plot
we observe that the solutions reach a boundary where @xmath416 @xmath185 and from there they move to the de sitter point @xmath308 .
the solid lines are for initial condition @xmath380 , the dashed lines for @xmath381 and the dotted lines for @xmath382 .
, height=264 ]
we have applied lie symmetry methods in order to extend the works of rubano and barrow @xcite and paliathanasis et al .
@xcite in scalar field cosmology for a general family of potentials , @xmath0 .
we have shown that there exists a unique connection between the lie point symmetries of the wdw equation and the conservation laws of the field equations .
we considered a general form of @xmath0 which contains hyperbolic functions for a scalar field with a perfect fluid and we have investigated the existence of lie point symmetries of the wdw equation .
this approach is more general than the application of noether point symmetries .
we recovered the result of @xcite , that is , that in scalar field cosmology amongst the variety of @xmath0 potentials the hyperbolic types play a key role because they admit conservation laws .
moreover , based on the lie point symmetries of the wdw equation , we have obtained the exact solutions of the field equations .
finally , we have performed a dynamical system analysis by studying the fixed points of the field equations in dimensionless variables .
we found various dynamical cases among which , if the current cosmological model is liouville integrable there is a unique stable point which describes the de - sitter universe as a late - time attractor for the dynamics .
ap acknowledges financial support of infn .
jdb acknowledges support from the stfc .
here we provide some details concerning the solutions of section [ sub51 ] . specifically , for a radiation perfect fluid , @xmath257 ( hence @xmath417 ( @xmath46 ) and @xmath418 eq.([ssf.21 ] ) gives @xmath419 \;.\ ] ] on the other hand , when @xmath420 ( or @xmath421 ) , the perfect fluid has an equation of state @xmath422 ( which can also be seen as the curvature term in a non - flat flrw spacetime ) ; then @xmath423 and:@xmath424 .\ ] ] this is a solution of the scalar field cosmology in a curved flrw spacetime .
in the general case where @xmath425 and @xmath426 we apply the following coordinate transformations,@xmath427 ^{-\frac { 1}{3\mu}}~,~\phi=\frac{2\sqrt{6}}{3\mu}\arctan h\left ( \frac{y}{x-\frac { y}{\alpha}}\right ) .\ ] ] in the new coordinate system the wdw equation becomes@xmath428 ^{\frac{1+\mu } { \mu}}\left [ -\left ( 1-\frac{1}{\alpha^{2}}\right ) \psi_{,xx}+\frac { 2}{\alpha}\psi_{,xy}+\psi_{,yy}+\left ( 2v_{0}^{\prime}\alpha^{-\frac{4}{\mu } -2}x^{-\frac{4}{\mu}-2}+2\rho_{m0}^{\prime}\right ) \psi\right ] = 0 , \label{ssf.281}\ ] ] where @xmath429 and the lie symmetry vector is ( [ ssf.04 ] ) @xmath166 therefore , the invariant solution of the wdw equation ( [ ssf.281 ] ) is @xmath430 where @xmath431 satisfies the following second - order ode@xmath432 = 0 . \label{ssf.2821}\ ] ] when @xmath433 which is the case of the exponential scalar field , from ( [ ssf.2821 ] ) we have @xmath434 furthermore , in the wkb approximation the wdw equation ( [ ssf.281 ] ) becomes the null hamilton - jacobi equation@xmath428 ^{\frac{1+\mu } { \mu}}\left [ -\left ( 1-\frac{1}{a^{2}}\right ) \left ( \frac{\partial \bar{s}}{\partial x}\right ) ^{2}+\frac{2}{\alpha}\left ( \frac{\partial \bar{s}}{\partial x}\right ) \left ( \frac{\partial\bar{s}}{\partial y}\right ) ^{2}+\left ( \frac{\partial\bar{s}}{\partial y}\right ) ^{2}+\left ( 2v_{0}^{\prime}x^{-\frac{4}{\mu}-2}+2\rho_{m0}^{\prime}\right ) \right ] = 0 \label{sf.021}\ ] ] with solution of the form @xmath435where@xmath436
and@xmath437 from the solution of the hamilton - jacobi equation ( [ sf.021 ] ) we can reduce the equivalent hamiltonian system of the field equation to the following system of first order equations@xmath438 \left [ \left ( x-\frac{y}{\alpha}\right ) ^{2}-y^{2}\right ] ^{\frac{1+\mu}{\mu}}~,\label{ssf.33}\\ \dot{y } & = \left [ \frac{1}{\alpha}p_{x}+p_{y}\right ] \left [ \left ( x-\frac{y}{\alpha}\right ) ^{2}-y^{2}\right ] ^{\frac{1+\mu}{\mu}}~ , \label{ssf.34}\ ] ] where @xmath439 and @xmath440 .
in order to make the reduced system simpler , we apply the transformation @xmath441and the system ( [ ssf.33])-([ssf.34 ] ) becomes@xmath442 for the exponential potential @xmath443 , from system ( [ ssf.37 ] ) we find the solution in closed form .
the solution is @xmath444@xmath445 where the scale factor is @xmath446 this is the solution of the exponential scalar field with matter in the einstein frame .
recall that the matter has an equation of state parameter of the form @xmath447 .
when @xmath153 and @xmath188 , from the dynamical system ( [ ssf.37 ] ) we have the solution @xmath448 ^{\frac{\mu}{2\left ( 1+\mu\right ) } } ~,~y\left ( \tau\right ) = -\frac{\alpha}{\alpha^{2}-1}x\left ( \tau\right ) -y_{0}.\ ] ] where @xmath449 and @xmath450 . in the case of @xmath145 the solution of the system ( [ ssf.37 ] ) is@xmath451 ( recall that for @xmath145 , we have @xmath452 .
hence , the solution for the scale factor is @xmath453 ^{\frac{1}{3}}\;. \label{ssf.44}\ ] ] furthermore , from the singularity condition @xmath454 , we have that @xmath455 which implies @xmath456where @xmath457 .
we obtain @xmath458 and @xmath459 $ ] .
therefore , we can express the hubble function in terms of the scale factor , i.e. @xmath460 where@xmath461 which means that the scalar field introduces an effective dark matter component , namely @xmath462 in the cosmic expansion .
m. tegmark _ et al .
_ , astrophys . j. * 606 * , 702 ( 2004 ) ; d. n. spergel _ et al .
_ , astrophys .
j. supplt .
* 170 * , 377 ( 2007 ) ; t. m. davis _ et al . _ ,
astrophys .
j. * 666 * , 716 ( 2007 ) ; m. kowalski _ et al .
_ , astrophys .
j. * 686 * , 749(2008 ) ; g. hinshaw _ et al .
_ , astrophys . j. supplt .
* 180 * , 225 ( 2009 ) ; j. a. s. lima and j. s. alcaniz , mon . not .
soc . * 317 * , 893 ( 2000 ) ; j. f. jesus and j. v. cunha , astrophys .
j. lett .
* 690 * , l85 ( 2009 ) ; s. basilakos and m. plionis , astrophys .
j. lett .
* 714 * , 185 ( 2010 ) ; e. komatsu e. et al . , 2011 ,
astrophys .
j. sup . , * 192 * , 18 ( 2011 ) ; g. hinshaw et al .
, astrophys .
j. sup . *
208 * , 19 ( 2013 ) ; o. farooq , d. mania and b. ratra , astrophys .
j. , * 764 * , 138 ( 2013 ) ; p. a. r. ade et al . , ( planck collaboration ) , astronomy and astrophysics * 571 * , a16 ( 2014 ) e. j. copeland , m. sami and s. tsujikawa , intern . journal of modern physics d , * 15 * , 1753,(2006 ) ; l. amendola and s. tsujikawa , _ dark energy theory and observations _ , cambridge university press , cambridge uk , ( 2010 ) . l. a. urena - lopez , t. matos , phys
d * 62 , * 081302 ( 2000 ) ; v. sahni and a. starobinsky , int .
d * 9 * 373 ( 2000 ) ; j.a.e .
carrillo , j.m .
silva and j.a.s .
lima , astr .
astroph . : new phenomena and new states of matter in the universe , proceedings of the third workshop ( iwara07 ) ( arxiv:0806.3299 ) s. capozziello , e. piedipalumbo , c. rubano and p. scudellaro , phys .
* 80 * 104030 ( 2009 ) ; m. szydlowski et al .
, * 38 * , 795 , ( 2006 ) ; yi zhang , yun - gui gong and zong - hong zhu , phys .
b. , * 688 * 13 ( 2010 ) ; m. sharif and i. shafique , phys .
d * 90 * 084033 ( 2014 ) ; d. momeni , r. myrzakulov and e. gdekli ( arxiv:1502.00977 ) ; l. collodel and g.m .
kremer , ( arxiv:1411.3580 ) a. paliathanasis , m. tsamparlis , s. basilakos and s. capozziello , phys .
d. * 89 * 063532 ( 2014 ) ; a. paliathanasis , s. basilakos , e.n .
saridakis , s. capozziello , k. atazadeh , f. darabi and m. tsamparlis , phys .
d. * 89 * 104042 ( 2014 ) ; a. paliathanasis and m. tsamparlis , phys .
d. * 90 * ( 2014 ) 043529 ; s. basilakos , s. capozziello , m. de laurentis , a. paliathanasis and m. tsamparlis , phys .
d. * 88 * 103526 ( 2013 ) | we give a general method to find exact cosmological solutions for scalar - field dark energy in the presence of perfect fluids .
we use the existence of invariant transformations for the wheeler de witt ( wdw ) equation .
we show that the existence of a point transformation under which the wdw equation is invariant is equivalent to the existence of conservation laws for the field equations , which indicates the existence of analytical solutions .
we extend previous work by providing exact solutions for the hubble parameter and the effective dark - energy equation of state parameter for cosmologies containing a combination of perfect fluid and a scalar field whose self - interaction potential is a power of hyperbolic functions .
we find solutions explicity when the perfect fluid is radiation or cold dark matter and determine the effects of non - zero spatial curvature . using the planck 2015 data ,
we determine the evolution of the effective equation of state of the dark energy . finally , we study the global dynamics using dimensionless variables .
we find that if the current cosmological model is liouville integrable ( admits conservation laws ) then there is a unique stable point which describes the de - sitter phase of the universe . |
the hera @xmath0 collider , located in hamburg , germany , was in operation in the years @xmath2@xmath3 .
protons with an energy up to @xmath4 gev were brought into collision with electrons or positrons of energy @xmath5 gev at two experiments , h1 and zeus , each of which collected a data sample with a total integrated luminosity of about @xmath6 fb@xmath7 .
the deep inelastic scattering ( dis ) interactions produced at hera , at a centre of mass energy @xmath8 of up to @xmath9 gev , provided an ideal environment to study rare processes , set constraints on the standard model ( sm ) and search for new particles and physics beyond the standard model ( bsm ) .
searches performed by the h1 collaboration for leptoquarks , lepton flavour violating leptoquarks and excited fermions are presented in the following .
analyses of other rare processes at hera are described elsewhere in these proceedings .
the hera data provide the unique possibility to search for the resonant production of new particles that couple directly to a lepton and a parton .
leptoquarks ( lqs ) , which appear naturally in various unifying bsm theories , are such an example . at hera
, lqs may be resonantly produced in the @xmath10-channel up to the centre of mass energy or virtually exchanged in the @xmath11-channel between the initial state lepton and a quark coming from the proton . beyond the centre of mass energy the production mechanism proceeds via contact interactions .
a search for lq production is performed by h1 in the @xmath12 lq framework of the buchmller , rckl and wyler ( brw ) model @xcite , using the full h1 hera data set .
lqs are classified in terms of fermion number @xmath13 , where @xmath14 and @xmath15 are the baryon and lepton number respectively . due to the more favourable density of quarks with respect to anti
quarks at high @xmath16 , the @xmath17 data are mostly sensitive to lqs with fermion number @xmath18 , whereas the @xmath19 data are more sensitive to @xmath20 lqs .
h1 results based on hera i data are presented in @xcite . in the search for first generation lqs ,
the processes @xmath21 and @xmath22 are considered .
these lq decays lead to final states similar to those of neutral current ( nc ) and charged current ( cc ) dis interactions , which constitute an irreducible sm background .
no signal is observed in the dis mass spectra and constraints on lq production are set , which extend beyond the previously excluded domains . for a yukawa coupling @xmath23 of electromagnetic strength , lq masses below @xmath24@xmath25 gev
are ruled out by h1 , depending on the lq type @xcite .
figure [ fig : lq ] presents the constraints on @xmath23 for the @xmath26 and @xmath27 lqs obtained by h1 as a function of lq mass , @xmath28 .
limits from lep @xcite are also shown , derived from indirect constraints on the process @xmath29 , as well as limits from the tevatron @xcite , where lqs would be pair produced via the strong interaction resulting in a production rate that is independent of @xmath23 . as mentioned above
, contact interactions can also be used to describe the effects of virtual lq production or exchange at hera in the limit of large lq mass @xmath30 and indirect lq limits from the zeus contact interaction analysis @xcite are also illustrated in figure [ fig : lq ] .
a search for lepton flavour violating ( lfv ) processes in @xmath31 collisions mediated by lqs is also performed by h1 . at hera , the lfv processes @xmath32 and @xmath33 may lead to final states with a muon or tau lepton , together with a hadronic jet .
h1 results based on hera i data are presented in @xcite , where searches for both second and third generation lqs are included .
a new search for second generation @xmath18 lqs using the complete h1 hera ii @xmath17 data is presented here @xcite .
events are selected with a high @xmath34 muon and large missing calorimetric @xmath34 . the resulting lq mass spectrum is shown in figure [ fig : lfv ] ( left ) .
two candidate data events are observed in agreement with the sm prediction , which , in contrast to the first generation , is negligible . in the absence of a lq signal
, limits are derived for the @xmath35 @xmath18 lq types as a function of @xmath28 . for yukawa couplings of electromagnetic strength , lfv lqs coupling to a muon
quark pair are ruled out for lq masses up to @xmath36 gev .
figure [ fig : lfv ] ( right ) presents the constraints for the coupling @xmath37 for the @xmath26 lq obtained by h1 as a function of @xmath28 . a direct limit obtained at the tevatron from leptoquark pair production
is also shown for comparison @xcite .
the existence of excited states of leptons and quarks is a natural consequence of models assuming composite fermions , and their discovery would provide convincing evidence of a new scale of matter .
the production and decay of such particles is described in gauge mediated ( gm ) and contact interaction ( ci ) models .
the excited electron decays @xmath38 , @xmath39 and @xmath40 with subsequent hadronic or leptonic decays of the @xmath41 and @xmath42 bosons are examined at hera in a search for excited electrons by h1 @xcite . in this search , which uses the complete h1 @xmath0 data sample ,
each decay channel is investigated and no indication of a signal is observed .
new limits on the production cross section of excited electrons are obtained within a gm model @xcite , where an upper limit on @xmath1 as a function of the excited electron mass is established for the specific relation @xmath43 between the couplings .
assuming @xmath44 and @xmath45 excited electrons with a mass lower than @xmath46 gev are excluded at @xmath47 c.l , as shown in figure [ fig : estar ] ( left ) . also shown
are direct @xcite and indirect @xcite exclusion limits obtained at lep , as well as a result from the tevatron @xcite . for the first time in @xmath31 collisions , gauge and four fermion contact interactions are also considered together for @xmath48 production and decays , assuming @xmath49 .
the ci term improves the limit on @xmath50 only slightly , demonstrating that the gm mechanism is dominant for excited electron processes at hera , as illustrated in figure [ fig : estar ] ( right ) .
a search for the production of excited neutrinos is performed by h1 , using the full @xmath17 data sample , integrated luminosity @xmath51 pb@xmath7 @xcite . due to the helicity dependence of the weak interaction and
given the valence quark composition and density distribution of the proton , the @xmath52 production cross section is predicted to be much larger for @xmath53 collisions than for @xmath54 .
the excited neutrino decay channels @xmath55 , @xmath56 and @xmath57 with subsequent hadronic or leptonic decays of the @xmath41 and @xmath42 bosons are considered and no indication of a @xmath58 signal is found .
upper limits on the coupling @xmath1 as a function of the excited neutrino mass are established within the gm model for specific relations between the couplings .
assuming @xmath59 and @xmath60 , excited neutrinos with a mass lower than @xmath61 gev are excluded at @xmath62% c.l .
, as shown in figure [ fig : nuqstar ] ( left ) . the best limit obtained at lep
is also indicated @xcite .
the excluded region is greatly extended with respect to previous results , based on only @xmath63 pb@xmath7 of @xmath17 data @xcite , and demonstrates the unique sensitivity of hera to excited neutrinos with masses beyond the lep reach .
finally , a search for excited quarks is performed using the full h1 @xmath0 data sample @xcite .
the electroweak decays of excited quarks @xmath64 , @xmath65 and @xmath66 with subsequent hadronic or leptonic decays of the w and z bosons are considered .
no evidence for excited quark production is found .
mass dependent exclusion limits on the ratio @xmath67 are derived within the gm model , as illustrated in figure [ fig : nuqstar ] ( right ) . assuming @xmath43 , no strong interactions @xmath68 and
@xmath69 , excited quarks with a mass lower than @xmath70 gev are excluded at @xmath47 c.l . the best limit obtained at lep also indicated @xcite , assuming that the branching ratio @xmath71(@xmath64 ) = @xmath72 .
these limits extend the excluded region compared to previous h1 excited quark searches @xcite . | recent results of searches for leptoquarks , lepton flavour violating leptoquarks and excited fermions ( electrons , neutrinos and quarks ) with the h1 experiment at hera are presented , which use up to the full @xmath0 h1 data set .
no evidence for the direct or indirect production of such particles is found .
the results are interpreted in terms of limits on the yukawa coupling of leptoquarks and lepton flavour violating processes and for excited fermions on the ratio of the coupling parameter to the compositeness scale @xmath1 , mainly in the framework of gauge mediated interactions .
the derived limits extend the excluded regions to higher masses than those reported in previous results . |
today , unprecedented efforts are devoted to the study of novel topological insulating phases , which exhibit remarkable properties at their boundaries.@xcite in two dimensions , these quantum states are characterized by propagating edge states , which carry dissipationless currents along one - dimensional boundaries . from a theoretical point of view
, these transport properties rely on the existence of large bulk gaps in the energy spectrum , which host robust and gapless edge excitations . different edge state structures and transport properties
can be realized according to the nature of the gaps : a magnetic field breaks time - reversal symmetry ( trs ) and leads to chiral edge states ( i.e. , particles propagating in a given direction ) and quantized hall currents,@xcite whereas a large spin - orbit coupling preserves the trs and produces helical edge structures.@xcite in the latter situation , referred to as the quantum spin hall ( qsh ) effect , edge states with opposite spins counterpropagate and contribute to a vanishing hall current , while producing a spin hall current.@xcite the origins of the quantum hall ( qh ) and qsh phases are fundamentally different , as the former is produced by an external field , whereas the latter relies on the intrinsic properties of the material .
however , one can interpret the qsh phase as being two opposite qh phases , one for each spin component .
this observation is easily shown through the kane - mele lattice model,@xcite which realizes the qsh effect by the inclusion of the intrinsic spin - orbit coupling ( iso ) , and corresponds to two copies of the ( qh ) haldane model.@xcite in recent works , the iso coupling has been combined with a magnetic field,@xcite or with a constant exchange term,@xcite in order to study the effect of the trs breaking on the qsh effect .
it has been shown that the qsh effect persists in the presence of magnetic field , leading to the `` trs - broken''@xcite or `` weak '' qsh
effect.@xcite the term `` weak '' refers to the absence of robustness against spin - flip scattering due to magnetic disorder . in absence of this scattering process ,
the hall and spin hall conductivities are protected and quantized . in this work ,
we present a different approach to study the interplay between qh and qsh physics .
contrarily to the general trend in the field of topological insulators , which is to include many terms and study complicated hamiltonians , here we investigate the minimal models that can produce topological phases .
we consider a tight - binding model of the honeycomb lattice under a perpendicular magnetic field , which includes the iso coupling , the rashba spin - orbit ( rso ) coupling , and the zeeman effect .
this spin-1/2 model was introduced in our recent paper,@xcite where we already studied the combined effects of some of these terms . here , instead
, we first investigate the effect of each of these terms individually on the qh phases generated by the magnetic field , in order to distill the problem .
in particular , we demonstrate the equivalence of the iso and zeeman terms in generating the weak qsh phase at zero energy in the presence of a magnetic field .
then , we show that the iso coupling can drive topological phase transitions between different topological insulating phases , an effect which totally relies on the combination of the iso coupling and the magnetic field .
secondly , we study the interplay between the terms incorporated in our model .
for example , we discuss the competition between the iso coupling and zeeman effect , motivated by the similarity of these terms for generating the weak qsh effect .
we show that the combination of these terms does not necessarily lead to a larger regime of parameters where the weak qsh effect is observed , which reveals the subtle competition between each of these terms and the magnetic field .
we also study the exquisite effects produced by the rashba coupling in the presence of a large exchange field , which is shown to generate several qh gaps , even in the absence of a magnetic field . from another perspective
, we also discuss the fate of the weak qsh phase , as the rso coupling is turned on , which introduces spin - flip terms into the hamiltonian , so that the spins are no longer aligned perpendicularly to the plane , and the spin hall conductivity is no longer protected .
an important motivation for the investigation of these topological phases and phase transitions is the rich variety of systems where they may be observed . despite its weak iso coupling
, graphene remains an appealing candidate for the observation of these phases , e.g. , by increasing the iso coupling artificially by adatoms.@xcite secondly , several experimental groups have reported on synthetic honeycomb lattices in a condensed matter system , for instance arrays of quantum dots on a gaas substrate,@xcite and `` molecular graphene '' , where the lattice is created by using the repulsive interactions of co molecules deposited in a triangular array on a cu(111 ) surface.@xcite these systems have larger lattice constants , which lead to a higher flux per plaquette at realistic magnetic field values , and allow for a more flexible control over the parameters than in real graphene .
a third type of experiments that motivates our studies is the recent realization of a honeycomb optical lattice for ultracold atoms.@xcite in cold - atom systems , large magnetic fields@xcite and spin - orbit couplings@xcite are produced synthetically by adjustable lasers ( see ref . for a review ) .
in particular , such artificial gauge potentials could be produced in optical lattices,@xcite thus leading to the possibility to probe the high - flux regime .
furthermore , we note that optical _
lattices subjected to well - designed gauge potentials can reproduce the properties of honeycomb lattice systems , exhibiting dirac - type physics.@xcite cold - atoms systems carry the advantages of high tunability , necessary for studying phase transitions , and allow for the study of topological phases in absence of interparticle interactions @xcite and disorder.@xcite these possible experimental realizations are explored in detail in the final part of this work .
this article is structured as follows .
we introduce the lattice model in sec .
[ sect_model ] and present the four terms characterizing our tight - binding hamiltonian : the usual hopping term , the two spin - orbit terms as well as the zeeman splitting .
then , in sec . [ sect_tools ] , we review the techniques used for our investigation and characterization of topological phases .
section [ sect_magnetic_field ] discusses the combined effect of the magnetic field and spin - orbit coupling .
then , the combined effects of spin - orbit couplings , zeeman splitting and uniform magnetic flux are explored in sec .
[ sect_competition_and_phasetrans ] .
the possible experimental realizations of our model , as well as the possibility to detect the effects presented in secs .
[ sect_magnetic_field ] and [ sect_competition_and_phasetrans ] , are reported in sec .
[ sect_experiment ] . finally , the conclusions are drawn in sec .
[ sect_conclusion ] .
the appendix provides further mathematical details of the calculations .
we consider a tight - binding model of spinful electrons in a two - dimensional honeycomb lattice , subjected to a uniform perpendicular magnetic field @xmath0 .
this model was introduced in ref . to investigate the effects of an external magnetic field on the kane - mele model ( cf .
refs . ) .
the magnetic field induces a zeeman effect , and causes the hopping terms to acquire a phase .
this phase is encoded by the so - called peierls substitution , where one replaces the momentum @xmath1 by @xmath2 , where @xmath3 is the gauge potential associated with the magnetic field .
thus , any hopping term from site @xmath4 to site @xmath5 picks up the phase factor @xmath6 , where @xmath7 . in the remainder of this text , we will express the magnetic field strength ( flux density ) in terms of the dimensionless quantity @xmath8 , defined as the flux per unit cell of the lattice , expressed in units of the elementary flux quantum @xmath9 . in our model
, we include four effects into the hamiltonian , written as @xmath10 , with @xmath11 here , @xmath12 is the annihilation operator for electrons at site @xmath4 .
the first term @xmath13 describes hopping between two nearest - neighbor ( nn ) sites @xmath5 and @xmath4 , where @xmath14 denotes the hopping amplitude .
the second term @xmath15 describes the zeeman effect , which is an on - site term that assigns different potentials to the two spin components via the pauli matrix @xmath16 .
its amplitude is proportional to the magnetic flux @xmath8 and the coefficient @xmath17 , which is related to the material s land @xmath18 factor by @xmath19 , where @xmath20 is the bohr magneton .
note that in a cold - atom emulation of this model , the parameters @xmath8 and @xmath17 are tuned individually ( cf . sec .
[ sect_experiment_optical_lattices ] ) .
the zeeman term lifts the degeneracy by shifting the two spin - degenerate copies of the spectrum up and down by an equal amount of energy .
the third term @xmath21 describes the iso coupling , which corresponds to a next - nearest - neighbor ( nnn ) hopping with amplitude @xmath22.@xcite here , the factor @xmath23 , where the sign depends on the value of @xmath24 , i.e. , the outer product of the two bond vectors connecting site @xmath4 to site @xmath5 via their unique common neighbor at site @xmath25 .
this hopping also involves the matrix @xmath16 , so that the sign of the hopping amplitude is opposite for the two spin components @xmath26 and @xmath27 .
thus , the two different spin species are effectively subjected to opposite local ( haldane - type @xcite ) magnetic flux . in absence of a magnetic field , this term opens a topologically nontrivial gap , and causes the qsh effect , characterized by so - called helical edge states , which are protected by trs.@xcite the final term @xmath28 is the contribution to the nn hopping due to the rso coupling .
this hopping has an amplitude @xmath29 , and involves the spin matrix @xmath30 , where @xmath31 , and @xmath32 is the vector connecting sites @xmath4 and @xmath5 .
the resulting spin matrix has only off - diagonal elements , so that this hopping involves a spin flip .
this term couples the two spin components , and as a consequence , @xmath16 is no longer a good quantum number for @xmath33 . in practice
, this means that spin states are no longer exclusively up or down , but may also point in different directions .
finally , we remark that the model considered here preserves inversion symmetry .
in particular , we disregard the effect of an additional staggered potential , which was shown to induce topological phase transitions between trivial and non - trivial phases in refs . .
the possible physical realizations of our model are discussed later in sec .
[ sect_experiment ] .
in the following sections , we will perform an analysis of the various topological phases produced by the terms in the model hamiltonian . before going into details , describing each term individually and their combined effects ,
we first present the framework and tools which we will use for this analysis .
readers who are already familiar with the notions of bulk and edge states , chern numbers , and ( spin ) hall conductivity may choose to skip this section . we will illustrate our framework by studying the spin - degenerate model , i.e. , we set @xmath34 in hamiltonian .
we assume @xmath35 throughout the whole text .
the bulk band structure can be computed numerically by applying periodic boundary conditions , namely by considering a toroidal geometry .
this requires to set @xmath36 , where @xmath37 and @xmath38 are integers , in which case the system is periodic in both spatial directions . under these conditions , and by choosing a proper gauge for the peierls phases @xmath6 , the system reduces to a @xmath39 magnetic unit cell .
( see the appendix for a description of the gauge structure and its spatial periodicity . ) by applying a fourier transform and invoking bloch s theorem , the schrdinger equation reduces to a @xmath40 eigenvalue problem , known as the discrete harper equation , @xmath41 where @xmath42 denotes the single - particle wave function at site @xmath43 , and where the @xmath44 matrices @xmath45 and @xmath46 are given in terms of the parameters defined in hamiltonian ( see the appendix for details ) .
the four components of the wave function @xmath47 are due to the two sublattice ( @xmath48,@xmath49 ) and the two spin ( @xmath50,@xmath51 ) degrees of freedom .
the @xmath40 harper problem in eq .
is analogous to the original ( spinless , square - lattice ) hofstadter problem,@xcite where the harper equation involves a @xmath52 matrix .
( color online ) hofstadter butterfly spectrum for the spin - degenerate model , hamiltonian with @xmath34 .
the bulk bands , i.e. , the regions where the density of bulk states is positive , are shown in black . in the bulk gaps ,
the colors and numbers indicate the quantized hall conductivity in units of @xmath53 . ]
solving the harper equation provides the bulk energy bands @xmath54 .
it is instructive to evaluate the sizes of the bulk gaps as a function of the magnetic flux @xmath8 , which results in a diagram known as the rammal - hofstadter butterfly,@xcite shown in fig .
[ fig_spectrum_spinless ] .
this figure may be obtained in the limit of large @xmath38 , where the bands become dispersionless .
the fractal structure , which is periodic in @xmath8 with a period of @xmath55 , is a result of the competition between two length scales : the magnetic length @xmath56 and the lattice spacing . here , @xmath8 parametrizes this competition , as the ratio between the area of the unit cell and that of the cyclotron orbits ( @xmath57 ) .
( color online ) ( a ) the cylinder geometry representing a ribbon with periodic boundary conditions in the @xmath58 direction .
( b ) spectrum at @xmath59 for the spin degenerate model ( @xmath34 ) .
the colors represent the expectation value @xmath60 : red and blue indicate edge states at the bottom and top edge , respectively , and gray represents bulk states .
the spin - degenerate edge states shown in ( a ) correspond to those in the qh gap at the fermi energy indicated by the dashed line in ( b ) . ] each bulk gap in the spectrum shown in fig .
[ fig_spectrum_spinless ] may be characterized by a topological invariant , which encodes the hall conductivity of this system when the fermi energy lies inside this gap .
the integer values of the hall conductivity are robust against external perturbations : they remain constant as long as the bulk gaps remain open . in the following , we recall the relation between this transport coefficient and the concept of edge states and topological invariants .
the edge - state analysis can be performed for a system featuring a boundary,@xcite such as a cylinder , as shown in fig .
[ fig_cylinder_edge_states](a ) .
when solving the harper equation on a cylinder , namely by applying periodic boundary conditions along one spatial direction only , the spectrum consists of _ bulk bands _ and _ topological edge states_.@xcite indeed , we typically find a few _ edge states _ within the bulk gaps , some of which _ cross _ the gap from one bulk band to the other ; see fig .
[ fig_cylinder_edge_states](b ) .
importantly , each edge state contributes @xmath53 ( one quantum of conductivity ) to the hall conductivity of the system .
the total hall conductivity is therefore equal to an integer number ( i.e. , the number of edge - state branches ) times @xmath53 .
the topological invariance of the hall conductivity is due to the fact that the number of edge states inside the bulk gap can not be altered unless the gap is closed .
this result can be used to derive the topological invariant associated to any of the bulk gaps , leading to the colors in the spectrum in fig .
[ fig_spectrum_spinless ] . for this spin - degenerate model ,
all topological invariants are multiples of @xmath61 because the system is twofold spin degenerate .
more precisely , in order to evaluate the hall conductivity of a bulk gap , we count the edge states whose dispersions intersect the fermi energy , taking into account their location and direction of propagation .
the location of each state is derived by computing its expectation value @xmath60 from the eigenstate , which provides the coloring of fig .
[ fig_cylinder_edge_states](b ) .
secondly , the direction of propagation can be obtained from the sign of its momentum derivative @xmath62 , where @xmath4 is the momentum parallel to the edge .
edge states with opposite directions contribute with opposite signs to the hall current and therefore to the hall conductivity , @xmath63 where @xmath64 and @xmath65 denote the number of left- and right - moving states , respectively .
the hall conductivity does not depend on the magnitude of the velocity ( @xmath66 ) , but instead is given by the number of edge - state channels.@xcite in this spin - degenerate model , the states are all either left- or right - moving , i.e. , they are _ chiral _ states .
the chirality of the edge modes provides protection of the hall conductivity against backscattering due to disorder in the presence of broken trs . remarkably , the number of edge states is directly related to topological invariants , the _ chern numbers _ , a quantity associated with each of the bulk bands .
the chern number is an integer topological index defined in the toroidal geometry,@xcite @xmath67 where @xmath68 is the berry connection associated to the eigenstate @xmath69 , and where bz denotes the ( toroidal ) brillouin zone .
the number of edge states @xmath70 inside the @xmath71th bulk gap , or equivalently the hall conductivity assigned to this gap ( cf .
discussion above ) , is then equal to the sum of all the chern numbers associated to the occupied bands,@xcite @xmath72 this identity is known as the _ bulk - boundary correspondence_,@xcite as it relates the topological indices associated to the bulk bands to the number of edge excitations .
this correspondence provides us with an important observation : although the dispersions of the edge states depend on the shape of the edges ( e.g. , zigzag , bearded , or armchair ) and on the system size , the number of edge - state branches does not : this number is a bulk property and therefore does not depend on the specific form of the edge . in the remainder of this paper , we will study systems with zigzag and bearded edges , without any loss of generality concerning the topological properties . in the next sections , we will characterize the topological phases by their associated hall conductivity and edge - state structures .
as described in the following , the edge states produced by the magnetic field are affected in the presence of strong perturbing terms , such as the spin orbit couplings and zeeman splitting . for the sake of simplicity
, we will show these effects using the edge - state analysis obtained from a cylindrical geometry , rather than focusing on the analysis of the bulk topological invariants .
an efficient method to compute the hall conductivity can be formulated in terms of the integrated density of states . inside a bulk gap , the integrated density of states @xmath73
is defined as the fraction of all states that lie below the gap ( which is an integer times @xmath74 ) .
the hall conductivity is then proportional to the @xmath8-derivative of this quantity , @xmath75 a result known as the steda formula.@xcite we have used this powerful formula to compute the hall conductivity , and generate the colorful butterfly spectra presented in the remainder of this paper . in the presence of terms involving the matrix @xmath16 in the hamiltonian , the spin @xmath76 symmetry is broken .
if spin - flip terms ( i.e. , linear combinations of @xmath77 and @xmath78 ) are absent , the spin in the @xmath79 direction is conserved , i.e. , it is a good quantum number . in this case , which happens
if the iso coupling or the zeeman splitting is present while the rso is not , the hamiltonian may be decomposed into the two spin components @xmath80 .
thus , one is able to define the component - wise chern numbers @xmath81 and @xmath82 and hall conductivities @xmath83 and @xmath84 .
the charge hall conductivity of the whole system is the sum of the two component - wise conductivities , @xmath85 , since both of them contribute equally to the hall current .
considering their contributions to the spin current , spin - up and spin - down edge modes have opposite `` spin charges '' @xmath86 .
thus , the spin hall conductivity is equal to the difference of the two component - wise conductivities , @xmath87 , where @xmath88 is the elementary quantum of spin hall conductivity . in the spin - degenerate model discussed earlier ,
the two component - wise hall conductivities are always equal , so that all spin hall conductivities vanish .
this is certainly different if the spin degeneracy is broken .
for instance , in the presence of iso , the qsh gap at zero magnetic field and at zero energy @xcite is characterized by @xmath89 . here ,
the contributions to the charge hall conductivity cancel , but the contributions to the spin hall conductivity add up to @xmath90 .
if the hamiltonian contains spin - flip terms , e.g. , the rashba coupling , then the spin is not conserved , and the aforementioned definition of the spin hall conductivity is no longer valid .
however , the related spin chern numbers , which equal @xmath91 if spin is conserved , remain well - defined topological invariants , even in absence of spin conservation.@xcite
zooming in on the butterfly of fig .
[ fig_spectrum_spinless ] near @xmath92 , we observe that the thickness of the bands decreases in the low - flux regime ( not shown ) .
thus , in this limit we can treat the bands as being infinitely thin , i.e. , as landau levels . in the low - energy regime , between the van hove singularities at @xmath93 , the band structure at low flux
coincides with the landau - level spectrum known from graphene,@xcite with energies @xmath94 ( where @xmath95 is the landau - level index ) .
moreover , computing the hall conductivity in this region , we find the `` anomalous '' hall - conductivity sequence @xmath96 ( @xmath97 ) , as observed in graphene.@xcite the breakdown of this structure around @xmath93 can be understood from the large density of states near the van hove singularities and from the chern numbers associated to these many bands .
@xcite the edge - state structures , hall conductivity plateaus and topological aspects of spinless electrons in a honeycomb lattice subjected to a magnetic field have been thoroughly described in ref . .
( color online ) hofstadter butterfly spectrum for the system with iso coupling ( @xmath98 ) . although the periodicity in @xmath8 is @xmath99 rather than @xmath55 ( see the appendix ) , we present only the range @xmath100 for the sake of comparison with the other terms .
the colors and numbers indicate the quantized hall conductivity in units of @xmath53 . ]
since the seminal work of kane and mele,@xcite it is known that the iso coupling opens a topologically nontrivial gap of size @xmath101 at zero magnetic field .
this bulk gap hosts two counterpropagating edge modes per edge , with opposite spins ( one kramers pair ) : these _ helical _ edge states are related by trs .
this topological phase is known as the qsh state , and may be regarded as two opposite qh phases ( i.e. , each spin performs the qh effect , with opposite chirality ) .
the system preserves trs ( due to absence of the magnetic field ) , which protects the qsh state against scattering processes caused by disorder .
although for any edge state a reversely propagating mode would be available to scatter to , the spin has to be flipped , and the two different ways to flip the spin interfere destructively with each other , due to a mutual phase factor of @xmath102.@xcite the combination of the iso coupling and a magnetic field leads to the breaking of the trs satisfied by the qsh phase .
however , due to the absence of spin flip terms in the hamiltonian , the helical edge - state structure persists . in terms of charge and spin hall conductivity , the resulting state is equivalent to the qsh state .
as announced in the introduction , this state is called the _ weak qsh _ phase@xcite or _ trs broken qsh _ phase.@xcite from the hofstadter butterfly in fig .
[ fig_spectrum_iso ] , we observe that the size of the weak qsh gap at zero energy tends to decrease if the flux is increased ( but not monotonically ) , and closes eventually . in fig .
[ fig_spectrum_iso ] , in which @xmath103 , the qsh gap at @xmath104 remains open in the range @xmath105 .
here we observe the competition between the iso coupling , which opens the ( weak ) qsh gap , and the magnetic field , which `` tries '' to destroy the weak qsh state by its trs breaking property . throughout the weak qsh gap ,
the edge states remain crossing at zero energy , which indicates that this phase is robust at least in the absence of magnetic disorder .
we finally note that the spectrum remains particle - hole symmetric , as in the spinless case .
( color online ) hofstadter butterfly spectrum for the model described by hamiltonian with zeeman effect , @xmath106 and @xmath107 .
the bulk bands are in black .
the colors and numbers indicate the quantized hall conductivity in units of @xmath53 .
the diagonal dashed lines indicate the outline of the butterfly ( i.e. , @xmath108 ) for the two spin components . ]
the zeeman term models an on - site spin - splitting effect that is always present in real materials subjected to magnetic fields . studying hamiltonian with @xmath109 and @xmath110
, we observe that the spin degeneracy is lifted by merely shifting the two spin states up and down in energy by @xmath111 .
the spectrum is otherwise left invariant . with respect to
the hofstadter butterfly represented in fig .
[ fig_spectrum_spinless ] , this means that the spin - up and spin - down copies are `` skewed '' in opposite directions , as shown in fig .
[ fig_spectrum_zeeman ] . because the lowest and highest bands of the spin - degenerate model are situated at approximately @xmath112 , the two spin components are completely separated above a certain flux value , approximately equal to @xmath113 .
furthermore , the zeeman shift breaks the periodicity of the butterfly spectrum along the @xmath8 axis .
as we already mentioned , the spectrum of the spin - degenerate model shows landau levels for low values of the flux ( cf .
[ sect_spindeg ] ) .
the hall conductivities in the gaps are @xmath114 .
if we concentrate on the bulk gaps around zero energy , we observe one and then three doubly degenerate edge states ( see fig .
[ fig_dispersions_zeeman_or_iso](a ) ) . when the zeeman splitting is nonzero , the two copies of these states shift up and down , as shown in fig .
[ fig_dispersions_zeeman_or_iso](b ) for @xmath115 . assuming that the splitting is sufficiently large
, a new gap opens at zero energy , where the original band was situated .
the spin - up and spin - down edge states now connect to bulk bands that have been shifted up and down , respectively . as a consequence , in the newly opened gap at zero energy
, we have @xmath116 and @xmath117 .
thus , the charge hall conductivity vanishes , and the spin hall conductivity is equal to @xmath90 . based on these values , we find that this state is a weak qsh phase . comparing this state to the one generated by iso ( see fig . [ fig_dispersions_zeeman_or_iso](c ) ) , we observe the similarity between them . in both cases ,
the landau levels for the two spin components shift in opposite directions .
the difference between the two terms lies in the different values of the shift : the zeeman shift @xmath118 is linear in the flux , so that it vanishes for @xmath119 , while the landau - level shift induced by the iso coupling equals @xmath120.@xcite a recent study shows that a flux - independent zeeman term ( called an exchange term ) may also generate the qsh state at zero magnetic field.@xcite the similarity between the zeeman and iso terms may be understood at a formal level , in terms of the respective linearized ( low - energy ) hamiltonians : the linearized zeeman term acts as @xmath121 , where the factors indicate the proper spin , sublattice pseudospin , and valley pseudospin , respectively . on the other hand , the iso coupling acts as @xmath122 .
_ a priori _ , these hamiltonians act differently on the eight - component field vector @xmath123 . in the lowest landau level ,
the four components @xmath124 vanish.@xcite substituting the remaining field components into the linearized hamiltonians yields equality up to a sign , @xmath125 which shows that the zeeman effect and the iso coupling act equivalently on the lowest landau level .
we remark that this reasoning is only valid at this specific landau level : for higher landau levels , all the eight fields are present , and the two terms become inequivalent . so far , most of the studies on the qsh effect have generally concentrated on the behavior of the system at zero energy . however , very interesting features also emerge at nonzero energies in the presence of external fields . for instance , in the low - flux regime , a zeeman gap is formed at the @xmath126 landau level ( at @xmath127 ) , as shown in fig .
[ fig_dispersions_zeeman_or_iso](b ) . similarly to the zero - energy gap , this gap also shows a difference between the number of edge states with spin up and spin down components , @xmath128 and @xmath129 , which may be understood from the values above and below the corresponding landau level of the spinless model .
again , the spin hall conductivity is nonzero , @xmath130 , but the difference with the weak quantum spin hall gap is that the charge hall conductivity is also nonzero , @xmath131 , as can be deduced from fig . [ fig_dispersions_zeeman_or_iso](b ) .
furthermore , the edge states in this gap all propagate in the same direction ( i.e. , they are chiral ) , thus providing robustness against disorder . here
, we refer to this phase as the _ spin - imbalanced quantum hall phase_.@xcite ( color online ) high - energy spectra for the model without spin - orbit couplings ( @xmath110 ) at @xmath115 .
( a ) spectrum for the spin - degenerate model ( @xmath132 ) , with quantum hall gaps corresponding to @xmath133 ( in units of the conductivity quantum ) .
( b ) spectrum with zeeman effect ( @xmath134 ) , showing the spin - filtered quantum hall state with @xmath135 and @xmath136 .
the other gaps show ( alternatingly ) quantum hall phases ( @xmath137 ) and spin - imbalanced quantum hall phases ( @xmath138 ) . ]
a special instance of the spin - imbalanced quantum hall phase may occur if edge states of one of the spin - components are absent , i.e. , either @xmath139 or @xmath140 , in which case one speaks about the _ spin - filtered quantum hall phase_. in this state , the magnitudes of the charge hall and spin hall currents are therefore equal when expressed in units of their respective conductance quanta .
the spin - filtered quantum hall phase appears at high energies and low flux in the presence of zeeman coupling , as we now explain . in the absence of zeeman coupling
, we observe equally - spaced landau levels , each of which has chern number @xmath141 , see fig .
[ fig_zeeman_high_e](a ) .
if the zeeman term is present , the gap that forms between the two copies of the original highest - energy landau level exhibits a spin - filtered quantum hall phase , characterized by the presence of edge states of only one spin component ; in this case @xmath128 and @xmath140 , see fig .
[ fig_zeeman_high_e](b ) .
the other gaps visible in fig .
[ fig_zeeman_high_e](b ) are spin - imbalanced and ordinary spin - degenerate quantum hall gaps , alternatingly .
let us mention that the spin - filtered and spin - imbalanced quantum hall phases are ubiquitous in systems with zeeman - split landau levels , e.g. , the quantum hall plateaus corresponding to odd - integer filling factors in gaas / algaas heterostructures,@xcite or in hgte quantum wells , which have strong zeeman effect and are thus ideal candidates for observation of these phases.@xcite ( color online ) hofstadter butterfly spectrum ( with rso coupling , @xmath142 ) .
the inset shows a magnification of the region surrounded by the dashed lines . in the inset
, the dashed line indicates the flux value @xmath143 discussed in the text .
the colors and numbers indicate the quantized hall conductivity in units of @xmath53 . ]
the rso coupling differs in an essential way from the iso coupling and the zeeman effect , in the sense that the hopping involves a spin flip . in other words ,
the spin matrices involved in the rashba hopping are off - diagonal ( i.e. , @xmath77 and @xmath78 ) , whereas the other terms in the hamiltonian involve diagonal spin matrices .
this has a profound effect on the spin structure of the system : if the rso interaction is nonzero , mixing occurs between the spin components and we find that the spin orientation of the eigenstates is generally site - dependent .
the spin direction is confined to the two - dimensional plane and is perpendicular to the momentum : aside from the out - of - plane component ( @xmath79-direction ) , the component perpendicular to the cylinder edge may also be nonzero , while the component parallel to the edge always vanishes . in other words ,
the helicity @xmath144 of the ( edge ) states remains zero .
this effect is caused by the nature of the rashba coupling in the hamiltonian , which involves the outer product of the spin vector @xmath145 with the bond vectors @xmath32 , so that the spin is always perpendicular to the hopping direction .
because the edge - state spins are no longer in the up or down state , the chern numbers for the respective spin components are no longer well - defined .
thus , the spin conductivity no longer takes quantized values .
nevertheless , the spin chern number remains a well - defined integer - valued topological invariant,@xcite which can be used as a tool to distinguish between trivial and nontrivial gaps .
the second difference with the iso coupling is that the zero - flux spectrum remains gapless and the spin degeneracy is lifted by the rso coupling ( considering @xmath146 and @xmath147 ) .
indeed , upon setting @xmath146 , the two dirac cones from the spinless model ( at the special points @xmath148 and @xmath149 ) are broken into four cones each : a central isotropic cone at @xmath148 or @xmath149 , and three anisotropic `` satellite '' cones around it .
this effect is known as trigonal warping.@xcite between the main cones and the satellite ones , there is a van hove singularity at low energy , which scales as @xmath150 for small @xmath151 .
the berry phases associated to the main and satellite cones are @xmath152 and @xmath153 , respectively , which add up to the berry phase @xmath154 of the dirac cones in the spin - degenerate model ( @xmath153 for each spin component ) . for energies
@xmath155 , the low - flux landau level spectrum is characterised by two different sets of landau levels with chern numbers @xmath61 and @xmath99 . outside this regime , there are only twofold - degenerate landau levels , because the satellite cones can no longer be resolved . around @xmath156 there is a crossover regime where the sixfold - degenerate bands split into three , each one being twofold degenerate .
adding a nonzero rashba term to the spin - degenerate model discussed in sec .
[ sect_harper_hofstadter ] will cause a spin - splitting of some bulk bands in the hofstadter butterfly , see fig .
[ fig_spectrum_rashba ] .
in particular , at low energies , the rashba term splits the fourfold - degenerate landau levels into twofold - degenerate landau levels . in the limit @xmath157
, we also note that the spectrum indeed remains gapless and particle - hole symmetric . however , this is not true for other dirac regimes in the spectrum .
for instance , at @xmath143 and @xmath158 ( see the inset of fig .
[ fig_spectrum_rashba ] ) , the rso coupling splits the central landau level , and the resulting gap exhibits two counterpropagating edge modes on each edge .
however , the spin hall conductivity is undefined because the spin direction varies as a function of the fermi energy
. nevertheless , at each energy inside the gap the spins of the two edge states are approximately equal , so that the spin current nearly vanishes . besides , the charge hall conductivity is exactly zero when the fermi energy is located in this gap .
the rso coupling has drastic effects on the edge states , as their spins no longer align perpendicularly to the sample plane , but get an additional in - plane component perpendicular to the edge . besides , the direction of the edge - state spins depends on the fermi energy .
this observation distinguishes the rashba effect from the zeeman effect in a tilted field , which puts all edge - state spins in the same direction .
in addition , the in - plane component is opposite for edge states at the opposite edges , whereas the perpendicular component is the same .
importantly , due to the dependence of the spin direction on the fermi energy , the latter can be conveniently used for spin manipulations ( in addition to the amplitude of the rashba coupling itself ) .
we illustrate this phenomenon around the rashba - split bulk band at @xmath143 and @xmath158 . in fig .
[ fig_dispersions_rashba](a ) , we show that in the limit of @xmath159 , the eigenstates are in - plane ( indicated as `` left '' and `` right '' in the figure , meaning that the spin direction is @xmath160 ) .
this result shows that for infinitesimal rso coupling , it is more natural to decompose the states in terms of the eigenstates of the pauli matrix @xmath78 than those of @xmath16 .
if the rashba coupling is then increased , the gap opens , and in the vicinity of the bulk bands , the edge states tend to rotate to the vertical direction , see fig . [ fig_dispersions_rashba](b ) .
in addition , we observe that not all spins have the same length .
this phenomenon occurs because the spins displayed in this plot represent the expectation values of the spin components ( see the appendix ) .
since the spin direction depends on the lattice position ( @xmath161 coordinate ) , the length of this expectation value may be less than unity .
we expect that the canting of the spins is most easily observed in gaps with a single edge state on each edge .
for instance , in ref . , it has been demonstrated that the rso coupling cants the spin in a spin - filtered gap generated by the zeeman effect . in this
setting , the spin textures may be controlled at will by tuning the coupling parameters and the fermi energy .
( color online ) spin splitting in an example at flux @xmath143 and near @xmath162 .
zeeman effect : ( a ) @xmath163 and ( b ) @xmath164 ; iso coupling : ( c ) @xmath165 and ( d ) @xmath166 ; rso coupling : ( e ) @xmath167 and ( f ) @xmath168 . ]
as shown before in the low - flux limit @xmath171 , the iso coupling and the zeeman effect affect the zero - energy modes equivalently in the sense that they both open a weak qsh gap .
indeed , it has been shown that for this mode , the iso coupling and zeeman effect are formally equivalent.@xcite on the other hand , the rso coupling does not open a gap at zero energy .
this fact hampers the comparison of the rso coupling to the aforementioned two terms in this energy regime . from the hofstadter butterfly shown in fig .
[ fig_spectrum_spinless ] , one observes that the landau - level structure in the low - flux , low - energy limit is not unique .
in fact , similar structures appear wherever the shape of the dispersion is characterized by dirac cones .
for example , at @xmath172 and for energies close to @xmath173 , the dispersion can be approximated by a dirac dispersion of the form @xmath174 , where @xmath175 is the momentum relative to the position of one of the two dirac cones .
the linear term is responsible for the square - root behavior of the ( fourfold - degenerate ) landau levels for flux values close to @xmath172 .
the quadratic term causes the asymmetry of the landau level spectrum with respect to the energy @xmath176 .
furthermore , we observe that the energy @xmath177 of the central landau level is linear in the flux rather than constant .
the approximate linear dependence of the spectrum @xmath178 at @xmath172 and @xmath179 suggests that the zeeman and iso terms should behave similarly in this regime , in direct analogy to the case @xmath180 discussed in sec .
[ sect_zeeman_flux_weakqsh ] .
although the energy of the lowest landau level is no longer constant in @xmath8 , the approximation holds , thus leading to a similar behavior with regard to the opening of the gaps , as shown in figs .
[ fig_splitting_examples](a d ) for the flux value @xmath143 which lies close to @xmath172 .
the zeeman effect and iso coupling split the central landau level at @xmath158 and open up a weak qsh gap , similar to the zero mode in the low - flux regime .
as discussed in sec .
[ sect_rashba ] , the rso coupling splits the central landau level , unlike the case for @xmath119 , see figs .
[ fig_splitting_examples](e , f ) .
although there are edge states inside the gap , it is a trivial state , because the spin hall conductivity approximately vanishes .
this comparison shows an important difference between zeeman coupling and iso coupling on one hand , and the rashba coupling on the other hand : the former two create the weak qsh phase , while the latter does not . in the following section ,
we show that in the presence of zeeman or iso coupling , the rashba coupling tends to destroy the weak qsh phases generated by the former effects .
after analyzing the effects of the three terms @xmath21 , @xmath28 , and @xmath15 in hamiltonian independently , a natural question arises : what effects emerge when these terms are combined ? here , we study the competition featuring two terms , by tuning the ratio between their corresponding coupling constants . tuning these coupling constants generally leads to opening and closing of gaps , so that the topological nature of the gaps may change . in that case , we deal with _ topological phase transitions_. we recall that kane and mele@xcite have studied the phase transition from a ( time - reversal symmetric ) qsh state to a trivial state , by increasing the ratio @xmath181 . here , we extend this study and we explore the more exotic phase transitions that are realized in the presence of a magnetic field and the zeeman effect .
let us analyze the interplay between the iso coupling and the magnetic field by investigating a gap at a fixed flux value , in absence of the zeeman effect and of the rashba coupling . by tuning the value of the iso coupling amplitude @xmath22
, we can close the gaps and obtain different topological phases at both sides of the transition . in fig .
[ fig_phase_transition_iso ] , we have shown the dispersions for @xmath59 , for values of @xmath22 below , at , and above the values where the gap around @xmath182 closes . around this energy ,
there are three bulk bands : two ( overlapping ) ones above and one below the gap . for values of @xmath183 , below the critical value
, we observe the weak qsh phase , i.e. , two counterpropagating edge modes with opposite spins on each edge , @xmath128 and @xmath184 .
the spin - down pair of edge modes connects the top and bottom bulk bands visible in this plot , and the spin - up pair connects the middle band to a lower band . for @xmath185 ,
the gap disappears : the bulk bands touch each other in three points at @xmath186 , and two of these points are connected by the spin - down edge states in the middle of the figure .
if the value of @xmath22 is increased further , a gap opens again , and a new pair of edge states ( with spin down ) appears .
moreover , the existing spin - down edge states have inverted their direction of propagation .
thus , for @xmath187 , we have a spin - imbalanced quantum hall state with @xmath128 and @xmath188 .
note that only the spin - down states are modified in this phase transition , while the spin - up states remain the same . in ref .
, this phase transition was investigated in the presence of a nonzero zeeman coupling in addition to the iso coupling . here , we emphasize that the zeeman effect is actually superfluous and that this transition is driven exclusively by the iso coupling .
variation of @xmath17 has no qualitative effect on the transition , but only shifts the energies of the bands and gaps . within a large range of values for @xmath17
, we obtain an identical phase transition as in fig .
[ fig_phase_transition_iso ] .
the critical value @xmath189 of the iso strength is independent of @xmath17 .
the phase transition presented here differs in a fundamental way from a crossing of landau levels . in the latter case ,
the chern numbers associated to the landau levels do not change at the transition , and as a consequence the number of edge states inside the gap between them remains unmodified as well . for the phase transition presented here , the number of edge states and consequently the chern numbers do change at the transition . focusing on the spin - down edge states , we observe that @xmath184 for @xmath190 and @xmath188 for @xmath187 , a difference of @xmath191 .
the chern numbers of the bands below and above the gap are @xmath55 and @xmath55 in the former case and @xmath141 and @xmath192 in the latter .
the sum of these chern numbers is unchanged , as required by the bulk - boundary correspondence , eq . .
the question arises as to whether the difference in the number of edge states at both sides of the transition may be predicted . _ a priori _ , this difference can not be predicted since it is the result of a complicated interplay of the iso coupling and the magnetic flux . however , the difference is always a multiple of the denominator @xmath38 of the flux @xmath36 .
this result may be understood from the fact that at a given flux value @xmath36 , the chern numbers of all the hofstadter bands obey @xmath193 . here
, @xmath194 is the _ modular multiplicative inverse _ of @xmath37 modulo @xmath38 , defined as the unique integer @xmath194 ( @xmath195 ) such that @xmath196 . for all
@xmath197 implies automatically that @xmath196 , assuming that the steda formula and the bulk - boundary correspondence hold .
these results are valid more generally for the model described by hamiltonian as long as the two spin components are decoupled . ]
consequently , the difference between two possible values of the chern numbers is always a multiple of @xmath38 , which proves the aforementioned claim .
this property may also be understood from the superlattice structure : the magnetic field produces @xmath38 copies of the original unit cell , and each of them reacts in the same way to the perturbation leading to the phase transition .
( color online ) phase diagram as a function of the iso coupling @xmath198 and the fermi energy @xmath199 for the fixed flux value @xmath59 . in the shaded regions ,
the system is metallic .
the colors red ( dark gray ) and blue ( light gray ) distinguish the spin - up and -down components , respectively .
in the white regions , the bulk is insulating , and the edges conduct .
there , the pairs of numbers indicate the number of edge states @xmath200 and @xmath201 for spin up and down , respectively . ] in fig .
[ fig_phase_diagram_vs_iso ] , we show the phase diagram as a function of the iso coupling @xmath198 and the fermi energy @xmath199 with the flux value @xmath59 held fixed . from this figure , we can easily read off the values of @xmath22 and the energies @xmath199 where the gaps close and the topological phase transitions occur .
a few points must be noticed : first of all , we clearly identify the weak - qsh phase at half - filling for @xmath202 .
the ( spin - degenerate ) qh phases are progressively destroyed for @xmath203 . for @xmath204 , we get spin - filtered qh [ e.g. , @xmath205 , spin - imbalanced qh [ e.g. , @xmath206 and the weak qsh phase . in addition , there are also more exotic phases where the edge states are neither chiral nor helical , e.g. , @xmath207 .
the spin - degenerate phases do not arise for @xmath204 .
secondly , the gaps are large ( of the order of @xmath14 ) in wide regions of the parameters , which is favorable for the detection of the transitions in cold - atoms experiments . in addition , the property that the number of edge states changes by a multiple of @xmath208 at each phase transition is clearly observed in fig .
[ fig_phase_diagram_vs_iso ] .
finally , the energies where the phase transitions occur may be shifted by tuning the strength of the zeeman effect .
since the spin - up and spin - down components are uncoupled , the zeeman effect will shift all red ( blue ) areas in fig .
[ fig_phase_diagram_vs_iso ] up ( down ) by a fixed amount of energy .
we note that the zeeman effect alone does not induce the same type of phase transitions as the iso coupling .
the zeeman effect can only close and open gaps between bands of different spin components , and does therefore not modify the chern numbers of the bands , whereas the iso coupling can also close gaps between two bands of the same spin component .
nevertheless , tuning the zeeman coupling strength @xmath17 allows one to modify the nature of the phase transitions driven by iso coupling .
( color online ) hofstadter butterfly spectrum with iso coupling ( @xmath98 ) and zeeman effect ( @xmath106 ) .
the colors and numbers inside the bulk gaps indicate the quantized hall conductivity in units of @xmath53 .
the vertical dashed line indicates the flux value where the two effects are comparable in strength . at this value , the crossover from the iso - dominated ( low @xmath8 ) to the zeeman - dominated ( high @xmath8 ) regime takes place . ]
if the zeeman effect and iso coupling are present simultaneously , the interplay between the two terms is governed by the value of the flux @xmath8 , because the amplitude of the former is linear in @xmath8 while that of the latter is constant .
thus , in the butterfly spectrum illustrated in fig .
[ fig_spectrum_zeeman_iso ] , we observe the iso - dominated regime at low flux ( cf . fig .
[ fig_spectrum_iso ] ) , and the zeeman - dominated regime at high flux ( cf .
[ fig_spectrum_zeeman ] ) .
the crossover between these two regimes takes place where the two terms are comparable in strength , approximately where @xmath210 , indicated by the vertical dashed line in fig .
[ fig_spectrum_zeeman_iso ] at @xmath211 .
we remark that this crossover effect is present more generally in the situation where the zeeman term competes with any other term with a constant amplitude : due to its linear @xmath8-dependence , the zeeman term will always dominate any other one at sufficiently high values of @xmath8 .
naturally , this phenomenon only occurs in systems where the zeeman splitting depends linearly on the magnetic flux @xmath8 . in the presence of magnetic impurities ,
the zeeman term may exhibit a nonlinear dependence on the applied magnetic flux , as discussed in ref . for mn - doped hgte quantum wells , but this subject is beyond the scope of this paper .
although the zeeman effect dominates over the iso coupling for @xmath212 , the influence of the iso coupling is still visible in the high - flux spectrum : at @xmath172 and @xmath213 , there are gaps at @xmath214 ( see fig .
[ fig_spectrum_zeeman_iso ] ) , which are absent in the butterfly without the iso coupling ( see fig . [ fig_spectrum_zeeman ] ) .
we note that the dispersions around these points are of dirac type for @xmath215 , consequently , the iso coupling opens these gaps in a way analogous to the gaps at zero flux and zero energy .
however , the gaps at @xmath172 and @xmath213 are spin - filtered qh phases and not qsh , due to the subtle competition between the iso coupling and the trs - breaking effects . in two recent works,@xcite
it has been shown that the combination of rso and an exchange field , similar in structure to the zeeman term , leads to chiral edge states at zero energy .
in addition , the problem has been investigated for a sodium / lithium iridate model which involves nn and nnn hopping terms.@xcite in the latter and more complex setup , nontrivial qh gaps appear also away from half filling . here ,
we consider the combined effects of the zeeman and rashba couplings , away from half - filling and in the absence of iso coupling .
we demonstrate that _ several _ qh gaps appear in the absence of nnn hopping terms , without including any gauge field ( i.e. , the peierls phases ) : namely , non - trivial qh phases are produced in our model by setting @xmath217 , @xmath33 and by applying a flux - independent zeeman , or exchange,@xcite term @xmath218 which differs from @xmath219 by the fact that the strength @xmath220 no longer depends on the flux @xmath8 ( which we now set to zero ) . we stress that it is the simple association of a constant exchange term with a nn rashba hopping term that leads to non - trivial topological phases , at half - filling but also at @xmath221 , as discussed in detail below .
( color online ) ( a - d ) energy spectra @xmath222 for @xmath217 .
the zeeman ( exchange ) and rashba coupling strength are indicated .
( e ) phase diagram in the parameter space defined by the coupling strengths @xmath223 and @xmath224 .
the number triplets indicate the charge hall conductivity in the lower , middle , and upper gaps , respectively .
the solid lines indicate the phase transitions for the gap at zero energy . at the dashed lines , the lower and upper gaps close and the trivial gap at zero energy remains open [ as in ( c ) ] , and at the dotted lines , all three gaps close [ as in ( a ) ] . at the dashed and dotted lines , the charge hall conductivities in the gaps do not change . ] when the exchange term is added to the tight - binding model with strength @xmath225 , the energy bands associated to the spin - up and spin - down components are shifted in opposite directions but overlap at half - filling .
in this situation , which is illustrated in fig . [
fig - rashba - exch](a ) , the system is a semi - metal for @xmath226 and a metal otherwise . when the rashba coupling is turned on , bulk gaps open not only at half - filling @xmath104 but also at @xmath227 ( cf .
[ fig - rashba - exch](b ) ) .
we have computed the edge - state structures as well as the topological indices associated to the gaps : we obtain that all these gaps are related to non - trivial qh phases .
more precisely , we find that the gaps at @xmath227 host a single edge state per edge , the spin orientation of which is given by @xmath228 : these gaps correspond to spin - filtered qh phases , with equal ( resp .
opposite ) charge ( resp .
spin ) hall conductivities .
we observe that the spin orientation of these isolated edge states is nearly vertical , which is due to the fact that the exchange term is important in this high - energy regime . at half - filling ,
the bulk gap hosts two edge states per edge , with same velocity and quasi - horizontal spin orientation .
this bulk gap corresponds to a topological qh phase with @xmath229 and a non - trivial spin orientation .
therefore the spin orientation of the edge states highly depends on the fermi energy .
we note that the gap openings occur for large exchange coupling , but for arbitrarily low rashba coupling .
when the exchange strength @xmath230 , the energy bands associated to the spin - up and spin - down components are well separated : a trivial gap opens at half - filling , while the dirac points are shifted to the energies @xmath231 , cf .
[ fig - rashba - exch](c ) .
in this situation , the rashba coupling still opens two non - trivial qh phases at @xmath232 , cf .
[ fig - rashba - exch](d ) .
again , these phases are characterized by spin - filtered edge states with nearly vertical spin orientation .
therefore , for sufficiently large exchange coupling , it is necessary to set the fermi energy away from half - filling in order to detect nontrivial topological phases .
this behavior is summarized in the phase diagram of fig .
[ fig - rashba - exch](e ) : for nonzero rashba coupling @xmath29 , the hall conductivity at zero energy is @xmath233 for @xmath234 and zero otherwise .
thus , topological phase transitions of the zero - energy gap take place at the gap closings at @xmath235 .
the lower and upper gaps ( at @xmath236 ) always exhibit a hall conductivity of @xmath237 , where the sign depends on that of @xmath238 but not on that of @xmath151 . inverting the sign of @xmath220 inverts the directions of the edge currents , as well as the @xmath79 component of the spins .
inverting the sign of @xmath29 inverts the in - plane component of the spins .
continuous variation of @xmath29 through zero changes the spin - direction continuously , where the spins tend to the @xmath79 direction in the limit @xmath239 .
thus , the gap closing at @xmath240 does not constitute a topological phase transition .
let us emphasize the fact that the chiral qh phases presented here are not the consequence of the external magnetic flux ( i.e. , peierls phases ) , since they are produced when @xmath216 , and solely in the presence of the exchange and rashba terms . when @xmath241 , new qh gaps open in the system , altering the exchange - rashba - induced phases presented in figs .
[ fig - rashba - exch](b ) and [ fig - rashba - exch](d ) .
this competition between qh gaps of different origins leads to complex and rich quantum phase transitions . in the low - energy regime , where the ( weak ) qsh phase is generated by either iso coupling or the zeeman effect
, the additional inclusion of the rashba coupling has several profound effects on the physics .
first of all , the rso coupling tends to decrease the size of the gap ( cf .
ref . ) , eventually completely closing it .
secondly , the combination of rso coupling with iso coupling and/or zeeman effect destroys the particle - hole symmetry of the spectrum .
finally , at finite magnetic field , the rso coupling opens up a gap ( i.e. , an avoided crossing ) between the edge states at their crossing around zero energy . around this crossing
, the spin direction rotates from down to up ( or vice versa ) , as illustrated in fig .
[ fig_dispersions_zeeman_rashba ] .
the size of this gap is approximately linear in @xmath8 and in @xmath29 ( in the limit @xmath171 and @xmath159 ) .
thus , the bulk gap is trivial in the sense that no edge states cross it from one bulk band to the other .
however , in the limit where the edge - state gap is small , one would still observe a weak qsh state . at finite temperatures ,
the thermal energy can `` bridge '' the edge - state gap , so that the charge carriers may tunnel from one edge state to another . as the edge - state gap size
is increased , the backscattering is enhanced , due to the decrease in the amount of tunneling across the edge - state gap . here , we expect that the value of the spin hall conductivity will deviate from its quantized value @xmath242 . the amount of deviation increases when the gap size increases .
( color online ) spectrum for the system with nonzero zeeman ( @xmath106 ) and rso coupling ( @xmath243 ) , for @xmath115 .
the inset shows a zoom of the region where a gap opens between the edge states at zero energy . for comparison with the spectrum in absence of rashba coupling , refer to fig .
[ fig_dispersions_zeeman_or_iso](b ) . ]
in this section , we will discuss the several experimental approaches that allow for the observation of the topological phases and phase transitions discussed in secs .
[ sect_magnetic_field ] and [ sect_competition_and_phasetrans ] . for each of them , we present the opportunities and challenges , and we discuss the range of parameters that can be probed .
condensed - matter systems typically have a lattice spacing in the order of @xmath244 , so that one flux quantum per unit cell ( @xmath213 ) corresponds to a magnetic field in the order of @xmath245@xmath246 .
thus , only the low - flux limit is relevant with respect to realistic condensed - matter realizations of our model .
the prime example of a two - dimensional system with a honeycomb lattice is graphene , which has received much attention both at the theoretical and experimental levels.@xcite in the presence of a magnetic field , graphene shows an anomalous quantum hall effect , where the landau - level energies@xcite depend on the magnetic field as @xmath247 , showing a radically different behavior than in semiconductor heterostructures ( e.g. , gaas / algaas ) . besides
, the effects of spin - orbit coupling have been extensively studied for graphene , leading to the concept of @xmath248 topological insulators.@xcite however , recent analysis has shown that graphene exhibits only a very weak iso coupling ( @xmath249@xmath250 , where @xmath251),@xcite which unfortunately prevents the observation of the qsh state in this material .
nevertheless , the iso interactions may be enhanced by putting heavy adatoms on top of the graphene surface,@xcite which would open up a route to build a topological heterojunction.@xcite the zeeman effect is of appreciable order of magnitude , @xmath252 , which corresponds to gaps of @xmath253 at a magnetic field of @xmath254.@xcite values of the rashba coupling as high as @xmath255 have been reported for graphene on a ni(111 ) substrate.@xcite adsorption of metal atoms on the graphene surface increases this coupling as well.@xcite alternatively , high values of the rashba coupling ( accompanied by the spin re - orientation at the edges ) may be induced by curving graphene nanoribbons.@xcite besides , it has been suggested that large rashba coupling leads to topological phases in bilayer graphene.@xcite in addition , large effective pseudomagnetic fields , greater than @xmath256 , have been recently realized in strain - induced graphene nanobubbles.@xcite in summary , the observation of the topological phases presented here remains elusive in graphene , but should not be definitively ruled out .
the qsh effect has nevertheless been observed in other two - dimensional compounds that have considerably larger values of @xmath22 , such as hg(cd)te quantum wells,@xcite where values of the order @xmath257@xmath258 have been reported.@xcite additionally , this material exhibits a strong zeeman effect as well ( @xmath259).@xcite the spin - filtered qh states have already been observed in these systems at low fields.@xcite hg(cd)te quantum wells doped with magnetic ions ( e.g. , mn ) exhibit an exchange interaction leading to the quantum anomalous hall effect at zero magnetic field,@xcite and to a nonlinear zeeman effect that causes reentrant behavior of the topological phases.@xcite the experimental limitation of the low flux values can be overcome by studying systems with a larger lattice spacing .
recently , several approaches have led to the engineering of artificial honeycomb lattices .
first of all , arrays of quantum dots on a gaas , arranged in a honeycomb pattern , have been engineered to simulate actual graphene.@xcite the lattice spacing in these systems is of the order of @xmath260 , so that high flux values ( @xmath261 ) are obtained already with small magnetic fields .
however , this large lattice constant leads to a small hopping amplitude , which has hampered the observation of the hofstadter butterfly until now .
patterned superlattices able to probe the hofstadter regime with usual magnetic field intensities have been realized in the past , but for square geometries.@xcite to probe the physics described here , one needs a superlattice with a smaller lattice constant , grown on a substrate that could lead to the generation of a strong rashba interaction , for example . in this case ,
chiral topological states could be probed also away from zero energy , as we discuss in sec .
[ sect_halfflux ] .
secondly , artificial graphene lattices have been created by manipulating co molecules with the tip of a scanning tunneling microscope and arranging them in a triangular lattice on a cu(111 ) substrate.@xcite these molecules repel the electrons on the cu surface , thus creating a honeycomb array of sites between which the electrons hop .
measurements of the density of states show that the electrons indeed exhibit a dirac dispersion .
the typical lattice constants of these systems are approximately @xmath262 times larger than in real graphene , thus allowing for flux values @xmath263 times larger than in graphene , with reasonable magnetic fields .
in addition , effective magnetic fields up to @xmath264 can be induced by triaxial strain of the artificial lattice .
it can be speculated that by using a heavier metal as substrate , it may be possible to engineer such a lattice with strong spin - orbit coupling , leading to the qsh phase .
although the parameters of the hamiltonian can be controlled to some extent in condensed - matter systems , the possibility to engineer the physics discussed here with ultracold atoms in optical lattices is very appealing .
the advantage of the ultracold - atom realization of this model is the flexibility to control the parameters separately across a large range , compared to condensed - matter systems where these parameters are generally fixed by the material properties and the sample geometry.@xcite a second major advantage of ultracold atoms over condensed - matter systems is that the magnetic fields are produced synthetically , i.e. , the peierls phases are not due to a real magnetic field , but induced by external fields acting on atomic internal states.@xcite in this way , large effective magnetic flux can be produced ( i.e. @xmath265 ) , which allows the system to be studied beyond the low - flux limit.@xcite the spin - orbit couplings can also be synthesized in optical lattices using similar methods.@xcite moreover , in these configurations , the zeeman coupling can be controlled independently , as it is not related to the synthetic magnetic fields@xcite ( see sec .
[ sect_rashba_exchange ] , where we study the extreme situation where the effect of the zeeman coupling is investigated in the absence of synthetic magnetic field , in which case this term is called an `` exchange term '' ) .
another important feature offered by optical - lattice setups is the possibility to control the interactions between the particles : by exploiting feshbach resonances,@xcite it is possible to reach the non - interacting regime and from there on , to include interactions in a controlled way to explore the robustness of topological order against interactions .
let us mention that these systems are usually free of disorder , although disorder can also be generated and tailored at will in optical lattices.@xcite these outstanding features allowcus to study the trs - broken qsh phase , namely the survival of the qsh phase in the presence of a trs - breaking perturbation.@xcite indeed , contrarily to condensed - matter experiments , the absence of ( magnetic ) disorder can stabilize the helical ( counterpropagating ) edge states even when trs is broken .
let us also comment on the fact that all the topological phases and edge - state excitations discussed here rely on the existence of robust bulk energy gaps .
therefore , when the temperature of the system approaches the size of the topological gaps of interest , one expects the effects to be strongly affected , and eventually to disappear . for the ultracold temperatures achieved in cold - atom laboratories , the gaps of interest should typically be of the order @xmath266 , where @xmath14 is the tunneling amplitude .
therefore , we anticipate that only the largest topological gaps presented in our figures should be considered from the cold - atom point of view . the quantum phase transitions presented in this work could be equally characterized either by a change in the topological invariants associated to the bulk gaps of interest , or by the modification of the edge - state structures .
although cold atoms do not offer the possibility to directly measure transport coefficients , several methods have been proposed to detect the topological invariants related to the quantized conductivities . @xcite these proposals are based on the fact that non - trivial band structures modify the atomic densities , leading to several signatures that can be directly seen in time - of - flight or density measurements . on the other hand ,
it is possible to directly probe edge state structures using bragg spectroscopy techniques.@xcite in a recent work,@xcite it was shown that a state - dependent light probe , focused on the edge of the cloud and transferring angular momentum to the atoms , clearly identifies the presence of chiral edge states in a qh optical lattice .
this method allows us to project the topological edge states on a dark background , making them visible in _ in situ _ density measurements . in principle
, this efficient scheme could be generalized to spin-1/2 systems in order to detect and distinguish between chiral and helical edge structures .
we have shown an in - depth analysis of the edge - state structures generated by the combination of the iso coupling , the rso coupling , and the zeeman effect , together with a uniform and perpendicular magnetic field . in sec .
[ sect_magnetic_field ] , we have studied each term individually in order to pinpoint their effect on the quantum hall states generated by the magnetic field . in the low - field limit ,
the iso coupling and zeeman effect are formally equivalent , and give rise to the weak qsh phase .
away from zero energy , we observe spin - filtered , spin - imbalanced , and ordinary quantum hall phases .
the rso coupling is different in the sense that it involves spin - flip terms , which causes the spin states to acquire an in - plane component .
in addition , the spin direction inside the gap depends on the fermi energy , and consequently the spin hall conductivity of the system is ill - defined . in the case of high magnetic fields ,
we observe dirac - type regimes away from zero energy . here , the zeeman and iso coupling open weak qsh gaps in a way analogous to the low - flux limit , whereas the rso coupling opens a trivial gap ( unlike the case at low flux , where the spectrum remains gapless ) .
in addition , the topological phase transitions reported in an earlier work@xcite have been analyzed in more detail here . a combination of the spin - orbit terms and the zeeman effect leads to subtle competitions , several examples of which have been discussed in sec .
[ sect_competition_and_phasetrans ] .
we have shown that , although the zeeman effect and iso coupling both give rise to the qsh phase , they do not always reinforce each other , which has been illustrated by the size of the parameter regime of the weak qsh phase in the low - flux limit . at high flux ,
the zeeman effect dominates over the iso coupling , but the effects of the latter are still well visible in the spectrum .
interestingly , a combination of the rashba term with an exchange coupling leads to nontrivial topological phases at zero magnetic field , away from half - filling .
finally , inclusion of the rashba coupling in the case where we have a weak qsh gap generated by the zeeman effect opens a gap at zero energy between the edge states , which provides an illustration of the breakdown of the qsh effect in the absence of trs and spin conservation .
this work emphasizes the various topological phase transitions and edge - state structures that could be observed in a wide range of materials and quantum emulators .
although we focused our analysis on the honeycomb lattice , where the spin - orbit couplings and the magnetic field configurations are expressed by specific tunneling terms and peierls phases , we note that our results do not rely on this specific geometry . the topological phases and
the phase transitions discussed here are universal , in the sense that they only rely on the inclusion of general ingredients , which can be found or engineered in many different contexts and geometries . indeed
, this universal property can be understood in the low - energy limit , in which various ( _ a priori _ , very different ) tight - binding models actually reduce to the same ( dirac - type ) equations . in general , we observe that the main ingredients needed to generate such physics are nontrivial gauge fields , which are coupled to the particles of interest ( e.g. , electrons in materials and atoms in optical lattices ) . in this work ,
we have aimed to emphasize the effects produced by each of these ingredients independently : we have investigated the effects of spin splitting , spin mixing , and gauge potentials leading to magnetic or haldane - type local flux .
we hope that our detailed analysis will deepen the understanding of topological insulating phases and that it will motivate further developments in this exciting and rapidly developing field .
we are grateful to m. o. goerbig and p. w. brouwer for useful discussions .
wb and cms acknowledge financial support by the netherlands organisation for scientific research ( nwo ) .
ng thanks the f.r.s - f.n.r.s for financial support .
geometrically , the honeycomb lattice is equivalent to a distorted square lattice where each unit cell has the shape of a rhombus .
since there are two inequivalent sets of sites , there are two sites per unit cell , and hence two sublattices , labeled @xmath267 and @xmath268 . the sites , labeled as @xmath5 and @xmath4 in the hamiltonian , may be uniquely identified by two integer labels @xmath197 and @xmath269 for the unit cell and @xmath267 or @xmath268 for the sublattice : in the following , we shall write @xmath270 ( where @xmath271 and @xmath272 ) to uniquely identify the sites .
( color online ) ( a ) the honeycomb lattice .
we indicated the nn vectors @xmath273 ( @xmath274 ) and the lattice vectors @xmath275 and @xmath276 .
the sites of the sublattices a and b are indicated with solid and open circles , respectively .
the dashed lines define boundaries of the unit cells , labeled as @xmath277 .
( b ) hopping phases @xmath278 for the unit cell @xmath277 .
the phases for the nn hopping are in green .
the phases for the nnn hopping are in red and blue for the sublattices @xmath267 and @xmath268 , respectively . ]
the coordinates of the sites are given as follows .
we choose the bond vectors @xmath279 , @xmath280 and @xmath281 as displayed in fig .
[ fig_lattice](a ) , where their length @xmath282 is chosen as the unit of length throughout this text .
subsequently , the lattice vectors are given by @xmath283 and @xmath284 , so that the coordinates of the sites are @xmath285 furthermore , the reciprocal space is spanned by the two vectors @xmath286 and @xmath287 , defined by @xmath288 .
we note that the results derived in this paper do not depend on the values of the vectors @xmath289 ; the only effect of a change of the @xmath289 is that the lattice vectors and the reciprocal lattice vectors will be different .
a convenient choice for the bond vectors is @xmath290 , @xmath291 and @xmath292 , effectively putting both sites of the unit cell @xmath277 on the coordinate @xmath277 .
the phases due to the magnetic field , associated with the hopping , are given in terms of the gauge potential @xmath3 as @xmath7 , where the integral is over the line between sites @xmath4 and @xmath5 .
we may employ the gauge freedom for the gauge potential to choose it in such a way that the phase @xmath293 for two sites @xmath270 and @xmath294 ( nns or nnns ) does not depend on the unit cell index @xmath269 .
the hopping phases resulting from such a choice are given in fig .
[ fig_lattice](b ) for hopping from the sites of the unit cell @xmath277 .
the reader may check that the phase picked up from hopping around a loop is proportional to the area of the loop .
in particular , when one considers one hexagon as the loop , the phase picked up is @xmath295 , which corresponds to an enclosed flux of @xmath8 ( in units of the flux quantum ) . here , we note that the area of one hexagon is @xmath296 , so that @xmath8 is related to the magnetic field strength @xmath49 as @xmath297 .
in addition to the gauge choice , let us furthermore assume that the flux value is a rational number , @xmath36 , where @xmath37 and @xmath38 are coprime integers .
then , the phase assigned to each hopping is periodic in the index @xmath197 with periodicity @xmath38 . in other words ,
the unit cell of the superlattice ( i.e. , the lattice together with the hopping phases ) is @xmath39 unit cells of the original lattice . with this observation
, we may invoke bloch s theorem , which states that we could write the electron wave functions @xmath298 as @xmath299 with this ansatz , the schrdinger equation which the fields have to satisfy reduces to a @xmath300-component matrix equation : the degrees of freedom are all fields @xmath42 in the unit cell of the superlattice , so that @xmath301 , and there are two sublattice and two spin components per unit cell of the original lattice .
this matrix equation is then merely an eigenvalue equation of a @xmath40 matrix , known as the harper equation,@xcite here , the @xmath303 matrices @xmath304 and @xmath305 encode hopping between unit cells with the same index @xmath197 and from unit cells with index @xmath306 to @xmath197 .
( note that the zeeman term is diagonal . )
thus , eq . can be written equivalently as eq . .
the @xmath303 matrices @xmath45 and @xmath46 may be written in terms of the sum of four contributions , that correspond to each of the terms in hamiltonian .
the contributions for the ordinary nn hopping are given by @xmath307 the zeeman effect is encoded by @xmath308 and @xmath309 .
the matrices for the iso coupling are @xmath310 where @xmath311 , @xmath312 , and @xmath313 denote the nnn vectors . here
, the phase factors due to the flux correspond to those in fig .
[ fig_lattice](b ) . due to the fact that the phases involve fractions of @xmath8 which are all multiples of @xmath314 , the @xmath8-periodicity of the butterfly spectra is @xmath99 , rather than the periodicity of @xmath55 which
is observed in absence of the iso coupling . finally , for the rso coupling
, the hopping matrices are given by @xmath315 where @xmath316 and @xmath317 .
the solution of the harper equation yields the dispersions of the bulk states , which are periodic over the so - called magnetic brillouin zone , @xmath38 times smaller than the original brillouin zone .
effectively , the geometry of the system is a torus , in this case .
however , to study edge states , a cylindrical geometry ( as displayed in fig .
[ fig_cylinder_edge_states ] ) is more convenient .
the edge - state spectrum is obtained from the harper equation similar to eq . , but with a matrix of dimension @xmath318 , where @xmath319 is the width of the ribbon in unit cells , and with the blocks in the top - right and bottom - left corner replaced by zeros ( i.e. @xmath320 ) . in this case
, the dispersion depends on just one component of @xmath175 , and its periodicity is the reciprocal lattice vector . to determine states as bulk states or edge states , we derive the density profile @xmath321 as a function of the unit cell index @xmath197 ( or equivalently , the coordinate component @xmath161 ) . for convenience ,
we choose @xmath322 to be the center of the ribbon , so that @xmath323 , where @xmath324 is the width of the ribbon .
the expectation value of this position , @xmath325 is used to distinguish edge states and bulk states .
edge states are characterized by a density profile that is sharply peaked at one edge , so that @xmath326 .
on the other hand , if the expectation value @xmath327 is close to zero , then it means that there is a significant contribution of the density away from the edge ; such a state is identified as a bulk state . in the edge state spectra of this paper , we have used this expectation value to color the states : red and blue colors indicate the two opposite edges , while gray is used for the bulk states . in the same way
, we use the spin expectation values @xmath328 ( @xmath329 ) to gain information about the spin states .
we note that this expectation is only reliable in the case that the spin is constant ( or almost constant ) where the density of the state is concentrated
. this may not be the case in the presence of rashba coupling .
in the case when the spin strongly depends on position , the length of the vector @xmath330 is less than unity .
87ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/revmodphys.82.3045 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.1057 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.45.494 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.146802 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.226801 [ * * , ( ) ] link:\doibase 10.1126/science.1148047 [ * * , ( ) ] link:\doibase 10.1126/science.1133734 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2015 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/97/23003 [ * * , ( ) ] link:\doibase 10.1103/physrevx.2.031004 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.066602 [ * * , ( ) ] link:\doibase 10.1103/physrevx.1.021001 [ * * , ( ) ] link:\doibase 10.1126/science.1204333 [ * * , ( ) ] link:\doibase 10.1038/nature10941 [ * * , ( ) ] link:\doibase 10.1038/nphys1916 [ * * , ( ) ] link:\doibase 10.1038/nature10871 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.043411 [ * * , ( ) ] link:\doibase 10.1038/nature08609 [ * * , ( ) ] link:\doibase 10.1038/nature09887 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.1523 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.255301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.175301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.255302 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/14/1/015007 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.010403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.145301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.130402 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.023404 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.035301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.190404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.165115 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1225 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.80.885 [ * * , ( ) ] link:\doibase 10.1103/physrevb.14.2239 [ * * , ( ) ] link:\doibase 10.1016/0022 - 1236(82)90094 - 5 [ * * , ( ) ] \doibase 10.1051/jphys:019850046080134500 [ * * , ( ) ] link:\doibase 10.1103/physrevb.48.11851 [ * * , ( ) ] link:\doibase 10.1103/physrevb.31.6207 [ * * , ( ) ] link:\doibase 10.1016/0003 - 4916(85)90148 - 4 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.71.3697 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.49.405 [ * * , ( ) ] link:\doibase 10.1088/0022 - 3719/15/22/005 [ * * , ( ) ] link:\doibase 10.1088/0022 - 3719/15/36/006 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.070402 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/12/3/033041 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.036808 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.125327 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.81.109 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.1193 [ * * , ( ) ] link:\doibase 10.1038/nature04233 [ * * , ( ) ] link:\doibase 10.1038/nature04235 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.205414 [ * * , ( ) ] link:\doibase 10.1063/1.3293411 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physrevb.85.195304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.165442 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.125435 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.045407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.161414 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.115439 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.073103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.121402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.176803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.107602 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.195444 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.115436 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.256801 [ * * , ( ) ] link:\doibase 10.1126/science.1191700 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.126803 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.035321 [ * * , ( ) ] link:\doibase 10.1143/jpsj.77.031007 [ * * , ( ) ] link:\doibase 10.1038/nphys1914 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.146802 [ * * , ( ) ] link:\doibase 10.1063/1.3493189 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.036802 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.136806 [ * * , ( ) ] link:\doibase 10.1038/nphys2259 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/97/33002 [ * * , ( ) ] link:\doibase 10.1038/nphys1507 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.235301 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.063629 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.053639 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.033622 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.255303 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.063614 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.115108 [ * * , ( ) ] | in this work , we explore the rich variety of two - dimensional topological phases that arise when considering the competing effects of spin - orbit couplings , zeeman splitting and uniform magnetic fields .
we investigate minimal models , defined on a honeycomb lattice , which clarify the topological phases stemming from the intrinsic and rashba spin - orbit couplings , and also from the zeeman splitting . in this sense
, our work provides an interesting path connecting the quantum hall phases , generally produced by the uniform magnetic field , and the quantum spin hall phases resulting from spin - dependent couplings .
first , we analyze the properties of each coupling term individually and we point out their similarities and differences . secondly , we investigate the subtle competitions that arise when these effects are combined .
we finally explore the various possible experimental realizations of our model . |
the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions form the backbone of interstellar chemistry and are important probes of the cosmic ray ionization rates in diffuse clouds , on the surfaces ( @xmath18 ) of molecular clouds , and ( with less reliability as we will discuss in this paper ) in the interiors ( @xmath19 ) of molecular clouds .
they are the backbone of chemistry because once h@xmath1 forms , cosmic ray ionization of h or h@xmath1 leads to the formation of these ions , and h@xmath2o@xmath0 recombination with electrons leads to oh and h@xmath1o .
reaction of oh with c or c@xmath0 leads to co. ion , which we do not discuss in detail in this paper although it is included in our models . ]
once these basic molecules are formed , many of the other polyatomic and rare species follow .
the pathway of cosmic ray ionization of hydrogen to these molecular ions follows two routes ( see figure 1 ) . in gas with significant h atoms
the ionization of h leads to h@xmath0 that then proceeds via a series of reactions ( see figure 1 top ) to oh@xmath0 , h@xmath1o@xmath0 and h@xmath2o@xmath0 ions .
we note that the charge exchange of o with h@xmath0 is slightly endothermic , so the reaction rate is proportional to exp(-230 k/@xmath20 ) ; this means that this reaction slows down at cooler temperatures , and a greater fraction of the cosmic ray ionizations of h are followed by recombination of h@xmath0 with neutral or negatively charged polycyclic aromatic hydrocarbons ( pahs or pah@xmath21 ) or electrons rather than proceeding to form o@xmath0 and then oh@xmath0 .
it is this atomic route to oh@xmath0 which is primarily important in diffuse clouds and in the @xmath22 surfaces of molecular clouds .
a second route dominates deeper in the opaque interiors of molecular clouds . here
the ionization of h@xmath1 leads to h@xmath23 , which then proceeds via a series of reactions ( see figure 1 bottom ) to oh@xmath0 , h@xmath1o@xmath0 and h@xmath2o@xmath0 .
a key competitor here to the formation of oh@xmath0 is the dissociative recombination of h@xmath24 with electrons , which has a high rate coefficient compared to the reaction of h@xmath24 with o. in addition , the reaction of h@xmath24 with co dominates that with o when the co abundance exceeds that of o. therefore , low electron abundances and high o abundances are needed to ensure that a large fraction of cosmic ray ionizations of h@xmath1 eventually produces oh@xmath0 . as we shall show in this paper ,
these two routes lead generally to two peaks in the abundances of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 as a function of depth or @xmath13 into a cloud .
the first peak ( at @xmath25 depending on the ratio of the incident fuv flux to the gas density ) occurs in atomic gas where the cosmic ray ionization of h begins the chemical chain .
] the second peak ( at @xmath4 , again dependent on the fuv flux / gas density ratio ) occurs in molecular gas where the cosmic ray ionization of h@xmath1 begins the chemical chain .
deeper in the cloud the gas phase oxygen freezes out as water ice on grain surfaces ( e.g. , hollenbach et al 2009 , hereafter h09 ) , and the gas phase abundances of the three ions drop . whereas significant columns of oh@xmath0 and h@xmath1o@xmath0 are produced in the first peak , most of the h@xmath2o@xmath0 column arises in the second peak . in warm ( @xmath26 k ) neutral gas with significant abundances of h@xmath1
there are other dominant pathways to form oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions than the two paths initiated by cosmic rays and shown in figure 1 .
one of the goals of this paper is to show under what conditions ( e.g. , fuv flux , gas density , cosmic ray ionization rate ) cosmic rays initiate the formation of the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions , and under what conditions other chemical routes dominate .
if cosmic rays dominate , then the observed columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 serve as a probe of the cosmic ray ionization rate .
the major goal of this paper is to show how these columns depend on the cosmic ray ionization rate , the gas density , the fuv flux , the total column or optical extinction @xmath27 through the cloud , and the molecular hydrogen abundance .
comparison with our models allow the cosmic ray ionization rate to be estimated from these ion columns and the atomic h column .
our method of estimating cosmic ray ionization rates from the abundance of eventual products of this ionization dates back to the 1970 s , when black & dalgarno ( 1973 ) pointed out the sensitivity of the oh and hd abundances to cosmic ray ionization rates .
a later comprehensive paper by van dishoeck & black ( 1986 ) showed how the abundance of oh in diffuse or translucent clouds could be used to estimate the cosmic ray rate .
one advantage of using oh@xmath0 or h@xmath1o@xmath0 to probe cosmic ray ionization rates over using oh is that oh is produced after a series of reactions following the formation of oh@xmath0 ( see figure 1 ) .
this makes the inference of cosmic ray ionization rate dependent on knowledge of the rate coefficients of these additional reactions .
recent observations strongly motivate the theoretical study of the @xmath28 , @xmath29 , and @xmath30 molecular ions . prior to 2010
, the only observations of these three ions within the interstellar medium ion has long been observed in cometary comae ( e.g. wehinger et al .
1974 ) , where it is produced by the photoionization of water by solar ultraviolet radiation . ]
consisted of a few detections of @xmath30 in rich interstellar sources of submillimeter line emission and absorption ( wootten et al 1991 ; phillips , van dishoeck & keene 1992 , hereafter pvdk92 ; goicoechea & cernicharo 2001 ; van der tak et al 2006 ) .
thanks largely to absorption - line spectroscopy with the hifi instrument on the _ herschel space observatory _ , together with additional ground - based detections of oh@xmath0 obtained with the apex telescope ( submillimeter rotational transition ) and the eso paranal observatory ( near - uv electronic transition ) ,
the observational picture has improved radically over the past two years .
all three molecular ions have now been detected in both the diffuse and dense galactic interstellar medium ( gerin et al .
2010 ; ossenkopf et al .
2010 ; wyrowski et al .
2010 ; neufeld et al .
2010 ; schilke et al .
2010 ; gupta et al . 2010 ; bruderer et al . 2010 ; benz et al . 2010 ; krelowski , beletsky , & galazutdinov 2010 ) and in external galaxies ( van der tak et al 2008 ; weiss et al .
2010 ; van der werf et al . 2010 ; gonzlez - alfonso et al . 2010
; aalto et al 2011 ) .
the most reliable column density determinations are obtained for foreground molecular clouds lying along the sight - lines to bright submillimeter continuum sources in the galactic disk . here , the absorbing material typically covers a wide range of line - of - sight velocities , arising in widely distributed material along the sight - line .
the total inferred column densities of the ions lie in the range @xmath31 , the largest values being attained for @xmath28 and the smallest for @xmath30 ( gerin et al .
2010 ; neufeld et al .
2010 ) . in their study of oh@xmath0 and h@xmath1o@xmath0 absorption along the sight - line to the luminous star - forming region w49n , neufeld et al .
( 2010 , hereafter n10 ) measured an average oh@xmath0/h@xmath1o@xmath0 abundance ratio of @xmath32 , with variations over the range @xmath33 .
these observed ratios are considerably larger than the value @xmath34 expected in fully molecular gas
. formation is followed by h@xmath1o@xmath0 formation and the destruction rate per ion is nearly the same ( due to reaction with h@xmath1 ) , leading to similar abundances for both ions ( n10 ) . ] by means of a simple analytical treatment of the chemistry of oh@xmath0 and h@xmath1o@xmath0 , confirmed by more detailed pure gas - phase models performed using the meudon pdr model ( le petit et al 2006 ; goicoechea & le bourlot 2007 ) , n10 concluded that the molecular fraction in the absorbing material lies in the range 2 8 @xmath35 .
this conclusion was supported by the observed distribution in velocity space of the oh@xmath0 and @xmath29 absorption , which proved similar to that of the atomic gas probed by 21 cm absorption studies , and quite dissimilar from that of the molecular gas traced by hf or ch absorption .
the analytic treatment introduced by n10 also allowed the cosmic ray ionization rate to be inferred from the abundances of oh@xmath0 and @xmath29 relative to atomic hydrogen .
the resulting estimate of the cosmic ray ionization rate in the range @xmath7= 0.6 2.4 @xmath36 ( primary ionization rate per h atom ) was broadly consistent with earlier values inferred quite independently from observations of the h@xmath24 molecular ion toward different sight lines ( indriolo et al .
one of the goals of the present study is to refine the n10 analytic treatment of diffuse cloud chemistry through detailed modeling , as an aid to interpreting the growing body of observational data now available .
this paper is organized as follows . in 2 we describe the chemical / thermal model of both an opaque molecular cloud illuminated by fuv radiation , or a diffuse cloud illuminated by the interstellar radiation field . in
3 we show the model columns and column ratios of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 as functions of the cloud gas density @xmath8 , the incident fuv flux @xmath37 ( in units of the local average interstellar field , see below ) , and the primary cosmic ray ionization rate @xmath7 per h atom for the case of opaque molecular clouds .
we also show the same results for diffuse clouds , but with the total column or @xmath27 through the cloud as an additional parameter . in
4 we compare our model results with previous h@xmath2o@xmath0 observations and recent oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 observations by _ herschel _ , and show how observations compared to models can constrain the cosmic ray ionization rates .
we summarize our results in 5 .
appendix a presents tables of key reaction rate coefficients and adopted abundances and grain parameters .
appendix b includes analytic expressions that explain the variation of oh@xmath0 with @xmath13 over the first peak , and the variation of h@xmath2o@xmath0 with @xmath13 over the second peak .
appendix c assesses the sensitivity of our results to certain chemical rate coefficients , the gas phase abundance of elemental oxygen and low ionization potential metals , and the freeze - out of species .
the numerical code we have developed to model the chemical and thermal structure of an opaque cloud externally illuminated by fuv flux is based on our previous pdr model described in h09 .
this 1d code models a constant density @xmath8 ( the hydrogen nucleus density ) slab of gas , illuminated from one side by an fuv flux of @xmath38 erg @xmath11 s@xmath15 incident perpendicular to the slab .
the unitless parameter @xmath37 is defined above in such a way that @xmath39 corresponds to the average local interstellar radiation field in the fuv band ( draine 1978 ) . corresponds to @xmath40 in the tielens & hollenbach ( 1985 ) units based on the habing ( 1968 ) local interstellar radiation field .
the shape of the fuv spectrum for @xmath41 is implicitly assumed to mimic that of the draine field , which is approximately that of a @xmath42 k star . ]
the code calculates the steady state chemical abundances and the gas temperature from thermal balance as a function of depth into the cloud .
it incorporates 63 chemical species , @xmath43 chemical reactions , and a large number of heating mechanisms and cooling processes .
the chemical reactions include fuv photoionization and photodissociation ; cosmic ray ionization ; neutral - neutral , ion - neutral , and electronic recombination reactions .
h@xmath1 self - shielding is included as described in h09 , and co self - shielding and the partial shielding of co by h@xmath1 is included as described in visser et al ( 2009 ) .
the code includes the photodissociation of molecules by
secondary " fuv photons produced ( ultimately ) by cosmic rays ( prasad & tarafdar 1983 ) .
we also include reactions with charged dust grains and pahs ; and the formation of h@xmath1 , oh , h@xmath1o , ch , ch@xmath1 , ch@xmath2 , and ch@xmath44 on grain surfaces .
the code treats the freezing of all condensable species to grain surfaces and three desorption processes : thermal desorption , photodesorption , and cosmic ray desorption .
the only significant difference in the desorption code used here versus the h09 code is the inclusion of the new ( higher by factor @xmath45 ) rate of photodesorption of co ( oberg et al 2009 ) .
the code does not include photodesorption by the secondary fuv photons ; this process is negligible at the peaks in the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 abundances .
the code has been used to model regions which lie at hydrogen nucleus column densities @xmath46 @xmath11 ( or @xmath47 ) from the surface of a cloud .
therefore , it applies not only to the photodissociated surface region , where gas phase hydrogen and oxygen are nearly entirely atomic and where gas phase carbon is mostly c@xmath0 , but also to regions deeper into the molecular cloud where hydrogen is in h@xmath1 molecules and carbon is in co molecules . even in these molecular regions ,
the attenuated fuv field can play a significant role in photodissociating h@xmath1o and o@xmath1 , in photodesorbing species adsorbed on grain surfaces , and in heating the gas .
however , the code is now sufficiently general that it finds the steady state solutions for abundances and temperature in any region of a molecular cloud , even where fuv is insignificant .
we emphasize that we present a steady - state model of chemical abundances as a function of depth into the cloud .
the chemical timescales can be quite long , which might suggest that time dependent models are more appropriate .
for example , the timescale to convert atomic h gas to fully molecular h@xmath1 gas is @xmath48 years , where @xmath8 is the hydrogen nucleus density in units of @xmath49 .
however , as we show below , the oh@xmath0 and h@xmath1o@xmath0 ions peak in abundance when @xmath50h@xmath51 , which occur typically at @xmath52 .
[ note , abundances in this paper are defined relative to hydrogen nuclei , so that @xmath50h@xmath53h@xmath54 and @xmath55h@xmath56h@xmath1 ) ]
. therefore , the timescales for even diffuse clouds of density @xmath57 @xmath49 to reach these abundances are less than @xmath58 years , which is shorter than the typical lifetime of a diffuse cloud ( wolfire et al 2003 ) .
steady state solutions therefore generally apply , at least for computing the columns of these ions .
the steady state models may somewhat underestimate the ion columns for low density diffuse clouds with @xmath59 , since the steady state solutions can lead to lower abundances of these two ions than time dependent solutions in the high @xmath13 regions .
this arises because the steady state abundances of h@xmath1 are higher than time dependent models which start with fully atomic gas .
higher h@xmath1 abundances lead to more destruction of oh@xmath0 and h@xmath1o@xmath0 in the high @xmath13 regions where h@xmath1 and not electrons dominate the destruction of these ions .
liszt ( 2007 ) provides a detailed analysis of the time dependent formation of h@xmath1 and oh@xmath0 as a function of @xmath13 in diffuse clouds with initial atomic conditions .
we also emphasize that our model does not include turbulent dissipation and heating of small pockets of gas along the line of sight ( e.g. , godard , falgarone , & pineau des forets 2009 and references therein ) .
we do , however , run pdr models with small fractions of the line of sight having either enhanced temperature or enhanced rates of ion - neutral drift to test the possible effects of turbulence , and find the effects on oh@xmath0 , h@xmath1o@xmath0 , h@xmath2o@xmath0 column densities are likely small . for simplicity
we assume constant h nucleus density @xmath8 in our models . at low @xmath13 and low @xmath60 , where much of the oh@xmath0 and h@xmath1o@xmath0
columns often arise , the temperature is quite constant so that constant density implies constant thermal pressure .
if thermal pressure ( and not turbulence ) dominates deeper into a high @xmath13 cloud , then the density will rise as one moves from its warmer pdr surface to its co - cooled molecular interior .
in addition , self gravity can raise the density of the interior regions at higher @xmath13 .
still another effect is the transition of atomic hydrogen to molecular hydrogen which can raise @xmath8 by a factor of 2 if thermal pressure is conserved .
we ignore the possible rise in density , which mainly affects the second peaks of the ions deep ( @xmath61 ) in the cloud .
if such a density enhancement occurs , it tends to depress the abundances of oh@xmath0 and h@xmath1o@xmath0 in the second peaks , but the abundance of h@xmath2o@xmath0 at the second peak is not sensitive to hydrogen density if pahs are present(see 3 ) .
of a given source there , then our models with that @xmath8 will give a good prediction of the behavior of the second peaks . ]
the second peak is dependent on the abundance of pahs , very small grains and low ionization potential metal ions there , which can control the electron density , @xmath62 .
the abundance of pahs and very small grains at high @xmath13 is uncertain due to their possible coagulation on larger grain surfaces .
one significant difference between the chemical code used in this paper and that used in h09 is the inclusion of pahs and very small grains ( vsgs , radius @xmath63 ) , which affect the ionization balance by enhancing the recombination of positive atomic ions . in the rest of this paper
we shall often use the term pah " to denote both pahs and vsgs .
pahs also affect the second peaks of the ions by destroying electrons in the reaction e + pah @xmath64 pah@xmath21 .
( the effects of pahs on ionization balance has been treated extensively before in the literature : e.g. , lepp & dalgarno 1988 , bakes & tielens 1998 , weingartner & draine 2001a , flower & pineau des forets 2003 , liszt 2003 , wolfire et al 2003 , 2008 ) .
h09 focused on the h@xmath1o and o@xmath1 peaks , which occur at relatively high @xmath65 .
although the pah abundances are not known deep in molecular clouds , h09 assumed that their abundances are low , due to coagulation of pahs on the surfaces of larger dust grains . here , however , we are mostly interested in the chemistry at low @xmath18 , and in particular in the chemistry of diffuse clouds .
these are the regions where pahs are observed to be present , and we use pah parameters derived from the literature . for simplicity
, we adopt a single size pah for the pah distribution , and assume that the standard pah has 100 c atoms , and a number abundance of @xmath66 with respect to h nuclei ( see wolfire et al 2008 for a discussion of the amount of carbon in pahs ; here we adopt 100 c atoms and not the 35 c atoms that wolfire et al adopted because of the result of draine & li 2007 , which suggested that the distribution in mass peaks at 100 c atoms ) .
one of the key impacts of pahs is in the recombination of h@xmath0 .
as shown in figure 1 ( top ) , cosmic ray ionization of h can lead to oh@xmath0 , but a competing route is the neutralization of h@xmath0 by an electron , pah , or pah@xmath21 .
pahs therefore lower the production of oh@xmath0 and the columns of all the ions in the first peak .
conversely , in the second peaks , the reduction in electron abundance caused by pahs cause fewer h@xmath24 ions to recombine with electrons , and lead to greater production rates of the ions as well as smaller destruction rates for h@xmath2o@xmath0 .
therefore , the columns of the ions increase with the presence of pahs in the second peaks .
we also discuss results with no pahs or very small grains at high @xmath13 . in appendix a we present table 1 which lists the rate coefficients of key reactions in the pathways to the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions , as well as reactions that are either new or changed since h09 .
of particular note here are the photoionization of oh and h@xmath1o , the photodissociation of oh@xmath0 and h@xmath1o@xmath0 , the treatment of pahs especially the photodetachment reaction rate for pah@xmath21 , the fine structure level population dependence in the charge exchange reaction of o with h@xmath0 , and some minor modifications in reaction rates key to determining the abundances of the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions , such as their dissociative recombination rates with electrons and their reactions rates with h@xmath1 . we have also included in table 1 rates that are important in producing h@xmath0 by chemical means rather than by cosmic ray ionization of h in pdrs .
there are two main chemical routes .
the first is initiated by the fuv photoionization of c to c@xmath0 .
the c@xmath0 reacts with oh to form co@xmath0 .
the co@xmath0 charge exchanges with h to form h@xmath0 .
the second is also initiated by fuv photoionization of c to c@xmath0 .
the c@xmath0 reacts with h@xmath1 to form ch@xmath0 .
the ch@xmath0 is photodissociated by fuv photons to produce h@xmath0 .
both these reaction chains are very much enhanced by high ( @xmath67 k ) temperatures in gas with appreciable h@xmath1 .
the first chain is enhanced because high gas temperatures lead to significant amounts of oh being formed by the reaction h@xmath1 + o @xmath64 oh + h. this reaction has an activation barrier of @xmath68 k and so is insignificant for low gas temperatures .
the second chain similarly is enhanced because the reaction c@xmath0 + h@xmath1 @xmath64 ch@xmath0 + h has an activation barrier of @xmath69 k ( see table 1 ) .
with c@xmath0 to form ch@xmath0 .
this reaction has no activation barrier and can be moderately important at low @xmath70 . ] in regions where either of these two chains dominate the production of h@xmath0 over cosmic ray ionization , the ions oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 only provide upper limits to the cosmic ray ionization rates .
other alternate routes to the production of oh@xmath0 and h@xmath1o@xmath0 ions include three routes which produce oh or h@xmath1o without the ions : the production of water on grain surfaces followed by photodesorption of the water to produce gas phase oh and h@xmath1o , the radiative association of o with h to form oh , and the reaction of fuv - pumped h@xmath1 in excited vibrational states with o to form oh . the gas phase oh and h@xmath1o can be photoionized by fuv photons to produce oh@xmath0 and h@xmath1o@xmath0 ( see table 1 ) . unlike the routes described in the preceeding paragraphs , these routes are never dominant in producing the ion column densities
. an analogous route to the formation of h@xmath1o on grain surfaces followed by photodesorption , but one not treated in this work , is the time - dependent evaporation of water ice which occurs around newly formed stars .
the rapid rise in embedded luminosity heats the dust grains above about 100 k , and the icy mantles on grains are then thermally evaporated .
this sudden injection of high abundances of water vapor into the gas is followed by reaction of the gas phase water with hco@xmath0 and h@xmath24 to form h@xmath2o@xmath0 .
eventually , the system relaxes to the steady state chemistry described in this paper , but for a short time , there might be a large enhancement in h@xmath2o@xmath0 ( millar , herbst , & charnley 1991 , pvdk92 ) .
if this release of h@xmath1o from the icy grain surface to the gas occurs in regions with elevated fuv fields , the photoionization of h@xmath1o could also result in enhanced h@xmath1o@xmath0 abundances ( gupta et al 2010 ) we also present in appendix a table 2 , which lists the gas phase elemental abundances , the pah properties , and the grain surface area per h nucleus adopted in our code . as noted above
, our opaque molecular cloud models invoke a constant density @xmath8 slab , illuminated from one side by a 1d normal fuv flux @xmath37 .
we solve for the gas temperature and the gas phase and ice abundances of the various species as a function of depth in the slab .
note that depth is synonymous with hydrogen nucleus column @xmath9 [ = @xmath9(h)+2@xmath9(h@xmath1 ) ] into the cloud or @xmath13 into the cloud .
we take @xmath71 @xmath11 .
we follow the chemistry to @xmath72 , beyond which there is little contribution to the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions .
we use the primary cosmic ray ionization rate per h atom as an input parameter , since our code calculates the secondary ionizations caused by cosmic rays and these depend on the h@xmath1 and electron abundances . in order to probe the range of cosmic ray ionization rates suggested in the literature , we include cases with primary ionization rates of @xmath7@xmath73 s@xmath15 and @xmath74 s@xmath15 per h atom , which correspond to total rates ( including secondary ionizations ) of about @xmath75 s@xmath15 and @xmath76 s@xmath15 respectively . note that the primary rate per h@xmath1 molecule is 2@xmath7 .
we also adopt @xmath7 as the primary rate for he atoms ; however , he has insignificant secondary ionizations . although there is evidence that the cosmic ray ionization rate may decrease with depth ( e.g. , rimmer et al 2011 , indriolo & mccall 2012 , and discussion later in this paper ) , for simplicity we assume the primary cosmic ray ionization rate does not vary with depth into a cloud in a given model .
the main effect of this assumption is that we may have overestimated the abundances and columns of the oh@xmath0 and h@xmath1o@xmath0 ions in the second ( deeper ) peak relative to the first peak .
however , since we vary @xmath7 in our parameter study , the reader can use lower @xmath7 for the columns in the second peak if so desired .
the abundance of h@xmath2o@xmath0 in the second peak is not sensitive to @xmath7 if pahs are present , as we will show below .
the main parameters that we explore for our one - sided , opaque molecular cloud models are the gas density @xmath8 , the incident fuv flux @xmath37 , and the primary cosmic ray ionization rate @xmath7 .
we study the parameter space @xmath77 @xmath49@xmath78 @xmath49 , 1@xmath79 , and @xmath80 s@xmath81@xmath7@xmath82 s@xmath15 .
we explore the sensitivity of the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions to assumptions about the pah chemistry , the elemental gas phase abundances of low ionization potential metals , and the rate coefficient for the formation of h@xmath1 on grain surfaces .
the diffuse cloud models treat a constant density slab of total thickness @xmath27 illuminated on both sides by a 1d normal fuv flux @xmath83 .
we explore the parameter space @xmath84 s@xmath85 @xmath7@xmath86 s@xmath15 , @xmath87 , @xmath88 @xmath49@xmath89 @xmath49 , 1@xmath90 , and @xmath91 where @xmath92 @xmath49 .
the models of wolfire et al ( 2003 ) provide a good estimate of @xmath93 in the galaxy by estimating @xmath37 from the star formation rate and the dust opacity as a function of galactocentric radius @xmath94 , and setting the thermal pressure to provide two phases , a cold diffuse cloud phase and a warm intercloud medium that fills most of the volume . assuming @xmath95 kpc as the solar location , wolfire et al
find @xmath96 , @xmath97 @xmath49 , and @xmath98 for diffuse clouds at @xmath95 kpc ; @xmath99 , @xmath100 @xmath49 , and @xmath101 at @xmath102 kpc ; @xmath103 , @xmath104 @xmath49 , and @xmath105 at @xmath106 kpc ; and @xmath107 , @xmath108 @xmath49 , @xmath109 at @xmath110 kpc
. therefore , our prime region for study is @xmath111 .
we are especially interested in how the ratio of the columns @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) and @xmath9(oh@xmath0)/@xmath9(h ) vary as functions of @xmath27 , @xmath12 and @xmath7/@xmath8 .
we note that the columns our models predict are columns perpendicular to the 1-sided ( molecular cloud ) slab or 2-sided ( diffuse cloud ) slab .
if the slabs are viewed at an angle @xmath112 with respect to the normal , the observed columns will increase by ( @xmath113 .
another effect that will obviously raise the columns is if there are more than one diffuse cloud ( in a given velocity range ) along the line of sight ( los ) , or if the molecular cloud is clumpy and fuv scattering then introduces several surfaces " along the los .
however , as we will show , @xmath7/@xmath8 in diffuse clouds can be estimated from our models from the ratios @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) and @xmath9(oh@xmath0)/@xmath9(h ) , and the ratios are independent of the geometric effects .
nevertheless , there is some degeneracy in the solution for @xmath7/@xmath8 , depending on the combination of the @xmath27 of a single cloud , @xmath12 , and the enhancement in columns created by the geometrical effects .
in order to understand the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions produced as a function of @xmath8 , @xmath37 , and @xmath7 , we first study the detailed chemical and thermal structure of a few specific ( standard ) cases . figure 2 shows the chemical abundances as a function of depth @xmath13 into the cloud for the case @xmath114 @xmath49 , @xmath115 , and @xmath7= @xmath116 s@xmath15 . this case
is chosen not only because it may be appropriate for the ambient interstellar radiation field ( isrf ) incident on a relatively low density gmc , but also because the surface ( @xmath117 ) structure ( i.e. , @xmath20 and chemical abundances as a function of depth or @xmath13 ) is illustrative of the depth structure of a diffuse or translucent cloud .
in addition , the cosmic ray rate may be appropriate to the interior of molecular clouds .
the main chemical result is that the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions all have a peak at @xmath118 , and then have a second peak at @xmath119 for this combination of @xmath8 and @xmath37 .
we first discuss the surface peaks at low @xmath13 .
oh@xmath0 peaks at @xmath120 , where the molecular hydrogen abundance is @xmath50h@xmath121 .
the h@xmath1o@xmath0 peaks slightly deeper , at @xmath52 , where @xmath50h@xmath122 .
finally , h@xmath2o@xmath0 peaks at @xmath123 , where @xmath50h@xmath124 .
appendix b provides an approximate analytic solution to the chemistry that explains the oh@xmath0 first peak , and its relation to @xmath50h@xmath125 . from @xmath126 to either @xmath127 ( oh@xmath0 ) or @xmath128 ( h@xmath1o@xmath0 ) or @xmath129 ( h@xmath2o@xmath0 ) the three ions rise in abundance with increasing @xmath13 because of the rise in the h@xmath1 abundance , which drives the cosmic ray - produced o@xmath0 to the molecular ions . at larger @xmath13 , but for @xmath130(h@xmath1 ) significantly less than its maximum value of @xmath131 , the abundance of oh@xmath0 and h@xmath1o@xmath0 plateau at the peak value because here both their formation and destruction rates are proportional to @xmath130(h@xmath1 ) .
h@xmath1o@xmath0 tends to peak at somewhat higher @xmath13 than oh@xmath0 because , even at the peak of oh@xmath0 abundance , not all cosmic ray ionizations lead to h@xmath1o@xmath0 and so its abundance continues to rise as the h@xmath1 abundance rises with increasing @xmath13 . finally , as @xmath130(h@xmath1 ) approaches 0.5 , two effects lead to a drop in the oh@xmath0 and h@xmath1o@xmath0 abundances with increasing @xmath13 .
one is that the formation rates of these ions saturate as nearly every cosmic ray ionization leads to their production , whereas the destruction rates still scale as @xmath130(h@xmath1 ) , which increases with increasing @xmath13 .
the other dominant effect is that the h abundance drops so that the oh@xmath0 formation rate via the top chain of reactions in figure 1 drops . the bottom chain at relatively low @xmath13 is not as efficient at producing oh@xmath0 , as the electrons are relatively abundant ( @xmath132 ) in these surface regions , and h@xmath24 recombines with electrons rather than forming oh@xmath0.(h@xmath1 ) approaches 0.5 .
the bottom chain dominates in the second peak , however , because the h@xmath1 abundance is high and the electron abundance is low . ]
the h@xmath2o@xmath0 is destroyed not by h@xmath1 , but by electrons , whose abundance stays quite constant ( supplied mostly by c@xmath0 but with possible contribution by h@xmath133 at high @xmath7/@xmath8 ) at the cloud surface .
the h@xmath2o@xmath0 starts to drop in abundance once the gas becomes predominantly h@xmath1 , due to the second effect described above .
we next discuss the second deeper peak in the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 ions .
as one moves deeper into the cloud ( typically , @xmath134 ) , the electron abundance starts to drop .
as this happens , a greater fraction of the cosmic ray ionizations of h@xmath1 leads to the production of the three ions , and their formation rates rise .
the destruction of oh@xmath0 and h@xmath1o@xmath0 is by h@xmath1 , which now has constant abundance ( the gas is fully h@xmath1 ) , so the destruction rates hold constant .
therefore , these two ions rise in abundance
. finally , they peak and fall in abundance for @xmath135 because gas phase oxygen freezes out as water ice , and again the oxygen reaction with h@xmath24 can not compete with h@xmath24 electronic recombination or its reaction with co. therefore , the formation rates of all three ions drop .
h@xmath2o@xmath0 behaves somewhat more dramatically , because its destruction is mainly by dissociative recombination with electrons .
thus , as the electron abundance drops , not only is the formation rate of h@xmath2o@xmath0 enhanced , but the destruction rate is suppressed .
therefore , h@xmath2o@xmath0 rises to much higher abundances than oh@xmath0 and h@xmath1o@xmath0 in the second peak
. it also drops at very high @xmath13 because gas phase elemental oxygen from which h@xmath2o@xmath0 forms freezes out as water ice . in this particular case
, there is a column of @xmath9(oh@xmath0)=@xmath136 @xmath11 in the first surface peak and @xmath137 @xmath11 in the deeper peak ; for h@xmath1o@xmath0 the columns are @xmath138 @xmath11 and @xmath136 @xmath11 ; and for h@xmath2o@xmath0 the columns are @xmath139 @xmath11 and @xmath140 @xmath11 .
the columns of oh@xmath0 and h@xmath1o@xmath0 are not detectable .
typically , columns of @xmath141 @xmath11 are needed for detection via absorption spectroscopy .
in particular , observations of oh@xmath0 often imply columns @xmath142 @xmath11 , which suggests that higher cosmic ray rates are required . therefore , for the rest of our standard cases we use @xmath7@xmath143 s@xmath15 .
most of the h@xmath2o@xmath0 column is produced in the second deeper peak , and in our model the predicted columns are detectable . for an absorption measurement a background submillimeter source behind or in a cloud with @xmath144 from observer to submillimeter source is required .
figure 3 shows the case @xmath145 @xmath49 , @xmath6 , and @xmath7= @xmath146 s@xmath15 : in other words , the same as figure 2 but with ten times the cosmic ray ionization rate .
the abundances of all three ions in the first peak rise in proportion to @xmath7 .
the oh@xmath0 and h@xmath1o@xmath0 ion abundances in the second peak also scale roughly with @xmath7 .
the column of @xmath9(oh@xmath0)=@xmath147 @xmath11 in the first peak and @xmath148 @xmath11 in the deeper peak ; for h@xmath1o@xmath0 the columns are @xmath149 @xmath11 and @xmath150 @xmath11 ; and for h@xmath2o@xmath0 the columns are @xmath151 @xmath11 and @xmath152 @xmath11 . because the electron abundance is low in the second peak , a significant fraction ( @xmath153 ) of cosmic ray ionizations of h@xmath1 lead to oh@xmath0 and h@xmath1o@xmath0 , and their destruction is by h@xmath1 , which does not change in abundance with varying @xmath7 .
thus , the scaling of oh@xmath0 and h@xmath1o@xmath0 abundances and columns with @xmath7 .. ] however , the h@xmath2o@xmath0 abundance in the second peak does not rise linearly with @xmath7 , but stays fairly constant , because although the formation rate scales with @xmath7 , the destruction rate also increases as @xmath7 is raised , due to the higher electron abundances produced by the enhanced cosmic ray flux . unlike oh@xmath0 and h@xmath1o@xmath0 , which are destroyed by h@xmath1 , h@xmath2o@xmath0 is destroyed by dissociative recombinations with electrons . in fact , to first order , we would expect the abundance of h@xmath2o@xmath0to scale as @xmath7/@xmath154@xmath7@xmath155 at the second peak .
the abundance of electrons @xmath156 deep in the cloud depends on the uncertain pah abundance deep in the cloud .
if the pah abundance remains as high as is indicated in diffuse clouds and cloud surfaces ( as we assume in our standard models ) , then we obtain the following result .
electrons are formed by cosmic ray ionization of h@xmath1 .
electrons are mainly destroyed by collisional attachment to neutral pahs , and the pahs are mostly neutral .
therefore , the abundance of electrons @xmath157@xmath7@xmath158 . as a result ,
if pahs are abundant in this deep peak , we predict that @xmath130(h@xmath2o@xmath0 ) @xmath159 @xmath7@xmath155 will be independent of both @xmath8 and @xmath7 !
note that figure 3 compared to figure 2 shows that the electron abundance in the second peak does scale roughly as @xmath7 , and that the abundance of h@xmath2o@xmath0 in the second peak does not change significantly as we increase the cosmic ray ionization rate by 10 .
appendix b presents an analytic solution for @xmath130(h@xmath2o@xmath0 ) near the second peak if pahs are present . since cosmic ray ionization of h@xmath1 is similar to x - ray ionization of h@xmath1 , this result implies that , _ if pahs are present _ , regions of enhanced x - ray ionization will not show enhanced h@xmath2o@xmath0 columns .
we discuss in 3.1.3 the case of no pahs at high @xmath13 .
figure 4 shows the same case as figure 3 , but plots a parameter @xmath160 , first discussed by n10 . here
, the parameter @xmath160 is defined as the rate per unit volume of formation of oh@xmath0 divided by the total ( not primary ) rate per unit volume of cosmic ray ionization of h and h@xmath1 . in effect , @xmath160 is an efficiency parameter in determining the formation of oh@xmath0 from cosmic rays . if pah , pah@xmath21 and electron abundances are relatively low , and h@xmath1 and gas phase o abundances are high , then @xmath160 is near unity .
essentially , one needs o@xmath0 to react with h@xmath1 before h@xmath0 reacts with e , pah , or pah@xmath21 .
although this qualitative limit is clear from the chemical pathways shown in figure 1 , we derive in appendix b an analytic formula for @xmath160 which makes this statement more quantitative . roughly , the condition for @xmath160 to be of order unity is : @xmath161\ ] ] however , if the reverse is true , then h@xmath0 can recombine with pah , pah@xmath21 , or electrons and disrupt the chain of reactions that lead to oh@xmath0 , leading to low ( @xmath162 ) values of @xmath160 .
figure 4 also plots the temperature and repeats the plots of the abundances of h@xmath1 and electrons , since they help determine the value of @xmath160 , as well as the abundance of oh@xmath0 ( see appendix b ) .
finally , we add the abundance of h@xmath24 to figure 4 since it is also used to estimate cosmic ray rates ( e.g. , indriolo et al 2007 , 2011 ) . because h@xmath24 is formed by the reaction of h@xmath1 with h@xmath23 and often destroyed by electrons ( see figure 1 ) , we see the h@xmath24 abundance rise with @xmath13 as the h@xmath1 abundance rises and the electron abundance falls . in general , the h@xmath24 probes the cosmic ray ionization rates at higher @xmath13 than the first peak in oh@xmath0
. figures 5 and 6 show the case @xmath163 @xmath49 , @xmath164 , and @xmath7= @xmath165 s@xmath15 .
this case is nearly identical to the standard case in h09 , and is representative of a giant molecular cloud ( gmc ) surface illuminated by an fuv field somewhat higher than the interstellar radiation field because of the presence of nearby o and b stars
. the total cosmic ray ionization rate ( @xmath166 s@xmath15 including secondary ionizations ) may represent values on the surfaces of gmcs , but may be somewhat high for the interior .
we see from figures 4 and 6 that the gas temperatures of the molecular interiors of these clouds at @xmath167 is of order 30 k for @xmath168 @xmath49 and 70 k for @xmath169 @xmath49 , due to cosmic ray heating when @xmath7= @xmath146 s@xmath15 .
typically , temperatures in molecular cloud interiors are observed to be @xmath170 k , suggesting that such high cosmic ray ionization rates may not be appropriate for molecular cloud interiors .
however , as we shall see , such high cosmic ray rates are required to explain observations of diffuse clouds , which should have the same cosmic ray rates as the surfaces ( @xmath22 ) of gmcs .
this suggests that @xmath7 may be higher on the surface of a molecular cloud than deep in its interior .
if chemistry is driven by fuv photoreactions and particle - particle reactions , such as h / h@xmath1 chemistry , then the chemical abundances mainly depend on the ratio @xmath171 .
therefore , one expects and sees that the h@xmath1 abundance of figure 5 closely matches that of figures 2 and 3 , which have the same @xmath171 ratio .
the h@xmath1 abundance in the higher @xmath37 case is a bit lower because of fuv photodissociation of h@xmath1 in fuv - pumped excited vibrational states of h@xmath1 .
however , the molecular ion abundances in the first peak depend to first order on the ratio @xmath7/@xmath8 ( see appendix b ) and are not too sensitive to @xmath37 .
thus , in this @xmath163 @xmath49 case , their abundances in the first peak drop by a factor of nearly 10 - 30 compared to figure 3 ( which has the same value of @xmath7 ) as the density rises by 100 from the @xmath114 @xmath49 assumed in figure 3 .
slight differences in electron abundances , h@xmath1 abundances , and @xmath20 explain the divergence from the expected @xmath172 dependence .
the second peaks of oh@xmath0 and h@xmath1o@xmath0 nearly scale as @xmath172 , as expected . however , the second peak of the h@xmath2o@xmath0 abundance ( @xmath173 ) is independent of @xmath8 , as predicted above by the scalings of the electron density with @xmath8 and @xmath7 if pahs are present .
figures 7 shows the case @xmath174 @xmath49 , @xmath175 , and @xmath7= @xmath146 s@xmath15 . this case may represent strongly illuminated pdrs such as may occur around embedded compact or ultracompact hii regions , or possibly embedded protostars illuminating the opaque walls of outflow cones . this
high density and high fuv flux case was chosen because most of the column of all three ions is produced not by cosmic ray ionization , but by other chemical reactions . in figure 7
we see an enormous enhancement of oh@xmath0abundance at @xmath176 . here , @xmath177 k and at the same time the abundance of h@xmath1 is moderately high , @xmath178 . at these elevated temperatures ,
as discussed in 2.1 , the h@xmath1 can react rapidly with o to form oh or with c@xmath0 to form ch@xmath0 , leading to reaction chains that make oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 .
one of the key heating mechanisms providing this high @xmath20 is the fuv pumping and the h@xmath1 formation pumping of excited vibrational levels of h@xmath1 , followed by collisional de - excitation of these levels which leads to gas heating ( e.g. , tielens & hollenbach 1985 ) .
the essential point is that the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 columns are not provided by cosmic ray ionization , and therefore can not diagnose the cosmic ray ionization rate , except to give an upper limit .
figure 8 shows the contours of the integrated ( from @xmath179 to @xmath180 ) columns of oh@xmath0 ions for a primary cosmic ray ionization rate of @xmath7=@xmath146 s@xmath15 per h atom , and plotted as functions of @xmath8 and @xmath37 .
note that the oh@xmath0 columns include both the first and second peaks .
we emphasize that these are columns perpendicular to the face of the pdr slab .
if clouds are observed obliquely , proportionately more column will be in the line of sight .
similarly , if we are observing a slab illuminated on both sides , the columns will be raised by a factor of 2 if the slab is quite optically thick ( @xmath181 ) so that first and second peaks occur on both sides .
the upper right hand portion of the figure is blacked out because radiation pressure , photoelectric emission , and photodesorption forces on dust grains , when @xmath182 , drives dust rapidly through the pdr ( weingartner & draine 2001b ) .
such high ratios of @xmath12 rarely occur in nature , and require a much more detailed pdr code .
figure 8 shows two main features .
first , at low @xmath5 or low @xmath183 @xmath49 , the columns of oh@xmath0 are roughly proportional to @xmath172 and almost independent of @xmath37 , as discussed above .
there is a weak dependence on @xmath37 for @xmath5 because higher @xmath37 leads to higher gas @xmath20 , and the o@xmath0 abundance rises with exp(-230 k/@xmath20 ) .
from this figure it is clear that since oh@xmath0 columns of at least @xmath184 @xmath11 are needed to make absorption observations feasible , low density ( @xmath185 @xmath49 ) clouds should be targeted if @xmath7= @xmath146 s@xmath15 .
it should also be noted , as is obvious in figures 2 , 3 and 5 , that both the first and second peaks contribute to the oh@xmath0 column so that there is substantial column at both @xmath186 and @xmath187 . the oh@xmath0 in this case is quite cold , @xmath188 k. second
, these relations completely break down in the upper right portion ( @xmath189 @xmath49 and @xmath190 ) of this contour plot . here , as discussed above , the high @xmath37 and @xmath8 ( high density brings the h / h@xmath1 interface closer to the surface , where the gas is warm ) lead to very warm ( @xmath191 k ) h@xmath1 near the surface ( @xmath22 ) of the cloud .
this warm h@xmath1 drives reactions which produce oh@xmath0 without the need for cosmic ray ionization . in this way
, observable columns of oh@xmath0 can be formed at these elevated values of @xmath37 and @xmath8 , and the columns occur at moderate @xmath192 . the oh@xmath0 in this case is quite warm , @xmath193 k , and its column is independent of @xmath7 , but depends on @xmath8 and @xmath37 .
figure 9 shows the contours of the integrated ( from @xmath179 to @xmath180 ) columns of h@xmath1o@xmath0 ions for a primary cosmic ray ionization rate of @xmath7=@xmath146 s@xmath15 per h atom .
this figure can be compared with figure 8 , which treats the same cosmic ray case for oh@xmath0 .
we focus on regions where these ions might be detectable , that is , for columns @xmath194 @xmath11 , and in the cosmic ray dominated zones of @xmath8 and @xmath37 . in these low density regions ,
comparison of figure 9 with 8 reveals that the column ratios are of order unity , as might be expected .
the formation rates of these two ions are nearly the same ( stemming from the cosmic ray ionization rate ) and the destruction rates of these two molecules are nearly the same ( via h@xmath1 in the regions where most of the columns are generated ) .
note that in figures 2 , 3 , and 5 the local oh@xmath0 abundance peaks at lower values of @xmath13 than the h@xmath1o@xmath0 abundance , and that figures 8 and 9 present integrated columns through a high @xmath13 slab . therefore , we can obtain higher column ratios of these two molecular ions if we truncate our slabs to small diffuse clouds with @xmath195 ( see below ,
3.2.2 ) .
we do not provide a contour plot similar to figures 8 and 9 but with different @xmath7 because the results are simple to describe .
for @xmath5 the oh@xmath0 and h@xmath1o@xmath0 columns scale as @xmath7 .
for @xmath196 the columns are independent of @xmath7 because cosmic rays do not produce the oh@xmath0 and h@xmath1o@xmath0 . forcing a fit with the expected ( to first order ) @xmath7/@xmath8 dependence in the cosmic ray - dominated region @xmath197 , we find roughly ( factor of 2 ) the simple expressions @xmath198 and @xmath199 we emphasize that these columns are the total oh@xmath0 and h@xmath1o@xmath0 columns in both the first and second peaks .
we do not present integrated ( from @xmath179 to @xmath180 ) columns of h@xmath2o@xmath0 ions because , as discussed above , most of their column arises at high @xmath200 and this second peak column is quite constant as a function of @xmath8 , @xmath37 , and @xmath7 if pahs are present at the same abundances as in diffuse clouds . with pahs present in the second peak , we find that the h@xmath2o@xmath0 column is slightly dependent on the gas density , varying from @xmath201 @xmath11 at @xmath202 @xmath49 to @xmath203 @xmath11 at @xmath204 @xmath49 . at densities of @xmath169 @xmath49 and with @xmath205 and @xmath7=@xmath146 s@xmath15 ,
we remind the reader ( see discussion of figure 3 ) that @xmath9(h@xmath2o@xmath0)@xmath206 @xmath11 in the first peak ; like oh@xmath0 and h@xmath1o@xmath0 the column of h@xmath2o@xmath0 in the first peak scales with @xmath7/@xmath8 so that even lower densities or higher @xmath7 are needed to make h@xmath2o@xmath0 detectable in the first peak ( e.g. , in a diffuse foreground cloud ) . pahs may not have a high abundance deep in the high @xmath13 regions of a molecular cloud , perhaps due to coagulation of pahs on larger grains ( see , for example , discussion in h09 ) .
we have run a number of cases with the pah abundance set to zero to see the effect on the second peaks . in these runs
, we also reduce the number of small grains by setting the minimum grain size to 100 , but we still refer below to this case as the `` no pah '' case for high @xmath13 cloud interiors .
the results depend on the gas phase abundance of metals , like s , si , mg , and fe ; see table 2 of appendix a for our assumed gas phase abundances of these species in the interstellar radiation field .
our code follows the freezeout of these species as a function of depth into the cloud , as photodesorption no longer is able to keep the metal atoms off the grain surfaces .
the photodesorption yields of these species are not known .
we find that for yields greater than 10@xmath207 , there are still sufficient gas phase metals present at the second peak to supply electrons and suppress the h@xmath2o@xmath0 abundance by factors of @xmath208 relative to the case with pahs .
however , if the photodesorption yields are very low so that the gas phase abundances of these metals at the second peak are @xmath209 , then we obtain h@xmath2o@xmath0 columns in the second peak comparable to the case with pahs .
for example , assuming @xmath210 and @xmath211 , we obtain @xmath9(h@xmath2o@xmath0)@xmath212 @xmath11 if @xmath7/@xmath213 s@xmath15 per h atom , and @xmath9(h@xmath2o@xmath0)@xmath214 @xmath11 if @xmath7/@xmath215 s@xmath15 per h atom . the main difference between the pah and no pah case lies in the electron loss mechanisms : with pahs the electrons are lost by encountering pahs ; with no pahs the electrons are lost by recombining with ions .
the atomic ions have slower rates of recombination than the molecular ions , so the presence of metal atomic ions leads to higher electron abundances and lower h@xmath2o@xmath0 columns .
the oh@xmath0 and h@xmath1o@xmath0 abundances and columns also decrease in the second peak , but by smaller factors ( @xmath216 ) , because their destruction is not by electrons but by h@xmath1 , whose abundance does not change . in appendix c
we discuss the lack of sensitivity of the results on certain key reaction rate coefficients , the assumed abundance of gas phase oxygen , and on the rate coefficient for h@xmath1 formation on grain surfaces .
we discuss the effect of _ raising _ the gas phase abundances of metals .
we also treat the increase in the h@xmath2o@xmath0 columns in the second peak if elemental o does not freeze ( as water ice ) at high @xmath13 .
diffuse clouds have relatively small columns , @xmath3 , whereas translucent clouds have columns intermediate between diffuse clouds and gmcs , @xmath217 .
although gmcs have sufficient column to incorporate both peaks of the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 abundances , diffuse and translucent clouds typically ( with the possible exception of the highest @xmath13 translucent clouds ) only contain the surface peak .
in fact , a diffuse cloud may have such small @xmath13 that the cloud may truncate the full peak of one or all of the ions .
therefore , the parameter @xmath27 , the total @xmath13 through the cloud becomes an important new parameter in determining , for example , ratios of the columns of the ions .
both diffuse and translucent clouds have insufficient column to be gravitationally bound , so generally they are in thermal pressure equilibrium with the interstellar medium .
the local typical value of the thermal pressure is @xmath218 @xmath49 k , and this pressure may rise by factors of 2 - 3 for clouds in the molecular ring of the galaxy ( wolfire et al 2003 ) . as discussed above ( 2.2 ) , depending on the galactocentric radius @xmath94 , they have densities 30 @xmath219 @xmath49 , temperatures @xmath220 k , and incident fuv fluxes characterized by 1@xmath221 . the ratio @xmath222 , critical to the photochemistry , varies typically from about 3 - 6 , although we explore a somewhat larger range 3 - 10 here . in the low density regime , @xmath223 @xmath49 ,
appropriate for diffuse and translucent clouds , there is a scaling of the results of thermochemical models .
the parameters which control the thermochemical structure of a cloud are just @xmath12 , @xmath7/@xmath8 , and @xmath27 .
we note that this scaling does not apply to the denser pdr models discussed in the previous section . at high density , @xmath224 @xmath49 ,
the cooling of the gas is suppressed by collisional de - excitation of the fine structure states of c@xmath0 and o. therefore , higher density gas will typically give higher gas temperatures , even when @xmath12 is held constant .
however , given that this scaling applies for diffuse and translucent clouds , we generally in this section plot our results as functions of @xmath12 , @xmath7/@xmath8 , and @xmath27 .
we note that this implies that the observations may give us a measure of @xmath7/@xmath8 , but to get an estimate of @xmath7 we will then have to estimate the gas density @xmath8 in the region observed .
figure 10 plots the column of oh@xmath0 , @xmath9(oh@xmath0 ) , as a function of @xmath27 for four values of @xmath7/@xmath225 ( recall , @xmath226 @xmath49 ) and for @xmath227 .
the dotted lines plot @xmath228h@xmath125 , the abundance of h@xmath1 at cloud center , and these values appear on the right of the figure .
recall @xmath27 is the total @xmath13 through the diffuse slab , measured perpendicular to the slab and the ion columns plotted are perpendicular to the slab .
figure 10 shows that @xmath9(oh@xmath0 ) increases with @xmath7/@xmath8 and with @xmath27 , although one sees a saturation of the column with @xmath27 at @xmath229 for the lower three cases of @xmath7/@xmath8 .
this saturation is due to the peaking of the local oh@xmath0 abundance at lower values of @xmath13 .
figure 10 shows that @xmath9(oh@xmath0 ) correlates with @xmath228h@xmath125 , and that moderate molecular hydrogen abundances @xmath228h@xmath230 are required to obtain substantial columns of oh@xmath0 .
observed oh@xmath0 columns in broad ( @xmath231 km s@xmath15 , see 4.1.1 ) velocity components in the ism are of order @xmath232 @xmath11 . with this value of @xmath233
, we see that it requires very high cosmic ray ionization rates , @xmath7/@xmath234 s@xmath15 and high @xmath235 for a single cloud , seen face on , to produce these columns .
therefore , it appears that geometrical effects such as seeing several clouds along the line of sight , or viewing the cloud obliquely , may be required to match observation .
in fact , as we discuss in 4.1.1 , many clouds may to required to cover the broad ( @xmath231 km s@xmath15 ) absorption components since single clouds would have much narrower velocity widths .
because of geometrical effects and variation in @xmath236 and @xmath12 , the observation of the column of oh@xmath0 is not sufficient to tightly restrict @xmath7/@xmath8 .
figure 11 plots @xmath9(oh@xmath0 ) as a function of @xmath27 for four values of @xmath7/@xmath225 and for @xmath237 : in other words , the same as figure 10 but with a ratio of @xmath12 that is 3.16 times higher . the main effect of raising @xmath12 is to push the h / h@xmath1 transition deeper into the cloud to higher @xmath13 .
this is seen in the plot of @xmath228h@xmath125 which rises to 0.01 at @xmath238 in this case , compared to @xmath239 in the previous case of @xmath227 . since h@xmath1 is required to make oh@xmath0 , this moves the peak of oh@xmath0 to higher values of @xmath13 , and thus pushes the rise of @xmath9(oh@xmath0 ) to higher values of @xmath27 than in figure 10
. however , this first peak has more column of oh@xmath0 compared with the case @xmath227 .
therefore , to produce the observed columns of oh@xmath0 now requires only @xmath7/@xmath240 s@xmath15 and high @xmath241 for a single cloud , seen face on .
the results on the total column of oh@xmath0 suggests that other observations must be brought into play to further restrict the cosmic ray ionization rate .
since h@xmath1o@xmath0 is often observed as well , we first consider the ratio @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) . as seen in figures 2 , 3 , and 5 , the h@xmath1o@xmath0 abundance peaks at higher values of @xmath13 than the oh@xmath0 abundance .
therefore , the ratio of the columns of these two ions may give us some measure of @xmath27 .
figure 12 plots the ratio @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) as a function of @xmath27 for four values of @xmath7/@xmath225 and for @xmath233 . as expected , at low values of @xmath243 , the ratio @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) is high , @xmath244 , as we have not yet reached high enough values of @xmath13 to include the h@xmath1o@xmath0 peak .
however , as @xmath27 increases , the ratio of the columns continues to drop until at @xmath245 , the ratio is approximately 2 - 3 , except in the case of the highest cosmic ray rate @xmath7/@xmath246 s@xmath15./@xmath247 s@xmath15 case is extreme ; cosmic ray ionization of h produces an electron abundance of @xmath248 . in this case , electrons and not h@xmath1 dominate the destruction of oh@xmath0 and h@xmath1o@xmath0 even when @xmath50h@xmath249 .
therefore , the fraction of cosmic ray ionizations that produce oh@xmath0 is higher than the fraction that produce h@xmath1o@xmath0 , even at the peaks of their abundance . hence we have much more oh@xmath0 than h@xmath1o@xmath0as seen in this figure , and , once we reach high enough @xmath27 where the ion chemistry shown in figure 1(top ) dominates , the ratio rises as a function of @xmath27 .
this holds as well in x - ray ionized regions with high ionization rates .
] once we have incorporated both the oh@xmath0 and the h@xmath1o@xmath0 abundance peaks , the column ratios are near unity , as noted in 3.1 .
figure 13 plots the ratio @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) as a function of @xmath27 for four values of @xmath7/@xmath225 and for @xmath237 : in other words , the same cases as figure 12 except the ratio @xmath12 is raised by a factor of 3.16 .
the rise in @xmath12 pushes the h / h@xmath1 transition and the peaks of oh@xmath0 and h@xmath1o@xmath0 to higher values of @xmath13 .
therefore , compared to figure 12 , we see the drop in the @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) ratio at higher values of @xmath27
. the observed ratios of @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) range from 3 - 15 .
if @xmath227 and if we assume that @xmath7/@xmath250 s@xmath15 , figure 12 suggests that @xmath251 to obtain these ratios .
however , if this is true , figure 10 implies that even with as high a value of @xmath7/@xmath252 s@xmath15 , we will need a geometrical factor " of @xmath32 in order to obtain @xmath9(oh@xmath0)@xmath253 @xmath11 .
this geometrical factor is the combination of many clouds along the line of sight , along with the enhancement in column due to viewing angle of the cloud .
the situation changes with @xmath237 , as seen in figure 13 . here , again assuming that @xmath7/@xmath250 s@xmath15 , the observed ratios can be obtained with @xmath254 . in this case , using figure 11 , the geometrical factor needs to be only @xmath255 to produce oh@xmath0 columns of @xmath201 @xmath11 when @xmath7/@xmath256 s@xmath15 .
observations of hi 21 cm are often available along the same lines of sight as the oh@xmath0 and h@xmath1o@xmath0 measurements . in this case
, the column of hi , @xmath9(h ) , along the line of sight associated with the velocity feature of the ions can be estimated . to be precise , the observations directly give @xmath9(h)/@xmath20 , where @xmath20 is an average temperature along the line of sight . in figures 14 and 15
we plot the ratio @xmath9(oh@xmath0)/[@xmath9(h)/@xmath257 on the vertical axis ( @xmath258 k ) and the ratio @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) on the horizontal axis for two fixed values of @xmath233 and @xmath259 .
the results of our 2 sided diffuse cloud models are shown as contours on this figure .
as noted above , higher average abundances of oh@xmath0 ( or @xmath9(oh@xmath0)/@xmath9(h ) ) requires higher @xmath7/@xmath8 . on the other hand ,
higher @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) requires either low @xmath27 or high @xmath12 . in the next section
we apply figures 10 , 11 , 14 , and 15 to the observational data to constrain @xmath7/@xmath8 along sightlines to w49n .
we have run additional models to estimate how turbulence might affect the production of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 .
it has been recognized for several decades that standard astrochemical models for diffuse molecular clouds greatly underpredict the observed column densities ( @xmath260 @xmath11 ) of ch@xmath0 ( e.g. , falgarone et al 2010a , b ; and references therein . )
the models presented here are no exception ; in our models with @xmath39 , @xmath202 @xmath49 , and @xmath261 , for example , we obtain a predicted ch@xmath0 column density @xmath262 @xmath11 , more than two orders of magnitude lower than that typically observed .
both shock waves with large scale order ( elitzur & watson 1978 , draine & katz 1986 , flower & pineau des forets 1998 ) , and interstellar turbulence ( e.g. joulain et al . 1998 ; lesaffre et al 2007 , godard et al 2009 , falgarone et al 2010a , b ) have been proposed as sources of heating that could enhance the ch@xmath0 abundance within some small fraction of the cloud volume ; in these models , ch@xmath0 is formed by the endothermic reaction of c@xmath0 with h@xmath1 .
it has also been long recognized ( e.g. draine & katz 1986 ) that models invoking ion - neutral drift either in c - type shock waves or in turbulent dissipation regions are most successful in simultaneously matching the observed column densities of ch@xmath0 and oh . in particular , while models in which the gas temperature is merely elevated without significant ion - neutral drift tend to overpredict the oh / ch@xmath0 ratio ( the neutral - neutral endothermic reaction @xmath263 being enhanced along with the reaction of c@xmath0 with h@xmath1 ) , models incorporating ion - neutral drift preferentially enhance the endothermic ion - molecule reactions that produce ch@xmath0 ( and sh@xmath0 ; godard et al . 2012 ) , _ without _ producing more oh than the observations permit .
we have crudely estimated the effects of turbulence by positing an enhanced rate of endothermic ion - neutral reactions in a small fraction of the cloud volume .
motivated by recent calculations performed by myers & mckee ( private communication ) , we adopted enhanced ion - neutral reaction rates - equal to the thermal rate coefficients at 1000 k - in 3% of the cloud volume : this leads to a predicted ch@xmath0 column in accord with what is typically observed .. ] we found that turbulence does not greatly enhance the oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 column densities through such a cloud .
there was a modest enhancement of these oxygen hydride ions within the region of ion - neutral drift , due to the faster reaction of h@xmath0 with o , but since that region occupies only a small fraction of the total volume , this effect did not affect the total column densities significantly . therefore , we conclude that turbulent dissipation or shocks in small regions of the clouds are unlikely to affect our conclusions concerning oh@xmath0 , h@xmath1o@xmath0 or h@xmath2o@xmath0 in the models presented here .
as discussed in 1 , recent _ herschel _ observations of bright galactic continuum sources have revealed oh@xmath0 and h@xmath1o@xmath0 absorption features arising in multiple foreground clouds along the targeted sight - lines ( gerin et al .
2010 ; n10 ; neufeld et al . 2011 ) .
the three crosses in figure 14 denote the @xmath264 and @xmath265 $ ] ratios for the two such sources for which results have been reported to date : g10.60.4 ( a.k.a .
w31c ) ( the lower limits ) and w49n ( 2 points to the right ) .
those given for g10.60.4 apply to all the material along the sight - line , while those given for w49n apply to two separate velocity intervals ( 30 - 50 km s@xmath15 and 50 - 78 km s@xmath15 ) .
the oh@xmath0 absorption in g10.60.4 is largely saturated , and thus the plotted values are lower limits . for w49n ,
the columns of h derived assumed @xmath266 k or @xmath267 ; the quoted values of @xmath9(h)/@xmath268 are @xmath269 @xmath11 for the @xmath270 km s@xmath15 feature and @xmath271 @xmath11 for the @xmath272 km s@xmath15 feature .
the total columns @xmath9(oh@xmath0 ) observed in these two velocity intervals are @xmath273 and @xmath274 @xmath11 , respectively .
the @xmath264 ratios measured toward w49n ( @xmath32 ) and the wide velocity range of the absorbing clouds suggest that the absorbing material is comprised of multiple clouds of small extinction .
the line of sight to w49n is approximately 11 kpc long , and passes as close as @xmath102 kpc to the galactic center . assuming @xmath233 , models with @xmath275 and @xmath276 best account for all the data ( figure 14 ) , with cosmic - ray ionization rates of @xmath277 and @xmath278 , respectively , for the @xmath279 km s@xmath15 and @xmath280 km s@xmath15 velocity intervals .
as seen in figure 10 for @xmath233 , the column of oh@xmath0 in a single cloud of size @xmath281 is @xmath9(oh@xmath0)@xmath282 @xmath11 and for @xmath283 is @xmath9(oh@xmath0)@xmath284 @xmath11 .
therefore , we either need @xmath285 clouds along the line of sight , or a smaller number but with geometrical enhancement effects .
note that a single cloud might have an oh@xmath0 absorption linewidth of only @xmath255 km s@xmath15 .
therefore , to cover the broad absorption features we are modeling ( @xmath286 or 28 km s@xmath15 ) , we need 10 - 30 clouds spread out along the line of sight so that their galactic rotational velocities coupled with their individual line widths cover this velocity range . therefore , this @xmath233 model has barely enough clouds to produce the relatively smooth absorption feature observed . assuming @xmath287 , models with @xmath288 and @xmath289 and with @xmath7/@xmath290 and @xmath291 s@xmath15 , respectively , best account for the data ( see figure 15 )
. with these values of @xmath27 and @xmath7/@xmath225 , inspection of figure 11 shows that we require @xmath292 clouds along the line of sight .
this value of @xmath93 has even fewer clouds to cover the broad absorption line observed .
moreover , the temperatures in clouds with @xmath287 are higher ( @xmath293 k ) than is typically observed near the solar neighborhood ( heiles & troland 2003 ) . and finally , the wolfire et al ( 2003 ) results suggest that along the line of sight to w49n , the average @xmath294 .
a similar modeling of the case @xmath295 , not shown in the figures , gives @xmath7/@xmath296 and @xmath297 s@xmath15 with @xmath298 and 0.08 , respectively .
again , because of the high cosmic ray rates , the gas temperatures are high in these clouds ( @xmath299 k ) . in addition , these cosmic ray rates seem improbably high .
this model is driven to high cosmic ray rates in part because otherwise ( with low @xmath93 ) the gas is cold , which reduces the rate that h@xmath0 can charge exchange with o. therefore , we finally conclude that our models suggest typical @xmath238 , @xmath300 , and @xmath7/@xmath301 s@xmath15 for the @xmath270 km s@xmath15 component , and @xmath7/@xmath302 s@xmath15 for the @xmath272 km s@xmath15 component . a better approach is to ask , what distribution of cloud @xmath27 or hydrogen nucleus column @xmath9 through the cloud will we encounter in traversing the @xmath303 kpc to w49n ? " hi 21 cm observational data suggest that @xmath304 , where @xmath305 is the number of clouds for @xmath306 @xmath11 , and proportional to @xmath307 for @xmath308 @xmath11 ( heiles & troland 2005 ) . is the column of h nuclei ; however , in the above range , generally @xmath309(h ) .
] we have crudely integrated a distribution of clouds with this dependence , with the constant of proportionality derived by requiring the integral to obtain a total column of @xmath9(h)@xmath310 @xmath11 , the observed atomic h column in each velocity component of w49n .
we use @xmath311 @xmath11 as an upper limit for the integration ; above this , we enter small number statistics since the total column is of this order .
the lower limit does not enter the integration significantly , as long as it is @xmath312 @xmath11 , since these clouds contain very little of the total column of any species .
the integration removes @xmath27 as a variable , and leaves only the parameters @xmath7/@xmath225 and @xmath93 as free parameters to match the observed oh@xmath0/h@xmath1o@xmath0 ratio and the total column of oh@xmath0 .
we find that with @xmath233 and @xmath7/@xmath313 s@xmath15 , we obtain @xmath9(oh@xmath0)@xmath314 @xmath11 and @xmath9(h@xmath1o@xmath0)@xmath315 @xmath11 .
these values match the velocity interval @xmath272 km s@xmath15 extremely well . increasing the cosmic ray rate to @xmath7/@xmath316 s@xmath15
, we obtain @xmath9(oh@xmath0)@xmath317 @xmath11 and @xmath9(h@xmath1o@xmath0)@xmath318 @xmath11 , a very good fit to the lower velocity component . considering the small number statistics , especially of the higher @xmath27 clouds ,
this is very good agreement .
interestingly , the main contribution to @xmath9(oh@xmath0 ) comes from near @xmath319 .
this is because of the dependence of @xmath9(oh@xmath0 ) on @xmath27 ( see figure 10 ) .
note that @xmath9(oh@xmath0 ) rises sharply with @xmath27 until about @xmath320 , and then it levels off .
therefore , the integral is weighted to that size cloud in the range of the turnover ( @xmath321 ) , which has the peak oh@xmath0 abundance ( averaged through the cloud ) .
in addition , the cloud distribution steepens when @xmath322 , thereby decreasing the contribution from higher @xmath27 clouds .
the total number of clouds in the range @xmath323 is approximately 20 .
we note that the h@xmath1o@xmath0 columns have greater contribution from higher @xmath27 clouds , which are less numerous , and therefore we predict the h@xmath1o@xmath0 absorption feature to have greater fluctuations .
we next examine the cases @xmath295 and 10 .
assuming @xmath324 , we find that we need cosmic ray rates of @xmath325 s@xmath15 and @xmath326 s@xmath15 to match the oh@xmath0 columns in the two velocity components , respectively .
however , the oh@xmath0/h@xmath1o@xmath0 column ratio is 12.1 , somewhat higher than observed , and again , the clouds are too warm ( @xmath327 k ) .
the high field pushes the peaks of oh@xmath0 and h@xmath1o@xmath0 deeper into the cloud , and our clouds tend to truncate the h@xmath1o@xmath0 column , leading to the high oh@xmath0/h@xmath1o@xmath0 ratio .
we need much higher cosmic ray rates than our solution above for @xmath287 with clouds of single @xmath328 because the cloud distribution leads to a significant column of h from clouds of low @xmath27 which have very little oh@xmath0 .
assuming @xmath329 , we find that we need cosmic ray rates of @xmath330 s@xmath15 and @xmath331 s@xmath15 to match the oh@xmath0 columns in the two velocity components , respectively .
however , the oh@xmath0/h@xmath1o@xmath0column ratio is 5.6 , lower than observed . with low @xmath93 ,
the clouds tend to contain both oh@xmath0 and h@xmath1o@xmath0 peaks , driving the ratio down .
we conclude that the best fit is with @xmath233 , giving @xmath7/@xmath332 s@xmath15 and @xmath333 s@xmath15 for the two velocity components , in good agreement with our analysis above that did not use the cloud distribution .
however , the cloud distribution gives added weight to the conclusion that @xmath334 .
the cosmic ray rates derived from our models are roughly 2 - 4 times larger than those inferred by n10 , who used a simple analytic treatment to infer cosmic ray ionization rates of @xmath335 and @xmath336 , where @xmath160 is the fraction of cosmic - ray ionizations that lead to oh@xmath0 . using pure gas - phase model results from the meudon pdr code ( le petit et al .
2006 ) , n10 found that @xmath160 lay in the range 0.5 1.0 for a wide variety of cloud conditions ; this would imply cosmic ray ionization rates in the range @xmath337 for the material along the sight - line .
the main source of the difference between our models and those of n10 is our inclusion of pah@xmath21
h@xmath0 recombination and the charge exchange of h@xmath0 with neutral pah in the present work , processes previously omitted that reduce the fraction of cosmic - ray ionizations that lead to oh@xmath0 production ( in our models @xmath338 in the regions where most of the oh@xmath0 lies )
. figures 16 and 17 repeat figures 14 and 15 except that the rates of pah and pah@xmath21 neutralization of h@xmath0 have been reduced by @xmath339 , so that they are negligible compared to the radiative recombination of h@xmath0 with electrons ( see appendix b ) .
note that this reduction could be achieved by either lower pah abundances or lower pah rate coefficients with h@xmath0 , or a combination of both .
another possibility is increased photoionization rates of pah@xmath21 and pah .
figure 16 shows that the best fit to the data for @xmath340 is @xmath341 and @xmath7/@xmath342 s@xmath15 , in complete agreement with n10 .
therefore , the pah chemistry is extremely important in deriving cosmic ray rates from the oh@xmath0 and h@xmath1o@xmath0 ions .
we consider the rates from figures 16 and 17 lower limits to the cosmic ray rates , whereas the rates from figures 14 and 15 could be considered upper limits , although they represent our best estimates of pah chemistry at this time .
the cosmic - ray ionization rates suggested by the predictions shown in figure 14 are also a factor @xmath343 larger than those typically inferred from observations of h@xmath24 in the galactic disk ( indriolo et al .
2007 , indriolo & mccall 2012 ) .
indriolo & mccall conclude , however , that cosmic ray rates vary by more than an order of magnitude over various sight lines , from @xmath7@xmath344 s@xmath15 to @xmath345 s@xmath15 , with a mean of @xmath346 s@xmath15 ( note that we have converted their total ionization rate per h@xmath1 to our primary ionization rate per h nucleus by dividing their values by @xmath347 ; we also note that their typical density was @xmath348 @xmath49 so their results imply @xmath7/@xmath225 that are two times smaller than these values .
their highest values of @xmath7/@xmath225 are still a factor of @xmath349 below what we find for w49n , so there appears to be a small discrepancy .
we point out that the cosmic ray rate derived from h@xmath24 observations is directly proportional to the assumed electron abundance .
indriolo & mccall assumed that the electrons were supplied by c@xmath0 , with an abundance of @xmath350 .
however , in our models of diffuse clouds , where typically the gas density is @xmath351 @xmath49 , the cosmic ray ionization of h can produce comparable or greater abundances of h@xmath0 than c@xmath0 .
therefore , we find for @xmath7/@xmath352 s@xmath15 , for example , that @xmath353 when @xmath130(h@xmath1)@xmath354 and @xmath355 when @xmath130(h@xmath1)@xmath356 .
this implies that the indriolo et al rates may need to be increased , by factors as much as @xmath349 , depending on @xmath7/@xmath225 .
given the small number of sources toward which @xmath357 and @xmath358 ratios have so far been reported , the proven variation in cosmic ray rates along different lines of sight , the lack of a common sight line to compare the cosmic ray rates derived by h@xmath24 versus by oh@xmath0and h@xmath1o@xmath0 , the possibility of higher electron abundances than assumed by indriolo and coworkers , and the lack of certainty in pah chemistry , it is not yet possible to draw firm conclusions about whether our results differ significantly from that of indriolo et al .
what is clear is that cosmic ray rates in the diffuse ism are higher than was previously thought .
_ herschel _ has also detected absorption by @xmath30 in foreground material along the sight - line to g10.6 0.4 ( gerin et al .
the _ average _ @xmath30/@xmath29 ratio along the sight - line is 0.7 , but the distribution of @xmath30 is clearly different from that of h@xmath1o@xmath0 and oh@xmath0 , with the @xmath30 concentrated within a single narrow velocity component where the h@xmath1 fraction is presumably largest .
this cloud is unassociated with the source , but is probably the foreground cloud with the largest @xmath13 along the line of sight .
it appears in the hf spectrum ( neufeld et al . 2010 ) and co spectrum ( c. vastel , private communication ; see also corbel & eikenberry 2004 ) , suggesting a high molecular fraction and therefore substantial @xmath13 .
the h@xmath2o@xmath0 column in this narrow component is @xmath359 @xmath11 , consistent with our predictions for the second peak in a high @xmath13 cloud .
pvdk92 searched for the 396 , 364 , and 307 ghz inversion - rotation lines of h@xmath2o@xmath0 in 20 galactic targets including gmcs , star forming regions , galactic center sources and a few evolved stars and reported definitive detections of h@xmath2o@xmath0 emission from the g34.3 + 0.15 hot core , two positions in sgr b2 ( a galactic center gmc ) , and the w3 irs5 , w3(oh ) , w51 m and orion kl high - mass star forming regions . in the case of orion and sgr b2
, h@xmath2o@xmath0 emissions had been previously discovered by wootten et al .
the h@xmath2o@xmath0 column densities derived by pvdk92 for the high - mass cores w3 irs5 , w3(oh ) , w51 m are fairly similar , ranging from a few @xmath360 to a few @xmath361 @xmath11 , and arguing for a similar environment and formation mechanism . given the large critical densities for the observed transitions , pvdk92 inferred relatively high densities ( @xmath362 @xmath49 ) and temperatures ( @xmath363 k ) for the h@xmath2o@xmath0emitting gas .
the h@xmath2o@xmath0 column densities observed in these high - mass cores are in good agreement with the predictions of our models for either the case in which pahs and/or very small grains are assumed to be present or the case where pahs and low ionization potential metals are highly depletedo@xmath0 column densities that are roughly an order of magnitude lower than the values inferred from the above observations , unless low ionization potential metal atoms are highly depleted at the second peak of h@xmath2o@xmath0 . ] ) ; the pah case predicts h@xmath2o@xmath0
column densities of @xmath364 @xmath11 , regardless of density @xmath8 or @xmath365 .
herschel s hifi instrument has allowed the detection of far - infrared and submillimeter h@xmath2o@xmath0 lines that are not accessible to ground - based observatories .
several h@xmath2o@xmath0 emission lines in the thz domain have been reported towards high - mass ysos in the w3 irs5 region ( benz et al .
a rotation diagram combining hifi data with lower frequency data from pvdk92 indicates a rotational temperature @xmath366 k , suggesting that the observed emission may arise either from hot and dense gas or is radiatively excited by continuum photons from hot dust grains .
benz et al . ( 2010 ) inferred @xmath367h@xmath2o@xmath0@xmath368 @xmath11 and interpreted the h@xmath2o@xmath0 emission as arising from the outflow walls heated and irradiated by the fuv radiation field from massive ysos .
their inferred h@xmath2o@xmath0 column densities and rotational temperatures are consistent with our models of dense gas illuminated by a relatively strong fuv field ( where large kinetic temperatures are achieved at relatively low @xmath369 ) . for sources where a strong infrared radiation field is present , submillimeter h@xmath2o@xmath0 line emission can be accompanied by far - infrared absorption within an extended envelope .
the h@xmath2o@xmath0 j@xmath370=@xmath371 and @xmath372 inversion - rotation and @xmath373 pure - inversion fir absorption lines at 100.577 , 100.869 and 181.054 @xmath374 m were detected by iso towards the optically thick fir continuum emission of sgr b2 ( goicoechea & cernicharo 2001 ) .
an h@xmath2o@xmath0 column density of @xmath375 @xmath11 was inferred in the warm and low density extended envelope of sgr b2 ( @xmath376 @xmath49 , @xmath377 from [ oi ] and [ cii ] observations ; goicoechea et al .
2004 ) . however , the p - h@xmath2o@xmath0 @xmath378 364 ghz emission line was subsequently mapped around the sgr b2(m ) core at 18 angular resolution with apex by van der tak et al .
( 2006 ) , who inferred column densities @xmath379 @xmath11 on the basis of a detailed excitation model .
this value is much larger than the estimates obtained from fir absorption lines , presumably because the latter only probe the outer envelope of the source whereas the submillimeter observations probe material to a much larger depth .
the h@xmath2o@xmath0 column densities inferred by van der tak et al .
( 2006 ) are much larger than our model predictions for externally - illuminated clouds ; this discrepancy likely reflects the presence of a strong internal luminosity source in sgr b2 that heats the dust grains within the interior and prevents the freeze - out of oxygen nuclei ( see appendix c ) .
thanks to the much improved sensitivity of space- and ground - based receivers and detectors , molecular ions can now be detected outside the milky way .
extragalactic h@xmath2o@xmath0 was first detected through jcmt observations of the p - h@xmath2o@xmath0 364 ghz emission line towards the prototypical ultraluminous infrared galaxy arp 220 , and towards m82 , an evolved starburst ( van der tak et al . 2008 ) .
h@xmath2o@xmath0 was subsequently detected towards ic342 , ngc253 , ngc1068 , ngc4418 and ngc 6240 , and upper limits obtained towards iras 15250 and arp 299 galaxies ( aalto et al . 2011 ) .
except for ic342 and m82 , the typical h@xmath2o@xmath0 column densities ( @xmath380 @xmath11 ) derived from extragalactic h@xmath2o@xmath0 detections are much larger than the predictions of our models for single clouds .
these observations are of h@xmath2o@xmath0 in emission and the authors assume @xmath381 to obtain these columns with error bars of factor of 2 . as in the case of sgr b2
, the discrepancy between our models and the observations may reflect the effect of enhanced dust temperatures that prevent the freeze - out of elemental oxygen .
alternatively , there may be a large number of clouds along the line of sight , for the starburst galaxies m82 and ic342 , however , where h@xmath2o@xmath0 column densities ( @xmath253 @xmath11 ) are inferred from the observations , enhanced dust temperatures or ionization rates are not apparently required .
for example , van der tak et al ( 2008 ) inferred @xmath9(h@xmath2o@xmath0)@xmath382 @xmath11 in m82 from an emission line which contained a combination of a broad ( @xmath383 km s@xmath15 ) and a narrow ( @xmath384 km s@xmath15 ) component .
in addition , a recent _ herschel_/hifi detection ( weiss et al . 2010 ) of the o - h@xmath1o@xmath0 @xmath385 ground - state line in absorption towards m82 revealed @xmath9(h@xmath1o@xmath0)@xmath386 @xmath11 , arising from a @xmath387 km s@xmath15 feature . the large width of the lines suggest numerous clouds in the beam .
overall , these numbers are consistent with our models with the h@xmath1o@xmath0 arising from the first ( and possibly second ) peak of numerous relatively low density clouds , whereas the h@xmath2o@xmath0 arises from the second peaks of many molecular clouds in the beam .
we model the production of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 in interstellar clouds , using a steady state photodissociation region code that treats the freeze - out of gas species , grain surface chemistry , and desorption of ices from grains .
the code includes pahs , which have an important effect on the chemistry . as a function of depth or @xmath13 into a molecular cloud ,
the ions tend to have two peaks , one at low @xmath3 and one at high @xmath200 , the exact value depending on @xmath12 . in the first peak
the ions are created by the cosmic ray ionization of h to h@xmath0 , followed by reactions with o and h@xmath1 ( see figure 1 top ) .
these peaks appear on the surfaces of molecular clouds as well as in diffuse clouds .
pahs can lower the production of the ions here , by neutralizing h@xmath0 and interrupting the reaction chain ( figure 1 top ) . at most ,
they depress the ion abundances and columns by a factor of @xmath45 . in molecular clouds a significant fraction of the column density of oh@xmath0and h@xmath1o@xmath0
is found in the first peak at the surface ( @xmath186 ) of the cloud .
for relatively low values of the incident far ultraviolet flux on the cloud ( @xmath5 ) , the columns of oh@xmath0 and h@xmath1o@xmath0 scale as the cosmic ray ionization rate divided by the gas density .
roughly , the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 in the first peak are @xmath388(@xmath7/@xmath146 s@xmath15 ) ( 100 @xmath49/@xmath8 ) @xmath11 , @xmath389 ( @xmath7/@xmath146 s@xmath15 ) ( 100 @xmath49/@xmath8 ) @xmath11 , and @xmath390(@xmath7/@xmath146 s@xmath15 ) ( 100 @xmath49/@xmath8 ) @xmath11 .
there is a second peak in oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 abundances at larger depths ( @xmath119 ) in molecular clouds , when the second route to oh@xmath0 formation , initiated by the cosmic ray ionization of h@xmath1 becomes dominant ( figure 1 bottom ) . here ,
lower electron abundances enhance the abundances of the ions by lowering the electronic recombination rate of h@xmath24 ( which raises the abundance of h@xmath24 and the formation rates of the three ions ) , and by lowering the electronic recombination of h@xmath2o@xmath0 ( its main destruction path ) . however , even deeper in the cloud , the oxygen freezes out as water ice on the grain surfaces , and the ion abundances fall as their formation rates fall , being starved for gas phase elemental oxygen
. if pahs or vsgs are present at these high values of @xmath200 , the electron abundance at the second peak is controlled by association with neutral pah or vsg . in this case , rather surprisingly , the column of h@xmath2o@xmath0 ( @xmath391 @xmath49 ) in the second peak and the peak abundance ( @xmath173 ) of h@xmath2o@xmath0 is independent of both @xmath7 and @xmath8 . raising the cosmic ray ionization rate increases the production rate of h@xmath2o@xmath0 , but it also increases the destruction rate by electrons by the same amount . if pahs and vsgs are not present at high @xmath13 , the column of h@xmath2o@xmath0 depends mainly on the gas phase abundance of elemental s , si , mg , and fe at the second peak .
if these abundances are low , @xmath209 , then columns comparable to the pah case are obtained . however , for higher abundances of the metals , the h@xmath2o@xmath0 column is reduced by a factor of roughly 10
. the columns of oh@xmath0 and h@xmath1o@xmath0 in the second peaks are usually of order of the columns in the first peaks .
the observed h@xmath2o@xmath0 columns of @xmath391 @xmath11 seen in many dense galactic molecular clouds therefore imply that either pahs or vsgs are present deep in molecular clouds or that the depletion of pahs and vsgs are accompanied by a very significant depletion of gas phase metals .
there are a few observations that imply much higher column ( @xmath379 @xmath11 ) of h@xmath2o@xmath0 , and these can only be explained in the context of our models as arising in very high @xmath13 clouds where high gas phase elemental o abundances are present throughout due to either desorption processes or time dependent effects .
we suspect that the grains in these sources may be so warm , @xmath392 k , that thermal desorption keeps water ice from depleting the oxygen reservoir . in the case of extragalactic sources with very broad velocity features ,
the columns may be increased by the presence of a large number of clouds along the line of sight . for high values of the incident far ultraviolet flux ( @xmath196 ) and high gas densities ( @xmath393 @xmath49 ) , producing warm ( @xmath26 k ) gas with significant h@xmath1 abundance at @xmath369 , chemical reactions initiated by the photoionization of carbon in the mainly atomic surface can form ionized hydrogen , which then leads to the formation of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 . in this case , their columns ( typically , of order @xmath394 @xmath11 ) are not related to the cosmic ray ionization rates .
h@xmath2o@xmath0 emission from w3 irs5 may be an example of such conditions .
we also model diffuse and translucent clouds in the interstellar medium , and show how observations of @xmath9(oh@xmath0)/@xmath9(h ) ( typically @xmath395 ) and @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) ( typically 3 - 15 ) can be used to estimate @xmath7/@xmath8 .
the ratio @xmath9(oh@xmath0)/@xmath9(h ) , which is essentially the average abundance of oh@xmath0 in all the clouds along the line of sight ( within the same absorption velocity component ) , is mainly a measure of @xmath7/@xmath8 .
the ratio of the oh@xmath0 column to the h@xmath1o@xmath0 column in diffuse clouds is mainly dependent on the ratio @xmath12 and @xmath27 ( i.e , the total hydrogen column through a single typical cloud ) .
if @xmath12 is known , or at least constrained to a narrow range such as @xmath396 typical of diffuse clouds in random locations along the line of sight in the galaxy , then observations of @xmath9(oh@xmath0)/@xmath9(h ) , @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) , and @xmath9(oh@xmath0 ) can determine @xmath7/@xmath8 , @xmath27 , and a geometrical factor which is a combination of the number of clouds along the line of sight and the typical angle that the line of sight makes as it passes through each diffuse cloud slab . using the observed distribution of @xmath27 in diffuse clouds
, the models can provide an estimate of @xmath93 .
we discuss the relation of our models to recent observations of oh@xmath0 and h@xmath1o@xmath0 by the herschel space observatory , and the ability to constrain the cosmic ray ionization rate through comparison of observations with these models .
we conclude that the cosmic ray primary ionization rates @xmath7 in the observed foreground diffuse clouds towards w49n have values of approximately @xmath7@xmath397 s@xmath15 , if our adopted pah chemistry is correct .
we find a hard lower limit of @xmath7/@xmath398 s@xmath15 by neglecting pah chemistry in w49n .
our best fit models suggest that @xmath334 in the diffuse clouds towards w49n .
our w49 models suggest that , in terms of producing oh@xmath0 and h@xmath1o@xmath0 column , the typical @xmath238 through a single cloud towards w49n . and that a diffuse cloud @xmath27 distribution measured by heiles & troland ( 2005 ) fits the data very well .
to produce the total column of oh@xmath0 and h observed in the two velocity components requires @xmath399 clouds in each component along the line of sight .
we discuss differences between our estimates and those of n10 and indriolo et al ( 2007 ) , pointing out the former neglected pah chemistry and the latter may have slightly underestimated the electron abundance in the diffuse foreground clouds with the highest cosmic ray ionization rates .
we look forward to further observations of oh@xmath0 and h@xmath1o@xmath0 along many sight lines to probe the cosmic ray ionization rates throughout the galaxy .
we would like to thank a. benz , j.h .
black , e. falgarone , b. godard , c. heiles , v. ossenkopf , and e. van dishoeck for many useful discussions .
we especially thank m. gerin , n. indriolo , and the anonymous referee for careful readings of the manuscript and many useful suggestions .
we gratefully acknowledge the support of a grant from the nasa herschel science center s theoretical research / laboratory astrophysics program .
partial support for mgw , mjk and djh was provided by a nasa long term space astrophysics grant nng05g46 g .
jrg is supported by a ramn y cajal research contract and thanks the spanish micinn for funding support through grants aya2009 - 07304 and csd2009 - 00038 .
table 1 presents some of the key reactions and/or some of the reaction rates that have changed since h09 .
we emphasize that this is not a complete listing of the @xmath43 chemical reactions in the pdr code described in h09 .
the pah rates are scaled by the factor @xmath400 , a scaling factor introduced in wolfire et al ( 2008 ) that incorporates the uncertainties in pah reaction rates , pah sizes , and pah abundances .
wolfire et al ( 2008 ) found from a semi - empirical fit of our pdr models to c , c@xmath0 , h , and h@xmath1 column densities in diffuse clouds that @xmath401 , and we have adopted that value in all of our pdr models .
table 2 presents the standard gas phase abundances and grain properties that we have adopted in most of our pdr models .
as discussed in the text , we have also run models with these values changed to test the sensitivity of the results to the standard values .
in particular , we have run models with @xmath130(pah)= 0 ; with @xmath130(si ) , @xmath130(fe ) , @xmath130(s ) , and @xmath130(mg ) all set to @xmath402 or @xmath403 ; and with @xmath130(o ) = @xmath404 .
ll @xmath405 & @xmath406 @xmath407 + @xmath408 & @xmath409 @xmath407 + @xmath410 & @xmath411 @xmath407 + @xmath412 & @xmath413 @xmath407 + @xmath414 & @xmath415 @xmath407 + @xmath416 & @xmath417 @xmath407 + @xmath418 & @xmath419 @xmath420 + @xmath421 & @xmath422 @xmath420 + @xmath418 & @xmath423 @xmath420 + @xmath421 & @xmath424 @xmath420 + @xmath425 & @xmath426 @xmath420 + @xmath427 & @xmath428 @xmath420 + @xmath429 & @xmath430 @xmath420 + @xmath431 & @xmath432 @xmath420 + @xmath433 & @xmath434 @xmath420 + @xmath435 & @xmath436 @xmath420 + @xmath437 & @xmath438 @xmath420 + @xmath439 & @xmath440 @xmath420 + @xmath441 & @xmath365 @xmath420 + @xmath442 & @xmath443 @xmath420 + @xmath444 & @xmath445 @xmath407 + @xmath446 & @xmath447 @xmath407 + @xmath448 & @xmath449 @xmath407 + @xmath450 & @xmath449 @xmath407 + @xmath451 & @xmath452 @xmath407 + @xmath453 & @xmath454 @xmath407 + @xmath455 & @xmath456 @xmath407 + @xmath457 & @xmath458 @xmath407 + @xmath459 & @xmath460 @xmath407 + @xmath461 & @xmath462 @xmath407 + @xmath463 & @xmath464 @xmath407 + @xmath465 & @xmath466 @xmath407 + @xmath467 & @xmath468 @xmath407 + @xmath469 & + @xmath470 & @xmath471 @xmath407 + @xmath472 & @xmath473 @xmath407 + @xmath474 & @xmath475 @xmath407 + @xmath476 & + @xmath477 & @xmath478 + @xmath479 & @xmath480 @xmath407 + @xmath481 & @xmath468 @xmath407 + @xmath482 & @xmath483 @xmath407 + @xmath484 & @xmath485 @xmath407 + @xmath486 & @xmath487 @xmath407 + @xmath488 & @xmath489 @xmath407 + @xmath490 & @xmath491 @xmath407 + @xmath492 & @xmath493 @xmath407 + lllc carbon & @xmath495 & @xmath496 & b + oxygen & @xmath497 & @xmath498 & c + silicon & @xmath499 & @xmath500 & d + iron & @xmath501 & @xmath502 & d + sulfur & @xmath503 & @xmath504 & d + magnesium & @xmath505 & @xmath506 & d + pahs & @xmath507 & @xmath508 & e + grain area & @xmath509 & @xmath510 & f + the results in 3.1 and 3.2 can be understood by a simple analytic chemical model that incorporates the main physics .
such a model , though approximate , has the advantage of allowing one to determine and understand the sensitivity to various model parameters , and serves to validate the numerical model .
_ analytic solution to @xmath130(oh@xmath0 ) and @xmath160 in the first peak ( @xmath18)_. the top panel of figure 1 describes the main chemical pathways to oh@xmath0 for @xmath18 , and table 1 lists the rate coefficients used for each of these reactions .
figure 1 omits the formation of oh@xmath0 by the photoionization of oh , which is only important at very low @xmath511 and never dominates the production of the column of oh@xmath0 for clouds of higher @xmath13 .
this minor route produces oh by the reactions o + h@xmath1(vibrationally excited ) , o + h , and the formation of water ice on grains followed by desorption leading to oh .
the photoionization of oh leads to a constant ( low ) oh@xmath0 abundance at the cloud surface .
we also neglect that route in the analytic solution , as it does not affect the peak abundance of oh@xmath0 , nor the values of @xmath160 near the peak .
in addition , we approximate @xmath160 for this peak by only including the rate of formation of oh@xmath0 by the reaction h@xmath1 + o@xmath0 @xmath64 oh@xmath0 + h. in the following we use rate coefficients from table 1 with two exceptions .
the stancil et al ( 1999 ) rate coefficients for the reaction of h@xmath0 with o and its reverse reaction are complicated , and we simplify them here , focussing on results for low density clouds .
we set @xmath512exp@xmath513 @xmath17 s@xmath15 as the rate coefficient for the reaction h@xmath0 + o @xmath64 o@xmath0 + h , and @xmath514 @xmath17 s@xmath15 as the rate coefficient for the reaction o@xmath0 + h @xmath64 h@xmath0 + o. let @xmath515 be the rate coefficient for the reaction h@xmath1 + o@xmath0 @xmath64 oh@xmath0 + h ; @xmath516 the rate coefficient for the recombination of h@xmath0 with electrons : @xmath517 the rate coefficient for the neutralization of h@xmath0 by pah@xmath21 ; and @xmath518 the rate coefficient for the charge exchange of h@xmath0 with pah .
let @xmath519 be the abundance of h@xmath1 with respect to h nuclei , @xmath156 the abundance of electrons , @xmath520 the abundance of neutral pahs , @xmath521 the abundance of pah@xmath21 , and @xmath522 the gas phase elemental abundance of o. the solution to @xmath160 , the efficiency of making oh@xmath0 from cosmic rays , is then : we utilized the fact that @xmath160 approaches unity when @xmath525 to set the condition in 3.1.1 ( eq .
1 ) on the abundances of o , h@xmath1 , electrons , pahs and pah@xmath21s which lead to @xmath526 .
one can immediately see that for low values of @xmath130(h@xmath1 ) , where @xmath527 , @xmath160 is proportional to @xmath130(h@xmath1 ) . here
, the dominant route of h@xmath0 destruction is by recombination with electrons , pahs , and pah@xmath21s .
the small fraction of the h@xmath0 created by cosmic rays that charge exchanges with o to form o@xmath0 that then reacts with h@xmath1 to form oh@xmath0 is proportional to the h@xmath1 abundance . substituting likely values of @xmath528 ( we focus on cases with high @xmath7/@xmath225 and here @xmath529 as well as c@xmath0 contribute electrons - see figure 4 ) , @xmath130(pah)@xmath530 , @xmath130(pah@xmath21)@xmath531 , and @xmath532 , we obtain @xmath533 \left[\left({x_e \over 3\times
10^{-4}}\right)t_2^{-0.75 } + 4.4\left ( { x({\rm pah^- } ) \over 1.5 \times 10^{-8}}\right)t_2^{-0.5 } + 2.7\left({x({\rm pah } ) \over 1.85 \times 10^{-7}}\right)\right]}},\ ] ] for higher values of @xmath537 , @xmath160 saturates at values near unity , as nearly every cosmic ray ionization leads to a production of oh@xmath0 . with standard values of @xmath156 , @xmath130(pah ) and @xmath130(pah@xmath21 ) equation ( 3 )
then becomes equation ( 3 ) shows that pah@xmath21 in particular can lower @xmath160 by neutralizing h@xmath0 .
we emphasize that equations ( 3 - 5 ) are valid only if the pah and pah@xmath21 rate coefficients we have adopted in table 1 are valid .
equations ( 4 - 5 ) assume the standard pah and pah@xmath21 abundances given above . rather than a simple saturated value of @xmath160 at high @xmath130(h@xmath1 ) , however , we see in the figures that the first peak in oh@xmath0 shows a value of @xmath160 that peaks and then falls with increasing depth , once most of the gas is h@xmath1 . this is because once the gas is more than half molecular ( @xmath541 ) , the cosmic ray ionization is mostly of h@xmath1 , which we have not included in the analytic treatment
. the route to oh@xmath0 via the ionization of h@xmath1 is not as efficient at the surface , because of the high abundance of electrons which react quickly with h@xmath24 .
thus , as @xmath130(h ) declines deeper into the cloud , the value of @xmath160 falls with increasing @xmath13 .
the analytic equation for @xmath50oh@xmath0@xmath542 requires the addition of reactions of oh@xmath0 with electrons and h@xmath1 that destroy oh@xmath0 .
let @xmath543 be the rate coefficient for the destruction of oh@xmath0 by h@xmath1 , and @xmath544 be the coefficient for destruction by electrons .
then , we find again , we can break this equation into two regimes . for @xmath50h@xmath1)@xmath548 , oh@xmath0
is mainly destroyed by dissociative recombination with electrons and the second term in the denominator can be ignored .
then , assuming standard parameters for the abundances of @xmath156 , @xmath130(pah ) , and @xmath130(pah@xmath21 ) , we obtain @xmath549 however , if @xmath550h@xmath1)@xmath551 , then h@xmath1 mainly destroys oh@xmath0 and , again using standard values for the parameters , we obtain the peak and plateau value of the oh@xmath0 abundance : for higher values of @xmath50h@xmath553 , the abundance of oh@xmath0 declines as @xmath160 saturates , and as @xmath130(h ) declines leading to the less efficient production of oh@xmath0 initiated by cosmic ray ionization of h@xmath1 . the column of oh@xmath0 in the first peak can then be estimated @xmath554 where @xmath555 is the width of the peak region .
this shows the important result that the peak abundance ( and approximately the column of oh@xmath0 in the first peak ) is proportional to @xmath556 . in order to find regions of high oh@xmath0 column that can be observed , and where cosmic rays are the ultimate cause of the ions , we therefore must look at low density regions with high cosmic ray ionization rates .
_ analytic solution to the second peak of h@xmath2o@xmath0 in the regime @xmath557 and with pahs present_. deeper in the cloud ( @xmath187 ) the dust and gas shield the fuv photons and carbon is no longer in the form of c@xmath0 but converts to co. here , the electron abundance drops rapidly with depth , and the gas is almost entirely molecular so that the route to h@xmath2o@xmath0 is started with the cosmic ray ionization of h@xmath1 .
we then define @xmath160(h@xmath2o@xmath0 ) so that the rate of formation of h@xmath2o@xmath0 per unit volume is @xmath558(h@xmath2o@xmath0 ) .
, then the total rate per h@xmath1 is roughly 2@xmath559 .
the cosmic ray ionization rate per unit volume is then 2@xmath560 in fully molecular zones .
] again @xmath160(h@xmath2o@xmath0 ) is equivalent to the fraction of cosmic ray ionizations that lead to h@xmath2o@xmath0 .
figure 1 shows the route to h@xmath2o@xmath0 from the ionization of h@xmath1 .
the key competition that determines @xmath160(h@xmath2o@xmath0 ) revolves around h@xmath24 . with all the hydrogen molecular , the routes via h@xmath1 dominate the rates via electrons .
thus , the only interruption in the chain that leads to h@xmath2o@xmath0 is the possibility that h@xmath24 will react with either electrons or co rather than with o to form either oh@xmath0 or h@xmath1o@xmath0 .
we then define @xmath561 as the rate coefficient for h@xmath24 reacting with o to form either oh@xmath0 or h@xmath1o@xmath0 ; @xmath562 is the rate coefficient for h@xmath24 to dissociatively recombine with electrons ; @xmath563 is the rate coefficient for h@xmath24 to react with co ; and @xmath564 is the rate coefficient for h@xmath2o@xmath0 to dissociatively recombine with electrons ( all channels ) . therefore , we find in our model runs that electrons are formed by the cosmic ray ionization of h@xmath1 and destroyed by attachment to neutral pahs .
most of the pahs deep in the cloud are neutral , and therefore using a neutral pah abundance of @xmath567 we find if either o or co dominate the destruction of h@xmath24 over electrons ( or @xmath569 ) , which we find is usually the case for most of our runs at the second peak of the h@xmath2o@xmath0 abundance , then we find , substituting @xmath156 into the equation for @xmath570 , at the peak , we generally find that most of the carbon is in co , so @xmath572 and that most of the remaining o is atomic , or @xmath573 .
therefore , we predict a peak h@xmath2o@xmath0abundance of @xmath574 this explains why the second peak abundance of h@xmath2o@xmath0 is independent of @xmath559 and @xmath8 .
the production rate per unit volume of h@xmath2o@xmath0 depends on @xmath575 but the destruction rate per unit volume depends on @xmath576 . since electrons are formed by cosmic ray ionization and destroyed by pahs , @xmath577 .
therefore , both production and destruction of h@xmath2o@xmath0 is proportional to @xmath575 , and the resulting abundance of h@xmath2o@xmath0 is independent of both parameters . as one goes to higher @xmath13 from the peak , the atomic o freezes out on grains as water ice , and thus the abundance of h@xmath2o@xmath0 declines .
the thickness of the region of peak h@xmath2o@xmath0 abundance is therefore of order @xmath578 , or a hydrogen column of @xmath579 @xmath11 . multiplying this by @xmath580
, we estimate columns of h@xmath2o@xmath0 of @xmath581 @xmath11 , as we found in our numerical runs .
there has been some indication ( rimmer et al 2011 ) that the cosmic ray ionization rate may decline with depth into a cloud . from the above result
, this should have little consequence on the h@xmath2o@xmath0 column .
we should emphasize that this prediction is dependent on the presence of pahs deep ( @xmath582 ) in the cloud .
if pahs coagulate on grain surfaces , then pahs do not remove free electrons from the gas phase . in this case , the electron abundance at the peak of the h@xmath2o@xmath0 can be higher than in the pah case if gas phase metal ions are present .
this suppresses the h@xmath2o@xmath0abundance at the peak .
therefore , the observation of high columns of h@xmath2o@xmath0 deep in molecular clouds may indicate either the presence of pahs lowering the electron abundances there , or , in the absence of pahs , may indicate a high depletion of gas phase metals at the second peak .
recently , klippenstein et al ( 2010 ) have provided theoretical rate coefficients for the reactions of h@xmath24 with o and with co that are 10 to 20% higher than the rate coefficients we adopted .
we have run our standard cases with these new coefficients and find that the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 increase by roughly 10% with the new coefficients .
in addition , cartledge et al ( 2004 ) suggest a gas phase oxygen abundance of 2.84@xmath583 , whereas jensen et al ( 2005 ) suggest a value of 4.2 to 4.7@xmath583 .
we have adopted 3.2@xmath583 but the variation in the literature suggests that we test the dependence of the ion columns on the gas phase abundance of elemental o. assuming an abundance of 4.5@xmath583 , an increase of 40% over our standard rate , increases the columns of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 by 10 - 20% .
therefore , the slight possible variations in these rate coefficients and/or the gas phase elemental o abundance have a negligible effect on our conclusions .
we have also tested our results for their sensitivity to the h@xmath1 formation rate coefficient for formation on grain surfaces .
we find that lowering this rate by a factor of 3 hardly affects the results for high density gas , and that the effect is most pronounced at low density . at @xmath169 @xmath49 , lowering the h@xmath1 formation rate by 3 lowers the columns of oh@xmath0 and h@xmath1o@xmath0 by factors of less than 2 in both peaks .
the h@xmath2o@xmath0 column in both peaks decreases by a bit less than a factor of 3 .
the main effect of lowering the h@xmath1 rate coefficient is to drive the h / h@xmath1 transition deeper into the cloud , which moves the first peaks of the ions to greater depths , but their columns and peak abundances do not change appreciably . in the text we discussed the effect of lowering the gas phase abundance of low ionization potential metal atoms such as s , fe , mg , and si if pahs are not present .
here , we examine the effect of raising their abundances . our standard gas phase abundances for these species are given in table 2 of appendix a. in our test cases , we raise the gas phase abundances of these species to 10@xmath584 , a factor of 10 - 100 times their standard values . for our standard case of @xmath163 @xmath49 and @xmath7@xmath585 s@xmath15 , but with no pahs , this raises the electron abundance at high @xmath13 in the gas from about 10@xmath207 to about @xmath403 .
this rise depresses the ( second ) peak abundance of h@xmath2o@xmath0 from about @xmath586 to about @xmath587 if the photodesorption yields of fe , mg , and si are @xmath588 , a likely upper limit ( h09 ) .
smaller yields lower the electron abundance , thereby changing the h@xmath2o@xmath0 less .
however , if pahs are present , metal ions recombine with pah@xmath21 and pah , which is a much faster process than with electrons .
hence , the electron abundance does not depend much on the metal abundances .
consequently , the ( second ) peak h@xmath2o@xmath0 abundance does not change significantly .
the first peaks of the ions are not affected , because the electrons are supplied by c@xmath0 , whose gas phase abundance is more than 10 times the ( high ) abundance of the metal ions .
section 4 discusses observations that indicate @xmath9(h@xmath2o@xmath0)@xmath589 @xmath590 @xmath11 in some clouds , possibly even as high as @xmath591 @xmath11 .
the only way our models can accommodate very large columns of h@xmath2o@xmath0 in a single cloud is if there is incomplete freeze - out of water ice on grains .
note that our models are steady state , so that time dependent effects might leave more elemental oxygen in the gas phase than steady state results would indicate .
another possibility is that the grains are sufficiently warm , @xmath392 k , to thermally evaporate water ice off the grains , or that cosmic ray rates are sufficiently high to desorb the ice mantles .
we have run our code fixing the grain temperature to be @xmath592 k , so that water ice can not form and elemental oxygen is plentiful in the gas phase at all @xmath13 . in the case of pahs , and for our standard case of @xmath163 @xmath49 and @xmath7@xmath143 s@xmath15 , we find that @xmath130(h@xmath2o@xmath0 ) plateaus at a value of about @xmath593 at about @xmath119 , and stays constant at that value for all higher @xmath13 .
( note that given the discussion in 3.1 , this result is independent of @xmath7 and @xmath8 ) .
therefore , a cloud with a hydrogen nucleus column of @xmath594 @xmath11 would have @xmath9(h@xmath2o@xmath0)@xmath595 @xmath11 , for example . without pahs , and assuming the low ionization potential metal atoms strongly deplete on grains , we would predict similar columns .
aalto , s. , costagliola , f. , van der tak , f. , & meijerink , r. 2011 , , 527 , a69 badnell , n. r. , et al .
2003 , , 406 , 1151 badnell , n. r. 2006 , , 167 , 334 bakes , e. l. o. , & tielens , a. g. g. m. 1998 , , 499 , 258 benz , a. o. , bruderer , s. , van dishoeck , e. f. , et al .
2010 , , 521 , l35 black , j. h. , & dalgarno , a. 1973 , , 184 , l101 boulanger , f. , falgarone , e. , puget , j. l. , & helou , g. 1990 , , 364 , 136 brian , j. , & mitchell , a. 1990 , phys . rep .
186 , 215 bruderer , s. , benz , a. o. , van dishoeck , e. f. , et al .
2010 , , 521 , l44 cartledge , s. i. b. , lauroesch , j. t. , meyer , d. m. , & sofia , u. j. 2004 , , 613 , 1037 corbel , s. , & eikenberry , s. s. 2004 , , 419 , 191 dalgarno , a. , yan , m. , & liu , w. 1999 , , 125 , 237 draine , b. t. 1978 , , 36 , 595 draine , b. t. , & katz , n. 1986 , , 310 , 392 draine , b. t. , & li , a. 2007 , , 657 , 810 draine , b. t. , & sutin , b. 1987 , , 320 , 803 elitzur , m. , & watson , w. d. 1978 , , 222 , l141 falgarone , e. , godard , b. , cernicharo , j. , et al . 2010 , , 521 , l15 falgarone , e. , ossenkopf , v. , gerin , m. , et al .
2010 , , 518 , l118 flower , d. r. , & pineau des forets , g. 1998 , , 297 , 1182 flower , d. r. , & pineau des forts , g. 2003 , , 343 , 390 gerin , m. , de luca , m. , black , j. , et al .
2010 , , 518 , l110 godard , b. , falgarone , e. , gerin , m. , et al .
2012 , arxiv:1201.5457 godard , b. , falgarone , e. , & pineau des forts , g. 2009 , , 495 , 847 godard , b. , falgarone , e. , gerin , m. , hily - blant , p. , & de luca , m. 2010 , , 520 , a20 goicoechea , j. r. , & cernicharo , j. 2001 , , 554 , l213 goicoechea , j. r. , rodrguez - fernndez , n. j. , & cernicharo , j. 2004 , , 600 , 214 goicoechea , j. r. , & le bourlot , j. 2007 , , 467 , 1 gonzlez - alfonso , e. , fischer , j. , isaak , k. , et al .
2010 , , 518 , l43 gupta , h. , rimmer , p. , pearson , j. c. , et al .
2010 , , 521 , l47 habing , h. j. 1968 , , 19 , 421 heiles , c. , & troland , t. h. 2003 , , 586 , 1067 heiles , c. , & troland , t. h. 2005 , , 624 , 773 hollenbach , d. , kaufman , m. j. , bergin , e. a. , & melnick , g. j. 2009 , apj , 690 , 1497 ( h09 ) indriolo , n. , geballe , t. r. , oka , t. , & mccall , b. j. 2007 , , 671 , 1736 indriolo , n. , geballe , t. r. , oka , t. , & mccall , b. j. 2011 , 66th international symposium on molecular spectroscopy , indriolo , n. , neufeld , d.a .
, gerin , m. , geballe , t. r. , black , j. h. , menten , k. m. , & goicoechea , j. r. 2012 , , submitted jensen , m. j. , bilodeau , r. c. , safvan , c. p. , seiersen , k. , andersen , l. h. , pedersen , h. b. , & heber , o. 2000 , , 543 , 764 jensen , a. g. , rachford , b. l. , & snow , t. p. 2005
, , 619 , 891 jochims , h. w. , baumgaertel , h. , & leach , s. 1996 , , 314 , 1003 joulain , k. , falgarone , e. , pineau des forets , g. , & flower , d. 1998 , , 340 , 241 klippenstein , s.j . , georgievskii , y. , & mccall , b.j .
2010 , j. phys .
, 114 , 278 kreowski , j. , beletsky , y. , & galazutdinov , g. a. 2010 , , 719 , l20 le petit , f. , nehm , c. , le bourlot , j. , & roueff , e. 2006 , , 164 , 506 lepp , s. , & dalgarno , a. 1988 , , 324 , 553 lesaffre , p. , gerin , m. , & hennebelle , p. 2007
, , 469 , 949 liszt , h. 2003 , , 398 , 621 liszt , h. s. 2007 , , 461 , 205 malloci , g. , joblin , c. , & mulas , g. 2007 , chemical physics , 332 , 353 mathis , j. s. , mezger , p. g. , & panagia , n. 1983 , , 128 , 212 mccall , b. j. , et al .
2004 , , 70 , e052716 meyer , d. m. , jura , m. , & cardelli , j. a. 1998 , apj , 493 , 222 millar , t. j. , herbst , e. , & charnley , s. b. 1991 , , 369 , 147 neau , a. , et al .
2000 , , 113 , 1762 neufeld , d. a. , goicoechea , j. r. , sonnentrucker , p. , et al .
2010 , , 521 , l10 ( n10 ) neufeld , d. , hollenbach , d. , kaufman , m. , et al .
2011 , iau symposium , 280 , berg , k. i. , van dishoeck , e. f. , & linnartz , h. 2009 , , 496 , 281 ossenkopf , v. , mller , h. s. p. , lis , d. c. , et al .
2010 , , 518 , l111 phillips , t. g. , van dishoeck , e. f. , & keene , j. 1992 , , 399 , 533 ( pvdk92 ) prasad , s. s. , & tarafdar , s. p. 1983
, , 267 , 603 rapacioli , m. , calvo , f. , joblin , c. , et al .
2006 , , 460 , 519 rimmer , p. , & herbst , e. 2011 , , 82 , 933 roberge , w. g. , jones , d. , lepp , s. , & dalgarno , a. 1991 , , 77 , 287 savage , b. d. , & sembach , k. r. 1996 , ara&a , 34 , 279 schilke , p. , comito , c. , mller , h. s. p. , et al .
2010 , , 521 , l11 sofia , u. j. , lauroesch , j. t. , meyer , d. m. , & cartledge , s. i. b. 2004 , apj , 605 , 272 stancil , p. c. , schultz , d. r. , kimura , m. , et al . 1999 , , 140 , 225 tielens , a. g. g. m. , & hollenbach , d. 1985 , , 291 , 722 van der tak , f. f. s. , aalto , s. , & meijerink , r. 2008 , , 477 , l5 van der tak , f. f. s. , belloche , a. , schilke , p. , et al .
2006 , , 454 , l99 van der werf , p. p. , isaak , k. g. , meijerink , r. , et al .
2010 , , 518 , l42 van dishoeck , e. f. 1988 , in rate coefficients in astrochemistry , eds .
millar & d. a. williams ( dordrecht : kluwer ) , p.49 van dishoeck , e. f. , & black , j. h. 1986 , , 62 , 109 van dishoeck , e. f. , jonkheid , b. van hemert , m. c. 2006 , faraday discussions , 2006 , 133 , 231 visser , r. , van dishoeck , e. f. , & black , j. h. 2009 , , 503 , 323 weingartner , j. c. , & draine , b. t. 2001a , , 563 , 842 weingartner , j. c. , & draine , b. t. 2001b , , 553 , 581 wei , a. , requena - torres , m. a. , gsten , r. , et al .
2010 , , 521 , l1 wolfire , m. g. , mckee , c. f. , hollenbach , d. , & tielens , a. g. g. m. 2003 , , 587 , 278 wolfire , m. g. , tielens , a. g. g. m. , hollenbach , d. , & kaufman , m. j. 2008 , , 680 , 384 wootten , a. , mangum , j. g. , turner , b. e. , et al .
1991 , , 380 , l79 wyrowski , f. , menten , k. m. , gsten , r. , & belloche , a. 2010 , , 518 , a26 wyrowski , f. , van der tak , f. , herpin , f. , et al .
2010 , , 521 , l34 | we model the production of oh@xmath0 , h@xmath1o@xmath0 , and h@xmath2o@xmath0 in interstellar clouds , using a steady state photodissociation region code that treats the freeze - out of gas species , grain surface chemistry , and desorption of ices from grains .
the code includes pahs , which have important effects on the chemistry .
all three ions generally have two peaks in abundance as a function of depth into the cloud , one at @xmath3 and one at @xmath4 , the exact values depending on the ratio of incident ultraviolet flux to gas density . for relatively low values of the incident far ultraviolet flux on the cloud ( @xmath5 ; @xmath6= local interstellar value ) , the columns of oh@xmath0 and h@xmath1o@xmath0 scale roughly as the cosmic ray primary ionization rate @xmath7 divided by the hydrogen nucleus density @xmath8 .
the h@xmath2o@xmath0 column is dominated by the second peak , and we show that if pahs are present , @xmath9(h@xmath2o@xmath0)@xmath10 @xmath11 independent of @xmath7 or @xmath8 . if there are no pahs or very small grains at the second peak , @xmath9(h@xmath2o@xmath0 ) can attain such columns only if low ionization potential metals are heavily depleted .
we also model diffuse and translucent clouds in the interstellar medium , and show how observations of @xmath9(oh@xmath0)/@xmath9(h ) and @xmath9(oh@xmath0)/@xmath9(h@xmath1o@xmath0 ) can be used to estimate @xmath7/@xmath8 , @xmath12 and @xmath13 in them .
we compare our models to _ herschel _ observations of these two ions , and estimate @xmath7@xmath14 s@xmath15 and @xmath16 @xmath17 for diffuse foreground clouds towards w49n . |
classical and weak - lined t tauri stars ( ctts and wtts ) are pre - main sequence stars showing vigorous x - ray emission with x - ray luminosities ( @xmath0 ) near the empirical saturation limit found for main - sequence ( ms ) stars , @xmath1 .
consequently , x - ray emission from t tauri stars has been attributed to solar - like coronal activity .
however , both accretion and outflows / jets may contribute to or alter x - ray production in the magnetic environment of strongly accreting young stars .
x - rays thus provide an important diagnostic for the detection and study of accretion and outflow activity .
support for x - ray _ suppression _ in ctts has been reported from x - ray photometry .
ctts are , on average , less x - ray luminous than wtts ( e.g. , ( * ? ? ?
strom & strom 1994 ) , ( * ? ? ?
* neuhuser 1995 ) ) .
this finding has been partly supported by recent , deep surveys of the orion nebula cluster ( ( * ? ? ?
* getman 2005 ) , ( * ? ? ?
* preibisch 2005 ) ) . on the other hand , x - rays may be _ generated _ by surface accretion shocks ( ( * ? ? ?
* lamzin 1999 ) ) .
if gas collides with the stellar photosphere in free fall , shocks heat it to a few million k. given the appreciable accretion rates , high shock densities of order @xmath2 @xmath3 are expected , as first reported for the ctts tw hya ( ( * ? ? ?
* kastner 2002 ) ) .
x - rays may similarly be produced in shocks forming in jets ( ( * ? ? ?
* raga 2002 ) ) . the _ xmm - newton extended survey of the taurus molecular cloud ( xest ) _ ( ( * ? ? ?
* gdel 2007a ) ) has provided new insights into these issues .
xest covers the most populated @xmath45 sq .
deg of the taurus star - forming region .
the average on - axis detection limit is @xmath5 erg s@xmath6 for lightly absorbed objects , sufficient to detect about half of the observed brown dwarfs ( ( * ? ? ?
* grosso 2007a ) ) .
we discuss x - ray results relevant for the star - disk interface in which accretion occurs and where ( parts of ) the jets may be accelerated .
@xcite present statistical x - ray studies of the xest ctts and wtts samples .
the wtts sample is essentially complete ( all but one of the 50 surveyed objects detected ) , and the ctts sample is nearly complete ( 85% of the 65 surveyed objects detected ) . the luminosity deficiency of ctts is confirmed .
more precisely , ctts are statistically less luminous by a factor of @xmath7 both in @xmath0 and in @xmath8 ( fig . [
fig : accretion]a ) while the @xmath9 distributions of the two samples are drawn from the same parent population . on the other hand , @xcite report the average electron temperature , @xmath10 , in the x - ray sources of ctts to be higher than in wtts , irrespective of the gas absorption column density ( @xmath11 ) . for wtts
, a trend also seen in ms stars is recovered , in the sense that the electron temperature increases with @xmath0 .
such trends are expected in stochastic - flare heated coronae ( ) , but a similar trend is absent in ctts in which the average temperature remains high for all activity levels .
in contrast to the apparent x - ray deficiency , a trend toward an ultraviolet excess is found in ctts based on the optical monitor ( om ) data ( ( * ? ? ?
* audard 2007 ) ) .
the uv excess supports an accretion scenario in which gas in accretion streams shock - heats near the surface to form hot spots ( ( * ? ? ?
* calvet & gullbring 1998 ) ) .
the om has also recorded a slow u - band flux increase in a brown dwarf , most likely due to an `` accretion event '' covering a time span of several hours ( ( * ? ? ?
* grosso 2007b ) ) .
accretion shocks may leave signatures in line - dominated high - resolution x - ray spectra . given the typical mass accretion rates on t tauri stars and their modest accretion hot spot filling factors of no more than a few percent ( ( * ? ? ?
* calvet & gullbring 1998 ) ) , densities of order @xmath12 @xmath3 or more are to be expected ( ( * ? ? ?
* telleschi 2007c ) , ( * ? ? ?
* gdel 2007c ) ) , and such densities are indeed indicated in the density - sensitive line ratios of ovii and neix triplets of some ctts ( e.g. , ( * ? ? ?
* kastner 2002 ) , ( * ? ? ?
* stelzer & schmitt 2004 ) ) .
xest has added relevant information on two further accreting pre - main sequence stars : t tau n ( ( * ? ? ?
* gdel 2007c ) ) and the herbig star ab aur ( ( * ? ? ?
* telleschi 2007b ) ) . however , in both _
xmm - newton _ rgs spectra , the ovii triplet line ratios are compatible with density upper limits of a few times @xmath13 @xmath3 ( fig .
[ fig : accretion]b ) , apparently not supporting the accretion - shock scenario .
how important really is accretion for ctts x - ray emission ?
-0.1truecm -5.2truecm fig .
[ fig : ttau ] compares _ xmm - newton _ rgs spectra of the active binary hr 1099 ( x - rays mostly from a k - type subgiant ; archival data ) the weakly absorbed wtts v410 tau ( ( * ? ? ?
* telleschi 2007c ) ) , the ctts t tau ( ( * ? ? ?
* gdel 2007c ) ) , and the old f subgiant procyon ( archival data ) .
hr 1099 and v410 tau show the typical signatures of a hot , active corona such as a strong continuum , strong lines of nex and of highly - ionized fe but little flux in the ovii line triplet .
in contrast , lines of c , n , and o dominate the soft spectrum of procyon , the ovii triplet exceeding the oviii ly@xmath14 line in flux .
t tau reveals signatures of a very active corona shortward of 19 but also an unusually strong ovii triplet .
because its @xmath11 is large ( in contrast to @xmath11 of v410 tau ) , we have modeled the intrinsic , unabsorbed spectrum based on transmissions determined in xspec using @xmath11 from epic spectral fits ( @xmath15 cm@xmath6 ; ( * ? ? ? * gdel 2007a ) ) . _
the ovii lines are the strongest lines in the intrinsic x - ray spectrum , _ reminiscent of the situation in procyon !
to generalize this finding , we plot in fig .
[ fig : softexcess ] the ratio between the intrinsic ( unabsorbed ) luminosities of the ovii @xmath17 line and the oviii ly@xmath14 line as a function of @xmath0 , comparing ctts and wtts with a larger ms sample ( ( * ? ? ?
* ness 2004 ) ) and ms solar analogs ( ( * ? ? ?
* telleschi 2005 ) ) .
the tts data are from @xcite , @xcite , @xcite , @xcite , and from our analysis of archival _ xmm - newton _ data of ru lup . for the tts sample given by @xcite , we have approximated @xmath18(ovii @xmath17 ) @xmath19(ovii ) ( ( * ? ? ?
* porquet 2001 ) ) .
the trend for ms stars ( black crosses and triangles , sect .
[ accretors ] ) is evident : as the coronae get hotter toward higher @xmath0 , the ratio of ovii @xmath17/oviii ly@xmath14 line luminosities decreases .
this trend is followed by the sample of wtts , while ctts again show a significant excess .
this is the essence of the * x - ray soft excess * in ctts first discussed by @xcite and @xcite : _ accreting pre - main sequence stars reveal a strong excess of cool ( 1 - 2 mk ) material , regardless of the overall x - ray deficiency at higher temperatures .
_ crucial xest results on x - ray production and accretion can be summarized as follows : 1 .
accreting ctts show a general deficiency in x - ray production when referring to the hot coronal gas recorded by x - ray ccd cameras ( ( * ? ? ? * telleschi 2007a ) ) ; 2 .
the average coronal @xmath10 is higher in ctts than in wtts ( ( * ? ? ?
* telleschi 2007a ) ) ; 3 . all ctts ( except the two flaring sources su aur [ also subject to high @xmath11 ] and dh tau ) show a * soft excess * defined by an anomalously high ratio between the fluxes of the ovii he - like triplet and the oviii ly@xmath14 line , when compared to wtts and ms stars .
the origin of the additional cool plasma in ctts is likely to be related to the accretion process .
accretion streams may shock - heat gas at the impact point to x - ray emitting temperatures .
this model is supported by high electron densities inferred from the observed ovii or neix triplets in some of the ctts ( e.g. , ( * ? ? ?
* kastner 2002 ) , ( * ? ? ?
* stelzer & schmitt 2004 ) ) , although high densities were not seen in the two xest accretors t tau ( ( * ? ? ?
* gdel 2007c ) ) and ab aur ( ( * ? ? ?
* telleschi 2007b ) ) . alternatively ,
the cool , infalling material may partly cool pre - existing heated coronal plasma , or reduce the efficiency of coronal heating in the regions of infall ( ( * ? ? ?
* preibisch 2005 ) , ( * ? ? ?
* telleschi 2007c ) , ( * ? ? ?
* gdel 2007c ) ) .
this model would at the same time explain why ctts are x - ray weaker than wtts ( ( * ? ? ?
* preibisch 2005 ) , ( * ? ? ?
* telleschi 2007a ) ) .
we can not assess what the relative importance of these processes is .
it seems clear , however , that the soft excess described here argues in favor of a substantial influence of accretion on the x - ray production in pre - main sequence stars .
the shock temperature in jets can be expressed as @xmath21 k ( for fully ionized gas ) where @xmath22 is the shock speed in units of 100 km s@xmath6 ( e.g. , ( * ? ? ? * raga 2002 ) ) .
jet speeds are typically of order @xmath23 km s@xmath6 ( ( * ? ? ?
* eislffel & mundt 1998 ) , ( * ? ? ?
* anglada 1995 ) , ( * ? ? ?
* bally 2003 ) ) , in principle allowing for shock speeds of similar magnitude .
faint , soft x - ray emission has been detected from a few protostellar hh objects ( ( * ? ? ?
* pravdo 2001 ) , ( * ? ? ?
* pravdo 2004 ) , ( * ? ? ?
* pravdo & tsuboi 2005 ) , ( * ? ? ?
* favata 2002 ) , ( * ? ? ?
* bally 2003 ) , ( * ? ? ?
* tsujimoto 2004 ) , ( * ? ? ? * grosso 2006 ) ) .
@xcite used a _
chandra _ observation to show that x - rays form within an arcsecond of the protostar l1551 irs-5 while the star itself is too heavily obscured to be detected .
strong absorption and extinction of protostars and their immediate environment make the launching region of their powerful jets generally inaccessible to optical , near - infrared , or x - ray studies .
however , a class of strongly accreting , optically revealed ctts also exhibit so - called micro - jets visible in optical lines ( ( * ? ? ?
* hirth 1997 ) ) , with flow speeds similar to protostellar jets .
ctts micro - jets have the unique advantage that they can - in principle - be followed down to the acceleration region both in the optical and in x - rays .
x - ray spectra of very strongly accreting , micro - jet driving ctts exhibit an anomaly ( fig .
[ fig : pinkspec ] , ( * ? ? ?
* gdel 2005 ) , ( * ? ? ? * gdel 2007b ) ) : the spectra of dg tau , gv tau , dp tau , cw tau , and hn tau are composed of two components , a cool component subject to very low absorption and a hot component subject to photoelectric absorption about one order of magnitude higher . for similar phenomenology , see also @xcite and @xcite .
the cool component shows temperatures atypical for t tau stars , ranging from @xmath24 mk , while the hot component reveals extremely high temperatures ( 10100 mk ) .
we discuss in the following the best example , the single ctts dg tau .
the hard spectral component of the dg tau point source is unusually strongly absorbed , with a gas column density ( @xmath25 @xmath26 ) higher by a factor of @xmath27 than predicted from the visual extinction @xmath28 of 1.53 mag if standard gas - to - dust ratios are assumed ( see ( * ? ? ?
* gdel 2007b ) and references therein ) . because the hard component occasionally flares , in one case being preceded by u band emission as in solar and stellar flares ( ( * ? ? ?
* gdel 2007b ) ) , it is most straightforwardly interpreted as coronal or `` magnetospheric '' .
the _ excess absorption _ is likely to be due to the heavy accretion streams falling down along the magnetic fields and absorbing the x - rays from the underlying corona / magnetosphere .
the excess absorption - to - extinction ( @xmath29 ) ratio then is an indicator of dust sublimation : _ the accreting gas streams are dust - depleted_. in contrast , @xmath11 of the soft x - ray component , @xmath30 @xmath26 ( 90% error range ) is _ lower _ than suggested from the stellar @xmath28 , @xmath31 @xmath26 .
a likely origin of these x - rays is the base of the jet .
such an origin is suggested by i ) the unusually soft emission not usually seen in t tauri stars ( ( * ? ? ?
* gdel 2007a ) ) , ii ) the low @xmath11 , and iii ) the explicit evidence of jets in the _ chandra _ image , as we will show below . a _ chandra _ x - ray image of the dg tau environment is shown in fig .
[ fig : jetim]b ( pixel size 0.49@xmath32 ) .
this image was produced by combining counts from a total of 90 ks of _ chandra _ exposure time .
also shown is a smoothed version . to suppress background and to emphasize the soft sources , only counts within the 0.61.7 kev range
are plotted .
there is clear evidence for a jet - like extension outside the stellar point spread function ( psf ) to the sw along a position angle of @xmath33 deg , but we also find a significant excess of counts in the ne direction ( pa @xmath34 deg ) .
this is coincident with the jet optical axis , which for the sw jet has been given as 217 - 237 deg ( ( * ? ? ?
* eislffel & mundt 1998 ) ) .
we verified , using raytrace simulations , that the jet sources are extended : a faint point source would occupy only a few pixels .
we also find the counter jet to be harder , with photon energies mostly above 1 kev , while the forward jet shows a mixture of softer and harder counts .
the spectral properties of the jet sources are reminiscent of the _ soft _ component in the `` stellar '' spectrum . 1 . a weakly absorbed , diffuse , soft component along the forward - jet axis ; 2 . an intermediately absorbed , diffuse , soft component along the counter - jet axis ; 3 . a weakly absorbed , compact , non - variable , soft component ( within the stellar psf ) ; 4 .
a strongly absorbed , compact , flaring , hard component ( within the stellar psf ) .
the low @xmath11 of the _ forward jet _ and of the _ soft stellar _ component suggests that the soft spectral emission originates from a region `` in front '' of the star .
we identify the soft component with x - ray emission from the base of the jets .
in contrast , the harder _ counter jet
_ suggests stronger absorption by the extended gas disk .
a determination of the gas - to - dust ratio is in principle possible by measuring the differential absorption and extinction of the two jets .
finally , the _ hard stellar _ component is attributed to a ( flaring ) corona , with the excess photoelectric absorption due to dust - depleted accretion gas streams . the combined power of the resolved jets and the unresolved soft spectral component is of order @xmath35 erg s@xmath6 , similar to the x - ray output of a moderate t tauri star .
this emission , distributed above the accretion disk , may be an important contributor to x - ray heating and ionization of gaseous disk surfaces ( ( * ? ? ?
* glassgold 2004 ) ) .
we speculate that protostellar jets in general develop the same kind of jet x - ray emission , but these sources remain undetected close to the star because of strong photoelectric absorption . this work has been supported by the international space science institute in bern , the swiss national science foundation ( at , ma , mg : grants 20 - 66875.01 , 20 - 109255/1 , pp002110504 ) , asi / inaf ( palermo group , grant asi - inaf i/023/05/0 ) , and nasa ( ma , ss , dp : grants nng05gf92 g , go6 - 7003 ) .
this research is based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and the usa ( nasa ) .
the cxc x - ray observatory center is operated by the smithsonian astrophysical observatory for and on behalf of the nasa under contract nas8 - 03060 . | we report on accretion- and outflow - related x - rays from t tauri stars , based on results from the `` xmm - newton extended survey of the taurus molecular cloud . ''
x - rays potentially form in shocks of accretion streams near the stellar surface , although we hypothesize that direct interactions between the streams and magnetic coronae may occur as well .
we report on the discovery of a `` soft excess '' in accreting t tauri stars supporting these scenarios .
we further discuss a new type of x - ray source in jet - driving t tauri stars .
it shows a strongly absorbed coronal component and a very soft , weakly absorbed component probably related to shocks in microjets .
the excessive coronal absorption points to dust - depletion in the accretion streams . |
for complex compressible flows involving multiphysics phenomenons like e.g. high - speed elastoplasticity , multimaterial interaction , plasma , gas - particles etc . , a lagrangian description of the flow is generally preferred . to achieve robustness , some spatial remapping on a regular mesh may be added .
a particular case is the family of the so - called lagrange+remap schemes @xcite , also referred to as remapped lagrange solvers that apply a remap step on a reference ( eulerian ) mesh after each lagrangian time advance .
legacy codes implementing remapped lagrange solvers usually define thermodynamical variables at cell centres and velocity at cell nodes ( see figure [ fig:1 ] ) . in poncet et al .
@xcite , we have achieved a multicore node - based performance analysis of a reference lagrange - remap hydrodynamics solver used in industry . by analyzing each kernel of the whole algorithm , using roofline - type models @xcite on one side and refined execution cache memory ( ecm ) models @xcite , @xcite on the other side ,
we have been able not only to quantitatively predict the performance of the whole algorithm with relative errors in the single digit range but also to identify a set of features that limit the whole performance .
this can be roughly summarized into three points : 1 . for typical mesh sizes of real applications ,
spatially staggered variables involve a rather big amount of communication to / from cpu caches and memory with low arithmetic intensity , thus lowering the whole performance ; 2 .
usual alternating direction ( ad ) strategies ( see the appendix in @xcite ) or ad remapping procedures also generate too much communication with a loss of cpu occupancy .
3 . for multimaterial flows using vof - based interface reconstruction methods
, there is a strong loss of performance due to some array indirections and noncoalescent data in memory .
vectorization of such algorithms is also not trivial . from these observations and as a result of the analysis
, we decided to `` rethink '' lagrange - remap schemes , with possibly modifying some aspects of the solver in order to improve node - based performance of the hydrocode solver .
we have looked for alternative formulations that reduce communication and improve both arithmetic intensity and simd ( single instruction
multiple data ) property of the algorithm . in this paper
, we describe the process of redesign of lagrange - remap schemes leading to higher performance solvers .
actually , this redesign methodology also gave us ideas of innovative eulerian solvers .
the emerging methods , named lagrange - flux schemes , appear to be promising in the extended computational fluid dynamics community .
the paper is organized as follows . in section [ sec:2 ] ,
we first formulate the requirements for the design of better lagrange - remap schemes . in section [ sec:3lag ]
we give a description of the lagrange step and formulate it under a finite volume form . in section [ sec:3 ]
we focus on the remap step which is reformulated as a finite volume scheme with pure convective fluxes .
this interpretation is applied in section [ sec:4 ] to build the so - called lagrange - flux schemes .
we also discuss the important issue of achieving second order accuracy ( in both space and time ) . in section [ sec:5 ]
we comment the possible extension to multimaterial flow with the use of low - diffusive and accurate interface - capturing methods .
we will close the paper by some concluding remarks , work in progress and perspectives .
starting from `` legacy '' lagrange - remap solvers and related observed performance measurements , we want to improve the performance of these solvers by modifying some of the features of them but under some constraints and requirements : 1 . a lagrangian solver ( or description ) must be used ( allowing for multiphysics coupling ) .
2 . to reduce communication , we prefer using collocated cell - centered variables rather than a staggered scheme .
3 . to reduce communication , we prefer using a direct multidimensional remap solver rather than splitted alternating direction projections
the method can be simply extended to second - order accuracy ( in space and time ) .
the solver must be able to be naturally extended to multimaterial flows . before going further ,
let us comment the above requirements .
the second requirement should imply the use of a cell - centered lagrange solver .
fairly recently , desprs and mazeran in @xcite and maire and et al .
@xcite ( with high - order extension in @xcite ) have proposed pure cell - centered lagrangian solvers based on the reconstruction of nodal velocities . in our study , we will examine if it is possible to use approximate and simpler lagrangian solvers in the lagrange+remap context , in particular for the sake of performance .
the fourth assertion requires a full multidimensional remapping step , probably taking into account geometric elements ( deformation of cells and edges ) if we want to ensure high - order accuracy remapping . to summarize , our requirements are somewhat contradictory , and we have to find a good compromise between some simplifications - approximations and a loss of accuracy ( or properties ) of the numerical solver .
as example , let us consider the compressible euler equations for two - dimensional problems . denoting @xmath0 @xmath1 , @xmath2 and
@xmath3 the density , velocity , pressure and specific total energy respectively , the mass , momentum and energy conservation equations read @xmath4 where @xmath5 , @xmath6 , @xmath7 , @xmath8 and @xmath9 . for the sake of simplicity , we will use a perfect gas equation of state @xmath10 , @xmath11 $ ] .
the speed of sound @xmath12 is given by @xmath13 .
for any volume @xmath14 that moves with the fluid , from the reynolds transport theorem we have @xmath15 where @xmath16 is the normal unit vector exterior to @xmath17 .
this leads to a natural explicit finite volume scheme in the form @xmath18 & & \phantom{aaaaa } - \delta t^n \sum_{a^{n+{\frac{1}{2}},l}\subset \partial k^{n+{\frac{1}{2}},l } } |a^{n+{\frac{1}{2}},l}|\ , { \bm{\pi}}_a^{n+{\frac{1}{2}},l}\cdot \nu_a^{n+{\frac{1}{2}},l}. \label{eq:2 } \ ] ] in , the superscript `` l '' indicates the lagrange evolution of the quantity . any eulerian cell
@xmath19 is deformed into the lagrangian volume @xmath20 at time @xmath21 , and into the lagrangian volume @xmath22 at time @xmath23 .
the pressure flux terms through the edges @xmath24 are evaluated at time @xmath21 in order to get second - order accuracy in time .
of course , that means that we need a predictor step for the velocity field @xmath25 at time @xmath21 ( not written here for simplicity ) . [ [ notations . ] ] notations .
+ + + + + + + + + + to simplify , in all what follows we will use the notation @xmath26 .
the remapping step considers the remapping the fields @xmath27 defined at cell centers @xmath22 on the initial ( reference ) eulerian mesh with cells @xmath19 .
let us denote @xmath28 a linear operator that reconstructs piecewise polynomial functions from a discrete field @xmath29 defined at cell centers of the lagrangian mesh @xmath30 at time @xmath23 .
the remapping process consists in projecting the reconstructed field on piecewise - constant function on the eulerian mesh , according to the intergral formula @xmath31 practically , they are many ways to consider the projection operation .
one can assemble elementary projection contributions by computing the volume intersections between the reference mesh and the deformed mesh .
but this procedure requires the computation of all the geometrical elements .
moreover , the projection needs local tests of projection with conditional branching ( think about the very different cases of a compression and expansion ) . thus the procedure is not simd and potentially leads to a loss of performance .
the incremental remapping can also be interpreted as a transport / advection algorithm , as emphasized by dukowicz and baumgardner @xcite that appears to be better suited for simd treatments .
let us now write a different original formulation of the remapping step .
in this step , there is no time evolution of any quantity , and in some sense we have @xmath32 , that we write @xmath33 we decide to split up this equation into two substeps : * backward convection : @xmath34 * forward convection : @xmath35 each convection problem is well - posed on the time interval @xmath36 $ ] .
let us now focus into these two steps and the way to solve them .
after the lagrange step , if we solve the backward convection problem over a time interval @xmath37 using a lagrangian description , we have @xmath38 actually , from the cell @xmath22 we go back to the original cell @xmath19 with conservation of the conservative quantities . for @xmath39 ( conservation of mass )
, we have @xmath40 showing the variation of density by volume variation . for @xmath41 ,
it is easy to see that both velocity and specific total energy are kept unchanged is this step : @xmath42 thus , this step is clearly computationally inexpensive . from the discrete field @xmath43 defined on the eulerian cells
@xmath19 , we then solve the forward convection problem [ eq:5 ] over a time step @xmath37 under an eulerian description .
a standard finite volume discretization of the problem will lead to the classical time advance scheme @xmath44 for some interface values @xmath45 defined from the local neighbor values @xmath46 .
we finally get the expected eulerian values @xmath47 at time @xmath23 .
+ notice that from and we have also @xmath48 thus completely defining the remap step under the finite volume scheme form .
we find that we no more need any mesh intersection or geometric consideration to achieve the remapping process .
the finite volume form is now suitable for a straightforward vectorized simd treatment . from
it is easy to achieve second - order accuracy for the remapping step by usual finite volume tools ( muscl reconstruction + second - order accurate time advance scheme for example ) .
let us note that the lagrange+remap scheme is actually a conservative finite volume scheme : putting into gives for all @xmath49 : @xmath50 that can also be written @xmath51 we recognize in pressure - related fluxes and convective fluxes that define the whole numerical flux .
the finite volume formulation is attractive and seems rather simple at first sight .
but we should not forget that we have to compute a velocity lagrange vector field @xmath26 where the variables should be located at cell nodes to return a well - posed deformation .
moreover , expression involves geometric elements like the length of the deformed edges @xmath52 . among the rigorous collocated lagrangian solvers ,
let us mention the glace scheme by desprs - mazeran @xcite and the cell - centered eucclhyd solver by maire et al @xcite .
both are rather computationally expensive and their second - order accurate extension is not easy to achieve . although it is possible to couple these lagrangian solvers with the flux - balanced remapping formulation , it is also of interest to think about ways to simplify or approximate the lagrange step without losing second - order accuracy .
one of the difficulty in the analysis of lagrange - remap schemes is that , in some sense , space and time are coupled by the deformation process .
below , we derive a formulation that leads to a clear separation between space and time , in order to simply control the order of accuracy .
the idea is to make the time step tend to zero in the lagrange - remap scheme ( method of lines @xcite ) , then exhibit the instantaneous spatial numerical fluxes through the eulerian cell edges that will serve for the construction of an explicit finite volume scheme .
because the method needs an approximate riemann solver in lagrangian form , we will call it a lagrange - flux scheme .
from the intermediate conclusions of the discussion [ sec:44 ] above , we would like to be free from any rather expensive collocated lagrangian solver .
however , such a lagrangian solver seems necessary to correctly and accurately define the deformation velocity field @xmath53 at time @xmath21 . in
what follows , we are trying to deal with time accuracy in a different manner .
let us come back to the lagrange+remap formula .
let us consider a `` small '' time step @xmath54 that fulfils the usual stability cfl condition .
we have @xmath55 by making @xmath54 tend to zero , ( @xmath56 ) , we have @xmath57 , @xmath58 , @xmath59 , @xmath60 , then we get a semi - discretization in space of the conservation laws .
that can be seen as a particular method of lines ( @xcite ) : @xmath61 we get a classical finite volume method @xmath62 with a numerical flux @xmath63 whose components are @xmath64 in , pressure fluxes @xmath65 and interface normal velocities @xmath66 can be computed from an approximate riemann solver in lagrangian coordinates ( for example the lagrangian hll solver , see @xcite ) . then , the interface states @xmath67 should be computed from an upwind process according to the sign of the normal velocity @xmath68 .
this is interesting because the resulting flux has similarities with the so - called advection upstream splitting method ( ausm ) flux family proposed by liou @xcite , but the construction here is different and , in some sense , justifies the ausm splitting . to get higher - order accuracy in space
, one can use a standard muscl reconstruction + slope limiting process involving classical slope limiters like for example sweby s limiter function @xcite : @xmath69 with @xmath70 $ ] for achieving second order accuracy . at this stage , because there is no time discretization , everything is defined on the eulerian mesh and fluxes are located at the edges of the eulerian cells .
this is one originality of this scheme compared to the legacy staggered lagrange - remap scheme that has to use variables defined on the lagrangian cells . to get high - order accuracy in time
, one can then apply a standard high - order time advance scheme ( rk2 , etc . ) . for the second - order heun scheme for example
, we have the following algorithm : a hll approximate riemann solver @xcite in lagrangian coordinates can be used to easily compute interface pressure and velocity . for a local riemann problem made of a left state @xmath88 and a right state @xmath89 , the contact pressure @xmath90 is given by the formula @xmath91 and the normal contact velocity @xmath92 by @xmath93 leading to simple formulas easily implementable in the second - order heun time integration scheme .
as example , we test the lagrange - flux scheme presented in section [ sec:4 ] on few one - dimensional shock tube problems .
we use a runge - kutta 2 ( rk2 ) time integrator and a muscl reconstruction with the sweby slope limiter given in .
[ [ sods - shock - tube ] ] sod s shock tube @xcite + + + + + + + + + + + + + + + + + + + + + + + + the initial data defined on space interval @xmath94 $ ] is made of two constant states @xmath95 and @xmath96 with initial discontinuity at @xmath97 .
we successively test the method on two uniform mesh grids made of 100 and 400 cells respectively .
the final computational time is @xmath98 and we use a cfl number equal to 0.25 and a limiter coefficient @xmath99 . on figure
[ fig : sod ] , one can observe a nice behavior of the euler solver , with sharp discontinuities and a low numerical diffusion into the rarefaction fan even for the coarse grid .
[ [ two - rarefaction - shock - tube ] ] two - rarefaction shock tube + + + + + + + + + + + + + + + + + + + + + + + + + + the second reference example is a case of two moving - away rarefaction fans under near - vacuum conditions ( see toro @xcite ) . it is known that the roe scheme breaks down for this case .
the related riemann problem is made of the left state @xmath100 and right state @xmath101 .
the final time of @xmath102 .
we again test the method of a coarse mesh ( 200 points ) and a fine mesh ( 2000 points ) .
numerical results are given in figure [ fig : cavit ] .
the numerical scheme appears to be robust especially in near - vacuum zones where both density and pressure are close to zero .
[ [ case - with - sonic - rarefaction - and - supersonic - contact ] ] case with sonic rarefaction and supersonic contact + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the following shock tube case with initial data @xmath103 and @xmath96 generates a sonic 1-rarefaction , a supersonic 2-contact discontinuity and a 3-shock wave .
the final time is @xmath102 and we use 400 mesh points , @xmath104 .
numerical results show a good capture of the rarefaction wave , without any non - entropic expansion - shock ( see figure [ fig:4 ] ) . [
[ case - of - shock - shock - hypersonic - shock - tube ] ] case of shock - shock hypersonic shock tube + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this last shock tube problem is a violent flow case made of two hitting fluids with @xmath105 and @xmath106 .
both 1-wave and 3-wave are shock waves , and the right state has a mach number of order 40 .
final time is @xmath102 , we use 400 grid points and the limiter coefficient @xmath107 is here @xmath108 ( equivalent to the minmod limiter ) .
one can observe of very nice behavior of the solver : there is no pressure or velocity oscillations at the contact discontinuity , and the numerical scheme preserves the positivity of density , pressure and internal energy ( see figure [ fig:5 ] ) .
in this section , we compare the new lagrange - flux scheme to the reference ( staggered ) lagrange - remap scheme in terms of millions of cell updates per second ( denoted hereafter as mcups ) .
tests are performed on a standard @xmath109 cores intel sandy bridge server e5 - 2670 .
each core has a frequency of 2.6 ghz , and supports intel s avx ( advanced vector extension ) vector instructions . for multicore support
, we use the multithreading programming interface ` openmp ` . in the reference
staggered lagrange - remap solver ( see @xcite ) , thermodynamic variables are defined at grid cell centers while velocity variables are defined at mesh nodes . due to this staggered discretization and the alternating direction ( ad ) remapping procedures , this solver is decomposed into nine kernels .
this decomposition mechanically decreases the mean arithmetic intensity ( ai ) of the solver . on the other hand ,
the lagrange - flux algorithm consists in only two kernels with a relative high arithmetic intensity which leads to two compute - bound ( cb ) kernels . in the first kernel , named `
predictionlagrangeflux ( ) ` , an appropriate riemann solver is called , face fluxes are computed and variables are updated for the prediction . the second kernel , named ` correctionlagrangeflux ( ) ` , is close in terms of algorithmic steps , since it also uses a riemann solver , computes fluxes and updates the variables for the correction part of the solver . in order to assess the scalability and absolute performance of both schemes , we present in table [ tab_perfs ] a performance comparison study .
first , we notice that the baseline performance e.g. the single core absolute performance without vectorization is quite similar for the two schemes , as can be seen in the first column .
however , we the lagrange - flux scheme has a better scalability , due to both vectorization and multithreading : our lagrange - flux implementation achieves a speedup of 31.1x with 16 cores and avx vectorization ( while ideal speed - up is 64 ) whereas the reference lagrange - remap algorithm reaches a speed - up of only 14.8x .
this difference is mainly due to the memory - bound kernels composing the reference lagrange - remap scheme .
indeed , speedups due to avx vectorization and multithreading are not ideal for kernels with relatively low intensity since memory bandwidth is shared between cores . [ table : multicore ] .performance comparison between the reference lagrange - remap solver and the lagrange - flux solver in millions of cell updates per second ( mcups ) , using different machine configurations .
scalability ( last column ) is computed as the speedup of the multithreaded vectorized version compared to the baseline purely sequential version .
tests are performed for fine meshes , such that kernel data lies in dram memory .
the lagrange - flux solver exhibits superior scalability , because it has by design better arithmetic intensity . [
cols="<,^,^,^,^ " , ]
although this is not the aim and the scope of the present paper , we would like to give an outline of the possible extension of lagrange - flux schemes to compressible multimaterial / multifluid flows , i.e. flows that are composed of different immiscible fluids and separated by free boundaries . + for pure lagrange+remap schemes , usually vof - based interface reconstruction ( ir ) algorithms are used ( young s plic , etc . ) .
after the lagrangian evolution , for the cells that host more than one fluid , fluid interfaces are reconstructed . during the remapping step
, one has to evaluate the mass fluxes per material . from the computational point of view and computing performance
, this process generally slows down the whole performance because of many array indirections in memory and specific treatment into mixed cells along with the material interfaces . if the geometry of the lagrangian cells is not completely known ( as in the case of lagrange - flux schemes ) , anyway we have to proceed differently . a possibility is to use interface capturing ( ic ) schemes , e.g. conservative eulerian schemes that evaluate the convected mass fluxes through eulerian cell edges .
this can be achieved by the use of antidiffusive / low - diffusive advection solvers in the spirit of desprs - lagoutire s limited - downwind scheme @xcite of vofire @xcite . in a recent work @xcite , we have analyzed the origin of known artifacts and numerical interface instabilities for this type of solvers and concluded that the reconstruction of fully multidimensional gradients with multidimensional gradient limiters was necessary .
thus , we decided to use low - diffusive advection schemes with a multidimensional limiting process ( mlp ) in the spirit of @xcite .
the resulting method is quite accurate , shape - preserving and free from any artifact .
we show some numerical results in the two next subsections .
let us emphasize that the interface capturing strategy perfectly fits with the lagrange - flux flow description , and the resulting schemes are really suitable for vectorization ( simd feature ) with data coalescence into memory .
let us first present numerical results on a pure scalar linear advection problem .
the forward - backward advection case proposed by rider and kothe @xcite is a hat - shaped function which is advected and stretched into a rotating vector field , leading to a filament structure .
then by applying the opposite velocity field , one have to retrieve the initial disk shape . in figure
[ fig : kr2 ] , we show the numerical solutions obtained on a grid @xmath110 for both the passive scalar field of variable @xmath111 $ ] and the quantity @xmath112 that indicates the numerical spread rate of the diffuse interface .
one can conclude the good behavior of the method , providing both stability , accuracy and shape preservation . 0.5 .,title="fig : " ] 0.5 .,title="fig : " ] 0.5 .,title="fig : " ] 0.5 .,title="fig : " ] 0.5 .,title="fig : " ] 0.5 .,title="fig : " ] we then consider our interface capturing method for multifluid hydrodynamics .
because material mass fractions are advected , that is @xmath113 one can use the advection solver of these variables but we prefer dealing with the conservative form of the equations @xmath114 in order to enforce mass conservation ( see also @xcite ) .
it is known that eulerian interface - capturing schemes generally produce spurious pressure oscillations at material interfaces ( @xcite ) .
some authors propose locally non conservative approaches @xcite to prevent from any pressure oscillations . here
we have a full conservative eulerian strategy involving a specific limiting process which is free from any pressure oscillation at interfaces , providing strong robustness .
this will be explained in a next paper .
the multimaterial lagrange - flux scheme is tested on the reference `` triple point '' test case , found e.g. in loubre et al .
this problem is a three - state two - material 2d riemann problem in a rectangular vessel .
the simulation domain is @xmath115 as described in figure [ fig : triple ] .
the domain is splitted up into three regions @xmath116 , @xmath117 filled with two perfect gases leading to a two - material problem .
perfect gas equations of state are used with @xmath118 and @xmath119 . due to the density differences , two shocks in sub - domains @xmath120 and @xmath121
propagate with different speeds .
this create a shear along the initial contact discontinuity and the formation of a vorticity .
capturing the vorticity is of course the difficult part to compute .
we use a rather fine mesh made of @xmath122 points ( about 1.8 m cells ) . on figure [ fig : triple1 ] , we plot the density , pressure , temperature fields respectively and indicate the location of the three material zones .
one can observe a nice capture of both shocks and contact discontinuities .
the vortex is also captured accurately .
0.49 points .
final time is @xmath123 . ]
0.49 points .
final time is @xmath123 . ]
0.49 points .
final time is @xmath123 . ]
0.49 points .
final time is @xmath123 . ]
this paper is primarily focused on the redesign of lagrange - remap hydrodynamics solvers in order to achieve better hpc node - based performance .
we have reformulated the remapping step under a finite volume flux balance , allowing for a full simd algorithm . as an unintended outcome
, the analysis has lead us to the discovery of a new promising family of eulerian solvers the so - called lagrange - flux solvers that show simplicity of implementation , accuracy , and flexibility with a high - performance capability compared to the legacy staggered lagrange - remap scheme .
interface capturing methods can be easily plugged for solving multimaterial flow problems .
ongoing work is focused of the effective performance modeling , analysis and measurement of lagrange - flux schemes with comparison of reference `` legacy '' lagrange - remap solvers including multimaterial interface capturing on different multicore processor architectures .
because of the multicore+vectorization scalability of lagrange - flux schemes , one can also expect high - performance on manycore co - processors like graphics processing units ( gpu ) or intel mic
. this will be the aim of next developments .
r. poncet , m. peybernes , t. gasc and f. de vuyst , performance modeling of a compressible hydrodnamics solver on multicore cpus , proceedings of the int .
conf . on parallel computing parco2015 , edinburgh , 2015 ( in press ) .
hirt , a.a .
amsden and j.l .
cook , an arbitrary lagrangian
eulerian computing method for all flow speeds .
journal of computational physics , 14:227253 ( 1974 ) .
d. benson , computational methods in lagrangian and eulerian hydrocodes , cmame , 99(2 - 3 ) , 235394 ( 1992 ) .
d. youngs , the lagrange - remap method , in _ implicit large eddy simulation : computing turbulent flow dynamics _ , f.f .
grinstein , l.g .
margolin and w.j .
rider ( eds ) , cambridge university press ( 2007 ) .
s. williams , a. waterman , and d. patterson : roofline : an insightful visual performance model for multicore architectures , commun .
acm , 52 , pp 6576 ( 2009 ) .
j. treibig and g. hager , introducing a performance model for bandwidth - limited loop kernels .
proceedings of the workshop `` memory issues on multi- and manycore platforms '' at ppam 2009 , lecture notes in computer science , 6067 , pp . 615624 ( 2010 ) .
h. stengel , j. treibig , g. hager and g. wellein , quantifying performance bottlenecks of stencil computations using the execution - cache - memory model .
ics15 , the 29th int
. conf . on supercomputing , 2015 ,
doi : 10.1145/2751205.2751240 .
p. colella and p.r .
woodward , the numerical simulation of two - dimensional fluid flow with strong shocks , j. comput .
phys.,54:115173 ( 1984 ) .
b. desprs and c. mazeran , lagrangian gas dynamics in two dimensions and lagrangian systems , arch .
rational mech .
178 ( 2005 ) 327372 .
maire , r. abgrall , j. breil and j. ovadia , a cell - centered lagrangian scheme for compressible flow problems , siam j. sci .
29 ( 4 ) ( 2007 ) 17811824 .
maire , a high - order cell - centered lagrangian scheme for two - dimensional compressible fluid flows on unstructured meshes , j. comput .
228 ( 2009 ) 23912425 .
j. k. dukowicz and j. r. baumgardner , incremental remapping as a transport / advection algorithm .
j. comput .
phys . , 160 , 318335 ( 2000 ) .
w. e. schiesser , the numerical method of lines , academic press , isbn 0 - 12 - 624130 - 9 ( 1991 ) .
toro , riemann solvers and numerical methods for fluid dynamics , 3rd edition , springer ( 2010 ) .
liou , a sequel to ausm : ausm+ , j. comp .
phys . , 129(2 ) , 364382 ( 1996 ) .
sod , a survey of several finite difference methods for systems of nonlinear hyperbolic conservation laws " .
j. comput .
27 : 131 ( 1971 ) .
sweby , high resolution schemes using flux - limiters for hyperbolic conservation laws , siam j. num .
21 ( 5 ) : 9951011 ( 1984 ) .
b. desprs and f. lagoutire , contact discontinuity capturing schemes for linear advection and compressible gas dynamics , j. sci .
comp . , 16(4 ) , 479524 ( 2001 ) .
b. desprs , f. lagoutire , e. labourasse and i. marmajou , an antidissipative transport scheme on unstructured meshes for multicomponent flows , ijfv , 3065 ( 2010 ) .
f. de vuyst , m. bchereau , t. gasc , r. motte , m. peybernes and r. poncet , stable and accurate low - diffusive interface capturing advection schemes , submitted to proc . of the multimat2015 conference wrsburg , special issue of the ijnmf ( 2015 ) .
j.s . park and c. kim , multi - dimensional limiting process for discontinuous galerkin methods on unstructured grids , chapter in computational fluid dynamics 2010 , springer , 179184 ( 2011 ) .
a.j . rider and d.b .
kothe , reconstructing volume tracking , j. comput .
phys . , 141(2 ) , 112152 ( 1998 ) .
a. bernard - champmartin and f. de vuyst , a low diffusive lagrange - remap scheme for the simulation of violent air - water free - surface flows , j. of comput .
physics , 274 , 1949 ( 2014 ) .
r. abgrall , how to prevent pressure oscillations in multicomponent flow calculations : a quasi conservative approach , j. comp .
phys . , 125(1 ) , 150160 ( 1996 ) .
r. saurel and r. abgrall , a simple method for compressible multifluid flows , siam j. sci .
comput . , 21(3 ) , 11151145 ( 1999 ) .
c. farhat , a. rallu and s. shankaran , a higher - order generalized ghost fluid method for the poor for the three - dimensional two - phase flow computation of underwater implosions , j. comp .
phys , 227(16 ) , 7640 - 7700 ( 2008 ) .
m. bachmann , p. helluy , j. jung , h. mathis and s. mller , random sampling remap for compressible two - phase flows , computers and fluids , 86 , 275283 ( 2013 ) .
r. loubre , p .- h .
maire , m. shashkov , j. breil and s. galera , reale : a reconnection - based arbitrary - lagrangian - eulerian method , j. comp .
phys . , 229 , 47244761 ( 2010 ) . | in a recent paper @xcite , we have achieved the performance analysis of staggered lagrange - remap schemes , a class of solvers widely used for hydrodynamics applications .
this paper is devoted to the rethinking and redesign of the lagrange - remap process for achieving better performance using today s computing architectures . as an unintended outcome
, the analysis has lead us to the discovery of a new family of solvers the so - called lagrange - flux schemes that appear to be promising for the cfd community . |
as a first deliverable within their scope and work programme , the mobile phone work group of the trusted computing group ( tcg mpwg ) has published a specification @xcite , which offers new potentials for implementing trust in mobile computing platforms by introducing a new , hardware - based trust anchor for mobile phones and devices .
this trust anchor , called a mobile trusted module ( mtm ) , has properties and features comparable to a trusted platform module ( tpm , see @xcite ) .
concurrently the mpwg issued a much more universal security architecture for mobile phones and devices on a higher abstraction level .
the pertinent specification is called _ tcg mobile reference architecture ( ra ) _ @xcite and abstracts a trusted mobile platform as a set of tamper resistant trusted engines operating on behalf of different stakeholders .
this architecture offers a high degree on flexibility and modularity in design and implementation of the trusted components to all participants in hard- and software development .
an important aspect of the _ tcg mobile reference architecture _ is the potential to virtualise significant parts of a trusted mobile platform as trusted software applications and services .
the trusted execution chain for this rests on the mtm .
the implementation of this chip depends on the security requirements of its specific use - case . for high levels of protection and isolation ,
an mtm could be implemented as a slightly modified trusted platform module ( tpm ) .
this enables cost - effective implementation of new security - critical applications and various innovative business models , in both the mobile and generic computing domain @xcite .
the present paper discusses the main structural features of the ra , highlighting the capabilities of the mtm as the main functional building block . after this technology review ,
we propose two basic methods for usage of the ra , namely the set - up of a trusted subsystem on a device by a remote owner , and its migration from one device to another .
this paper is organised as follows . in section [ section :
tcg_mpwg_architecture ] , we explore the significant parts of the _ mpwg reference architecture_. it is divided into four parts .
subsection [ section : tcg_mpwg_architectural_overview ] gives an overview of the security architecture , and subsection [ section : mobile_trusted_module ] details the concepts of the proposed architectural approach for an mtm and the requirements to virtualise its functionality , whereby a high security and isolation level is maintained .
furthermore , we propose a model for remote stakeholder take - ownership in [ section : ro_takeownership ] and migration of trusted subsystems in [ section : migration_remote_stakeholder_subsystems ] . in section
[ section : design_trusted_engines ] , we show how such an architecture can be implemented on trustworthy operating platforms .
the tcg mpwg has developed an architecture on a high level of abstraction for a trusted mobile platform , which offers numerous variations for design and implementation . in this section ,
we reflect essential parts of this architecture and an overview of significant platform components in terms of our objective .
a trusted mobile platform is characterised as a set of multiple tamper - resistant engines , each acting on behalf of a different stakeholder .
broadly , such an platform has several major components : trusted engines @xmath0 , trusted services @xmath1 customised by trusted resources @xmath2 .
a general trusted mobile platform is illustrated in figure [ fig : tcg_mpwg_trusted_mopile_platform_architecture ] .
we define a trusted subsystem @xmath3 as a logical unit of a trusted engine together with its interrelated hardware compartment .
a @xmath4 of a stakeholder @xmath5 can formally described by a tuple @xmath6 in each trusted subsystem @xmath4 either a remote or local entity acts as a stakeholder , who is able to configure its own subsystem and define his security policy @xmath7 and system configuration @xmath8 within an isolated and protected environment .
the _ mpwg reference architecture _ specifies the following principal entities : the local stakeholders _ device owner @xmath9 _ and _ user @xmath10 _ ; and the remote stakeholders _ device manufacturer @xmath11 _ , and more general _ remote owners @xmath12 _ ( e.g. a communication carrier , or service provider ) .
the functionality of a @xmath4 is either based on dedicated resources of an embedded engine @xmath13 , or may be provided by trusted services @xmath14 of external engines .
each subsystem is able to enforce its security policy @xmath7 and subsystem configuration @xmath8 . as a consequence
, the functionality of a trusted subsystem @xmath15 is constrained by the available resources @xmath16 with their derived trusted services @xmath17 , by the offered functionality of external trusted services @xmath14 , by the security policy @xmath7 , and finally the system configuration @xmath8 of an engine s stakeholder .
all internal functions executed inside @xmath18 are isolated from other subsystems by the underlying security layer and is only accessible if a proper service interface is defined and exported .
a @xmath18 relies on the reputation of the stakeholder @xmath5 as basis for that trust .
therefore , each stakeholder issues a security policy @xmath7 and a set of credentials belonging to embedded trusted components of its subsystem @xmath18 .
this policy contains reference measurements ( rim ) , quality assertions and security - critical requirements . the most important concept within the _ mpwg reference architecture
_ is that of trusted engines . the purpose of a trusted engine is to provide confidence in all its embedded services ,
which are internally or externally provided by the engine .
it is a protected entity on behalf of a specific stakeholder that has special abilities to manipulate and store data , and to provide evidence of its trustworthiness and the current state of the engine .
figure [ fig : tcg_mpwg_generic_engine ] shows a generic trusted engine . in general , each engine has at least following abilities : * implement arbitrary software functionalities as trusted and/or normal services , * provide the evidence for its trustworthiness , * report the evidence of its current state , * obtain and use endorsement keys ( ek ) and/or attestation identity keys ( aik ) , * access a set of trusted resources , and * import and/or export services , shielded capabilities and protected functionality . in order to establish a definite categorisation
, the mpwg differentiates engines according to their functional dispensability .
therefore , an engine is either dedicated to a mandatory ( of @xmath9 or @xmath11 ) or a discretionary domain ( of @xmath9 ) . engines inside a mandatory domain are permanently located on a trusted platform and hold security - critical and essential functionality .
all essential services of a trusted mobile platform should be located inside the mandatory domain , which does _ not permit a local stakeholder to remove a remote owner from the engine_. mandatory engines have access to a _ mobile remote owner trusted module ( mrtm ) _ to guarantee that a valid and trustworthy engine state is always present .
non - essential engines and services are replaceable by the device owner @xmath9 and should be located inside the discretionary domain .
engines inside the discretionary domain are controlled by the device owner @xmath9 .
discretionary engines are required to be supported by a _
mobile local owner trusted module ( mltm)_. as illustrated in figure [ fig : tcg_mpwg_generic_engine ] , an internal trusted service has access to several trusted resources .
the tcg calls these resources _ root - of - trusts ( rot ) _ representing the trusted components acting on base of the trusted execution chain and providing functionality for measurement , storing , reporting , verification and enforcement that affect the trustworthiness of the platform .
the following rots are defined for the mobile domain : * root of trust for storage ( rts ) , * root of trust for reporting ( rtr ) , * root of trust for measurement ( rtm ) , * root of trust for verification ( rtv ) , and * root of trust for enforcement ( rte ) each rot vouches its trustworthiness either directly by supplied secrets ( ek , aik ) and associated credentials , which are only accessible by authenticated subjects of the stakeholder , or indirectly by measurements of other trusted resources .
these resources are only mutable by authorised entities of a stakeholder . in this paper
, we group several logically self - contained rots to simplify the presentation of interfaces and the communication layer . in a typical arrangement ,
the rts and rtr represent one unit , while the rtm and rtv build another unit within an @xmath19 .
however , note that the rtv and the rtm depend on protected storage mechanisms , which are provided by the rts .
thus , it is also plausible to implement all rots together as a common unit within an engine .
* rts / rtr * are the trusted resources that are responsible for secure storage and reliable reporting of information about the state of trusted mobile platform .
an rts provide pcrs and protected storage for an engine and stores the measurements made by the rtm , cryptographic keys , and security sensitive data .
an rtr signs the measurements with cryptographic signature keys of @xmath20 .
* rtm / rtv * in general , an rtm is a reliable instance to measure software components and provide evidence of the current state of a trusted engine and its embedded services . in the mobile domain , to avoid communication costs , this functionality is extended by a local verifier , which checks the measurements against a given _ reference integrity metrics _ ( rim ) .
this process can be done instantly as the measurements are performed employing a combination of rtm and rtv .
figure [ fig : tcg_mpwg_measurement_and_verification ] depicts such a _
measure@xmath21verify@xmath21extend _ process .
an * rte * is required if an engine uses allocated resources and services . in this case , such rot acts as a trusted boot loader and ensures the availability of all allocated trusted resources and services within that trusted subsystem .
a trusted engine integrates all functionality by customising available platform resources as software services .
such a service offers computation , storage , or communication channels to other internal or external services and applications based on dedicated or allocated resources .
the mpwg categorises them into : trusted , normal , and measured services .
a trusted service customises trusted resources .
thus , a trusted service is implicitly supplied with an @xmath22 or @xmath23 in order to attest its trustworthiness .
trusted services are intended to provide reliable measurements of their current state and to provide evidence of the state of other normal services or resources .
normal services are customising normal resources and implement functionality , but are not able to provide evidence of their trustworthiness by own capabilities .
however , normal services can access internal trusted services to use their provided functionality .
therefore , an internal normal services is able to vouch its trustworthiness by associated integrity metrics that have been measured by a trusted service . the generic term _
mobile trusted module ( mtm ) _ refers to a dedicated hardware - based trust - anchor .
it is typically composed of an rts and rtr and has characteristics comparable to a tpm .
according to their design objective the mpwg distinguishes between mrtm and mltm .
both must support a subset of tpm commands as specified in @xcite . additionally , an mrtm has to support a set of commands to enable local verification and specific mobile device functionality .
the _ tcg mpwg reference architecture _ does not exclude to utilise a tpm v1.2 ( or even a tpm v1.1 ) as an mtm , if an appropriate interface consisting of a set of commands conforming to the mpwg specification and associated data structures are provided .
although it is possible to implement this architecture upon a standard tpm , we here focus on a more sophisticated solution based on a trustworthy computing platform such as emscb / turaya @xcite . in this context , we expect three different solutions for isolation , key management and protection of @xmath19 .
a * standard tpm - based model * uses a non - modified standard tpm to build the trusted computing base .
the secret keys are stored into a single key - hierarchy on behalf of @xmath9 as specified in @xcite . in this case ,
an adversary or malicious local owner may be able to access the secret keys of a remote stakeholder and take control of a remote owner compartment .
a @xmath9 can also disable the whole mtm or corrupt mandatory engines of remote stakeholders .
a * software - based mtm - emulation model * uses a software - based allocated @xmath24-emulation with an isolated key - hierarchy .
all sensitive and security - critical , such as @xmath22 or @xmath25 , are only protected by software mechanisms outside of the tamper - resistant environment of a dedicated mtm @xcite . *
generic mtm - based model supporting multiple stakeholders and virtual mtms . * in order to circumvent resulting drawbacks and mitigate attacks , we favour a solution with a higher level of security .
for this reason , we adopt the proposed secure co - processor variant of @xcite and describe a generic mtm with support for multiple stakeholder environments . in a cost - efficient scenario ,
the trusted mobile platform is implementable based on a single generic mtm and several virtualised mtms for each trusted engine .
hence , at least one dedicated mtm has to be available and additionally a unique vmtm has to be instantiated in each trusted subsystem @xmath26 . in such case ,
a physically bounded mtm in the platform acts as a master trust anchor and offers mrtm and mltm functionality with respect to its specific use case .
a trusted software layer offers a _ vmtm proxy service _ to all embedded trusted engines @xmath27 .
the main task of this service is to route mtm commands from a @xmath27 to its dedicated instance @xmath28 .
the advantage is that all security - critical mtm commands are tunnelled to @xmath28 and are executed within the protected environment of the dedicated mtm .
figure [ fig : tcg_mpwg_vmtm_virtualization ] illustrates the architecture of a generic mtm with isolated vmtm compartments .
this architecture requires a slightly modified tpm .
mainly , we add a trusted component , the _ vmtm instance manager _ , which is responsible to separate vmtm instances from each other .
this includes administration , isolated execution , memory management and access control for each stakeholder compartment .
thus , a vmtm instance is able to hold an autonomous and hardware - protected key hierarchy to store its secrets and protect the execution of security - critical data ( e.g. signature and encryption algorithms ) .
the take - ownership operation establishes the trust relationship between a stakeholder and trusted mobile platform .
currently , the _ mpwg reference architecture _ does not define how a remote owner is to perform this initial setup and take - ownership of its @xmath29 .
hence we propose a method in this section .
the main idea behind our procedure is to install and instantiate a blank trusted subsystem @xmath30 containing a pristine engine @xmath31 with a set of generic trusted services @xmath32 .
this subsystem is then certified by a remote owner , if the platform is able to provide evidence of its pristine configuration and policy conformance with respect to @xmath12 .
figure [ fig : tcg_mpwg_vmtm_takeownership_sequence ] illustrates this process , which we now descirbe . * platform and protocol precondition . * in a preliminary stage , the trusted mobile platform has carried out the boot process and has loaded the trusted computing base and the engine @xmath33 with its trusted services .
the trusted platform has checked that the installed hardware and running software are in a trustworthy state and configuration .
it is able to report and attest this state , if challenged by an authorised entity .
* remote stakeholder take - ownership protocol .
* in the first phase , the trusted engine @xmath33 carries out a take - ownership preparation for the remote stakeholder .
a blank engine @xmath31 is installed and booted by the @xmath34 , and a clean @xmath35 instance is activated inside the dedicated @xmath24 .
an initial setup prepares the pristine engine @xmath31 .
a endorsement key - pair @xmath36 is generated within @xmath35 with a corresponding certificate @xmath37 and @xmath38 . ] .
next , @xmath31 performs an attestation of its current state .
the attestation can be done by the local verifier @xmath39 inside the @xmath40 using @xmath41 certificates of the remote stakeholder @xmath12 .
if no suitable @xmath41 and corresponding @xmath41-certificate are available for an pristine engine , alternatively a remote attestation with an associated privacy ca is also possible .
@xmath31 creates a symmetric key @xmath42 and encrypts the public portion of the endorsement key @xmath36 , the corresponding certificate @xmath37 , attestation and purpose information .
next , @xmath31 encrypts @xmath42 with a public key @xmath43 and sends both messages to the remote owner . after reception by the remote stakeholder
, the messages are decrypted using the private portion of key @xmath43 .
we assume that this key is either available through a protected communication channel or pre - installed by the device manufacturer . in a next step ,
@xmath12 verifies the attestation data and checks the intended purpose of @xmath31 .
if the engine and device attestation data is valid and the intended purpose is acceptable , the @xmath12 generates an individual security policy @xmath44 .
the @xmath12 signs the @xmath37 and creates @xmath41 certificates for local verification of a complete @xmath19 .
furthermore , @xmath12 creates a setup configuration @xmath45 , which enforces the engine to individualise its services and complete its configuration with respect to the intended purpose and given security policy . in this step , @xmath12 encrypts the messages with the public portion of the @xmath46 and transfers this package to the engine @xmath47 .
finally , the trusted engine @xmath31 decrypts the received package and installs it inside the @xmath48 and thus completes its instantiation . if a stakeholder wants to move a source @xmath49 to another mtm - enabled platform , for instance to port user credentials from device to device , all security - critical information including the storage root key ( srk ) has to be migrated to the target @xmath50 . in our scenario
, we assume the same remote owner ( e.g. mobile network operator ) on both subsystems @xmath51 and @xmath52 .
to be able to securely migrate the srk , we suggest a modification of the current mpwg specification to allow _ inter - stakeholder - migration _ of a complete isolated key hierarchy . thus , an isolated key hierarchy is ( 1 ) migratable between environments of identical stakeholders , ( 2 ) if and only if an entitling security policy on both platforms exists .
the advantage of migration between identical stakeholder subsystems is that the migration process does nt require a trusted third party .
we only involve the owner in combination with local verification mechanisms of the @xmath53 to migrate the trusted subsystem ( including the srk ) to another platform .
this enables for instance direct , device - to - device porting of credentials , e.g. using short - range communication .
we here propose a complete , multilateral and secure migration protocol , which is illustrated in figure [ fig : tcg_mpwg_vmtm_migration_protocol ]
. * platform and protocol precondition . * similar to section [ section : ro_takeownership ] , the trusted mobile platform has carried out the same initial steps as mentioned above .
furthermore , the remote owner has performed an remote take - ownership procedure as described in [ section : ro_takeownership ] .
* trusted subsystem migration protocol . * at the beginning of the migration protocol , the device owner @xmath54 of the source platform @xmath55 initialises the migration procedure and requests an appropriate migration service of @xmath51 .
next , the trusted platform @xmath55 is instructed by @xmath53 to establish a secure channel to the target platform @xmath56 . after the connection is available
, @xmath51 activates the corresponding migration service of @xmath52 to perform the import procedure .
thereupon , the target subsystem @xmath50 performs a local verification of @xmath51 .
if revoked , it replies with an error message and halts the protocol .
otherwise @xmath57 requests an confirmation from the local owner @xmath58 .
next , the target subsystem @xmath52 generates a nonce @xmath59 . in order to provide evidence of its trustworthiness
, @xmath52 sends all necessary information to the source subsystem @xmath51 .
this includes the current state @xmath60 , a certificate of @xmath52 , security policy @xmath61 and the nonce @xmath59 .
having received the target subsystem s message , @xmath51 verifies the state of @xmath52 .
if the target system is in a trustworthy state and holds an acceptable security policy and system configuration , the current state of @xmath51 is locked to nonce @xmath59 . the @xmath51 generates a symmetric migration key @xmath62 , serialises its instance and encrypts it with the migration key , which is bound to an acceptable configuration of @xmath52 .
next , the key - blob and the encrypted instance are sent to the destination @xmath52 . in particular , this includes the whole isolated key - hierarchy @xmath63 with @xmath64 , the security policy @xmath65 , and the required subsystem configuration @xmath66 . finally , the target subsystem @xmath52 decrypts the received blob and uses @xmath67 as its own @xmath25 .
the subsystem verifies the obtained security policy @xmath65 and the subsystem configuration @xmath66 . with this information
, @xmath52 rebuilds the internal structure of the source .
the source system should then be notified of the success of migration and ultimately delete the migrated key hierarchy ( or even do it before sending the migration package as indicated for simplicity in figure [ fig : tcg_mpwg_vmtm_migration_protocol ] ) .
otherwise one obtains replicated trusted subsystems , by themselves indistinguishable to the remote owner . but
this may depend on the policies to be enforced in the particular use case .
a prototypical implementation of the trusted engines and the specified trusted services was realised as an extension to the existing emscb / turaya computing platform .
turaya is an implementation of the emscb security architecture .
it provides fundamental security mechanisms and a protected and isolated execution environment , which meet the requirements of the _ mpwg reference architecture _ @xcite .
figure [ fig : mtm_vsim_emscb_platform ] illustrates our model , in which a hypervisor / microkernel executes a legacy operating system in coexistence with a running instance of the emscb - based security architecture .
the latter controls a virtual machine with several trusted engines and services compliant to the mpwg requirements @xcite . in the following paragraphs
, we outline the significant platform layers concerning our approach .
the * hardware layer * of our model includes a generic mtm as described in section [ section : mobile_trusted_module ] , in addition to conventional hardware components .
this mtm acts as a dedicated master trust anchor for the complete trusted mobile platform .
the * virtualisation layer * provides generic hardware abstraction , between the physical hardware of a trusted mobile platform and the _ trusted software layer _ below .
the emscb project supports microkernels of the l4-family @xcite such as hypervisors @xcite . in general , all solutions provide mechanisms for resource management , inter - process - communication ( ipc ) , virtual machines , memory management and scheduling . in our specific case ,
the virtualisation layer includes also a fully functional device driver for a dedicated generic mtm .
furthermore , it is responsible for instantiation of both the trusted software layer and the legacy operating system .
the * trusted software layer * provides security functionality and is responsible for isolation of embedded applications and software compartments .
it also implements the _ vmtm proxy service _ as described in section [ section : mobile_trusted_module ] .
currently , emscb / turaya provide an excellent foundation by its security services ( trust manager , compartment manager , storage manager ) , which are required by the rtr and rtv , protected storage and trusted engines management agent of @xmath68 .
therefore , it is reasonable to build the significant parts of the device manufacturer engine @xmath68 within this layer . trusted engines @xmath27 within the * application layer
* are implemented as parallel and isolated l4linux compartments @xcite on behalf of different stakeholders .
each compartment has access to its vmtm instance through an embedded client - side device driver .
this driver constrains the functionality with respect to its specific use case ( mrtlm or mltm ) .
furthermore , @xmath27 has an @xmath69 , which is responsible for building all required allocated trusted resources and services depending of its specific system configuration @xmath70 and security policy @xmath71 .
we have introduced the trusted engines and mtms in terms of our objective . in this context , we have exposed significant parts of the mpwg reference architecture and how it can be implemented on a ( very slightly modified ) tpm trust - anchor .
we have shown how to deploy trusted virtualised compartments on devices and exhibited basic operations required in the mobile domain , such as migration . using a vmtm in lieu of a subscriber identity module ( sim )
as a trusted and protected software allows expansion to a much wider field of authentication and identification management systems even on standard pc platforms @xcite .
supporting online transactions by authentication via credentials held in a vmtm may be one attractive use case . however , there are some privacy and security challenges associated with this implementation on a desktop computer , which require further research
. finally , replacing sims / usims by multi - purpose vsims may be attractive even for genuine mobile devices . trusted computing group , _ tcg mobile reference architecture_. specification version 1.0 , revision 1 . 12 june 2007 .
nicolai kuntze and andreas u. schmidt .
transitive trust in mobile scenarios . in gnter
mller , editor , _ proceedings of the international conference on emerging trends in information and communication security ( etrics 2006 ) _ , volume 3995 of _ lecture notes in computer science ( lncs ) _ , pages 7385 .
springer - verlag , 2006 .
kuntze , n. , schmidt , a.u .
: trusted computing in mobile action . in venter , h.s .
, eloff , j.h.p . , labuschagne , l. , eloff , m.m .
, eds . : proceedings of the information security south africa ( issa ) conference ( 2006 ) nicolai kuntze and andreas u. schmidt . trusted ticket systems and applications . in _ to appear in : new approaches for security , privacy , and trust in complex systems .
proceedings of the ifip sec2007 .
sandton , south africa 14 - 16 may 2007_. springer - verlag , 2007 . | in its recently published _ tcg mobile reference architecture _ , the tcg mobile phone work group specifies a new concept to enable trust into future mobile devices . for this purpose
, the tcg devises a trusted mobile platform as a set of trusted engines on behalf of different stakeholders supported by a physical trust - anchor . in this paper , we present our perception on this emerging specification .
we propose an approach for the practical design and implementation of this concept and how to deploy it to a trustworthy operating platform .
in particular we propose a method for the take - ownership of a device by the user and the migration ( i.e. , portability ) of user credentials between devices . |
since the discovery of high - temperature superconductivity , doped antiferromagnets ( af ) have been studied intensively .
it is widely accepted that many properties of the superconducting cuprates are determined by the hole - doped cuo@xmath3 planes .
a number of experiments @xcite indicate that the cuprates are near a quantum - critical point of antiferromagnetic instability : the undoped materials are known to be antiferromagnetic mott - hubbard insulators , whereas hole doping destroys the antiferromagnetic long - range order ( lro ) and leads to superconductivity .
the investigation of the interplay between long - range magnetic order and quantum disorder is therefore of great theoretical interest .
a model system which shows a quantum transition between an ordered and a disordered magnetic phase is the @xmath4 bilayer antiferromagnet @xcite . here each of the two planes is composed of a nearest - neighbor heisenberg model with coupling constant @xmath5 .
the spins of corresponding sites of each layer are coupled antiferromagnetically with a coupling constant @xmath6 . in the limit of small @xmath7
the model describes two weakly coupled af planes . at @xmath8 this system possesses af long - range order and gapless goldstone excitations . in the opposite case of large @xmath7 , pairs of spins interacting via @xmath6 form spin singlets being weakly coupled by @xmath5 . then the spin excitations are gapped triplet modes , there is no magnetic lro . at a critical ratio @xmath9 a quantum transition between the two phases occurs which is believed to be of the o(3 ) universality class @xcite .
the bilayer antiferromagnet has been studied by a number of numerical and analytical techniques .
quantum monte carlo calculations @xcite and series expansions @xcite yield an order - disorder transition point of @xmath10 . a similar result has also been obtained analytically using a diagrammatic approach to account for the hard - core interaction between triplet excitations @xcite .
bond - operator mean - field theory has recently been applied to the bilayer heisenberg af @xcite and gives a transition point of @xmath11 .
note that schwinger boson mean - field theory @xcite predicts a very large value of @xmath12 , and also self - consistent spin - wave theory @xcite , which yields @xmath13 , fails to reproduce the numerical results .
chubukov and morr @xcite have pointed out that this discrepancy is due to the neglect of longitudinal spin fluctuations in the conventional spin - wave approach . in this paper
we discuss the bilayer antiferromagnet at zero temperature with hole doping as an additional degree of freedom .
we consider the standard @xmath1 model on a bilayer square lattice consisting of @xmath14 sites per plane , so the total number of lattice sites is @xmath15 .
each pair of corresponding sites in different planes is considered to form a rung , so we have @xmath14 rungs . more precisely , the hamiltonian reads @xmath16 here and in the following @xmath17 and @xmath18 denote @xmath19 of the bilayer lattice , @xmath20 denotes a summation over all pairs of nearest - neighbor rungs .
@xmath21 is the plane index , so @xmath22 denotes a lattice site .
@xmath23 is the electronic spin operator and @xmath24 the electron number operator at site @xmath22 .
the electron operators @xmath25 exclude double occupancies , .
we choose the @xmath26-axis along the direction of the rungs , i.e. , perpendicular to the planes .
first we want to comment the zero - temperature phase diagram of the doped bilayer @xmath1 model which has to our knowledge not been systematically studied up to now .
the following non - thermal control parameters can be considered : the ratio @xmath7 , the doping level @xmath27 , and the relative hopping strength @xmath28 . the doped single - layer antiferromagnet is known to exhibit a strong dependence of magnetic properties on the hole concentration @xmath27 . with increasing hole concentration the staggered magnetization decreases and vanishes at a critical hole concentration @xmath29 of a few percent where the system becomes paramagnetic @xcite .
( this is consistent with experiments on high-@xmath30 superconductors @xcite . ) on the other hand , in the undoped limit a large interplane coupling @xmath6 also destroys the antiferromagnetic lro as discussed above .
thus it is likely that the system shows ( i ) af lro accompanied by spontaneous symmetry breaking and the existence of massless goldstone modes in the region of very small doping and small @xmath6 .
otherwise it is paramagnetic ; here one has to distinguish between ( ii ) a spin - gapped phase which occurs at small @xmath27 , small hopping @xmath31 , and large interplane coupling , and ( iii ) a gapless phase at larger doping and small interplane coupling . the gapped phase ( ii )
is dominated by interplane singlet pairs ( incompressible spin fluid ) whereas the gapless phase ( iii ) shows only weak interplane coupling and behaves mainly like a single layer system at @xmath32 ( compressible spin fluid ) .
( issues like pseudogap features and transport properties will not be adressed here . ) in the present work we restrict ourselves to the zero - doping limit ( one hole ) .
so the magnetic ( bulk ) properties will not be affected by the doped hole , i.e. , we probe the transition between the antiferromagnetic ( i ) and the gapped ( incompressible ) paramagnetic phase ( ii ) .
the dynamical properties of holes in high-@xmath30 superconductors have been subject to a lot of experimental and theoretical work .
angle - resolved photoemission ( arpes ) experiments at zero@xcite or small doping@xcite indicate a quasiparticle band with a small bandwidth providing evidence for strong electronic correlations in the high-@xmath30 compounds .
a large number of numerical and analytical studies reveal that a single hole in an af spin background has non - trivial properties : the spectral function consists of a pronounced coherent peak at the bottom of the spectrum and a incoherent background at higher energies .
the coherent peak can be associated with the motion of a dressed hole , i.e. , a hole surrounded by spin defects ( `` spin polaron '' ) @xcite .
the hole motion in bilayer antiferromagnets has been studied in a few papers and only for a small parameter range : using self - consistent born approximation ( scba ) it has been shown @xcite that small interplane coupling only weakly affects the hole properties .
exact diagonalization studies @xcite have indicated a larger effect of the interplane coupling at finite doping
. however , the system sizes accessible to numerical methods ( @xmath33 in ref . )
do not allow for the study of small hole densities and arbitrary momenta .
in this paper we prefer an analytical approach to the one - hole problem which is based on the picture of the spin - bag quasiparticle ( qp ) or magnetic polaron @xcite .
the magnetic background for the hole motion will be modelled within a modified bond - operator representation of the spins on each rung . in the ordered phase one type of triplet bosons condenses which will be included in the modified basis operators .
so our excitation operators continuously interpolate between the triplet excitations of a singlet product state and the transverse and longitudinal excitations of a nel - ordered state .
the deviations of the spin background caused by the hole motion are described by a set of path operators @xcite .
the one - hole spectral function will be evaluated using a cumulant version @xcite of mori - zwanzig projection technique@xcite .
the paper is organized as follows : in sec .
ii we propose a ground - state wave function for the undoped bilayer antiferromagnet . in sec .
iii we develop the hamiltonian for the doped system in this new operator representation and discuss the various hole motion processes .
iv focuses on the motion of a single hole in an otherwise half - filled bilayer antiferromagnet .
the one - hole spectral function for this case corresponds directly to the result of an arpes experiment on an undoped system . after a brief sketch of the projection technique
we investigate the spectral properties in dependence of the magnetic background , especially when crossing the phase boundary between gapped paramagnet and antiferromagnet .
the results show that the character of the one - hole spectrum is mainly determined by the local spin environment of the hole whereas the spectrum changes only weakly when crossing the order - disorder transition .
a conclusion will close the paper .
to begin with , we consider the case of half - filling . using bond operators we derive a wave function which describes the ground states in both the quantum disordered phase ( at large @xmath6 ) and the nel phase ( at small @xmath6 ) .
the wave function will be constructed within a cumulant formalism @xcite .
it bases on an expansion around a product of rung states which are singlets in the disordered phase and sums of singlets and @xmath26-type triplets in the ordered phase .
this wave function will be used in sec .
iv as `` magnetic background '' to calculate dynamical properties of a single hole . in the limit of vanishing intraplane coupling , @xmath34 , pairs of spins on each rung
form a singlet ground state .
the excitations are localized triplets with an energy gap @xmath6 .
switching on @xmath5 leads to an interaction of triplets on neighboring rungs and to a dispersion of the triplet excitations . to describe this dimerized phase we employ a bond operator representation @xcite for the spins per rung . for each rung
@xmath17 containing two @xmath35 spins @xmath36 , @xmath37 we introduce operators for creation of a singlet and three triplet states out of the vacuum @xmath38 : @xmath39 2 then the following representation holds @xcite : @xmath40 the algebra of the operators @xmath41 has to be specified to reproduce the correct algebra for the spin operators . here
bond operators satisfying the usual bosonic commutation relations will be chosen @xcite . in order to ensure that the physical states are either singlets or triplets one has to impose the condition :
@xmath42 the singlet product ground state for @xmath34 can be written as @xmath43 we can now substitute the operator representation ( [ spin_rep ] ) into the heisenberg hamiltonian for the bilayer and obtain the following hamiltonian @xmath44 in the bond ( rung ) operator representation @xcite : @xmath45 the sum @xmath20 runs over pairs of neighboring rungs of the lattice .
the ground state of @xmath46 is given by the product state @xmath47 defined in eq .
( [ singlet_prod ] ) .
approaching the quantum critical point the triplet excitations will become gapless at momentum @xmath48 , see e.g. refs . .
the magnetic quantum phase transition to an antiferromagnetically ordered state includes spontaneous symmetry breaking and can be described via the condensation of triplet bosons in one particular direction ( which has to be induced by an infinitesimal staggered field ) . neglecting fluctuations a symmetry - broken state with a condensate of @xmath26-type triplet bosons
can be written as @xmath49 where the spontaneous staggered magnetization points in @xmath26-direction .
the exponential term accounts for boson condensation , the condensation amplitude is given by @xmath50 .
@xmath51 means rotational symmetry in spin space ( then @xmath52 transforms the singlet product state @xmath47 into a nel state .
@xmath53 is the af ordering wave vector .
the basic idea for a proper description of the excitations of the product state @xmath54 ( [ phi_symmb ] ) is now to transform the basis states on each rung .
we replace the basis operators @xmath55 for singlets and @xmath56 for @xmath26-triplets by @xmath57 the new basis per rung consists of the operators @xmath58 which still satisfy bosonic commutation relations . for @xmath51
this basis reproduces the usual bond - boson basis . for non - zero @xmath50
the state @xmath59 interpolates between the rung singlet and a ( local ) nel state , the state @xmath60 is orthogonal to @xmath59 .
the product state @xmath54 ( [ phi_symmb ] ) can now be written as @xmath61 for @xmath62 this state is an antiferromagnetic nel state .
its excitations are given by @xmath63 , @xmath64 , and @xmath65 . here
, @xmath66 create one spin flip in one of the planes and can be regarded as transverse spin fluctuations .
the operator @xmath65 flips both spins on one rung which can be interpreted as longitudinal spin fluctuation .
the basis parameter @xmath50 has to be determined separately which will be discussed in the next subsection .
below it is shown that in the ordered phase @xmath50 varies continuously between @xmath67 for @xmath68 and @xmath51 at the transition to the disordered phase .
note that the idea of a condensate for one type of bosons has also been employed by chubukov and morr @xcite who applied a modified version of non - linear spin - wave theory to the problem of the undoped bilayer antiferromagnet . using ( [ basis_def ] )
we can reformulate the hamiltonian for the ( undoped ) bilayer heisenberg model in terms of the new basis operators @xmath58 .
note that the splitting of the hamiltonian is different from ( [ h_undop_diso ] ) .
@xmath69 2 @xmath70 and @xmath71 denote the sublattices of the square lattice , i.e , @xmath72 for @xmath73 or @xmath71 , respectively . for @xmath51
the hamiltonian ( [ h_undop_o ] ) is identical to ( [ h_undop_diso ] ) derived with the usual bond operators . for @xmath74 the ground state of @xmath75 defined in ( [ h_undop_o ] ) is given by the modified product state @xmath76 ( [ condens_prod ] ) .
it can be seen that @xmath77 in ( [ h_undop_o ] ) contains creation , hopping , and conversion terms of the three types of excitations @xmath78 .
contrary to the hamiltonian @xmath79 in ( [ h_undop_diso ] ) where the triplet excitations can only occur in pairs , here also single @xmath80-type excitations can be created and destroyed , e.g. , by @xmath81 .
the effect of these terms is directly related to the basis parameter @xmath50 and will be used to determine its value .
furthermore it can be seen that all terms creating @xmath80-type excitations out of @xmath54 vanish for @xmath82 .
this means that the ground state in the limit of vanishing interplane coupling @xmath6 does not contain single longitudinal spin fluctuations .
they will become , however , important for describing the destruction of the antiferromagnetism with increasing @xmath6 .
note that @xmath80-type bosons formally occur together with transverse fluctuations even at @xmath82 .
this happens if a transverse fluctuation created by @xmath5 hits a second transverse fluctuation being already present at the same rung , in other words , if spins on corresponding sites in each of the planes are flipped independently by @xmath5 .
these processes have to be included in an exact treatment of the ground - state wave function ; within the approximation of independent bosons ( which will be described in sec .
ii c ) they are neglected . to obtain an approximation for the ground - state wave function we adopt a cumulant approach @xcite which has been successfully applied to a variety of strongly correlated systems .
one starts from a splitting of the hamiltonian into @xmath44 where eigenstates and eigenvalues of @xmath46 are known .
the ground state @xmath83 of @xmath84 can be constructed from the unperturbed ground state @xmath47 of @xmath46 by application of a so - called wave operator @xmath85 which contains the effect of @xmath79 .
it has been shown @xcite that @xmath85 can be written in an exponential form , @xmath86 , where @xmath87 introduces fluctuations into @xmath47 .
the ground - state energy @xmath88 is calculated @xcite according to @xmath89 where @xmath90 denotes a cumulant expectation value with respect to @xmath47 . for the disordered phase of the system it is suitable to start from the singlet product state @xmath47 given in eq .
( [ singlet_prod ] ) and to include fluctuations in form of pairs of triplet excitations .
an appropriate ansatz for the wave function is @xmath91 the first term in the exponential contains pairs of triplets ( with arbitrary distance ) .
the operators @xmath92 in the second term represent higher - order fluctuations , e.g. , four - triplet operators .
they will be used in sec .
iv . for the determination of the coefficients @xmath93 and @xmath94 the cumulant method
provides a set of non - linear equations @xcite , @xmath95 the dot @xmath96 indicates that the quantity inside @xmath97 has to be treated as a single entity in the cumulant formation .
the equations ( [ coeff_eq_diso ] ) are derived from the requirement that @xmath98 is an eigenstate of @xmath84 . from symmetry
it follows that the coefficients @xmath99 only depend on the difference vector @xmath100 ( but not on @xmath101 ) .
so the state @xmath83 obeys rotational invariance in spin space .
when evaluating the terms in ( [ coeff_eq_diso ] ) the exponential series terminates after a few terms since the exponent contains only operators which create ( but not destroy ) fluctuations .
the resulting non - linear coupled equations for @xmath102 can be partially decoupled by a fourier transformation to momentum space .
the remaining set of equations has to be solved self - consistently . for the ordered phase the treatment is similar
, however , one has to take care of the broken rotational symmetry .
starting point is the product wave function @xmath76 ( [ condens_prod ] ) . in the operator @xmath85
we now explicitely distinguish between pairs of transverse ( @xmath103 ) and longitudinal ( @xmath104 ) fluctuations : @xmath105 2 the operators @xmath92 again contain higher - order fluctuations and will be discussed later .
the parameters @xmath102 , @xmath106 , and @xmath94 have to be determined in analogy to ( [ coeff_eq_diso ] ) by the equations @xmath107 which again reduce to ( [ coeff_eq_diso ] ) in the case of @xmath51 .
for @xmath51 we expect the transverse and longitudinal fluctuations to become equivalent , @xmath108 . for @xmath62 the amplitude of the longitudinal fluctuations in the ground state vanishes , @xmath109 , if one neglects higher - order processes in @xmath5 as noted above
. note that @xmath85 in ( [ omega_o ] ) does not contain operators for the creation of single @xmath80-type excitations @xmath110 although such operators are present in the hamiltonian ( [ h_undop_o ] ) .
the reason is that such operators change the density of the condensate of the @xmath26 triplet bosons and refer to the same degree of freedom as the parameter @xmath50 in the product state @xmath76 .
the aim here is , however , to describe the condensation of the @xmath26-type bosons completely via the introduction of @xmath50 and the transformed basis states . in this way
@xmath85 contains only multi - spin fluctuations .
now we discuss the parameter @xmath50 which is expected to be zero throughout the disordered phase and to vary continuously from 0 to 1 with decreasing @xmath6 in the ordered phase . within the cumulant method @xmath50 can be determined consistently using the identity @xmath111 this can be understood as the condition that @xmath85 contains * no * additional term like @xmath112 in the exponential , i.e. , the boson condensation is fully accounted for by the transformed basis state @xmath76 . within the cumulant formalism
this can be seen from the equality @xmath113 here we have used the fact that a cumulant expectation value does not change when the operator @xmath114 is subject to cumulant ordering instead of being applied directly to the wave function @xmath47 ( see ref . ) .
the operator @xmath115 can be interpreted as new wave operator ( applied to @xmath47 ) .
it now contains the additional parameter @xmath50 which has to be determined from @xmath116 which is equivalent to eq .
( [ lambda_eq ] ) .
if we restrict the fluctuation operators in the exponential of @xmath85 to excitation pairs as described above ( i.e. neglect further operators @xmath92 ) and take fully into account the constraint ( [ constraint ] ) , we find a critical coupling of @xmath117 for the order - disorder transition . at this level of approximation
the magnetization in the limit of decoupled planes ( @xmath118 ) is around 86 % of its classical value being larger than predicted by monte - carlo and spin - wave calculations .
the described approximation can be systematically improved by including higher - order fluctuations . this behavior is also known from calculations within the coupled - cluster method for the heisenberg model @xcite .
it can be shown that in the limit of an infinite set of operators the cumulant method yields exact results .
for instance , if we include two additional operators @xmath92 with three and four fluctuations on adjacent sites into the operator @xmath85 ( [ omega_o ] ) the transition point shifts to @xmath119 . however , the analytical and numerical effort to set - up and solve the non - linear equations equivalent to ( [ coeff_eq_diso],[coeff_eq_o ] ) increases drastically with including higher - order fluctuations .
the solution of the non - linear equations ( [ coeff_eq_diso],[coeff_eq_o ] ) shows that the absolute values of all coefficients in @xmath85 are small compared to 1 . in the disordered phase
the maximum is taken by @xmath120 ( the coefficient for nearest - neighbor triplet pairs ) at the point of the phase transition . in the ordered phase
the maximum is at @xmath121 with @xmath122 . with increasing distance between the triplets the values decay fastly
; higher - order fluctuations acquire a coefficient being one or more orders smaller in magnitude .
the fact that the coefficients are small is equivalent to the statement that the total triplet density even at the transition point is small ( @xmath123 ) , see ref . . in fig
[ fig_gsprop ] we show results for the ground - state energy @xmath88 and the staggered magnetization @xmath124 for the approximation using triplet pairs only as described above and for the spin - wave approach ( see below ) .
the data are compared with recent series expansions from ref . .
it can be seen that all approximations well describe the behavior of the magnetization which first increases with increasing @xmath6 and then drops to zero at the critical coupling .
= 8 truecm another approximation we want to discuss briefly is to neglect the constraint ( [ constraint ] ) completely @xcite and also the quartic triplet terms in the hamitonians ( [ h_undop_diso ] ) and ( [ h_undop_o ] ) .
the parameter @xmath50 is simply chosen so that the prefactors of the terms creating single @xmath80-bosons out of @xmath76 vanish .
one can formally set @xmath125 ( condensation of @xmath126-bosons ) ; @xmath76 can be considered as `` vacuum '' . neglecting
the constraint ( [ constraint ] ) implies that the three types of excitations @xmath78 are treated as independent bosons . after a fourier transformation one arrives at @xmath127 2 with @xmath128 and @xmath129 in the ordered phase ( @xmath130 ) and @xmath51 otherwise .
the hamiltonian ( [ h_sw ] ) can be easily diagonalized by a bogoliubov transformation . in this case
the ground - state wave functions ( [ omega_diso ] ) and ( [ omega_o ] ) ( with pairs of triplet excitations ) become exact .
the described harmonic approximation can be considered as linear spin - wave theory for the bilayer problem * with * inclusion of longitudinal spin fluctuations @xcite .
the critical coupling here becomes @xmath131 . in the limit of decoupled planes the results are equal to that of the linear spin - wave approach , e.g.
, the magnetization takes 60.5 % of its classical value , see fig .
further details on the ground - state calculations will be published elsewhere .
if we remove one electron from rung @xmath17 a single hole state on this rung is created .
let us denote by @xmath132 the creation operator of a hole with spin @xmath133 on rung @xmath17 , where the index @xmath21 represents the number of the plane .
this operator is defined by the equation @xmath134 where @xmath135 .
this is a fermionic operator which satisfies the usual anticommutation relations .
the hole on the rung @xmath17 interacts with the triplet excitations on its nearest - neighbor sites .
the interaction hamiltonian can be easily found by calculating all possible one - hole matrix elements of the initial hamiltonian ( [ h_tj ] ) ( see ref . ) .
for the disordered phase ( @xmath51 ) the result is @xmath136+h.c .
\right)\nonumber\\ & -&t_\parallel \sum_{\langle ij\rangle } \left(i \ { \overline { \bf s}}_{j , i } \left[{\bf t}_{i}^{\dag}\times { \bf t}_{j}\right ] + h.c.\right ) \nonumber\\ & -&{j_\parallel \over2}\sum_{\langle ij\rangle } \left(i \ { \overline { \bf s}}_{i , i } \left[{\bf t}_{j}^{\dag}\times { \bf t}_{j}\right ] + h.c.\right )
. \label{hh}\end{aligned}\ ] ] following refs .
we have introduced the notations @xmath137 for the triplet vector and @xmath138 for generalized hopping operators where @xmath139 is the vector of pauli matrices .
in addition we have to impose the constraint that a hole state and one of the doubly occupied states can not coexist on the same rung : @xmath140 the several terms in the hamiltonian ( [ hh ] ) have been discussed by eder @xcite .
they contain hopping without changing the spin background ( direct hopping , second and third term ) as well as spin - fluctuation assisted hopping ( fourth and fifth term ) and exchange processes where the hole remains on its rung ( last two terms ) . together with the hamiltonian for the doubly occupied rungs ( [ h_undop_diso ] )
the hamiltonian ( [ hh ] ) contains the complete one - hole dynamics for the disordered phase . for the case of more than one hole
additional hole interaction processes would have to be considered .
however , in the present work we restrict ourselves to the discussion of the one - hole problem . if we work in the magnetically ordered phase it is convenient to use the generalized basis @xmath141 from sec .
ii to represent the states of the doubly occupied rungs . from the hamiltonian ( [ hh ] ) and the definition ( [ basis_def ] ) of the basis operators
one can construct a hamiltonian for an ordered background state .
it again contains processes proportional to @xmath31 where the hole hops to a neighboring rung and exchange terms propertional to @xmath5 . to be short we only state some hopping matrix elements of the resulting hamiltonian . for
that we symbolically introduce two - rung states @xmath142 where @xmath143 and @xmath144 are the states on two adjacent rungs with @xmath73 and @xmath145 .
@xmath146 the first two terms represent processes of hole motion without desturbing the spin background ( so - called direct hopping ) whereas the last two terms are examples for hole hopping with creation of a `` spin defect '' .
it can be seen that direct hopping is impossible in the limit of @xmath62 , i.e. , in a nel - ordered background .
to investigate the hole motion we consider a one - particle green s function describing the creation of a single hole with momentum @xmath147 at zero temperature : @xmath148 where @xmath26 is the complex frequency variable , @xmath149 , @xmath150 .
the quantity @xmath151 denotes the liouville operator defined by @xmath152_-$ ] for arbitrary operators @xmath70 ; @xmath153 is the hole momentum with @xmath154 or @xmath155 for the bonding or antibonding band , respectively .
@xmath83 is the full ground state of the undoped system as described in sec .
the effect of the hole creation operator @xmath156 applied to @xmath83 can be related to the fermion operators @xmath157 introduced in the last section .
one obtains for example for @xmath158 sublattice @xmath159 and analogous relations for other basis states and for @xmath160 .
the hole motion processes will be described in the concept of path operators @xcite which create strings of spin fluctuations attached to the hole . for the application of projection technique
we define a set of path operators @xmath161 which couple to a hole and create local spin defects with respect to the undoped ground state @xmath83 .
the first operator @xmath162 is the unity operator , the second one @xmath163 moves the hole by one lattice spacing creating one spin excitation and so on .
we are interested in calculating dynamical correlation functions for the operators @xmath164 : @xmath165 the green s function ( [ hole_gf ] ) for the hole is then given by @xmath166 .
using cumulants the correlation functions @xmath167 can be rewritten as @xcite : @xmath168 the brackets @xmath169 denote cumulant expectation values with @xmath47 .
the `` unperturbed '' ground state at half - filling , i.e. , one of the product states ( [ singlet_prod ] ) or ( [ condens_prod ] ) for the disordered or ordered system , respectively , will be denoted in the following by @xmath47 for both phases .
the operator @xmath85 has been described above , it transforms @xmath47 being the ground state of @xmath46 into the full ground state @xmath83 of @xmath170 at half filling . in the following , @xmath85
is approximated by an exponential ansatz as described in sec .
ii . since the expression ( [ kumkorr ] )
contains both @xmath85 and @xmath171 , the expectation values can be calculated only in the lowest non - trivial order of the fluctuations in @xmath85 , i.e. , the expansion will be restricted to linear fluctuation terms . in order to correctly describe the short - range spin fluctuations in the system ,
operators with up to 4 triplet excitations with a maximum distance of 4 lattice spacings have been employed . for the evaluation of their coefficients
we have linearized the set of equations ( [ coeff_eq_diso],[coeff_eq_o ] ) .
this is possible since the absolute values of the coefficients are small compared to 1 as discussed in sec .
ii c. therefore quadratic and higher terms in the fluctuations operators can be neglected . by comparing the coefficients of the short - range fluctuations obtained in this way with the ones from the full ( non - linear ) ground - state calculation of sec .
ii c we have numerically verified that the error in the matrix elements ( [ qafm_matel ] ) introduced by this linearization is smaller than a few percent . since the linearized equations for
the coefficients do not yield a phase transition point we fix it at the value @xmath172 known from series expansions @xcite and monte carlo calculations @xcite .
( in the limit of an infinite set of operators @xmath173 within the cumulant method the solution of the non - linear equations has to yield the `` exact '' transition point which we assume to be at @xmath172 . ) in the ordered phase we set @xmath174 which is suggested by the results obtained in sec .
ii . using mori - zwanzig projection technique
@xcite one can derive a set of equations of motion for the dynamical correlation functions @xmath175 . neglecting the self - energy terms
it reads : @xmath176 @xmath177 and @xmath178 are the static correlation functions and frequency terms , respectively .
they are given by the following cumulant expressions : @xmath179 these terms describe all dynamic processes within the subspace of the liouville space spanned by the operators @xmath164 .
the use of cumulants ensures size - consistency , i.e. , only spin fluctuations connected with the hole enter the final expressions for the one - hole correlation function . in the present calculations we have employed up to 1600 projection variables with a maximum path length of 3 .
the neglect of the self - energy terms leads to a discrete set of poles for the green s functions , so the present approach can not account for linewidths . in all figures
we have introduced an artificial linewidth of @xmath180 to plot the spectra .
for details of the calculational procedure see e.g. ref . .
now we turn to the discussion of the final results .
first we consider the case of vanishing interplane hopping @xmath181 , i.e. , the hole motion is restricted to one plane .
( in this case @xmath182 can be dropped . ) = 8.3 truecm = 8.3 truecm the one - hole spectral function @xmath183 for @xmath181 , @xmath184 and various values of @xmath7 is shown in figs .
[ fig_spec0_1 ] - [ fig_spec0_3 ] . for zero and small @xmath6 ,
i.e. , in the limit of decoupled planes , we observe a pronounced quasiparticle ( qp ) peak and a weak background at higher energies .
the qp peak follows a dispersion with minima at @xmath2 .
these features are well - known from the one - hole problem in the single - layer antiferromagnet . with increasing @xmath6
the dispersion is `` washed - out '' because the antiferromagnetic correlations in the background state are weakened . at @xmath185 a cross - over to a dispersion with minima at @xmath48 and @xmath186 occurs .
note that this cross - over point still lies * inside * the antiferromagnet phase .
near the phase transition the spectral weight of the lowest pole at @xmath186 becomes small .
most of the spectral weight can be found in a band with minimum at @xmath48 and maximum at @xmath186 .
the weak band visible at the bottom of the spectrum around momentum @xmath186 can be considered as `` shadow band '' originating from the antiferromagnetic background , i.e. , it is obtained by shifting the band with large weight by @xmath187 .
= 8.3 truecm further increasing @xmath6 drives the system into the gapped paramagnetic ( pm ) phase . at the phase transition
it can be seen that shadow bands being present in the af phase disappear in the pm phase . at large @xmath6
the hole behaves like a free fermion with a dispersion proportional to @xmath188 .
now we consider the effect of interplane hopping .
we fix the ratio of the parameters to @xmath189 which follows from the derivation of the @xmath1 model from a hubbard model for the bilayer system with on - site repulsion @xmath190 .
spectra for different parameter sets with @xmath191 are shown in figs .
[ fig_spec100_u25 ] - [ fig_spec400_u25 ] .
an interplane coupling of @xmath192 has already changed the character of the bands compared to @xmath181 ( cf . fig . [
fig_spec0_1]a ) .
further increasing @xmath193 to 1.5 ( fig .
[ fig_spec225_u25 ] ) forms a pronounced band with a free - fermion form at the bottom of the spectrum in the bonding channel ( @xmath154 ) .
in contrast , the antibonding spectrum ( @xmath194 ) becomes incoherent since spectral weight is transferred to higher energies . since we are still in the antiferromagnetic phase the lowest peaks in the antibonding channel appear at the same energies as in the bonding channel shifted by momentum @xmath195 , i.e. , these can again be considered as shadow bands . at @xmath196
the af order is destroyed . as shown in fig .
[ fig_spec400_u25 ] for @xmath197 the shadow bands are disappeared , and in the antibonding channel a pronounced free - fermion band at high energies is formed .
higher values of @xmath198 will completely suppress the incoherent weight left in the antibonding channel .
= 8.3 truecm = 8.3 truecm varying the ratio @xmath199 ( which is 0.4 for all figures presented here ) does not change the picture qualitatively .
larger values of @xmath5 suppress the incoherent background in all spectra .
for very small values of @xmath5 the spectral functions become incoherent since the hole is dressed with an increasing number of spin fluctuations , i.e. , the radius of the spin polaron increases .
furthermore , the bandwidth of the quasiparticle dispersion at small @xmath6 is mainly controlled by @xmath5 ( instead of @xmath31 ) .
this means that the bandwidth in units of @xmath31 decreases with decreasing @xmath199 , see also next subsection .
the location of the cross - over between the two dispersion forms depends weakly on @xmath199 which will be discussed below .
= 8.3 truecm next we are going to examine the properties of the low - lying bands .
again we first discuss the case of vanishing interplane hopping , @xmath181 . for @xmath200 one
finds a narrow qp band shown in fig .
[ fig_disp1 ] ( top panel ) .
it has minima at @xmath2 and a bandwidth of approximately @xmath201 .
it corresponds to the coherent motion of a dressed hole , i.e. , a spin - bag quasiparticle .
( note that a larger set of path operators would give a larger bandwidth of around @xmath202 which is known from the calculations for a single - layer antiferromagnet @xcite .
however , paths of length 4 and more are hard to access for the bilayer system due to the increasing numerical effort . ) the main contribution to the hole motion in the af case can be understood as follows : the hopping hole locally destroys the antiferromagnetic spin order leaving behind a string of spin defects .
quantum spin fluctuations can repair pairs of frustrated spins which leads to a coherent motion in one of the antiferromagnetic sublattices ( spin - fluctuation - assisted hopping ) . for @xmath203
the bandwidth of this coherent hole motion is of order @xmath5 because the spin - flip part of the intraplane heisenberg interaction @xmath5 is necessary to remove the spin defects caused by hopping .
with increasing @xmath6 the bandwidth of the dispersion first decreases ; at a value of @xmath185 the minima move to @xmath48 and @xmath186 .
slightly larger values of @xmath6 lead to an increase of the bandwidth , and spectral weight is transferred to a band which has nearly tight - binding form .
so the lowest pole at @xmath204 has very small weight around momentum @xmath186 .
this shadow band is shown as dashed line in fig .
[ fig_disp1 ] . in the paramagnetic phase for larger @xmath6 the shadow band disappears .
the qp peak has a dispersion being nearly the one of a free fermion but with a reduced bandwidth .
this behavior follows from the existence of the direct hopping term with prefactor @xmath205 in @xmath206 ( [ hh ] ) .
it leads to a disperison proportional to @xmath31 ; the saturation bandwidth equals half of the bandwidth of the uncorrelated system ( @xmath207 ) .
= 6.5 truecm = 6.5 truecm the results for non - zero interplane hopping @xmath198 are shown in fig .
[ fig_disp2 ] . again , for small @xmath198 ( small @xmath6 ) the dispersion character is `` antiferromagnetic '' whereas for large @xmath198 ( large @xmath6 ) the bands have a simple tight - binding form . for @xmath208 and @xmath209 their dispersion is given by @xmath210 the bandwidth for the in - plane motion is reduced by a factor of 2 compared to a free fermion since the relevant hopping matrix element is @xmath211 .
= 8 truecm in fig .
[ fig_bandw1 ] we finally show the bandwidth of the qp dispersion for the case of vanishing interplane hopping and @xmath212 , i.e. , for the dispersions shown in fig . [ fig_disp1 ] .
it can clearly be seen that the cross - over from an `` antiferromagnetic '' to a simple tight - binding dispersion occurs at @xmath185 and is connected with a pronounced minimum in the qp bandwidth . for @xmath213 the bandwidth increases and saturates for large @xmath6 at @xmath214 as discussed above .
results for other values of @xmath199 are qualitatively similar .
the minimum in the dispersion ( i.e. the cross - over point ) shifts with @xmath199 : for @xmath215 it is located at @xmath216 , for @xmath217 it is found at @xmath218 .
this in turn means that the cross - over in the shape of the single - hole dispersion can be driven by varying @xmath199 at * fixed * @xmath7 . in order to understand the behavior of the qp dispersion and its bandwidth we illustrate in more detail the connection between the in - plane hole motion and the short - range in - plane spin correlations .
the basic ingredience for the observed behavior of the qp dispersion is the competition between direct hopping which dominates in the disordered phase and spin - fluctuation - assisted hopping known from the single - layer af .
the relevant matrix element for direct hopping between states without spin deviations is @xmath219 where @xmath220 are nearest - neighbor sites and @xmath83 denotes the undoped background state .
this term can be expressed @xcite by the static in - plane nearest - neighbor spin correlation @xmath221 with @xmath222 as @xmath223 spin - fluctuation - assisted hopping is more complicated since it involves two hopping steps and one spin - fluctuation process .
the matrix element for the spin - fluctuation process ( being the most important for not too small values of @xmath31 ) can be written as where @xmath224 moves the hole by two hopping steps ( creating a path of spin defects ) and @xmath220 are now next - nearest neighbors . transforming the expectation value into spin correlation functions one obtains @xcite @xmath225 where the average over the possible paths of length 2 has already been performed .
= 7.5 truecm the in - plane dispersion shape can be described by the energy difference @xmath226 .
values @xmath227 correspond to an `` antiferromagnetic '' dispersion whereas @xmath228 occurs for a nearest - neighbor tight - binding dispersion . neglecting longer paths and more complicated contributions to the hole motion
the qp dispersion is given by the sum of a nearest - neighbor and a next - nearest neighbor dispersion originating from the two processes described above .
then @xmath229 can be roughly estimated from the matrix elements @xmath230 and @xmath231 : @xmath232 the prefactor of @xmath230 arises from the number of nearest neighbors ; for the @xmath231-prefactor the influence of two hopping steps has to be kept in mind , a fit to numerical results for the single - layer problem yields a value of order 1 . from this
we see that the cross - over phenomenon can be understood in terms of the interplay between short - range spin correlations and the ratio of @xmath199 .
[ fig_spincorr1 ] shows the values of @xmath233 for nearest and next - nearest neighbor sites obtained from the present calculation . with increasing @xmath7 the absolute values of the in - plane spin correlation functions decrease from their single - layer values and
are weakened within the antiferromagnetic phase . at the transition
point the magnitude has dropped by a factor around @xmath234 compared to @xmath68 which again coincides with the fact that the density of spin ( triplet ) excitations is small in the disordered phase even at the transition point @xcite .
we have also plotted the quantity @xmath229 estimated from eq .
( [ delta_approx ] ) for two different values of @xmath199 .
one notices a semi - quantitative agreement of the zero in @xmath229 and the dispersion minimum described in sec .
in this paper we have presented for the first time a systematic analytical study of the one - hole dynamics on both sides of a magnetic ordering transition in a low - dimensional antiferromagnet .
the system under consideration was a bilayer antiferromagnet described by a @xmath235-@xmath236 hamiltonian in the limit of zero doping .
the magnetic background state has been modelled with modified bond operators . in the disordered phase
these operators describe the singlet ground state and the triplet excitations . in the af ordered phase they account for the condensation of one type of triplet bosons and describe transverse as well as longitudinal fluctuations . using the spin - polaron concept which describes the one - hole states in terms of local spin deviations
we have calculated the one - hole spectral function for the whole range of magnetic couplings . for the disordered background ( large @xmath7 ,
gapped spin excitations ) spin fluctuations around the hole are suppressed .
the hole motion is dominated by direct hopping processes , i.e. , hopping without disturbing the spin background . in the ordered phase for very small interplane coupling
we recover the results known from the single - layer hole motion : the spectrum consists of a coherent qp peak at the bottom and an incoherent background .
the qp can be associated with a mobile hole dressed by spin fluctuations .
the bandwidth of its coherent motion is controlled by @xmath5 .
the cross - over between these two scenarios occurs * inside * the ordered phase where the antiferromagnetic short - range correlations become weakened .
the cross - over is located between @xmath237 depending on @xmath238 ( for @xmath239 it is found at @xmath185 ) .
note that the cross - over can also be driven by variation of the hopping strength @xmath31 at a fixed value of @xmath7 well in the antiferromagnetic phase ( e.g. 1.3 ) .
this behavior follows from the competition between direct nearest - neighbor hopping and spin - fluctuation - assisted next - nearest - neighbor hopping .
in contrast , when crossing the magnetic phase boundary at @xmath240 there are no drastic changes in the spectrum ( and therefore in the arpes response of such a system ) . the only differences between the spectra in both phases near the phase transition are weak shadow bands in some regions of the brillouin zone in the af phase .
note that our approximation which describes the hole motion processes in terms of short - range spin fluctuations is questionable in a region in the close vicinity of the phase transition due to the existence of long - ranged critical fluctuations .
however , it has been argued recently@xcite that the influence of the critical modes suppresses the quasiparticle weight only in the limit of vanishing hopping @xmath241 . in contrast , for finite hopping the number of spin fluctuations near the hole remains finite even at the transition point .
so we expect the picture presented in this paper to be valid at least in the regions away from the transition , i.e. , @xmath242 , which is supported by the fact that our spectral functions on both sides of the phase transition ( e.g. at @xmath243 and 2.7 ) do not show major differences .
so the main result of the present paper can be summarized as follows : a magnetic phase transition has only weak influence on the arpes spectrum of a doped antiferromagnet .
the properties of the spectrum are , however , dominated by the short - range environment of the hole .
the statement is expected to hold also for the doping - induced phase transition in the single - layer af . here
the antiferromagnetic correlations are strong even in the paramagnetic phase , i.e. , from the above considerations one expects to find an `` antiferromagnetic '' hole dispersion on both sides of the transition .
this is exactly what was observed in recent work @xcite : analytical investigations of one hole in the af phase as well as numerical studies of finite systems ( which have a singlet ground state without long - range order ) both show a coherent hole motion with a dispersion of width @xmath244 and minima at @xmath2 .
it should be pointed out that the above discussion applies to intermediate and high energy scales ( order @xmath235 , @xmath236 ) only .
of course there exist low - energy properties of the spectrum which are expected to be influenced by quantum criticality , e.g. , the linewidth in a finite - temperature photoemission experiment should show scaling behavior in the quantum - critical region associated with the transition @xcite .
these features as well as the hole dynamics in the bilayer system at low but finite doping are beyond the scope of the present study and will be subject of future research .
y. uemura , phys .
lett . * 66 * , 2665 ( 1991 ) , t. imai @xmath245 @xmath246 , phys .
* 70 * , 1002 ( 1993 ) , m. takigawa , phys
b * 49 * , 4158 ( 1994 ) , g. aeppli , t. e. mason , s. m. hayden , h. a. mook , and j. kulda , science * 278 * , 1432 ( 1997 ) .
a. v. chubukov and d. k. morr , phys .
b * 52 * , 3521 ( 1995 ) . a. w. sandvik and d. j. scalapino , phys .
lett . * 72 * , 2777 ( 1994 ) .
a. w. sandvik , a. v. chubukov , and s. sachdev , phys .
b * 51 * , 16483 ( 1995 ) . c. n. a. van duin and j. zaanen , phys . rev
. lett . * 78 * , 3019 ( 1997 ) .
w. h. zheng , phys .
b * 55 * , 12267 ( 1997 ) . v.
n. kotov , o. p. sushkov , w. h. zheng , and j. oitmaa , phys .
lett . * 80 * , 5790 ( 1998 ) .
k. hida , j. phys .
. jpn . * 59 * , 2230 ( 1990 ) .
a. j. millis and h. monien , phys .
* 70 * , 2810 ( 1993 ) , phys .
b * 50 * , 16606 ( 1994 ) .
y. matsushita , m. p. gelfand , and c. ishii , j. phys .
68 * , 247 ( 1999 ) , d .- k .
yu , q. gu , h .- t .
wang , and j .- l .
shen , phys .
b * 59 * , 111 ( 1999 ) .
j. igarashi and p. fulde , phys . rev .
b * 45 * , 12357 ( 1992 ) , g. khaliullin and p. horsch , phys .
b * 47 * , 463 ( 1993 ) , a. belkasri and j. l. richard , phys . rev .
b * 50 * , 12896 ( 1994 ) .
m. vojta and k. w. becker , ann .
( leipzig ) * 5 * , 156 ( 1996 ) , phys .
b * 54 * , 15483 ( 1996 ) .
b. o. wells @xmath245 @xmath246 , phys .
74 * , 964 ( 1995 ) .
d. s. dessau @xmath245 @xmath246 , phys .
lett . * 71 * , 2781 ( 1993 ) .
d. s. marshall @xmath245 @xmath246 , phys .
lett . * 76 * , 4841 ( 1996 ) .
y. nagaoka , phys .
147 * , 392 ( 1966 ) , w. f. brinkman and t. m. rice , phys .
b * 2 * , 1324 ( 1970 ) , b. i. shraiman and e. d. siggia , phys .
* 60 * , 740 ( 1988 ) .
s. a. trugman , phys .
b * 37 * , 1597 ( 1988 ) .
j. r. schrieffer , x. g. wen , and s. c. zhang , phys .
lett . * 60 * , 944 ( 1988 ) , e. dagotto and j. r. schrieffer , phys . rev .
b * 43 * , 8705 ( 1991 ) . a. nazarenko and e. dagotto , phys . rev .
b * 54 * , 13058 ( 1996 ) , w .- g .
yin and c .- d .
gong , phys .
b * 56 * , 2843 ( 1997 )
. r. eder , y. ohta , and s. maekawa , phys .
b * 51 * , 3265 ( 1995 ) , phys . rev .
b * 52 * , 7708 ( 1995 ) . r. f. bishop , theor . chim .
acta * 80 * , 95 ( 1991 ) , r. f. bishop , r. g. hale , and y. xian , phys .
73 * , 3157 ( 1994 ) , c. zeng , d. j. j. farnell , and r. f. bishop , j. stat
. phys . * 90 * , 327 ( 1998 ) . | the two - layer square lattice quantum antiferromagnet with spins @xmath0 shows a magnetic order - disorder transition at a critical ratio of the interplane to intraplane couplings .
we investigate the dynamics of a single hole in a bilayer antiferromagnet described by a @xmath1 hamiltonian . to model the spin background we propose a ground - state wave function for the undoped system which covers both magnetic phases and includes transverse as well as longitudinal spin fluctuations .
the photoemission spectrum is calculated using the spin - polaron picture for the whole range of the ratio of the magnetic couplings .
this allows for the study of the hole dynamics of both sides of the magnetic order - disorder transition . for small interplane coupling
we find a quasiparticle with properties known from the single - layer antiferromagnet , e.g. , the dispersion minimum is at @xmath2 . for large interplane coupling
the hole dispersion is similar to that of a free fermion ( with reduced bandwidth ) .
the cross - over between these two scenarios occurs inside the antiferromagnetic phase which indicates that the hole dynamics is governed by the local environment of the hole . 2 |
in recent years both the applications for strongly correlated quantum impurity models and the number of successful approaches to harvest their physical results have grown enormously .
those models were introduced to describe the effects of magnetic transition metal impurities immersed in metallic hosts @xcite .
originally they were derived to capture remarkable physical properties like the resistance minimum @xcite at a specific temperature scale @xmath0 @xcite or the anomalous magnetic susceptibility and specific heat of such materials .
today a whole realm of applications for quantum impurity models has opened .
they describe the physics of quantum dots and wires @xcite as well as molecular electronics @xcite .
applications range from nanoelectronics all the way to quantum information processing @xcite .
their properties are essential for today s technological applications in single electron transistors @xcite exhibiting the coulomb blockade effect @xcite or in devices dominated by rkky interaction @xcite .
the behavior of various magnetic phenomena and the fascinating branch of heavy fermion physics is described by strongly correlated quantum impurity models @xcite .
recent studies have shown that the remarkable material graphene exhibits kondo physics @xcite which may be investigated theoretically by virtue of quantum impurity models .
these models have further been applied to understand the adsorption of atoms onto surfaces @xcite .
in addition , they are of theoretical importance as solvable models of quantum field theories @xcite .
a renewed interest in understanding and calculating dynamic quantities of these models was created with the advent of dynamical mean - field theory ( dmft ) @xcite . in the foundations of this theory quantum impurity models
have to be solved as an auxiliary problem . +
a wide range of methods and approximations have been suggested for the solution of quantum impurity models .
they however prove to be a very delicate subject because standard perturbative approaches diverge @xcite .
prominent methods to gain physical conclusions include a self consistent perturbative expansion @xcite and bethe ansatz techniques @xcite for one dimensional problems .
the low energy physics is very well described by numerical renormalization group ( nrg ) @xcite and in some limits also by functional renormalization group ( frg ) @xcite , and density matrix renormalization group ( dmrg ) @xcite .
there is a range of slave particle methods @xcite available as well as methods based on hubbard s x - operator technique @xcite and calculations using variational wave functions @xcite .
valuable physical insight has been gained by using equation of motion techniques applying different approximation schemes @xcite . for moderate
system sizes the hirsch - fye quantum monte carlo ( qmc ) @xcite algorithm has proven to achieve good results . in the past years different approaches to continuous time qmc
@xcite have been applied very successfully to solve quantum impurity models especially in application with dmft . in this context
exact diagonalization ( ed ) methods have been explored to solve small systems @xcite . + as of today some limits of quantum impurity models are understood with great precision but there appear several gaps to be bridged .
the low energy properties of these models are reproduced very well by renormalization group based approaches ( i.e. nrg ) .
these approaches in general have trouble to capture the high energy parts of the spectrum .
the same may be said about qmc methods which if applicable yield dynamic quantities in imaginary time .
the analytic continuation to the real energy axis is ill conditioned .
spectra obtained by , for example , the maximum - entropy method @xcite have a large uncertainty for higher energies .
exact diagonalization methods , in principle , grant access to low as well as high energy parts of the spectrum at the same time . due to the prohibitively large hilbert space however only small systems ( about ten to twenty sites ) may be treated with this method , whose low - energy behavior is expected to deviate from the one of the infinite lattice significantly .
nevertheless , the advantage consists in the fact that the spectral properties may be determined directly on the real energy axis . besides the issue of the low energy scale ,
also the flexibility to adapt to various impurity configurations and geometries is limited in many methods .
nrg has been successfully applied only to the one and two impurity case so far .
qmc approaches may suffer from the sign problem for more complex multiband models @xcite .
the region of large interaction strength is naturally difficult to treat in standard perturbative / diagrammatic approaches ( i.e. diagrammatic perturbation theory or frg ) .
+ in the present work we test cluster perturbation theory ( cpt ) @xcite and the variational cluster approach ( vca ) @xcite on the single impurity anderson model @xcite .
the great flexibility and versatility of cpt / vca allows for obtaining approximate single - particle dynamic quantities and static expectation values in all parameter regions of any lattice impurity model with local interactions .
however these many body cluster methods can not be expected to describe the low energy excitations as accurately as specifically tailored methods do .
it is however interesting to see whether the correct low energy behavior may be reproduced at least to some extent .
cpt as well as vca bare several advantages @xcite : i ) they yield spectra directly on the real axis and ii ) also the high energy incoherent part of the dynamics becomes available .
iii ) they are applicable in all parameter regions and also at high interaction strengths .
iv ) they have the advantage of comparatively low computational cost for a required resolution .
our main goal in studying the well understood single impurity is to benchmark cpt / vca for future application to the not so well understood case of multiband impurity models in various spatial geometries .
this publication also sets the foundations for a future extension to non equilibrium problems .
+ the text is organized as follows .
the single impurity anderson model is introduced in sec .
[ sec : model ] .
a short review on cpt and vca in this context is given in sec .
[ sec : method ] .
a self consistent formulation of vca previously introduced in the context of non equilibrium problems @xcite is presented in sec .
[ sec : sc ] .
some remarks about the choice of variational parameters are provided in sec . [
sec : varpar ] . in sec .
[ sec : omega ] we discuss the grand potential @xmath1 for infinite fermionic systems in relation with the vca .
results for the single - particle dynamics of the siam are provided in sec .
[ sec : results ] . in this section also
the quality of the low energy kondo physics is compared to benchmarking results from nrg , dmrg , ct - qmc , hartree - fock and bethe ansatz calculations .
finally , we summarize and conclude our findings in sec .
[ sec : conclusion ] .
we consider the single impurity anderson model ( siam ) @xcite in real space , in one dimension @xmath2 a tight binding band of non interacting s - electrons with nearest neighbor @xmath3 hopping is described by @xmath4 where @xmath5 is the on - site energy of the particles , @xmath6 is the overlap integral between nearest neighbor orbitals and @xmath7 where @xmath8 is eventually taken to be infinity . the operators @xmath9 and @xmath10 , respectively , create and annihilate electrons in orbital @xmath11 with spin @xmath12 .
the impurity hamiltonian consists of a single f - orbital with local coulomb repulsion @xmath13 , @xmath14 with @xmath15 creating an electron with spin @xmath12 and on - site energy @xmath16 located at the impurity .
the particle number operator is defined as @xmath17 .
finally the coupling between a non interacting s - orbital and the impurity f - orbital is given by @xmath18 where @xmath19 is the hybridization matrix element between the s- and the f - orbital of the impurity atom ( see fig.[fig : siam ] for illustration ) . + and on - site energy @xmath5 .
the impurity f - orbital is subjected to a local on - site energy @xmath16 and local coulomb interaction @xmath13 and is hybridized with one of the s - orbitals ( here the one at the beginning of the chain ) via a hybridization matrix element @xmath19 .
this maps the impurity f - orbital onto site @xmath20 and the impurity s - orbital onto site @xmath21 in this geometry , the rest of the conduction band s - electron orbitals are mapped onto sites @xmath22 to @xmath23 .
the semi - infinite non interacting chain is truncated at some site @xmath24 .
this decomposes the model into two clusters : an interacting cluster of variable size including the interacting impurity f - orbital and a semi - infinite chain of non interacting s - orbitals . in cpt
/ vca these decomposed systems are coupled via a hopping element @xmath6.,scaledwidth=48.0% ] we have set the chemical potential @xmath25 to the center ( @xmath5 ) of the conduction electron density of states and choose @xmath26 .
the resonance width @xmath27 is defined as @xmath28 for the model defined in eq.([eq : hconduction ] ) the local density of states of the conduction electrons @xmath29 is given by @xmath30 . in the forthcoming discussion we refer to the particle - hole symmetric case when we furthermore set @xmath31 .
all calculations are performed with @xmath32 and @xmath33 which yields @xmath34 .
all reported results , except for the ct - qmc data in sec .
[ sec : ctqmc ] , are for zero temperature .
a handle on dynamic single - particle correlations and expectation values is given by the single - particle green s function @xmath35 which we calculate within cluster perturbation theory ( cpt ) @xcite as well as the variational cluster approach ( vca ) @xcite .
cpt and vca have been previously applied inter alia to the fermionic hubbard model and vca also with great success to bosonic systems @xcite with broken symmetry phases .
the groundwork of vca lies within cluster perturbation theory which is a cluster extension of strong coupling perturbation theory , valid to first order in the inter - cluster hopping .
the main result of cpt is that the green s function of the physical system @xmath36 ( which we call full green s function throughout this text ) may be obtained by a dyson - like equation in matrix form @xmath37 here @xmath38 denotes the green s function of a cluster which comes about by tiling the lattice of the physical system into smaller , numerically exactly solvable patches .
this tiling is done by removing the hoppings between sites connecting such clusters
. therefore the matrix @xmath39 contains all single - particle terms connecting clusters ( i.e. the inter cluster hopping which will be referred to as @xmath40 below ) .
the subscript @xmath41 denotes the non interacting green s function . to apply this approach to the siam we start by splitting the physical model under consideration ( eq.([eq : hsiam ] ) ) into appropriate pieces . here
we consider a cluster decomposition consisting of two parts .
one part , consisting of a cluster of size @xmath24 , which contains the interacting impurity f - orbital @xmath42 and a second , infinitely large part , the environment , which contains the rest of the conduction band @xmath43 the original hamiltonian eq.([eq : hsiam ] ) , defined on the semi - infinite lattice , may now be rewritten as @xmath44 here @xmath40 is the part of @xmath45 describing the hopping from the interacting cluster to environment `` cluster '' , which is the only term not included in the two clusters . for the siam the two bare green s functions @xmath46 and @xmath47 ( , which correspond to @xmath48 ( eq.([eq : cluster ] ) ) and @xmath49 ( eq.([eq : environment ] ) ) , ) needed for eq.([eq : dyson ] ) may be evaluated separately .
this is a bit different from the usual application of cpt to translationally invariant systems which normally leads to a single cluster having discrete spectra , embedded in a superlattice .
therefore the application of cpt to this problem does obviously not suffer from issues arising due to periodization prescriptions for the green s function or self - energy @xcite .
we are dealing with two fundamentally different clusters , where one has a discrete ( interacting cluster eq.([eq : cluster ] ) ) and the other a continuous spectrum ( environment eq.([eq : environment ] ) ) . due to the continuous spectrum of the environment a numerically favorable representation of the green s function of the physical system @xmath36 in terms of the lehmann representation ( see for example ref . )
is not possible . for evaluating quantities from the green s function @xmath36 one
therefore has to use a direct numerical integration , which works best on the matsubara axis .
+ the cluster green s function @xmath46 is determined by exact diagonalization of eq.([eq : cluster ] ) .
we apply the lanczos algorithm @xcite to find the ground state and a band lanczos method to obtain the green s function .
the band lanczos method is initialized with a set of all annihilation and creation operators under consideration applied to the ground state .
thereby we obtain the so - called q - matrices @xcite which are used to calculate the green s function @xmath50 essentially this is the lehmann representation for zero temperature green s functions .
the sum over @xmath51 denotes a sum over a possibly @xmath52-fold degenerate set of ground states .
the sum over @xmath53 is over a set of orthonormal basis - states having one particle more than the ground state ( particle part ) and one particle less than the ground state ( hole part ) .
+ the green s function of the environment @xmath47 is given analytically by the green s function of a semi - infinite tight binding chain @xcite @xmath54 where @xmath55 is the retarded / advanced green s function of the infinite tight binding chain if the infinitesimal imaginary part ( @xmath56 ) of @xmath57 is positive / negative .
+ vca , the variational extension of cpt , is based on the self - energy functional approach ( sfa ) @xcite . in the sfa
one considers the legendre transformed luttinger - ward @xcite functional @xmath58 $ ] , which is a universal functional of the self - energy , i.e. it does not depend on @xmath59 .
@xmath60 generates the green s function , i.e. @xmath61}{\delta \sigma } & = -{\text{g}}[\sigma]\;\mbox { , } \label{eq : vca1}\end{aligned}\ ] ] where @xmath62 denotes the inverse temperature . introducing the ( non - universal ) self - energy functional @xmath63 & = f[\sigma ] - \text{tr}\ln { \left ( \left(-{\text{g}}^{-1}_0 + \sigma\right){\text{g}}_\infty\right ) } \,\mbox { , } \label{eq : vca3}\end{aligned}\ ] ] ( see ref . for a definition of @xmath64 ) , one recovers dyson s equation at its stationary point @xmath65}{\delta \sigma } & = -{\text{g}}[\sigma ] + \left ( { \text{g}}^{-1}_0 - \sigma \right)^{-1 } \stackrel{!}{= } 0\,\mbox{. } \label{eq : vca2}\end{aligned}\ ] ] eq.([eq : vca2 ] ) is an equation for the physical self - energy @xmath66 given the luttinger - ward functional @xmath58 $ ] and the free green s function @xmath59 .
the trace tr is short for @xmath67 , the sum is over fermionic matsubara frequencies and the small form trace tr denotes a sum over lattice sites and spin .
the idea is that , due to its universality , @xmath58 $ ] ( and thus @xmath68 $ ] ) can be evaluated exactly by exploiting a different system ( so called `` reference system '' ) which differs from the physical system by single - particle terms only .
this reference system @xmath69 is simpler , and thus exactly solvable .
it is defined on a cluster tiling of the original lattice and has the same interaction as the original system @xmath70 .
the vca reference system is chosen to be a cluster decomposition of the original lattice , as the one introduced for cpt above .
comparing eq.([eq : vca3 ] ) for the full and the reference system yields @xmath71 & = \omega'[\sigma , { \text{g}}_0 ] + \text{tr}\;\text{ln}\left ( -\left ( { \text{g}}_0^{-1 } -\sigma \right ) \right ) \\ & - \text{tr}\;\text{ln}\left ( -\left ( { \text{g}}_0^{-1 } -\sigma \right ) \right)\;\mbox { , } \label{eq : vca2}\end{aligned}\ ] ] where lowercase @xmath38 denote green s functions of the reference system .
thus the sfa / vca approximation consists in solving eq.([eq : vca2 ] ) in a restricted range of self - energies @xmath66 , i.e. the ones produced by the reference system . in this way
, the space of allowed @xmath66 is spanned by the set of single - particle parameters of the reference system , @xmath72 .
this means that the functional @xmath68 $ ] ( eq.([eq : vca2 ] ) ) becomes a function of those parameters @xmath73 the stationarity condition determining the physical parameters eq.([eq : vca2 ] ) is then given by @xmath74 the green s function of the physical system is obtained by the cpt relation eq.([eq : dyson ] ) .
the matrix @xmath39 ( eq.([eq : dyson ] ) ) in vca contains all single - particle terms not included in the reference system ( i.e. @xmath40 ) as well as , in addition , the deviation introduced by vca , @xmath75 of the single - particle parameters of the reference system @xmath72 with respect to the ones of the original system @xmath76 . in the following we fully adopt the zero temperature formalism , in which according expressions for the grand potential and related quantities may be readily evaluated . + in ref .
we explored an alternative version of vca whereby the variational parameters @xmath72 were determined by a suitable self consistent criterion , instead of looking for the stationary point of the grand potential @xmath1 eq.([eq : stationarity ] ) .
this alternative approach was introduced to treat systems out of equilibrium , although it can equally be adopted in equilibrium .
the advantage of this approach is that the solution of a self consistent equation is numerically easier than the search for a saddle point .
the idea of this self consistent approach is to use a reference system which resembles the full system best . + the strategy is to find those values @xmath72 for the set of parameters of the reference system which let the expectation values of their corresponding single - particle operators @xmath77 coincide with those of the full system @xmath78 .
here , the angle brackets denote expectation values in the cluster and the full system coupled by cpt or vca respectively .
consider the on - site energies @xmath79 and @xmath80 as variational parameters .
we will look for those cluster parameters @xmath79 and @xmath80 which fulfill the relations @xmath81 the sum is over all non interacting sites included in the cluster .
this amounts to solving a system of non - linear equations in each step of the self consistency cycle . in general
it is possible to vary each single - particle parameter individually . for reasons of keeping the numerics
tractable we use one @xmath80 only , which we take to be the same for each orbital in the chain .
extension to a larger number of @xmath80 is straightforward . to fix this parameter
we require the average particle density on the non interacting sites to fulfill the condition eq.([eq : sccond ] ) .
this corresponds to the condition presented in ref .
( 13 ) therein ) . in some situations , ( see below , ) we will alternatively consider the hybridization matrix element @xmath82 and the intra - cluster hopping @xmath83 as variational parameters , and proceed in an analogous way . specifically ,
the particle number expectation values in eq.([eq : sccond ] ) are replaced by hopping expectation values .
again for @xmath83 we use a single variational parameter for hopping between all uncorrelated sites and fix it by requiring the mean value of hopping in the cluster and the full system to coincide .
a discussion of this self consistency condition in connection with ( cluster ) dmft is given in ref . .
we use an improved multidimensional newton - raphson algorithm to find the roots of the system eq.([eq : sccond ] ) . in some parameter regions
no solution is found . + a comparison between results obtained via the usual sfa , i.e. as stationary points of the grand potential @xmath1 , which we will now refer to as vca@xmath84 , to the ones obtained by the above - mentioned self consistent condition ( vca@xmath85 ) will be given in the results section ( sec .
[ sec : results ] ) .
+ in vca one can , in principle , optimize all possible single - particle parameters which are present in the original model , as well as additional ones . by adding bath sites not present in the original model ,
one includes dynamical contributions to the cluster green s function @xcite .
the numerical difficulty increases with the number of variational parameters .
for the vca@xmath85 case a multidimensional root finding algorithm has to be adopted . for the vca@xmath84 case , a saddle point in many dimensions has to be located . since the allowed set of variational parameters limits the search space for the self - energies one will find a solution in this restricted space only .
it is therefore desirable to vary as many single - particle quantities as possible .
a balance has to be found between a large space of available self energies and numerically feasible multidimensional algorithms .
many works have addressed the question of which parameters are the most important to vary and how the choice of variational parameters will influence or limit the results @xcite . as discussed in refs .
, it is important to include an overall chemical potential as a variational parameter in order to preserve thermodynamic consistency . as a compromise
, we will take two variational parameters @xmath86 , which cover the overall chemical potential .
note that this amounts to shifting an overall on - site energy in the whole cluster plus an extra independent shift at the correlated site . for the variation of on - site energies we observe the grand potential @xmath1 to be maximal at the stationary point which is in agreement with results for other models .
further parameters in the siam are the hopping @xmath6 and the hybridization @xmath19 . as discussed for example in ref . , the variation of hopping parameters is not straightforward . for the vca@xmath87 approach ,
we observe a maximum of @xmath1 at @xmath88 in the center of two symmetric stationary points .
the two symmetrically lying minima are equivalent due to the fact that the self - energy is an even power of @xmath19 .
as one tunes the parameters away from particle - hole symmetry this stationary point is lost in the crossover region from the kondo plateau to a doubly or unoccupied impurity ( see sec . [
sec : efscan ] ) .
in this parameter region the hopping @xmath6 and hybridization @xmath19 are probably not appropriate to be used as variational parameters within vca@xmath87 .
+ in the following , we always choose the set @xmath89 or @xmath90 for calculations at particle - hole symmetry , which also includes @xmath91 , since the variation of on - site energies will always yield zero deviations from the physical parameters and thus reproduce the cpt result here .
for all other parameter regions it is sufficient to consider @xmath86 as variational parameters .
the reference system consists of two parts , a finite interacting system and a non interacting system of infinite size , the environment .
@xmath92 is given by the sum of the grand potentials of the interacting cluster ( @xmath93 ) and of the environment ( @xmath94 ) ( , which correspond to @xmath48 ( eq.([eq : cluster ] ) ) and @xmath49 ( eq.([eq : environment ] ) ) ) .
here we outline how to determine the grand potential for such kinds of reference systems . for the green s function within the cpt / vca approximation the dyson equation
is given in eq.([eq : dyson ] ) .
the green s function and have the block structure @xmath95 up to this point all matrices involving environment indices have infinite size .
as far as the green s function itself is concerned this is no problem as we are primarily interested in @xmath96 for which the dyson equation reduces to @xmath97 and therefore @xmath98 a bit more tedious is the elimination of the explicit dependence on the environment part of g , as far as the grand potential eq.([eq : vca3 ] ) is concerned .
we start out from a form of the grand potential functional given by snchal @xcite @xmath99 in app .
[ sec : appendixomeganew ] it is shown that @xmath100 can be expressed solely in terms of cluster quantities @xmath101 along the lines outlined in ref .
, the resulting integral can be regularized and expressed as @xmath102 the quantities @xmath103 are the grand potentials of the uncoupled reference system .
the constant infinite contribution @xmath94 is absorbed into the definition of @xmath1 .
it plays no further role as it does not depend on the variational parameters .
this integral may be evaluated as suggested in ref . by integrating from 0 to @xmath104 , from @xmath104 to @xmath105 and from @xmath105 to @xmath23 .
@xmath104 and @xmath105 represent two characteristic scales in the problem ( for example the smallest / largest eigenvalue of the hamiltonian matrix ) . for the last part of the integral a substitution @xmath106 is performed .
we use an adaptive gauss legendre integrator for the evaluation .
we have evaluated several benchmarking dynamic quantities of the siam . in the following , results for the impurity density of states will be presented and compared to ed , nrg and dmrg data .
we will elaborate on the strengths and weaknesses of the methods as well as the comparison of cpt to vca .
furthermore , we will discuss the relation between vca@xmath85 , where the variational parameters are determined self consistently via eq.([eq : sccond ] ) and vca@xmath84 , where the variational parameters are defined at the stationary point of the grand potential .
we will show that the kondo resonance is reproduced within the framework of cpt / vca and that the variational results fulfill certain analytic relations like the friedel sum rule ( eq.([eq : fsr ] ) ) .
the method will be shown to provide reasonably accurate results in a wide range of parameter regimes of the model .
low energy properties related to the kondo temperature @xmath0 will be discussed in context with renormalization group results .
the imaginary frequency green s function and self - energy will be compared to ct - qmc results .
cpt / vca rely on the green s function of an interacting cluster of size @xmath24 which is obtained by exact diagonalization . due to this fact , it is unavoidable that some effects of the finite size cluster influence the solution of the full system .
( except in the case of vanishing interaction strength , i.e. @xmath107 . ) therefore suitable clusters have to be chosen on a basis of physical results .
some aspects of this are discussed by balzer _
et al . _
@xcite in the context of dmft and vca and by hand _
et al . _
@xcite in the context of dmrg . in this work
we consider interacting clusters of even size only .
for these systems the ground state does in general not suffer from spin degeneracy .
furthermore , the spatial position of the impurity is important .
this can be inferred from the bath s density of states , which vanishes for @xmath108 at every second site .
it may also be seen in the structure of the ground state , for which the size of the degenerate sectors alternates with the geometrical size of the cluster . throughout this work
we position the impurity f - orbital at the beginning of the infinite chain , although essentially the same results are achieved by attaching it to an s - orbital at site two , four , etc . inside the chain .
+ the single - particle spectral function @xmath109 is obtained from the retarded green s function @xmath110 ( see e.g. ref . ) @xmath111 the diagonal element at the impurity f - orbital @xmath112 describes the impurity density of states @xmath113 .
a physical property of the siam which poses a challenge to numerical methods is the kondo - abrikosov - suhl resonance often referred to as kondo peak @xcite .
it arises in the parameter regime where the magnetic moment of the impurity is screened by the conduction electrons to form a singlet state @xcite .
the particle - hole symmetric model lies in the center of this kondo region .
this quasi particle excitation is for example not captured in mean field approaches . with increasing interaction strength @xmath13 the numerical solution becomes increasingly challenging .
+ in this section we elaborate on the results for the density of states in the particle - hole symmetric case . + a comparison of the local single - particle spectral function at the impurity f - orbital as obtained by exact diagonalization and vca@xmath84 is shown in fig.[fig : ed_cpt_vca ] . the ed result is for a ten site system with open boundary conditions .
the vca@xmath84 result is for an infinite reference system , where the interacting part of the reference system was taken to be of size @xmath114 .
an ed treatment of a finite - size siam can not reproduce the low energy resonance at zero energy in the single - particle spectral function ( see also app .
[ sec : appendixfsr ] ) .
it consists , in the particle - hole symmetric case , for an even number of sites ( and open boundary conditions ) , of symmetrically lying excitations which shift closer to zero for increasing system size and represent a large energy scale . for an odd number of sites
a pole in the local single - particle green s function of the impurity f - orbtial will be present at @xmath108 .
cpt as well as vca are able to reproduce finite spectral weight at @xmath108 ( even for @xmath115 ) , since these methods are formulated for an infinite system . the finite - size structure in the high energy incoherent part of the spectrum , owing from the excitations of the interacting part of the reference system , is strongly reduced in vca@xmath84 . + ( cyan ) .
vca@xmath84 was used with two variational parameters : the hopping @xmath6 and the hybridization @xmath19 considering a length of the interacting part of the reference system of @xmath114 .
all data shown is for an interaction strength of @xmath116 .
all results have been obtained for a large numerical broadening @xmath117 .
the inset shows a zoom to the low energy region.,scaledwidth=48.0% ] results for the single - particle spectral function eq.([eq : spectral ] ) of the impurity f - orbital are shown in fig.[fig : spectraphsym ] for four different interaction strengths @xmath118 . as a reference ,
the spectra obtained with nrg and dmrg from peters @xcite are plotted .
renormalization group approaches like nrg are especially suited to reproduce the low energy quasi particle excitations of this model and therefore serve as a reference for our data .
the spectra of peters were obtained for a flat conduction electron density of states , which was mapped by linear discretization in energy space onto the corresponding orbitals of a semi - infinite chain .
our model is based on a semi - circular density of states of the conduction electrons .
the low energy part of the spectra is comparable because we have chosen the only relevant parameter for the low energy part of the spectrum : @xmath27 accordingly .
this parameter fully determines the influence of the conduction electrons on the impurity f - orbital for low energies and therefore the low energy part of the spectrum .
the high energy part of the spectrum may deviate slightly and is not directly comparable but yields a crude reference .
in addition , we have chosen here a very large numerical broadening of @xmath117 for reasons of comparison only .
this value was used in the dmrg calculations and is needed there to obtain spectra using the correction vector method .
this influences the width and the height of the kondo resonance , located at @xmath108 .
the cpt spectral weight at @xmath108 appears too broad in the plot in comparison with the nrg result .
this is only partly due to a large numerical broadening . due to the nature of the cpt method
we can not expect it to reproduce the low energy spectrum as well as rg calculations do .
the height of the kondo resonance appears too small in this figure because of the large @xmath56 .
it converges with @xmath119 to the result predicted by scattering theory ( see fig.[fig : spektralinset ] and sec . [
sec : efscan ] ) .
the high energy incoherent parts of the spectrum located at @xmath120 and @xmath121 develop more and more with increasing length of the interacting part of the reference system @xmath24 . a comparison of the center of gravity of the high energy incoherent part of the spectrum of the @xmath122 site cpt result and the fifty site dmrg result is in reasonable agreement .
there are spurious structures in the spectral density , originating from the cluster green s function of the finite system , preventing continuous spectra to form .
we would like to note that the accurate determination of the green s function of the reference system is of prime importance .
an inaccuracy in pole - positions or pole - weights for very small but non - vanishing weights will yield spurious artifacts in the spectra in the vicinity of @xmath108 .
+ to improve on the result of cpt we considered the hopping matrix element @xmath6 and the hybridization matrix element @xmath19 as variational parameters .
the parameters used for the evaluation of the reference system were determined with two different methods .
vca@xmath84 results are depicted in the plot for a length of the interacting cluster of @xmath114 . as shown in the figure
this method strongly reduces the finite size peaks in the high energy incoherent part of the spectrum .
the width of that part of the spectrum is reproduced correctly for high values of @xmath13 where the full width at half maximum ( fwhm ) within vca is given by @xmath123 .
this comes very close to the expected @xmath124 @xcite of the high energy atomic excitations .
vca improves the spectral properties of the kondo resonance with respect to cpt , bringing it closer to the fifty site dmrg result .
the data obtained using the self consistent vca approach vca@xmath85 agree very well with the result based on vca@xmath84 on the position of the spectral features .
the respective weight however disagrees for low values of interaction strength @xmath13 , which is due to the different values predicted for the variational parameters by the two procedures .
one should note that the two broad lorentzian high energy peaks ( in vca@xmath84 as well as vca@xmath85 ) consist of many excitations which will be revealed upon repeating this calculation with smaller @xmath56 .
+ ) is shown resolved in real space .
the spectrum was obtained using cpt on a @xmath122 site interacting cluster .
the impurity parameters were @xmath116 , @xmath125 and the numerical broadening was set to @xmath117 .
the spectrum shown in fig.[fig : spectraphsym ] ( c ) corresponds to the data shown for the impurity f - orbital located at site @xmath20 in the plot .
the density plot is shown with a logarithmically scaled coloring from blue indicating zero to red indicating high values.,scaledwidth=48.0% ] a more detailed look on the spectral region of the kondo resonance is provided in fig.[fig : spektralinset ] .
the cpt / vca@xmath87 data is compared to nrg and frg data as well as results obtained from a restricted hartree - fock calculation from karrasch _ et al . _
the cpt / vca results are plotted for lengths of the interacting part of the reference system @xmath126 for two different sets of parameters .
the results for higher @xmath24 are always located towards the center of the figure .
the results corresponding to the resonance at @xmath108 were obtained for the particle - hole symmetric model .
for this set of parameters we used the hybridization @xmath19 as a variational parameter .
the second peak shown centered around @xmath127 corresponds to a parameter set right at the border of the kondo region .
the variational parameters used away from particle - hole symmetry are @xmath86 .
one can see that the cpt result is not converged for @xmath114 site interacting clusters yet .
in contrast , the vca@xmath87 result seems to converge much faster .
although in the plot it looks like the vca result does not improve much upon a restricted hartree fock calculation , we will show in the following that cpt / vca yields results in all parameter regimes of the siam which can not be reproduced within a mean field treatment .
+ corresponds to the particle - hole symmetric model : @xmath128 , @xmath129 , while the resonance away from zero corresponds to a set of parameters right at the edge of the kondo region : @xmath128 , @xmath130 . for comparison
we show nrg ( yellow ) and frg ( dark brown ) data as well as results obtained from a restricted hartree - fock calculation ( blue ) from karrasch _ et al . _
@xcite ( the nrg results are partially hidden by the frg results . ) .
the cpt result ( cyan ) is shown for lengths of the interacting part of the reference system @xmath131 .
results for higher @xmath24 are always located towards the center of the plot . in the particle - hole symmetric case
vca@xmath84 ( magenta ) was performed with variational parameters @xmath89 for @xmath126 .
away from particle - hole symmetry vca@xmath84 was performed with variational parameters @xmath86 for the same lengths of the interacting part of the reference system @xmath24 . for the cpt / vca calculations a numerical broadening of @xmath132
the inset shows a zoom to the top region of the peaks.,scaledwidth=48.0% ] the variational parameters obtained for the two sets of parameters shown in fig.[fig : spektralinset ] are presented in fig.[fig : varpar ] .
in addition to the vca@xmath87 parameters , which were used for the results above , the variational parameters obtained in vca@xmath85 are also depicted .
we plotted the difference of the parameter of the reference system @xmath72 to the physical parameter @xmath76 : @xmath133 .
all deviations @xmath133 appear to converge to zero with increasing length @xmath24 of the interacting part of the reference system .
notice that the self consistent approach always leads to a @xmath133 of greater magnitude with respect to vca@xmath84 . remarkably , the spectra obtained by vca@xmath84 and vca@xmath85 for the parameter set @xmath86 are in very good agreement even though the variational parameters are rather different .
the most striking difference is that the self consistent approach yields a negative @xmath134 while the @xmath1 based vca yields a positive @xmath134 .
this is however compensated by the different @xmath135 . using the hybridization @xmath19 as a variational parameter , the @xmath136 obtained by vca@xmath87 and vca@xmath85 agree rather well .
remarkably , the resulting density of states is very different , which shows that the calculation is extremely sensitive to this parameter .
+ . shown is the difference of the parameters of the reference system @xmath72 to the physical parameter @xmath76 : @xmath137 .
parameters obtained by vca@xmath84 ( crosses ) are compared to those obtained by vca@xmath85 ( circles ) .
the variational parameters @xmath134 ( dark brown ) and @xmath135 ( yellow ) correspond to the calculation away from particle - hole symmetry in fig.[fig : spektralinset ] while the parameter @xmath136 ( olive ) corresponds to the calculation at particle - hole symmetry .
lines are only guides to the eye.,scaledwidth=48.0% ] a spatially resolved image of the spectral function , calculated with cpt , for the parameter set used in fig.[fig : spectraphsym ] ( c ) is shown in fig.[fig : spatialspectrum ]
. the qualitative picture would be the same in vca , merely the structures are slightly shifted .
this view reveals how the perturbation , introduced by the impurity , is fading away slowly in an alternating fashion . at every second site away from the impurity
a dip at @xmath108 is present , which is usually referred to as fano dip .
+ the low computational effort of cpt / vca proofs advantageous for calculating spectra .
the vca procedure ( for a twelve site interacting cluster ) usually converges in minutes to hours on a standard workstation pc , while more demanding numerical methods often need days to a week to converge .
furthermore , the spectra are exactly determined from the lehmann representation and no ill - posed analytical continuation is required in comparison to methods working in imaginary time or imaginary frequency space . to our knowledge
the most accurate spectra available for this model so far are published in ref . .
+ overall one may conclude that cpt , vca@xmath84 and vca@xmath85 reproduce a kondo resonance , which fulfills the friedel sum rule ( eq.([eq : fsr ] ) ) for @xmath138 .
the vca results improve drastically upon the cpt data which may be seen in a much faster convergence in @xmath24 and a suppression of finite size effects especially in the high energy part of the spectrum which in addition has the expected width within vca .
vca@xmath84 and vca@xmath85 agree rather well on the position of the spectral features .
however they assign very different spectral weight to them at low values of interaction strength @xmath13 .
+ the occupation of the impurity f - orbital is given at @xmath139 by @xmath140 this integral may be evaluated from the imaginary frequency green s function , which in turn is directly accessible within cpt / vca . + to see whether cpt / vca are good approximations in all parameter regions of the siam , we vary the on - site energy of the impurity @xmath16 at fixed interaction strength @xmath13 .
the local impurity density of states at the chemical potential ( @xmath141 ) and the impurity occupation number are plotted for various lengths of the interacting part of the reference system @xmath142 for the same model parameters .
the vca@xmath84 result is shown in fig.[fig : efscanvca ] , a vca@xmath85 calculation in fig.[fig : efscanscn ] and the cpt data in fig.[fig : efscancpt ] .
+ we start out by discussing the vca@xmath84 result ( fig.[fig : efscanvca ] ) .
the variational parameters @xmath76 used within vca@xmath84 are the on - site energy of the impurity @xmath16 and the on - site energies of the uncorrelated cluster sites @xmath5 .
the density of states @xmath143 displays a pronounced plateau which is related to the existence of a quasi particle peak ( kondo resonance ) pinned at the chemical potential .
the parameter regions leading to an empty ( @xmath144 ) or to a doubly occupied ( @xmath145 ) impurity do not show a pinning of the kondo resonance at the fermi energy , as expected . in the half filled region which lies in between , virtual spin fluctuations lead to a pronounced quasi particle peak at the chemical potential .
we observe that the result converges with increasing length of the interacting part of the reference system @xmath24 to the physically expected result .
due to the variational parameters considered , the deviations of the results as a function of @xmath24 are rather small as compared to cpt where the results change significantly with increasing size of the reference system ( see fig.[fig : efscancpt ] ) .
we expect cpt calculations in the empty or doubly occupied regions to converge rather fast ( within a few sites ) while calculations in the kondo regime , and particularly in the crossover region , may fully converge only at very large ( i.e. exponentially ) sizes of the reference system @xcite .
this is inferred from the spin - spin correlation function in the cluster which is observed to decay sufficiently fast outside the kondo plateau ( i.e. it is effectively zero at the boundary of the cluster ) but shows long range correlations inside the plateau .
the vca@xmath85 results are obtained with one variational parameter @xmath146 .
the reason for not using @xmath86 again is that the result is almost the same as the one obtained with vca@xmath84 ( see fig.[fig : efscanvca ] ) .
however in some ( small ) parameter regions the numerical evaluation becomes difficult .
the vca@xmath85 data shown in fig.[fig : efscanscn ] shows a clear improvement as compared to cpt but does not reach the quality of the vca@xmath84 result in terms of convergence in system size .
+ the friedel sum rule ( fsr ) @xcite provides an exact relation between the extra states induced below the fermi energy by a scattering center and the scattering phase shift .
it also holds true for interacting systems .
this gives a relation between the f - orbital occupation @xmath147 , and the density of states at the fermi energy : @xmath148 in our case the mean occupation in the kondo regime is @xmath149 .
since both , the occupancy of the f - orbital and the magnitude of the local density of states at the fermi energy , can be evaluated independently , we can check the validity of the friedel sum rule in our approximation .
results are shown in fig.[fig : efscanvca ] applied to the @xmath150 site vca@xmath84 data .
the vca@xmath84 results fulfill the friedel sum rule almost in the whole kondo region . at the crossover to an empty or doubly occupied impurity
the friedel sum rule is not fulfilled exactly any more but approximated very well .
further outside the agreement is again excellent .
the variational parameters of vca are crucial to fulfill the friedel sum rule as can be seen from a cpt calculation ( fig.[fig : efscancpt ] ) which violates it in all parameter regions .
it appears that vca@xmath84 with variational parameters @xmath86 naturally drives the system to fulfill this condition .
the vca@xmath85 result ( fig.[fig : efscanscn ] ) violates the sum rule too .
this is not a feature of vca@xmath85 in general but has to do rather with the choice of variational parameters , which was just @xmath146 in this case .
the vca@xmath85 result for two variational parameters @xmath86 looks qualitatively like the respective vca@xmath84 result .
+ at @xmath108 and average occupation of the impurity ( dashed lines ) for different lengths of the interacting part of the reference system @xmath151 ( blue , green , red and cyan ) as a function of the impurity on - site energy @xmath16 . the coulomb interaction @xmath13 is kept constant at @xmath128 .
the numerical broadening used is @xmath132 .
the set of single - particle parameters considered for variation within vca@xmath84 is @xmath86 .
note that here the point @xmath152 corresponds to the particle - hole symmetric case .
the friedel sum rule ( eq.([eq : fsr ] ) ) was applied to the @xmath150 result ( dotted - violet ) .
it is fulfilled to a very good approximation in the kondo region and far outside of it .
small deviations from the friedel sum rule arise at the crossover region to an empty or doubly occupied impurity .
the inset shows a zoom to the kondo plateau.,scaledwidth=48.0% ] at @xmath108 and average occupation of the impurity ( dashed lines ) for different lengths of the interacting part of the reference system @xmath151 ( blue , green , red and cyan ) as a function of the impurity on - site energy @xmath16 .
the coulomb interaction @xmath13 is kept constant at @xmath128 .
the numerical broadening used is @xmath132 .
the set of single - particle parameters considered for variation within vca@xmath85 is @xmath146 .
note that here the point @xmath152 corresponds to the particle - hole symmetric case .
the friedel sum rule ( eq.([eq : fsr ] ) ) was applied to the @xmath150 result ( dotted - violet ) .
it is fulfilled in a region of @xmath153 and @xmath154.,scaledwidth=48.0% ] and average occupation of the impurity ( dashed lines ) for different lengths of the interacting part of the reference system @xmath151 ( blue , green , red and cyan ) as a function of the impurity on - site energy @xmath16 .
the coulomb interaction @xmath13 is kept constant at @xmath128 .
the numerical broadening used is @xmath132 .
note that here the point @xmath152 corresponds to the particle - hole symmetric case .
the friedel sum rule ( eq.([eq : fsr ] ) ) was applied to the @xmath150 result ( dotted violet ) .
it is drastically violated .
however the results are far from converged for the small lengths of the interacting part of the reference system considered here.,scaledwidth=48.0% ] scanning the interaction strength @xmath13 at fixed impurity on - site energy @xmath16 confirms the presence of the kondo behavior . shown in fig.[fig : uscan ] are results obtained with vca@xmath84 using the same variational parameters @xmath86 as above . in the weakly correlated part ( @xmath155 )
the density of states at the chemical potential is low .
the intermediate region ( @xmath156 ) signals the crossover to the kondo regime . for larger
@xmath13 the kondo regime is reached with an impurity occupation of @xmath157 , which may be inferred from the friedel sum rule . in the inset of the figure ,
the cpt results for the same lengths of the interacting part of the reference system @xmath24 are shown .
the cpt results are by far not converged for the interacting cluster sizes considered here .
this emphasizes the importance of the variational parameters . + for different lengths of the interacting part of the reference system @xmath142 ( dark brown , cyan , olive and magenta ) as a function of the interaction strength @xmath13 .
the impurity on - site energy @xmath16 is kept constant at @xmath129 .
the numerical broadening is chosen to be @xmath132 .
the set of single - particle parameters considered for variation within vca@xmath84 is @xmath86 .
the inset shows the cpt results.,scaledwidth=48.0% ] our results in fig.[fig : efscanvca ] and fig.[fig : uscan ] agree very well with those of calculations based on x - operator technique exercised by lobo _
et al . _
@xcite . in their work
a strong coupling perturbation theory is applied starting from the anderson molecule as a basis and using the friedel sum rule as a condition to fix the position of an infinitely narrow conduction band .
+ analytic considerations ( see app . [
sec : appendixfsr ] ) allow insight into the behavior of the friedel sum rule in ed , cpt and vca .
there it is shown that ed always has to violate the friedel sum rule while cpt always fulfills it in the particle - hole symmetric case .
this comes about in the first place because the height of the kondo resonance at @xmath108 does not depend on the self - energy .
a pinning of the kondo resonance however can only be achieved via the improved self - energy contributions obtained within vca .
+ the results of this section clearly show that vca is able to capture the basic physics of the siam in every parameter region .
the improvement obtained by going over from cpt to vca is crucial to fulfill exact analytic relations .
moreover we have shown that cpt / vca is clearly superior to ed calculations .
vca is , even at small @xmath24 , capable of fulfilling the fsr also away from particle - hole symmetry due to a pinning of the kondo resonance at the fermi energy .
this pinning can be attributed to the better approximation of the self - energy of vca with respect to cpt .
+ to delve into the cpt / vca results for the whole parameter range of the siam a `` phase diagram '' is presented in this section .
this should be understood to be a mere scan of the parameters @xmath13 and @xmath16 because the model does not undergo a phase transition .
the density of states of the impurity at the chemical potential @xmath143 is shown in fig.[fig : phasediagramrho ] in a density plot .
this figure essentially shows the height of the kondo resonance as a function of interaction strength and on - site energy of the impurity .
the different regimes of the siam , as obtained by an atomic limit calculation , are indicated as black lines .
these lines divide the physics into regions where the impurity is doubly , singly or not occupied . in the singly occupied region ( @xmath158 ) local moments and their screening are expected to appear .
this region , which bestrides the cone enclosed by black lines , is the region where kondo physics may take place within this approximation .
the parameter regions where the impurity is empty or doubly occupied lie above and below this cone .
more sophisticated methods will lead to a smearing out of the border of these regions and introduce a crossover area with competing effects . a boundary expected between a single resonance and a spurious local moment behavior where the single resonance is split into two for spin up and spin down respectively
is obtained by mean field theory @xcite . in
the mean field approach the density - density interaction of the impurity hamiltonian is replaced by a spin dependent density exposed to the mean contribution of the other spins density .
the mean field boundary is then obtained by replacing the mean field parameters @xmath159 and @xmath160 by the particle number @xmath161 and magnetization @xmath162 .
setting the magnetization to @xmath163 in the self consistent equations yields the implicit result @xmath164 and @xmath165 .
the plot shows that the kondo plateau is reproduced very well by vca@xmath87 .
the results appear almost converged for lengths @xmath166 of the interacting part of the reference system . increasing @xmath24 yields better results in the crossover region .
results obtained by means of cpt do not reproduce the kondo plateau very well for small @xmath24 .
+ at @xmath108 ( i.e. the height of the kondo resonance ) .
the results are obtained with vca@xmath84 for a set of variational parameters @xmath86 , @xmath167 and @xmath132 .
the black line indicates the different regions obtained from an atomic limit calculation . in the right cone
local moments are to be expected . while in the upper region
the impurity is expected to be empty and in the lower half to be doubly occupied . the blue curve shows the onset of a spurious magnetic state as obtained by a mean field treatment ( see text).,scaledwidth=48.0% ] the average impurity occupation for the same parameter region is shown in fig.[fig : phasediagramn ] .
the result obtained with vca@xmath87 clearly shows the kondo plateau where the impurity is singly occupied .
the parameter regions of a doubly occupied or empty impurity lead to a density of states in the impurity which is zero at the chemical potential ( compare to fig.[fig : phasediagramrho ] ) .
+ as a function of @xmath13 and @xmath16 for the same parameters as in fig.[fig : phasediagramrho].,scaledwidth=48.0% ] the results of this section have been obtained using vca@xmath84 with variational parameters @xmath86 .
it should be noted that using only @xmath146 already yields good results . as mentioned in sec .
[ sec : efscan ] , cpt needs very large values @xmath24 to yield the same quality of the results as vca does with much smaller values of @xmath24 .
+ in this section we examine the low energy properties of the symmetric siam . in the strong coupling limit a single scale , the kondo temperature @xmath0 ,
governs the low energy physics @xcite .
+ the kondo temperature @xmath0 is known from bethe ansatz results for the particle - hole symmetric siam @xcite @xmath168 this scale , which is inversely proportional to the spin - flip rate of the impurity , divides the physics of the siam into two regions : a local moment behavior of the impurity , where the spin is free , and a low temperature region where the local spin and the conduction electrons become entangled and form a singlet state @xcite . + quantities which depend inversely on @xmath0 are the effective mass @xmath169 and the static spin susceptibility @xmath170 .
the kondo temperature may furthermore be extracted from the width or weight of the kondo resonance in the local density of states of the impurity f - orbital .
we investigate and compare the results for the scale @xmath0 obtained from the direct determination of @xmath0 ( from the fwhm and the spectral weight of the kondo resonance ) and the inverse quantities @xmath169 and @xmath170 .
we find that the results of all four measurements turn out to yield the correct qualitative behavior in vca@xmath84 .
however , in a region where the dependence of @xmath0 is exponentially dependent on the interaction strength @xmath13 the exponential prefactor is not predicted correctly .
therefore we introduce a scaling factor @xmath53 ( eq.([eq : tk ] ) ) which turns out to be the same for all four ways of determining @xmath0 . in particular
this factor is independent of the set of model parameters used .
the scaling factor may be calculated semi - analytically for a reference system consisting of a two site interacting cluster and the semi - infinite environment within vca@xmath84 and vca@xmath85 ( @xmath89 ) .
the calculation for vca@xmath84 leads to an integral expression for the stationary point of the grand potential @xmath1 with respect to @xmath136 from which the optimal @xmath136 can be obtained numerically ( see app . [
sec : appendixtwositeanalytical ] ) . the kondo scale may be determined from the so obtained values of @xmath171 by @xmath172 this leads to a perfect exponential behavior as defined in eq.([eq : tk ] ) with @xmath173 the issue of obtaining an exponential scale but not the correct exponent for the functional dependence on @xmath13 is common to various approximate methods ( for example variational wave functions where the issue was cured by introducing an extended ansatz by schnhammer @xcite , saddle - point approximations of a functional integral approach @xcite or frg @xcite ) .
a faint analogy may be drawn here to gutzwiller approximation , where an exponential energy scale in @xmath13 arises by a renormalized hybridization parameter @xmath19 @xcite , which is also the case for vca@xmath84 .
+ the self consistent calculation for vca@xmath85 also leads to an integral expression for the determination of @xmath136 .
this expression is obtained by requiring the expectation values of the hopping from the impurity f - orbital to the neighboring site in the reference system to be the same as the expectation value in the physical system .
this procedure does not yield an exponential scale in @xmath13 .
the optimal cluster parameter @xmath82 shows spurious behavior as a function of @xmath13 .
we conclude that vca@xmath85 with @xmath89 can not reproduce the low energy properties of the siam even qualitatively , while vca@xmath84 yields the correct behavior apart from an exponential factor .
the effective mass @xmath169 is defined as the quasi particle renormalization @xcite @xmath174}{d\,\omega}\bigg|_{\omega=0^+ } \\
\nonumber & = \frac{d[\im{\text{m}}\,{\text{g}}^\sigma_{ff}(i\omega , u)]}{d\,\omega}\bigg|_{\omega=0^+}\,\times\\ & \left(\frac{d[\im{\text{m}}\,{\text{g}}^\sigma_{ff}(i\omega,0)]}{d\,\omega}\bigg|_{\omega=0^+}\right)^ { -1}\;\mbox { , } \label{eq : meff}\end{aligned}\ ] ] where we introduced the dependence on the interaction strength @xmath13 explicitly . in the kondo regime , this quantity becomes inversely proportional to the kondo temperature .
+ we want to answer the question whether the kondo scale is approximately captured by cpt / vca or not .
therefore we compare the functional form and the exponent obtained from the effective mass and the analytic result for @xmath0 , eq.([eq : tk ] ) .
the result for the effective mass obtained within vca@xmath84 is shown in fig.[fig : mstar ] .
the variational parameter used was @xmath89 .
the functional form is reproduced well by vca@xmath84 ( i.e. it starts out quadratically and goes over to an exponential behavior in the kondo region ) .
however the exponent @xmath175 is not reproduced correctly .
vca@xmath84 yields a lower exponent of @xmath176 .
the factor @xmath53 is defined in app .
[ sec : appendixtwositeanalytical ] , determined from a semi - analytical calculation of @xmath0 within vca@xmath84 .
this additional factor is the same for all initial parameters ( within the kondo regime ) , it is particularly independent of @xmath27 . if plotted over a scaled @xmath13 axis , @xmath177 , the vca@xmath84 result would lie on top of the nrg data .
however using larger sizes of the interacting part of the reference system does not lead to much better results regarding @xmath53 .
it is to be expected , that a significant improvement can only be obtained using exponentially large @xmath24 .
the cpt result shows a very different convergence behavior in @xmath24 which is rather slow .
+ an attempt was made to extrapolate the cpt data to @xmath178 by a simple @xmath179 scaling .
it is interesting to observe that this extrapolated curve coincides nicely with the vca result ( @xmath180 ) in the low @xmath13 region .
however , we expect this extrapolation based on small @xmath24 to be insufficient to capture the exponential scaling of the data in @xmath24
. note that since in cpt the self - energy is taken from the cluster , the cpt results for the effective mass coincide with ed results for systems of size @xmath24 . + of the kondo resonance eq.([eq : meff ] ) as a function of interaction strength @xmath13 .
we plot cpt results for lengths of the interacting part of the reference system @xmath181 ( magenta ) , the to @xmath178 extrapolated cpt result ( olive ) , as well as vca@xmath84 results ( blue ) .
the data points for the cpt result in the low @xmath13 region are not shown to avoid messing up the plot .
the variational parameter used for the vca@xmath84 result was @xmath89 .
the vca@xmath84 data was obtained for @xmath167 . for cpt as well as vca@xmath84 , we used a numerical broadening of @xmath132 . for comparison the results obtained by nrg ( yellow ) and frg ( dark brown )
are shown @xcite.,scaledwidth=48.0% ] since the height of the kondo resonance is fixed by the friedel sum rule eq.([eq : fsr ] ) the width and the weight ( area ) of the peak are proportional to the kondo temperature @xmath0 . obtaining the spectral weight or fwhm of the kondo resonance from the spectrum introduces a large uncertainty .
nevertheless we made an attempt , to get an idea of the behavior of @xmath0 .
we fixed the spectral weight by the first minimum to the left and to the right of the central peak ( see also ref . ) .
in general the effective mass and static spin susceptibility will yield more reliable results but it is instructive to compare these four ways of determining @xmath0 . + shown in fig.[fig : tkspectweightfwhm ] is the evolution of the spectral weight and the fwhm of the kondo resonance with increasing interaction strength @xmath13 .
the data was acquired using vca@xmath84 with a variational parameter @xmath89 for the particle - hole symmetric siam . within the uncertainty ,
the same exponential behavior for the kondo temperature @xmath0 is obtained as by calculating the effective mass in vca@xmath84 .
+ results for the spectral weight ( blue ) and full width at half maximum ( olive ) of the kondo resonance as a function of interaction strength @xmath13 .
the variational parameter used was @xmath89 .
a length of the interacting part of the reference system of @xmath167 sites and a numerical broadening of @xmath132 were used for this calculation .
data points marked with a circle were used for the fit of the exponential function in the kondo region .
the black line shows the kondo temperature @xmath0 as obtained by bethe ansatz calculations eq.([eq : tk]).,scaledwidth=48.0% ] the static spin susceptibility @xmath170 is given by the linear response to an applied magnetic field @xmath182 in z direction @xmath183 in the kondo regime this quantity too becomes inversely proportional to the kondo temperature .
for the calculations in this section we introduce an additional spin dependent term in the impurity hamiltonian ( eq.([eq : himpurity ] ) ) @xmath184 the static spin susceptibility @xmath170 as obtained with vca@xmath84 is shown in fig.[fig : chim ] .
the variational parameter used was @xmath89 .
as a reference the results of nrg and frg @xcite are shown .
the behavior of the vca result is good for small interaction strength @xmath13 .
the vca result shown for @xmath180 appears already converged while the cpt result would require much larger systems .
+ eq.([eq : chim ] ) is shown as a function of interaction strength @xmath13 .
the variational parameter used was @xmath89 .
the data was obtained for @xmath167 sites and @xmath132 . for comparison the results obtained by nrg ( blue ) and frg ( green )
are shown @xcite.,scaledwidth=48.0% ] we would like to highlight that vca@xmath84 reproduces an energy scale @xmath0 .
results from direct calculation of @xmath0 , calculation of the effective mass @xmath169 and the static spin susceptibility @xmath170 yield the correct functional form but not the right exponent .
+ in this section we compare cpt / vca results to qmc data .
we obtained the monte carlo results using the continuous time quantum monte carlo ( ct - qmc ) code of the triqs @xcite toolkit and its implementation of the hybridization expansion ( ct - hyb ) @xcite algorithm using legendre polynomials @xcite .
this method enables access to very low temperatures and is especially suited to obtain low energy properties @xcite .
the ct - qmc data provides statistically exact and reliable results to test our data .
+ all ct - qmc calculations were done for a single impurity orbital at @xmath185 and @xmath186 .
we used a semicircular hybridization function with half bandwidth @xmath187 and @xmath33 .
this setup corresponds to the same model under investigation here .
the value for the interaction strength @xmath185 was chosen because of the relatively low expected kondo temperature of @xmath188 .
for all calculations @xmath189 mc updates where conducted , with a sweep size of @xmath190 updates , plus a @xmath191 thermalization period .
+ to ensure that the kondo resonance is correctly reproduced by ct - qmc we evaluated the matsubara green s function for various values of inverse temperature @xmath62 .
the height of the kondo resonance is given by the friedel sum rule eq.([eq : fsr ] ) to be @xmath192 for the parameters used here ( @xmath34 ) . to obtain @xmath193
we extrapolate twice , first in @xmath194 for each @xmath62 , then we use these results and extrapolate to @xmath195 . the extrapolation to @xmath194 is done linearly using the first two matsubara frequencies . the imaginary part of @xmath196 and the extrapolated value to @xmath194 are shown in the inset of fig.[fig : ctqmcextrapolation ] for @xmath197 $ ] .
those extrapolated values are plotted as a function of temperature ( fig.[fig : ctqmcextrapolation ] ) .
these data points are then extrapolated to @xmath198 using a fit by a rational model function .
the result clearly shows the onset of the kondo resonance when the temperature is lowered below the kondo temperature @xmath0 .
the extrapolation to @xmath139 shows very good agreement ( @xmath199 ) with the result expected from the friedel sum rule within the uncertainty .
it is important to note that the ct - qmc results converge very nicely in @xmath62 .
although for higher @xmath62 lower matsubara frequencies become available , the overall shape of the green s function does not change significantly . + .
an extrapolation to zero temperature is attempted , which yields a good agreement with the result predicted by the friedel sum rule ( green circle ) within the uncertainty ( red triangle ) .
the inset shows the imaginary part of the impurity green s function for various @xmath62 ( see legend ) and the extrapolated points at @xmath200.,scaledwidth=48.0% ] therefore we may compare the @xmath139 cpt / vca results for the green s function and self - energy to the ct - qmc data .
the matsubara green s functions of the impurity f - orbital @xmath201 obtained by ct - qmc ( @xmath202 ) , cpt and vca are shown in fig.[fig : ctqmcvscpt_g ] .
we use @xmath202 as a compromise between low temperatures and still reliable ct - qmc results ( within manageable computation time ) .
the @xmath202 result was obtained using @xmath203 legendre coefficients .
a detailed analysis has shown that this number is sufficient to get high frequency moments of the self - energy @xmath66 accurately .
the vca@xmath84 results were obtained with one variational parameter @xmath89 for @xmath185 , @xmath34 and @xmath132 in the particle - hole symmetric case . for the cpt calculation we used the same parameters . for both methods we considered lengths of the interacting part of the reference system of @xmath181 .
the vca result lies near the ct - qmc data but underestimates the slope of the curve at low @xmath204 .
the vca result provides a huge improvement upon cpt for the lengths of the interacting part of the reference system shown here .
the real part of @xmath201 is exactly zero within cpt / vca as it is supposed to be .
note that the value of @xmath205 which is fixed by the friedel sum rule is exactly reproduced within cpt and vca for the particle - hole symmetric case .
obtained by ct - qmc ( @xmath202 ) , cpt and vca@xmath84 .
the cpt / vca results were obtained for @xmath181 .
the real part shown in the lower part of the figure is zero .
the friedel sum rule prediction of @xmath206 is fulfilled by all methods .
the legend of this figure serves as well as legend for fig.[fig : sigma_large_iwcpt ] and fig.[fig : ctqmcvscpt_sigma ] .
that is why the last entry ( large @xmath204 exp .
( see eq.([eq : sigmalargeiw ] ) ) ) is displayed in the legend but is missing in the graph of this figure.,scaledwidth=48.0% ] the same is shown for the self - energy of the impurity f - orbital @xmath207 in fig.[fig : ctqmcvscpt_sigma ] . from the imaginary part of @xmath207
one can infer the convergence of the cpt / vca result with larger length of the interacting part of the reference system @xmath24 .
the real part of the self - energy ( @xmath208 is again exactly reproduced within cpt / vca .
+ in the following , we discuss the self - energy @xmath209 for the two interesting cases of very low and very high matsubara frequency .
we start out by conducting an expansion of the self - energy @xmath210 for high matsubara frequencies ( @xmath211 ) which shall be outlined here briefly .
the self - energy matrix is defined by @xmath212 here @xmath45 is the one - particle part of the hamiltonian . in the particle - hole symmetric case considered here
it contains all the hoppings as well as the on - site energy of the impurity @xmath31 .
we conduct a series expansion in powers of @xmath213 of @xmath210 . apart from the real constant @xmath214
all @xmath215-dependent terms of @xmath216 are anti - symmetric in @xmath215 .
therefore even powers in @xmath217 vanish .
expanding the green s function @xmath218 yields for the self - energy @xmath210 @xmath219 where @xmath220 and @xmath221 is the ground - state energy of @xmath70 . collecting powers of @xmath215 yields a cumulant - like expansion for the self - energy @xmath210 @xmath222 here we consider the zeroth and first order in @xmath213 only and obtain for @xmath209 @xmath223 where the self - energy at the impurity f - orbital @xmath224 is the only non - vanishing matrix element of @xmath225 .
this result is plotted as a reference in fig.[fig : ctqmcvscpt_sigma ] .
due to the nature of the cpt / vca approximation these methods always yield the exact self - energy for high matsubara frequency as shown in fig.[fig : sigma_large_iwcpt ] .
times energy @xmath226 obtained by ct - qmc ( @xmath202 ) , cpt and vca@xmath84 .
the cpt / vca results were obtained for @xmath181 .
cpt as well as vca@xmath84 become exact for high matsubara frequencies .
an expansion of @xmath209 for large @xmath204 eq.([eq : sigmalargeiw ] ) is additionally shown ( straight line at @xmath227 ) .
the legend for this figure is the same as for fig.[fig : ctqmcvscpt_g ] and is displayed there.,scaledwidth=48.0% ] obtained by ct - qmc ( @xmath202 ) , cpt and vca@xmath84 .
the cpt / vca results were obtained for @xmath181 .
an expansion of @xmath209 for large @xmath204 eq.([eq : sigmalargeiw ] ) is shown in addition ( magenta line which diverges at zero ) .
cpt / vca always reproduces the exact self - energy for high matsubara frequencies .
the legend for this figure is the same as for fig.[fig : ctqmcvscpt_g ] and is displayed there.,scaledwidth=48.0% ] the low energy properties examined in the previous section depend basically on the slope of the matsubara green s function at @xmath228 .
the results shown in fig.[fig : ctqmcvscpt_g ] and fig.[fig : ctqmcvscpt_sigma ] show that this slope is underestimated by cpt / vca in comparison to ct - qmc , at least at the small lengths of the interacting part of the reference system available .
+ the above results suggest a possible application of vca as an impurity solver for zero temperature dmft .
the results would not suffer from a bath truncation error as in exact diagonalization based dmft
. a big advantage would be the low demand on computational power of vca as well as the approximate reproduction of the main features of the local density of states ( i.e. kondo resonance and high energy incoherent part of the spectrum ) .
+ we explore the possibility to improve the vca results achieved by varying the internal single - particle parameters of the model by introducing a symmetry breaking `` spin flip field '' at the impurity f - orbital .
the term added to the impurity hamiltonian eq.([eq : himpurity ] ) @xmath229 explicitly breaks the conservation of spin in the cluster solution .
we are interested in the model with a physical parameter @xmath230 so this variable may only attain a finite value as a variational parameter @xmath231 in the reference system .
we investigate the particle - hole symmetric model at @xmath33 and @xmath32 .
our findings indicate that any finite value of @xmath231 splits the kondo resonance and has thus to be discarded on physical grounds for the system under investigation . +
while this prevents the application of this field to improve the vca results , it gives very nice insight in the physics of the siam as described by cpt / vca .
we find that a critical interaction strength @xmath232 depending on the length of the interacting part of the reference system exists which separates solutions which would prefer a finite @xmath231 from those which would prefer @xmath233 .
the critical interaction strength for @xmath234 is given by @xmath235 .
the grand potential @xmath236 is plotted for various interaction strengths @xmath13 in fig.[fig : bx ] . for an analogous calculation for @xmath180 site
interacting clusters a value of @xmath237 is achieved .
the mean field result would yield a critical interaction strength @xmath238 for the parameters used here .
we interpret this value as a signature of the onset of local moment behavior .
the values for @xmath232 are of course not to be taken literally , they depend very much on the finite size of the cluster under investigation . + ( eq.([eq : omegaresult ] ) ) as a function of the interaction strength @xmath239 ( see legend ) .
the data was obtained by studying a @xmath234 site interacting cluster coupled to a semi - infinite lead .
the numerical broadening used was @xmath132 .
the crosses indicate the respective minimum of the grand potential .
there exists a critical @xmath235 above which a finite @xmath231 is preferred by the system.,scaledwidth=48.0% ] the splitting of the kondo resonance caused by a non - zero variational field @xmath231 is shown in fig.[fig : bxspectra ] .
the value of @xmath116 used for this calculation lies in the region above @xmath232 where the system prefers a nonzero field @xmath231 . + .
the plots were obtained using vca ( i.e. the physical field @xmath240 is always zero ) . instead of taking the parameter @xmath231 at the stationary point of the grand potential ( this value would be @xmath241 for the parameters used ) we explicitly plug in a fixed value for @xmath231 .
the length of the interacting part of the reference system used was @xmath180 for the model parameters @xmath116 .
the numerical broadening used was @xmath132.,scaledwidth=48.0% ]
in this work we have applied the variational cluster approach ( vca ) to the single impurity anderson model .
we devised a cluster tiling applicable to this non - translationally invariant model which leads to a cluster with a discrete spectrum and an environment having a continuous spectrum .
we have derived an expression for the change of the grand potential originating from the coupling of the impurity to the semi - infinite bath .
+ we have compared results for the single - particle dynamics to data obtained by exact diagonalization and cluster perturbation theory ( cpt ) .
we found that the variational extension made by the vca is vital for a good reproduction of the expected behavior of the siam .
the cpt / vca spectra both yield a kondo resonance in the impurity density of states with the correct height as predicted by the friedel sum rule
. a close look at the kondo resonance shows that the vca is able to reproduce the resonance and the functional behavior for the kondo temperature in a remarkable way .
the kondo temperature is expected to show exponential behavior in interaction strength in the kondo regime .
vca yielding an exponential behavior however tends to underestimate the exponent .
comparison of dynamic quantities to continuous time quantum monte carlo solidifies the origins of this behavior .
the high energy incoherent part of the spectrum shows strong finite size effects within cpt which are partly removed by virtue of the vca .
vca furthermore reproduces the expected position and width of the high energy part of the spectrum . for the asymmetric model ,
the friedel sum rule is fulfilled in all parameter regions implying that the kondo resonance is pinned at the chemical potential in the kondo region .
in addition a self consistent formulation of the vca , previously introduced in the context of non equilibrium problems @xcite , was explored .
results obtained by the self consistent approach show agreement with results obtained by vca based on the grand potential for the density of states of the impurity f - orbital . thereby the positions of the spectral features agree very well with the traditional vca result , while the spectral weight distribution may deviate especially for small values of interaction strength .
comparison to results obtained from bethe ansatz , renormalization group approaches and data obtained from x - operator based calculations show reasonable agreement for all quantities investigated .
+ in conclusion , while there are certainly more accurate methods to deal with a single quantum impurity model @xcite , especially at low energies , our work shows that vca is a flexible and versatile method which provides reasonably accurate results with modest computational resources .
here , the vca self consistency condition proves to be crucial .
this allows to obtain the same accuracy that cpt would provide with a much larger , inaccessible , interacting cluster size .
one of the advantages is the flexibility of the method , i.e. it is straightforward to extend it to many impurities , non equilibrium problems @xcite , etc . in the spirit of nrg
@xcite one could improve on the present results by carrying out an appropriate unitary transformation on the bath such that the bath hoppings decay with increasing distance from the impurity . in this work
, we do not include such an improvement and choose a constant hopping sequence ( eq.([eq : hconduction ] ) ) , since our goal is to benchmark vca / cpt only .
nevertheless , a hybrid approach combining nrg and vca would be an interesting extension of the present work .
+ we are grateful to r. peters for providing his dmrg and nrg data shown in fig.[fig : spectraphsym ] and to c. karrasch for providing his frg and nrg data shown in fig.[fig : spektralinset ] , fig.[fig : mstar ] and fig.[fig : chim ] .
mn wants to thank k. schnhammer , t. pruschke , m. knap and p. dargel for fruitful discussions .
we thank o. parcollet and m. ferrero for important , critical feedback concerning the convergence of our ct - qmc data .
we made use of the ct - qmc code of the triqs @xcite toolkit .
this work is partly supported by the austrian science fund ( fwf ) p18551-n16 .
mn acknowledges financial support by the frderungsstipendium of the tu graz .
here we outline the proof of eq.([eq : aux3 ] ) .
we start out from eq.([eq : aux0 ] ) , i.e. @xmath242 taylor expansion yields @xmath243_{cc } + { \text{tr}\,\bigg } [ \big({\text{t}}{\text{g}}\big)^{n}\bigg]_{ee}\right\}\;\mbox{.}\end{aligned}\ ] ] due to @xmath244 each term @xmath245 in the first trace occurs only in the form @xmath246 the expressions in the second part of eq.([eq : aux1 ] ) can be modified by a cyclic permutation of the factors in the argument of the trace @xmath247 in which @xmath245 again only occurs in the dressed form @xmath248 .
if we replace all occurrences of @xmath249 by @xmath248 then there are no matrices @xmath250 and @xmath251 respectively , left .
hence we can as well introduce in eq.([eq : aux0 ] ) the following replacements @xmath252 as it leads in the series expansion to the same expressions . the argument in eq.([eq : aux0 ] ) then assumes the form @xmath253 with the abbreviation @xmath254 .
prompted by the schur complement , the matrix can be factorized into upper and lower triangular block matrices @xmath255 such that the determinant is easily computed , since the determinant of the second matrix is @xmath21 and the determinant of the first matrix is simply the product of the determinants of the diagonal blocks , resulting in @xmath256 the final result for eq.([eq : aux0 ] ) reads @xmath257
it is possible to gain a somewhat deeper understanding of the behavior of the fsr within ed / cpt / vca by considering the local green s function at the impurity f - orbital @xmath258 where @xmath259 is the contribution due to the single - particle terms of the hybridization and @xmath210 the self - energy due to the local interaction ( for details see app . [
sec : appendixtwositeanalytical ] ) . in the following
we consider the particle - hole symmetric case .
we are interested in the behavior of the retarded green s function @xmath260 at the fermi energy ( @xmath108 ) , which we investigate by taking the limit on the matsubara axis @xmath261 .
we expand the self - energy @xmath262 up to linear order in @xmath263 and rewrite the expression using the definition of the effective quasi particle mass @xmath169 ( see eq.([eq : meff ] ) ) @xmath264 inserting into eq.([eq : gfsrstart ] ) we obtain @xmath265 - \re{\text{e}(\gamma(i\nu ) } \bigg)^{-1}\;\mbox{. } \label{eq : gfsr2}\end{aligned}\ ] ] we will now investigate two separate , general cases of an ed and a cpt / vca treatment of the green s function . + from eq.([eq : gfsr2 ] ) it is easy to see that the remnant @xmath169 of the impurity self - energy and therefore the self - energy itself does not contribute to the friedel sum rule . in outlining how to notice this
, we simultaneously show that a discrete spectrum of the conduction band ( as obtained for example in ed ) will not fulfill the fridel sum rule .
consider an arbitrary discrete spectrum of the conduction electrons with hybridization @xmath266 with excitation energies @xmath267 .
splitting into real and imaginary parts and inserting into eq.([eq : gfsr2 ] ) gives @xmath268 upon neglecting the weak dependence of @xmath269 and @xmath270 on @xmath263 one obtains , in this case , in the limit @xmath271 : @xmath272 if all @xmath273 or @xmath274 if any @xmath275 .
we would like to further illustrate this for the specific model considered in this work ( i.e. an impurity coupled to a semi - infinite chain with open boundary conditions ) @xmath276 where we supressed spin indices .
we obtain by using the equation of motion for the local impurity f - orbital green s function @xmath277_+\rangle - < < [ c_f^{{\phantom{\dag}}},\mathcal{h}]_-;c_f^\dag>>\;\mbox{,}\end{aligned}\ ] ] for an even number of sites ( including the impurity ) @xmath278 and for an odd number of sites @xmath279 where @xmath280 is a rational function which is well behaved upon taking the limit @xmath281 ( i.e. it approaches a constant @xmath282 ) and the real part is always zero . upon inserting into eq.([eq : gfsr2 ] ) one can easily verify that for an even number of sites this yields zero spectral weight at the fermi energy , while for an odd number of sites it yields @xmath283 ( i.e. there is a pole exactly at @xmath108 ) .
therefore the ed results alternate with even / odd system size between @xmath284 and @xmath285 .
this result shows that the fsr is always violated in ed because a finite value of the impurity density of states at the fermi energy may only be obtained using an artificial numerical broadening .
it furthermore shows that all terms involving @xmath169 go to zero and can not contribute to the sum rule .
+ now we turn to the case of cpt / vca , where the conduction electron hybridization takes the form ( see eq.([eq : env ] ) ) @xmath286 for the model considered in eq.([eq : hsiam ] ) ( i.e. a semi - circular density of states of the conduction electrons ) . the cpt / vca green s function of the physical system is then given upon insertion of this @xmath287 into eq.([eq : gfsr2 ] ) @xmath288 which yields upon expansion of the square root up to linear order in @xmath263 @xmath289 and when the limit @xmath290 is taken @xmath291 which is exactly the value predicted by the fsr ( eq.([eq : fsr ] ) ) @xmath292 this result is independent of the size of the reference system .
away from particle - hole symmetriy ( occupation of the f - orbital not one ) the calculation becomes more tedious .
the numerical vca calculations however show that a pinning of the kondo resonance at the fermi energy is obtained in contrast to cpt at small @xmath24 .
to gain a better understanding of the behavior of the low - energy properties of the siam we solve a small system semi - analytically .
we obtain the scaling of the kondo temperature @xmath0 with interaction strength @xmath13 within vca@xmath84 as well as vca@xmath85 . a reference system consisting of a two site cluster and an infinite environment is used .
+ the kondo scale will be determined from the effective mass eq.([eq : meff ] ) using the self energy of a two site cluster eq.([eq : sigmatwosite ] ) which leads to @xmath293 the kondo scale @xmath0 is inversely proportional to @xmath294 eq.([eq : tktwosite ] ) .
therefore within this approximation the behavior of the optimal cluster parameter @xmath82 governs the low energy physics .
+ to determine the optimal hopping @xmath82 the green s function of the reference system @xmath38 is calculated @xmath295 the cpt / vca green s function @xmath36 @xmath296 is obtained by eq.([eq : dyson ] ) using @xmath297 here @xmath298 eq.([eq : gfftwosite ] ) corresponds to the impurity f - orbital , @xmath299 the second site in the interacting part of the reference system and @xmath300 eq.([eq : env ] ) the semi - infinite environment .
the cluster parameter @xmath82 is given by the physical parameter @xmath19 plus the variation @xmath136 .
note that in here we work with the reduced expressions for @xmath1 and @xmath36 justified in app .
[ sec : appendixomeganew ] .
schnhammer and brenig calculated the green s function of the correlated orbital for this model perturbativeley and showed that their expression becomes exact in the limit of vanishing bandwidth @xcite .
this is exactly the case considered here , where the impurity f - orbital is coupled to a single non interacting site providing a bath with vanishing bandwidth .
they obtained @xmath301 where the hybridization @xmath302 in our case is given by @xmath303 and the self - energy @xmath304 is given by @xmath305 from this all elements of the cluster green s function may be obtained by the equation of motion technique .
+ to be able to calculate the grand potential @xmath1 , the ground state energy of the interacting part of the reference system , @xmath306 needs to be obtained ( which can be done for example by diagonalization of the hamiltonian matrix or by an integral over the green s function ) @xmath307 using the expression for the cpt / vca green s function @xmath38 and the ground state energy and taking the derivative of the grand potential eq.([eq : omegaresult ] ) with respect to @xmath136 one is able to obtain an integral expression which allows to determine @xmath82 within vca@xmath84 : @xmath308 the resulting @xmath309 is shown in fig.[fig : twosite ] ( left ) and is used to calculate the effective mass eq.([eq : effmasstwosite1 ] ) shown right in the figure . the effective mass shows exponential behavior but the exponent does not match the bethe ansatz result as discussed in sec .
[ sec : lowenergy ] .
+ next we attempt to obtain the vca@xmath85 solution for the two - site problem .
the only variational parameter is @xmath19 and therefore we determine the expectation value of @xmath310 self consistently . here
@xmath21 denotes the impurity s s - orbital . since we are considering a spin - symmetric model we sum over both spin directions and denote this expectation value as @xmath311 in the following .
the hopping expectation value is given by @xmath312 evaluation of this expectation value in the cluster yields @xmath313 evaluation of this expectation value in the total system gives @xmath314 upon requiring the two expectation values to coincide @xmath315 the optimal value of @xmath136 is obtained numerically .
the resulting @xmath309 is plotted in fig.[fig : twosite ] ( left ) and is used to calculate the effective mass eq.([eq : effmasstwosite1 ] ) shown right in the figure .
the effective mass does not show exponential behavior in vca@xmath85 . of the reference system as obtained by the semi - analytical equations for vca@xmath84 eq.([eq : twositeomega ] ) and vca@xmath85 eq.([eq : twositesc ] ) .
as a reference the @xmath316 data of our numerical simulation is shown too .
right : the effective mass eq.([eq : effmasstwosite1 ] ) obtained by the optimized parameter @xmath82 of the reference system ( see left figure ) .
additionally shown is the bethe - ansatz @xcite eq.([eq : tk ] ) and nrg @xcite result as a reference.,scaledwidth=48.0% ] | we study the single impurity anderson model by means of cluster perturbation theory and the variational cluster approach ( vca ) .
an expression for the vca grand potential for a system in a non interacting bath is presented .
results for the single - particle dynamics in different parameter regimes are shown to be in good agreement with established renormalization group results .
we aim at a broad and comprehensive overview of the capabilities and shortcomings of the methods .
we address the question to what extent the elusive low energy properties of the model are reproducible within the framework of vca .
these are furthermore benchmarked against continuous time quantum monte carlo calculations .
we also discuss results obtained by an alternative , i.e. self consistent formulation of vca , which was introduced recently in the context of non equilibrium systems . |
the one - dimensional ( 1d ) antiferromagnetic heisenberg chain @xmath3 is one of the most thoroughly investigated paradigms of interacting many - body systems .
it is well - known that for zero magnetic field the low - energy physics depends very much on the spin length @xmath4 . as haldane predicted in 1983,@xcite integer @xmath4 chains
possess an energy gap @xmath5 above the ground state , the size of which vanishes exponentially as @xmath6 , whereas half - integer @xmath4 chains are gapless .
the presence of the haldane gap in integer spin chains implies the occurrence of a critical field @xmath7 beyond which the system gets magnetized .
as the gap collapses at the critical field the ground state structure changes adequately , and in an already magnetized state there remains no conceptual difference in the low - energy properties between integer and half - integer chains .
partially magnetized antiferromagnetic heisenberg chains at low energies are expected to be one - component luttinger liquids@xcite ( lls ) irrespective of the spin length @xmath4 .
this is known rigorously in the @xmath1 case where the model can be analyzed using the bethe ansatz.@xcite the validity of the luttinger liquid description for @xmath8 chains and coupled spin-1/2 ladders has been investigated and confirmed by analytical@xcite and numerical@xcite methods . although there is a theoretical possibility of finding gapful behavior ( magnetization plateau ) at special values of the magnetization , such a scenario does not seem to be realized in pure nearest - neighbor heisenberg chains ( see ref . for @xmath9 ) .
neither would we expect multi - component lls , which can otherwise occur for more general ( higher - order ) couplings.@xcite the ll concept bears a strong relationship to @xmath10 conformal field theory ( cft).@xcite in fact the ll is hardly more than the cft adapted to the situation with two fermi points @xmath11 momentum apart .
the one - component ll is a three parameter theory.@xcite in the present context the first parameter , the location of the fermi points @xmath12 , is determined by the magnetization through @xmath13 where @xmath14 is the ( bulk ) magnetization .
the second parameter , the fermi velocity @xmath15 is just an energy scale , whereas the third parameter , to be called the `` dressed charge '' @xmath0 , determines the universality class and the critical exponents .
( in the bosonization literature the notation @xmath16 is standard . ) the traditional ll parameters , the velocities for charge and current excitations can be expressed with @xmath0 as @xmath17 , @xmath18.@xcite all three parameters are functions of the actual magnetization @xmath19 and the spin length @xmath4 .
the ll parameters of the @xmath2 chain in a magnetic field have already been determined to some extent using numerical methods .
sakai and takahashi diagonalized small finite chains up to @xmath20 with periodic boundary condition and used the prediction of cft on finite size energy spectra to estimate the critical exponents.@xcite a similar method was used by usami and suga@xcite for the @xmath1 ladder which behaves as a haldane - gap system for strong ferromagnetic interchain interaction .
more recently campos venuti _
_ used the dmrg to compute directly the transverse two - point function @xmath21 on chains with @xmath22 and fitted these data to the expected asymptotic form of @xmath23.@xcite in this paper we apply an alternative method .
we calculate precise numerical estimates to the critical exponents and by this to the ll dressed charge @xmath0 via a direct numerical determination of the magnetization profiles in finite open chains .
the magnetization profile is , by definition , the positional dependence of the local magnetization @xmath24 in an open chain with some well - defined boundary condition .
the form of the profile depends on the applied boundary condition , and may involve surface exponents as well whenever the boundary condition on the left and right ends differ . in order to simplify the expected behavior
we will only consider cases where the boundary condition is identically open ( free ) on both ends ; in this case only bulk exponents come into play.@xcite in a semi - infinite chain , far from the chain s end , the magnetization profile is expected to decay to its bulk value @xmath19 algebraically as @xmath25 where @xmath26 is a non - universal amplitude , @xmath27 is a phase shift , @xmath28 is the ( bulk ) critical exponent , defined through the translation invariant longitudinal two - point function @xmath29 , and @xmath30 is determined as a function of @xmath19 by eq .
( [ m_kf ] ) .
the form of the _ friedel oscillation _ in eq .
( [ semiinf ] ) contains indirect information about the ll parameter @xmath0 , and this can be exploited in a numerical procedure to calculate precise estimates .
this program has been carried out for the @xmath1 chain and ladders in ref .
, for the kondo lattice model in ref . , for the hubbard model in ref . , and for the @xmath31 model in ref
while the algebraic decay with @xmath32 is a standard consequence of criticality,@xcite the @xmath11 oscillation is a special ll feature , which stems from the fact that we work with two families of ( chiral ) cft operators living around the two fermi points in @xmath33-space . without the @xmath11 oscillation
it is a standard exercise in cft to derive the shape of the magnetization profile in a strip geometry of width @xmath34 by applying the logarithmic mapping.@xcite this transformation , together with the proper account for the @xmath11 ll term yields a prediction for the magnetization profile of a finite luttinger liquid segment of length @xmath34 with open boundary condition on both ends@xcite @xmath35 \left [ { l\over \pi } \sin\!\left(\pi\frac{r}{l}\right)\right]^{-\eta_z/2 } \!\!\!.
\label{profile1}\ ] ] note that there is no need to introduce explicitly a phase shift @xmath27 as in the semi - infinite case eq .
( [ semiinf ] ) , since the symmetry of the profile with respect to @xmath36 implies that @xmath37 . the predicted ll profile in eq . ( [ profile1 ] )
is based on the continuum limit . however ,
for finite lattice systems such as the heisenberg chain , we expect corrections .
conformal invariance and thus eq . ( [ profile1 ] ) is only expected to be valid asymptotically in the large @xmath34 limit with @xmath38 satisfying @xmath39 .
clearly , in a strict sense the magnetization can not diverge at the chain ends since @xmath40 .
phenomenologically this natural cut - off at the boundary acts as an effective impurity put into the cft model at the system ends .
this defines an effective `` scattering length '' , @xmath41 , associated with the boundary , suggesting that we should replace the naive system size @xmath42 in eq .
( [ profile1 ] ) with an `` effective '' system size @xmath43 . based on this intuitive argument in the following
we will work with a slightly modified ansatz @xmath44 \nonumber\\ & & \times \left [ { n+2\,\delta n\over \pi } \sin\!\left(\pi\frac{n-1/2+\delta n}{n+2\,\delta n}\right)\right]^{-\eta_z/2 } \label{profile2}\end{aligned}\ ] ] where @xmath45 is a free fitting parameter .
when @xmath46 eq .
( [ profile2 ] ) is simply the discrete version of eq .
( [ profile1 ] ) ( note that now the system is defined for @xmath47 ) . in the lack of a consistent scheme to calculate corrections , the above amendment to the fitting formula is not rigorous .
justification can stem from a direct comparison with exact results . in the following we first discuss in detail the exactly solvable @xmath1 chain as a benchmark case .
we demonstrate that the ansatz in eq .
( [ profile2 ] ) is capable of reproducing the ( numerically ) exact @xmath1 critical exponents to high precision even with @xmath46 .
when @xmath45 is also fitter for , the accuracy improves by an order of magnitude .
the @xmath1 heisenberg chain in a magnetic field can be solved exactly by the bethe ansatz ( ba ) method.@xcite the ba itself can not yield the correlation functions and critical exponents directly , but assuming conformal invariance the operator content can be read off from the structure of the low - energy excitations above the ground state.@xcite these latter can be determined by a systematic calculation of the finite size corrections to the @xmath48 ba equations ( assuming periodic boundary condition).@xcite the energy and momentum of the lowest energy ( primary ) states ( with respect to those of the ground state , @xmath49 and @xmath50 ) can be cast into the form @xmath51 where @xmath52 is a shorthand for two integer topological quantum numbers labeling the states .
the @xmath53 momentum term is @xmath54 with @xmath30 defined in eq .
( [ m_kf ] ) .
the `` conformal dimensions '' @xmath55 read @xmath56 ^ 2 , \label{deltashalf}\ ] ] where @xmath0 is the dressed charge .
the topological quantum numbers have a direct physical interpretation in the ll representation : @xmath57 ( @xmath58 ) denotes the number of fermions added to ( transferred from the left fermi point to the right in ) the band .
chain ; @xmath59 denotes results using the naive fit @xmath46 , @xmath60 denotes fits where @xmath45 is optimized .
solid line is the bethe ansatz result .
inset shows the critical exponents and the numerically determined amplitude @xmath26 . at @xmath61
the dressed charge is @xmath62 . ]
the critical exponents appearing in the one- and two - point correlation functions can be expressed with the conformal dimensions @xmath55 . in general a physical operator such as
@xmath63 decomposes into all operators which are not forbidden by conservation ( selection ) rules .
in particular , for @xmath63 the number of fermions in the band is conserved , thus necessarily @xmath64 .
the asymptotic decay is determined by the operator which has the lowest critical exponent , i.e. , @xmath65 giving @xmath66 and an oscillation @xmath11 by eq .
( [ qhalf ] ) .
the transverse critical exponents @xmath67 is determined by the operator @xmath68 , i.e. , @xmath69 with the associated oscillation frequency @xmath70 .
equations ( [ etaz ] ) and ( [ etax ] ) imply @xmath71 . in the ba approach the dressed charge @xmath72
is determined by a set of integral equations for the density of rapidities @xmath73 , and the dressed charge function @xmath74,@xcite @xmath75 where @xmath76 the limits of integration are defined implicitly through the constraint @xmath77 the coupled integral equations in eqs .
( [ rhalf ] ) , ( [ zhalf ] ) and ( [ bhalf ] ) are to be solved numerically.@xcite first , assuming @xmath78 is given , eq .
( [ rhalf ] ) is solved for @xmath79 .
this is inserted into eq .
( [ bhalf ] ) to find the associated value of @xmath19 .
when @xmath19 is given , as in our case , this procedure can be iterated to find @xmath78 as a function of @xmath19 at arbitrary precision .
finally eq .
( [ bhalf ] ) is solved for the function @xmath74 and the dressed charge @xmath72 is determined .
the numerically determined dressed charge as a function of the magnetization is plotted in fig .
[ fig : zhalf ] . in the next section
this will serve as a reference curve to check the accuracy of the fitting ansatz in eq .
( [ profile2 ] ) .
in order to test the fitting formula eq .
( [ profile2 ] ) we calculated the magnetization profile in the ground state of finite chain segments with open boundary condition using the density matrix renormalization group method.@xcite the chain length was @xmath80 and we kept @xmath81 states .
the truncation error was found to be in the range @xmath82 .
we applied the finite lattice algorithm with four iteration cycles which was found sufficient for convergence .
an example of the magnetization profile as determined by the dmrg is depicted in fig .
[ fig - pa ] .
chain with @xmath80 , @xmath81 and magnetization @xmath83 . only half of the chain shown - the other half is mirror symmetric . ]
it is worth discussing the fitting procedure itself in some detail .
the ansatz in eq .
( [ profile2 ] ) has five parameters : @xmath84 and @xmath28 . although @xmath85 is a conserved quantity , and thus can be set to a given value in the dmrg , due to the open boundary condition the control over the exact ( bulk ) value of @xmath19 is lost .
while it is true that for long enough chains @xmath19 will be close to @xmath86 , the finite size deviation should be tracked during the fitting procedure .
similarly , although @xmath30 is a well - specified function of @xmath19 in the bulk [ see eq .
( [ m_kf ] ) ] , it was found advantageous to keep it as an independent fit parameter , and only use eq.([m_kf ] ) a posteriori as a consistency check . on the other hand , it is better to avoid fitting on @xmath45 directly . instead the best working alternative seems to
be making a four - parameter fit on @xmath87 and @xmath28 , while @xmath45 kept fixed .
the optimal value of @xmath45 is the one which yields the highest stability with respect to local fits , i.e. , calculating the four fitting parameters from a small number of sites at different locations in the chain .
an example of this procedure is shown in fig.[fig : fitting ] .
we found that the optimal value of @xmath45 is a weak function of @xmath19 , being about @xmath88 for @xmath61 , and decreasing monotonically to @xmath89 for @xmath90 .
this is more or less consistent with the value @xmath91 used in the bosonization approach of ref . . , @xmath92 , @xmath93
local fit parameters at position @xmath94 are defined by fitting for sites @xmath95 , with @xmath96 the wavelength of the oscillations .
the highest stability of the parameters is achieved for @xmath89 .
equation ( [ m_kf ] ) is satisfied up to @xmath97 . ]
having obtained @xmath28 using the above fitting procedure , the fundamental quantity of the theory , the dressed charge @xmath0 , can be calculated from eq .
( [ etaz ] ) .
figure [ fig : zhalf ] shows the numerically determined dressed charge as a function of @xmath19 .
the relative error of the fitting procedure is under @xmath98 , except for very small @xmath19 values where logarithmic corrections to the fitting formula and @xmath0 are expected , and the fitting procedure loses stability ( see the error bar in the figure at @xmath99 ) .
otherwise , the agreement with the exact values is very good and stable for @xmath100 . for comparison fig .
[ fig : zhalf ] also shows the estimate of @xmath0 when the naive scattering length @xmath46 is used .
even with this the error is within @xmath101 , except for very small @xmath19 .
we conclude that for @xmath1 the ll ansatz for the magnetization profile , eq .
( [ profile2 ] ) , is a very efficient tool in calculating the critical exponents ( dressed charge ) .
the final result is highly accurate already with @xmath46 , but an additional increase in precision can be achieved by adding the scattering length as an extra fit parameter .
encouraged by the success of the fitting procedure for @xmath1 , in this section we apply it to the @xmath2 heisenberg chain .
the forthcoming analysis is not intended to prove that the @xmath2 chain in its magnetized regime is a luttinger liquid - this has been done convincingly already.@xcite instead , our starting point is the assumption that we have a one - component luttinger liquid , and then use ll theory and the numerically calculated magnetization profiles to determine the value of @xmath0 and the critical exponents with high accuracy . although the spin-1 problem can not be treated rigorously , there are two limits where clear theoretical predictions exist . in the high magnetization limit near saturation
the physics can be understood by regarding the system as a dilute gas of magnons created in the ferromagnetic vacuum.@xcite magnons behave as bosons with short - range repulsive interaction . in the dilute limit
the exact form of the interaction is irrelevant and an essentially hard - core boson description becomes valid .
in 1d hard - core bosons are equivalent to free spinless fermions , which imply an ll parameter @xmath102 with correlation exponents @xmath103 and @xmath104 as @xmath105 .
there is a similar argument in the @xmath106 limit . below the critical magnetic field @xmath107 , where @xmath108 is the haldane gap , the elementary excitations are massive spin-1 bosons.@xcite at @xmath109 these bosons condensate , but due to the interboson repulsion the magnetization only increases gradually .
near @xmath110 the boson density is low and by the above argument we again obtain @xmath102 , @xmath103 and @xmath104 . note that this argument only works for @xmath2 with a haldane gap . for @xmath1 for which the nonmagnetized ground state is massless , the elementary excitations form a strongly interacting dense gas with @xmath111 as is given by the bethe ansatz , and the su(2 ) symmetry at @xmath61 .
alternative theories for @xmath2 , based on either a majorana fermion representation@xcite or on the bosonization of spin-1/2 ladders@xcite lead to a similar conclusion in this limit .
although the magnetized @xmath2 chain is a luttinger liquid , this classification only applies at low energies and long distances .
indeed , at higher energies and shorter distances the @xmath2 chain produces features which can not be understood within the framework of ll theory .
these features , absent in @xmath1 chains , stem from the additional degrees of freedom staying massive for @xmath112 .
these degrees of freedom have signature both in the energy spectrum and correlations , and manifest themselves in the numerical , finite chain calculations . their origin can be easily understood in the low - magnetization limit . at zero magnetic field
the system possesses an energy gap , the haldane gap .
the lowest excited states form a triplet branch with a minimum energy at momentum @xmath113 .
the operator @xmath114 has large matrix elements between the ground state and the @xmath115 component of this triplet .
this leads to an exponentially decaying alternating ( antiferromagnetic ) behavior in the longitudinal correlation function .
when the magnetic field is switched on , the zeeman energy splits the triplet branch , and at @xmath116 the @xmath117 component at @xmath113 crosses over with the ground state .
however the @xmath115 component remains in the spectrum ( at energy @xmath118 ) and still contributes to the short - range longitudinal correlation functions . as a consequence the two - point function shows a crossover from a seemingly exponential decay on short distances caused by the massive mode to an algebraic decay determined by the soft , ll mode on longer distances .
there is a similar effect in the one - point function we consider here . near the chain s end
there is an exponentially localized effective @xmath1 degree of freedom , the so - called `` end spin '' , which also survives in the magnetized regime , at least when the magnetization is not too high.@xcite its presence produces a crossover from exponential decay to algebraic decay in the one - point correlation function as is illustrated in fig .
[ fig - pb ] . as the magnetization increases the massive modes rise in energy , and have a less and less significant impact on the low - energy physics .
at the same time the end spins gradually dissolve and disappear in the bulk as was observed by yamamoto and miyashita.@xcite chain with @xmath80 , @xmath81 at the bulk magnetization @xmath119 . only half of the chain shown - the other half is mirror symmetric . the ll oscillation is clearly visible above @xmath120 .
solid line is the ll fit .
inset shows the end of the chain on a log - lin scale featuring the exponentially decaying initial oscillation which crosses over to ll behavior at @xmath121 . ] in order to measure the critical exponent @xmath28 we used the dmrg algorithm with @xmath80 and @xmath81 with five iteration cycles .
the truncation error varied in the range @xmath122 , the calculation being more precise in the high magnetization regime . a limited number of runs with @xmath123 , @xmath124 was done to check the numerical precision and to obtain results at points where longer systems were needed .
we applied the fitting procedure described above and care was taken to stay in the bulk of the chain sufficiently far from the ends . since for @xmath61 the correlation ( localization ) length of the end spin is about 6 lattice sites , which becomes even shorter for @xmath125 , a chain length @xmath80 was found sufficient .
we observed that the fitting procedure is somewhat less accurate here than for @xmath1 , meaning that finite size corrections to the luttinger liquid profile eq.([profile2 ] ) are more important in for @xmath2 .
the fitting procedure becomes especially unreliable below @xmath126 and around @xmath127 . in the former case the increasing wavelength , which becomes comparable to the system size , whereas in the latter case
the vanishing prefactor @xmath128\to 0 $ ] can be blamed for the numerical difficulty . far enough from these problematic regions
the relative error of the calculated exponent was estimated to be less than @xmath101 . beyond measuring the local magnetization profile @xmath129 , for @xmath2
there is another , independent quantity whose profile can also be measured easily .
this is the local quadrupole moment @xmath130 .
note that for @xmath1 this quantity is redundant and thus carries no additional information .
for @xmath2 , however , the quadrupole profile provides us an alternative way to measure the critical exponent which in many cases became even more precise than the one obtained through the magnetization profile .
the quadrupole profile is expected to behave according to the same scaling form eq .
( [ profile2 ] ) with the replacement @xmath131 , @xmath132 . determined numerically for the @xmath2 chain .
open symbols denote results for @xmath92 , closed symbols for @xmath133 .
the estimated error is comparable to the symbol size .
solid line is only a guide to the eye .
inset shows the longitudinal and transverse critical exponents @xmath28 and @xmath67 , resp . ]
figure [ fig : expon ] shows the dressed charge determined using eq .
( [ etaz ] ) from the measured @xmath28 exponent of magnetization and quadrupole profiles .
the inset also shows the corresponding critical exponents @xmath28 and @xmath67 as a function of the bulk magnetization .
we see that for @xmath105 the predicted value @xmath102 is reached very rapidly .
we analyzed the @xmath19-dependence close to @xmath134 and found it to be describable with a power law with a rather large exponent around 4 , although a scaling faster than any power law can not be excluded either .
due to the lack of sufficient numerical precision near @xmath134 we were unable to resolve this question reliably .
for @xmath106 the conclusion is also somewhat vague because the @xmath19-dependence is very steep and due to numerical difficulties we were unable to approach this limit closely enough .
however , a @xmath102 value predicted by the theory seems highly consistent with our data . assuming this , our numerical values seem to indicate an @xmath19-dependence @xmath135 , with @xmath136 . knowing that @xmath19 scales above the gap as @xmath137,@xcite this would imply @xmath138 for small @xmath139 with @xmath140 . between the two limits @xmath0 is larger , reaching its largest value @xmath141 at @xmath142 .
this behavior should be compared to that of @xmath1 , where @xmath143 in the whole regime ( see fig .
[ fig : zhalf ] ) .
earlier data on the critical exponents available in the literature , which were determined by other methods such as the finite - size scaling analysis of the energy spectrum@xcite or the staggered structure factor,@xcite are consistent with our results . chain .
open symbols denote results for @xmath92 , closed symbols for @xmath133 .
inset shows the best estimate of @xmath45 . ]
the numerically determined critical amplitudes are depicted in fig .
[ fig : ampli ] .
the amplitude of the magnetic moment fluctuations @xmath26 increases monotonically as a function of @xmath19 . however , the rate of increase is not smooth as it is seen in the figure .
@xmath26 changes sign somewhere close to @xmath90 .
where @xmath144 the observable fluctuations are governed by the next smallest critical exponent , and thus a precise measurement is beyond our method .
in contrast with this , the amplitude of the quadrupole fluctuations @xmath145 remains positive in the whole magnetized regime .
we observe that it decreases for small @xmath19 , reaching its minimal ( still positive ) value at @xmath142 . above this
it increases and saturates for @xmath105 . finally , it is interesting to note that the optimal value of the scattering length @xmath45 changes considerably as @xmath19 varies . as the inset of fig .
[ fig : ampli ] shows that @xmath45 is around @xmath146 for small @xmath19 , then decreasing to @xmath147 at @xmath148 , then increasing again to @xmath149 at @xmath134 .
there is a relatively large uncertainty in the optimal value determined , but this imprecision has little impact on the estimated value of @xmath0 and the correlation exponents .
in this paper we have analyzed the critical fluctuations in open @xmath1 and @xmath2 heisenberg chains in their magnetized regime .
the low energy physics is a one - component luttinger liquid in both cases .
we have calculated the ll s characteristic dressed charge and the amplitude of the leading critical fluctuations . our method consisted of determining numerically the magnetization and quadrupole operator profiles and applying a fitting procedure based on conformal invariance .
the method has been thoroughly tested on the bethe ansatz solvable @xmath1 chain , confirming its reliability and high precision . in the @xmath2 case ,
where the exact solution is unknown the method provided high - precision estimates of the critical exponents , justifying and complementing earlier results derived by alternative methods .
we also determined critical amplitudes which has much less been studied so far for these systems . beyond calculating the characteristic parameters with high accuracy
our results also allow us to make a detailed comparison between the @xmath1 and @xmath2 chains .
although both are luttinger liquids , the respective dressed charges as functions of the magnetization differ quite considerably .
the only limit where the two systems become equivalent is the full saturation limit , @xmath150 , where @xmath151 . otherwise , for @xmath1 the dressed charge is @xmath143 , while for @xmath2 it is @xmath152 .
there are interesting differences in the behavior of the critical amplitudes and the scattering lengths at the chain ends , as well .
we thank f. igli and j. slyom for valuable discussions .
this work was partially supported by the hungarian scientific research found ( otka ) under grant nos .
t30173 , t43330 and f31949 , and by the bolyai research scholarship program .
the numerical calculations were done at the niif supercomputing facility . | antiferromagnetic heisenberg spin chains in a sufficiently strong magnetic field are luttinger liquids , whose parameters depend on the actual magnetization of the chain . here
we present precise numerical estimates of the luttinger liquid dressed charge @xmath0 , which determines the critical exponents , by calculating the magnetization and quadrupole operator profiles for @xmath1 and @xmath2 chains using the density matrix renormalization group method .
critical amplitudes and the scattering length at the chain ends are also determined .
although both systems are luttinger liquids the characteristic parameters differ considerably . |
the theoretical description of mesons is problematic because , unlike @xmath5 mesons , the masses and widths of the and states @xcite are not in agreement with potential model calculations based on hqet @xcite .
theoretical explanations for the discrepancy invoke @xmath6 molecules @xcite , chiral partners @xcite , unitarized chiral models @xcite , tetraquarks @xcite , and lattice calculations @xcite , but a satisfactory description is still lacking ( see @xcite for more details ) . improved measurements of the meson parameters can lead to a better understanding of these states .
+ in this analysis a precise measurement of the mass and decay width is performed based on a high statistics data sample @xcite .
the meson , referred to as the below , was first seen in -continuum reactions @xcite , and more recently in @xmath7 decays .
the current world average mass value published by the particle data group is based on measurements with large statistical and systematic uncertainties : @xmath8 @xcite ; the mass difference between the and the meson has been measured to be @xmath9 @xcite .
an upper limit on the decay width ( @xmath10 at @xmath11 confidence level ) , and a measurement of the spin - parity of the meson ( @xmath12 ) , have been obtained , but based on low - statistics data samples only @xcite .
the mixing between the meson and the other @xmath12 state was investigated in ref .
@xcite . + this analysis is based on a data sample corresponding to an integrated luminosity of @xmath13 recorded at the resonance and @xmath14 recorded @xmath15 below that resonance with the detector at the asymmetric - energy collider pep - ii at the slac national accelerator laboratory . in this analysis ,
mesons are reconstructed from continuum events in the @xmath16 channel ; those originating from @xmath7 decays are rejected .
+ the detector is described briefly in sec .
[ sec : det ] .
the principal criteria used in the reconstruction of the @xmath16 mass spectrum and the selection of -candidates are discussed in sec .
[ sec : reco ] .
the relevant monte carlo ( mc ) simulations are described in sec .
[ sec : mc ] , while the detector resolution parametrization is considered in sec .
[ sec : res ] .
measurements of the mass and total width for the state are obtained from a fit to the @xmath16 invariant mass distribution as discussed in sec .
[ sec : datafit ] .
decay angle distributions are studied in sec .
[ sec : angdist ] , where the implications for the spin - parity of the state are also discussed .
sources of systematic uncertainty are described in sec .
[ sec : syst ] , and the results of the analysis are summarized in secs .
[ sec : results ] and [ sec : summ ] .
the detector is described in detail elsewhere @xcite . charged particles
are detected , and their momenta measured , with a combination of five layers of double - sided silicon microstrip detectors ( svt ) and a 40-layer cylindrical drift chamber ( dch ) , both coaxial with the cryostat of a superconducting solenoidal magnet that produces a magnetic field of @xmath17 .
charged particle identification is achieved by measurements of the energy loss @xmath18 in the tracking devices and with an internally reflecting , ring - imaging cherenkov detector .
the energy of photons and electrons is measured with a csi(tl ) electromagnetic calorimeter , covering @xmath11 of the @xmath19 solid angle in the @xmath20 rest frame .
the instrumented flux return of the magnetic field is used to identify muons and @xmath21 s .
the is reconstructed via its decay mode @xmath16 , with @xmath22 and @xmath23 .
the is reconstructed through two decay modes , @xmath24 and @xmath25 , which will be labeled @xmath26 and @xmath27 , respectively , in the following . to improve the mass resolution , the mass difference @xmath28 is examined .
+ events are selected by requiring at least five charged tracks , at least one of which is identified as a charged kaon . also , at least one neutral kaon candidate is required .
each track must approach the nominal interaction point to within @xmath29 in the transverse direction , and to within @xmath30 in the longitudinal ( beam ) direction .
kaon candidates are identified using the normalized kaon , pion and proton likelihood values ( @xmath31 , @xmath32 and @xmath33 ) obtained from the particle identification system , by requiring @xmath34 and @xmath35 .
furthermore , the track must be inconsistent with the electron hypothesis or have a momentum less than @xmath36 .
tracks that fulfill @xmath37 and @xmath38 are selected as pions .
+ candidates for the decay are formed by selecting all @xmath24 pairs ( @xmath25 combinations in the second mode ) that have an invariant mass within @xmath39 of the nominal mass @xcite .
candidates for the decay are formed by adding a to the , such that the mass difference between and is less than @xmath40 .
a candidate consists of a pair with invariant mass within @xmath41 of the nominal mass @xcite .
a kinematic fit is applied to the complete decay chain , constraining the candidate vertex to be consistent with the interaction region .
mass constraints are not applied to intermediate states .
those candidates with a @xmath42 fit probability greater than @xmath43 are retained . to suppress combinatorial background and events from @xmath7-decays
, we require the momentum @xmath44 of the in the center - of - mass ( cm ) frame to exceed @xmath45 . + the @xmath46 and @xmath47 mass spectra for accepted candidates , shown in figs .
[ fig : recores](a ) and [ fig : recores](d ) , are fitted with a signal function consisting of a sum of two gaussians with a common mean value , and a linear background function .
the width of the signal regions for @xmath48 , @xmath49 and @xmath50 candidates is defined as twice the full width at half maximum ( fwhm ) of the signal line shapes . a signal window of @xmath51 for the @xmath26 ( @xmath27 ) mode around the mean mass of @xmath52 obtained from the fit
is used to select candidates . for these candidates ,
the @xmath53 mass difference distributions are shown in figs .
[ fig : recores](b ) and [ fig : recores](e ) .
these are fitted with the sum of a relativistic breit - wigner signal function and a background function consisting of a polynomial times an exponential function .
a signal region of @xmath54 around the fitted mean value of @xmath55 is chosen for both decay modes . to further reduce the background , the angle between the flight direction of the candidate and the line connecting the interaction point and the decay vertex is required to be less than @xmath56 . for candidates passing these selection criteria , the candidate invariant mass distributions ( figs . [
fig : recores](c ) and [ fig : recores](f ) ) are fitted with the sum of a signal function , consisting of the sum of two gaussians , and a linear background function . a signal window of @xmath57 around the fitted mean mass of @xmath58 is selected for both decay modes .
+ in the case of an event with multiple candidates , the candidate with the best fit probability is chosen .
the candidate spectra after all selection criteria are shown in figs .
[ fig : ds1mass](a ) and [ fig : ds1mass](b ) .
the fits to these spectra use a double - gaussian signal function and a linear background function . note that for this preliminary fit the intrinsic width and the resolution are not taken into account .
the fwhm values obtained are @xmath59 and @xmath60 , respectively , with corresponding signal yields of about @xmath61 and @xmath62 entries .
+ ( 0,0 ) ( -12.,9.3)(a ) ( -12.,4.2)(d ) ( -6.5,9.3)(b ) ( -6.5,4.2)(e ) ( -1.,9.3)(c ) ( -1.,4.2)(f ) invariant mass distributions in data after applying all selection criteria for the ( a ) @xmath26 and ( b ) @xmath27 mode . a double - gaussian signal function and a linear background function
are used to describe the data in a preliminary fit.,title="fig:",scaledwidth=30.0% ] invariant mass distributions in data after applying all selection criteria for the ( a ) @xmath26 and ( b ) @xmath27 mode . a double - gaussian signal function and a linear background function
are used to describe the data in a preliminary fit.,title="fig:",scaledwidth=30.0% ] ( 0,0 ) ( -1.,9.3)(a ) ( -1.,4.2)(b )
monte carlo events are generated for @xmath63 , with @xmath64 and @xmath65 , by evtgen @xcite .
the detector response is simulated using the geant4 @xcite package . for each decay mode , and for each of the corresponding decays , @xmath66 events are generated .
the line shape is generated using a non - relativistic breit - wigner function having central value @xmath67 and intrinsic width @xmath68 ( this sample is labeled @xmath69 in the following ) .
the range of generated masses is restricted to values between @xmath70 and @xmath71 .
the masses of the daughter particles are taken from ref .
+ in order to test the mass resolution model , a second set of mc samples with @xmath72 events for each decay mode is generated using a breit - wigner width of @xmath73 ( @xmath74 sample ) .
in addition to these signal mc samples , separate @xmath48 and @xmath50 samples are created from data and generic mc simulations without requiring a or .
they are used mainly for resolution studies . + the mc and data are in good agreement for the transverse momentum distributions of pions , kaons , @xmath5 and @xmath75 mesons , and for the number of svt coordinates of pions and kaons . the agreement is worse for the number of dch coordinates , where the data show systematically fewer coordinates than the mc , giving rise to a resolution that is about @xmath76 smaller in the mc than in data .
this is illustrated in fig .
[ fig : resrat ] , which shows the @xmath77 and @xmath78 dependence of the ratio between the fwhm of the resolution functions in mc and data , where @xmath79 is the momentum in the cm frame .
this effect will be discussed further in sec . [
sec : syst ] .
there is also disagreement between the number of signal entries in mc and data as a function of ( fig .
[ fig : pstardiff ] ) .
this effect will be addressed in secs .
[ sec : res ] and [ sec : syst ] .
-dependence of the ratio between the fwhm of the resolution functions from -mc and data .
( a ) @xmath22 .
( b ) @xmath80 .
the solid line shows the fitted mean ratio with a value of @xmath81.,title="fig:",scaledwidth=30.0% ] -dependence of the ratio between the fwhm of the resolution functions from -mc and data .
( a ) @xmath22 .
( b ) @xmath80 .
the solid line shows the fitted mean ratio with a value of @xmath81.,title="fig:",scaledwidth=30.0% ] ( 0,0 ) ( -4.5,9.3)(a ) ( -4.5,4.2)(b ) ( -5.6,8 . ) 90 @xmath82 ( -5.6,2.8 ) 90 @xmath82 and ( b ) @xmath27 decay modes.,title="fig:",scaledwidth=30.0% ] and ( b ) @xmath27 decay modes.,title="fig:",scaledwidth=30.0% ] ( 0,0 ) ( -4.3,9.4)(a ) ( -4.3,4.3)(b )
the resolution model is derived from the signal mc by studying the difference @xmath83 between the reconstructed and generated mass values .
the multi - gaussian ansatz @xmath84 is found to accurately model the mass resolution spectra .
this represents a superposition of gaussian distributions with the same mean value @xmath85 but variable width @xmath86 , starting from minimum width @xmath87 and increasing to maximum width @xmath88 .
the fwhm of the distribution is numerically calculable once @xmath87 and @xmath89 are known .
the mass resolution for the different particles depends on the cm momentum . therefore , the parameter @xmath87 of eq .
( [ eq : mulg ] ) is obtained as a function of .
+ figures [ fig : res_rfit](a ) and [ fig : res_rfit](b ) show @xmath83 distributions for the full range . from these plots
the value of the parameter @xmath89 is determined to be @xmath90 and @xmath91 for the @xmath26 and @xmath27 modes , respectively .
events are divided into 30 intervals from @xmath92 to @xmath93 and the fit repeated for each interval , resulting in -dependent @xmath87 values ( figs .
[ fig : res_sigma](a ) and [ fig : res_sigma](b ) ) .
the corresponding -dependent fwhm of the resolution functions is shown in figs .
[ fig : res_fwhm](a ) and [ fig : res_fwhm](b ) .
+ in order to validate this resolution model , the -dependent resolution function with the corresponding parameters @xmath94 and @xmath89 is convolved with a non - relativistic breit - wigner function and fitted to the signal mc distribution ( mc sample @xmath69 ) .
the results are shown in figs .
[ fig : mctestfit](a ) and [ fig : mctestfit](b ) , and the reconstructed values for mean and width are listed in table [ tab : testfit ] .
the corresponding generated values for both decay modes are @xmath95 for the mean and @xmath96 for the width .
the small deviations between generated and reconstructed values are discussed in sec .
[ sec : syst ] .
.reconstructed values for and ( fit to mc sample @xmath69 ) .
the resolution model used is derived from mc sample @xmath69 . [
cols="<,^,^",options="header " , ] [ tab : systcomb ]
in this paper , precision measurements of the mass and decay width of the charmed - strange meson via the decay @xmath97 are presented .
two decay modes are analyzed , with the that originates from the decaying either through @xmath24 or @xmath25 .
+ the results include the first significant measurement of the total decay width of the .
this width is determined to be @xmath98 compared to the @xmath11 confidence level upper limit of @xmath99 given in ref .
the mass of the is measured to be @xmath100 and the @xmath101 mass difference to be @xmath102 the result for the @xmath101 mass difference represents a significant improvement compared to the current world average of @xmath103 @xcite .
+ based on a decay angle analysis , the @xmath104 assignment for the meson is confirmed .
we are grateful for the extraordinary contributions of our 2 colleagues in achieving the excellent luminosity and machine conditions that have made this work possible .
the success of this project also relies critically on the expertise and dedication of the computing organizations that support . the collaborating institutions wish to thank slac for its support and the kind hospitality extended to them
this work is supported by the us department of energy and national science foundation , the natural sciences and engineering research council ( canada ) , the commissariat lenergie atomique and institut national de physique nuclaire et de physique des particules ( france ) , the bundesministerium fr bildung und forschung and deutsche forschungsgemeinschaft ( germany ) , the istituto nazionale di fisica nucleare ( italy ) , the foundation for fundamental research on matter ( the netherlands ) , the research council of norway , the ministry of education and science of the russian federation , ministerio de educacin y ciencia ( spain ) , and the science and technology facilities council ( united kingdom ) .
individuals have received support from the marie - curie ief program ( european union ) and the a. p. sloan foundation .
99 b. aubert _ et al . _
( collaboration ) , phys .
lett . * 90 * , 242001 ( 2003 ) .
b. aubert _ et al . _
( collaboration ) , phys .
lett . * 93 * , 181801 ( 2004 ) .
v. mikani _ et al . _
( belle collaboration ) , phys .
lett . * 92 * , 012002 ( 2004 ) .
b. aubert _ et al . _
( collaboration ) , phys .
d * 74 * , 032007 ( 2006 ) .
d. benson _ et al .
_ ( cleo collaboration ) , phys . rev .
d * 68 * , 032002 ( 2003 ) .
p. krokovny _ et al .
_ ( belle collaboration ) , phys . rev .
91 * , 262002 ( 2003 ) .
r.n . cahn and j.d .
jackson , phys .
d * 68 * , 037502 ( 2003 ) .
t. barnes , f.e .
close , and h.j .
lipkin , phys .
d * 68 * , 054006 ( 2003 ) .
bardeen , e.j .
eichten , and c.t .
hill , phys .
d * 68 * , 054024 ( 2003 ) .
nowak , m. rho , and i. zahed , acta phys .
polon . * b35 * , 2377 ( 2004 ) .
e. van beveren and g. rupp , phys .
lett . * 91 * , 012003 ( 2003 ) .
kolomeitsev and m.f.m .
lutz , phys .
lett . b * 582 * , 39 ( 2004 ) .
l. maiani , f. piccinini , a.d .
polosa , and v. riquer , phys .
d * 71 * , 014028 ( 2005 ) .
k. terasaki , phys .
d * 68 * , 011501 ( 2003 ) .
a. dougall , r.d .
kenway , c.m . maynard and c. mcneile ( ukqcd collaboration ) ,
b * 569 * , 41 ( 2003 ) .
bali , phys .
d * 68 * , 071501 ( 2003 ) .
p. colangelo , f. de fazio , and r. ferrandes , mod . phys .
lett . * a19 * , 2083 ( 2004 ) . e.s . swanson , phys
. rep .
* 429 * , 243 ( 2006 ) .
the use of charge conjugated reactions is implied throughout the text .
h. albrecht _ et al .
_ ( argus collaboration ) , phys .
b * 230 * , 162 ( 1989 ) . k. nakamura _ et al .
_ ( particle data group ) , j. phys . * g*37 * * , 075021 ( 2010 ) . j. alexander _ et al . _
( cleo collaboration ) , phys .
lett . b * 303 * , 377 ( 1993 ) .
b. aubert _ et al . _
( collaboration ) , phys .
d * 77 * , 011102(r ) ( 2008 ) .
v. balagura _ et al . _
( belle collaboration ) , phys .
d * 77 * , 032001 ( 2008 ) .
b. aubert _ et al . _
( collaboration ) , nucl .
instrum .
methods phys .
a * 479 * , 1 ( 2002 ) .
d. lange , nucl .
instrum .
methods phys .
, sect . a * 462 * , 152 ( 2001 ) . s. agostinelli _ et al . _
( geant4 collaboration ) , nucl .
instrum .
methods phys .
a * 506 * , 250 ( 2003 ) . f. von hippel and c. quigg , phys .
d * 5 * , 624 ( 1972 ) .
s. chung , cern yellow report 71 - 8 ( 1971 ) . c. amsler and c. bizot , comput .
commun . * 30 * , 21 ( 1983 ) .
j. richman , calt-68 - 1148 ( 1984 ) .
r.d . field and s.
wolfram , nucl .
phys . b * 213 * , 65 ( 1983 ) .
b. aubert _ et al . _
( collaboration ) , phys .
d * 72 * , 052006 ( 2005 ) .
l. lyons , d. gibaut , and p. clifford , nucl .
instrum .
methods phys .
res . , sect .
a * 270 * , 110 ( 1988 ) . | the decay width and mass of the meson are measured via the decay channel @xmath0 using @xmath1 of data recorded with the detector in the vicinity of the resonance at the pep - ii asymmetric - energy electron - positron collider .
the result for the decay width is @xmath2 . for the mass
, a value of @xmath3 is obtained .
the mass difference between the and the is measured to be @xmath4 , representing a significant improvement compared to the current world average .
the unnatural spin - parity assignment for the meson is confirmed . |
the case of a non - relativistic quantum particle confined in a one dimensional box with moving walls subject to dirichlet boundary conditions has been investigated in great detail in @xcite . in this paper
we consider more general boundary conditions and study the geometric phases that emerge .
the boundary conditions we focus on are those consistent with the unitarity of the dynamics as well as with dilation symmetry .
geometric phases were investigated by berry and wilkinson @xcite who considered the behaviour of the eigenfunctions of the laplacian in a region with a triangular boundary with dirichlet boundary conditions , when the shape of the region was varied adiabatically .
this study revealed the existence of `` diabolical points '' , shapes which have an accidental degeneracy in the spectrum .
varying the shape of the region in a small circuit around the diabolical point led to a reversal in the sign of the eigenfunction .
similar effects were also noticed earlier in molecular physics @xcite as explained in the book by shapere and wilczek @xcite .
these sign reversals were an early example of a geometric phase . in these problems
the geometric phase is essentially of topological origin .
this is because of the time reversal symmetry of the problem , wave functions can be chosen real and this constrains all geometric phases to be @xmath0 or @xmath1 . in a later work by berry , the time reversal symmetry was broken by the introduction of magnetic fields and this led to the discovery of the full geometric phase @xcite , which has been subsequently studied and generalized in many directions @xcite and widely applied @xcite . in this paper
we consider a particle in a box subject to general boundary conditions , which ( apart from some special cases ) violate time reversal symmetry .
our hamiltonian operator is the laplacian .
the location of the boundaries is adiabatically varied by translations and dilations , which gives us a two parameter space of variations .
we find that there is a geometric phase and compute the two - form on the parameter space .
it turns out that this two - form is similar to the area two - form on the poincar upper half - plane . in section [ boundcond ]
we describe the most general boundary conditions that make the laplacian self - adjoint and we focus on a subset of these which are invariant under dilations . in section [ sec : movfix ] we show how one can reduce the problem of moving walls into a fixed domain . in section [ sec : berry ] , we compute the geometric phase two - form , which measures the extent of anholonomy in a closed circuit .
this calculation involves some subtleties which require a renormalization scheme .
section [ sec : berry ] gives an alternative perspective on the renormalization .
section [ sec : degen ] deals with the two special cases in which the boundary conditions do not break time reversal symmetry and section [ sec : conclusion ] is a concluding discussion .
in this section we use the powerful technique of _ boundary triples _
@xcite ( see [ appendix ] ) to classify the self - adjoint extensions of the laplacian on an interval .
we are going to use this approach in order to find all possible boundary conditions which preserve unitarity and are invariant under dilation .
let us consider a quantum spinless particle of mass @xmath2 confined in a one dimensional box @xmath3 $ ] .
the hamiltonian , describing the kinetic energy of the particle , is ( @xmath4 ) @xmath5 where @xmath6 is the space of test functions , i.e. the infinitely differentiable functions with compact support in @xmath7 .
the adjoint operator , @xmath8 , has the same functional form of the operator ( [ eq : ham ] ) but acts on a larger space , namely @xmath9 , the space of square integrable functions on @xmath10 $ ] whose first and second ( distribution ) derivatives are square integrable .
this hamiltonian operator is symmetric but certainly not self - adjoint , @xmath11 , and thus it can not be associated to an observable of the physical system .
the hamiltonian s deficiency indices , determined by the equation @xmath12 are equal to 2 , so that , by von neumann s theorem ( see for example @xcite ) , the self - adjoint extensions of the operator ( [ eq : ham ] ) are in a one - to - one correspondence with the unitary operators on @xmath13 .
unfortunately , this is a non - constructive theorem and one needs to find other ways of working with self - adjoint extensions . with this end in view , we define the following maps from @xmath14 to the space of boundary data @xmath13 ( see [ appendix ] ) :
@xmath15 these are well defined since @xmath16 , and the following identity holds @xmath17 where @xmath18 is the boundary form defined in ( [ eq : boundf ] ) , which measures the `` lack of self - adjointness '' of the operator @xmath19 . here
@xmath20 is the canonical scalar product of @xmath21 , while @xmath22 denotes the scalar product of @xmath23 .
given these maps , we have by ( [ eq : bt ] ) that @xmath24 is a boundary triple ( see [ appendix ] ) for the hamiltonian ( [ eq : ham ] ) , and all self - adjoint extensions of @xmath19 are given by ( [ eq : exts ] ) , which reads @xmath25 where @xmath26 is a unitary 2 @xmath27 2 matrix .
this is the result obtained in @xcite , which expresses all possible self - adjoint extensions of the laplacian in terms of unitaries on the hilbert space of boundary data @xmath13 .
the choice of particular unitary matrices gives rise to some well - known boundary conditions , for example : @xmath28 @xmath29 being the first pauli matrix .
now we would like to extract and parametrize a particular subset of boundary conditions which are invariant under dilations and will be useful in the following .
the set we are looking for is made up by all those boundary conditions which do not mix functions with derivatives at the boundary , that is of the form @xmath30 where @xmath31 .
it is easy to show that the conditions that have to be satisfied by the former four parameters in order to represent a self - adjoint extension of the hamiltonian on the interval @xmath10 $ ] are @xmath32 if we set @xmath33 , the desired boundary conditions read @xmath34 and the unitary matrix in ( [ eq : bc ] ) associated to this self - adjoint extension is provided by @xmath35 some comments are in order . if @xmath36 we obtain periodic ( [ eq : periodic ] ) and antiperiodic ( [ eq : antiperiodic ] ) boundary conditions , while for @xmath37 or @xmath38 mixed dirichlet and neumann conditions arise .
however , pure dirichlet or neumann conditions can not be reached by our parametrization .
thus , the family in ( [ eq : non ] ) , which we denote by @xmath39 , where @xmath40 , does not exhaust the whole set of dilation - invariant boundary conditions , which is instead provided by @xmath41 moreover , it is worth noticing that the set @xmath42 does not form a subgroup of @xmath43 . from a physical perspective
the boundary conditions in ( [ eq : mybc ] ) are nonlocal , since they connect the value of the wave function at one end of the interval with its value at the other end .
a physical realization of them require that the interval be bended into a ring with the two ends forming a tunneling junction through which the wave function can acquire a phase given by ( [ eq : non ] ) .
see fig .
[ fig : bendingfn ] . this can be experimentally implemented by means e.g. of superconducting quantum interference devices , where the properties of the josephson junction are suitably chosen to give the required phase @xcite .
we start by generalizing the problem of a particle of mass @xmath2 in a one dimensional box with moving walls subject to dirichlet boundary conditions ( extensively discussed in @xcite ) to a larger class of boundary conditions , which we picked out in ( [ eq : family ] ) . for convenience we parametrize the one dimensional box by @xmath44,\ ] ]
so that @xmath45 is the center of the interval , and @xmath46 is its length , and consider the hamiltonian ( kinetic energy ) @xmath47 where @xmath48 is a fixed complex number representing particular boundary conditions ( [ eq : mybc ] ) , and @xmath49 is the sobolev space of square integrable functions on @xmath50 , whose first and second derivatives are square integrable functions .
some comments are necessary . in the previous section we proved that the above boundary conditions yield a good self - adjoint extension of the hamiltonian on an interval .
as already remarked these do not mix the values of the functions at the border with their derivatives . in what follows
we will see that these are the only ones which are invariant under dilations , a crucial property for what we are going to investigate .
next we take into account the dynamics of this problem by taking smooth paths in the parameter space @xmath51 : @xmath52 and @xmath53 .
clearly we are translating the box by @xmath54 and contracting / dilating it by @xmath55 .
as underlined in @xcite determining the quantum dynamics of this system is not an easy problem to tackle with , since we have hilbert spaces , @xmath56 , varying with time and we need to compare vectors in different spaces .
the standard approach is to embed the time - dependent spaces into a larger one , namely @xmath57 , extend the two - parameter family of hamiltonians ( [ eq : myham ] ) to this space and try to unitarily map the problem we started with into another one , with a family of time - dependent hamiltonians on a fixed common domain .
with this end in view we embed @xmath58 into @xmath57 in the following way @xmath59 where @xmath60 is the complement of the set @xmath61 , so that we can consider the extension of the hamiltonians defined in ( [ eq : myham ] ) as @xmath62 where the embedding and the direct sum obviously depend on @xmath63 and @xmath64 . following @xmath65 we recall how to reduce this moving walls problem into a fixed domain one .
the composition of a translation @xmath66 and of a subsequent dilation @xmath67 maps the interval @xmath50 onto @xmath68,\ ] ] which does not depend on @xmath64 and @xmath63 .
next we need to define a unitary action of both groups on @xmath57 .
a possible choice is @xmath69 and both @xmath70 and @xmath71 form one - parameter ( strongly continuous ) unitary groups .
the factor @xmath72 is consistent with the physical expectation that @xmath73 transforms as the square root of a density under dilation . in order to make the expression @xmath74 meaningful , from now on
we are going to identify @xmath63 with a pure number given by the ratio of the actual length of the box and a unitary length .
the infinitesimal generator of the group of translations is the momentum operator @xmath75 so that spatial translations are implemented by the unitary group @xmath76 similarly , the generator of the dilation unitary group is given by the virial operator over its maximal domain : @xmath77 where @xmath78 denotes the closure of the operator @xmath79 .
dilations on @xmath57 are , thus , implemented by @xmath80 next we define the two - parameter family of unitary operators on @xmath57 , which are going to fix our time - dependent problem @xmath81 by this unitary isomorphism we are mapping @xmath82 into @xmath83 where we have used the identity @xmath84 the operators in ( [ eq : mynewham ] ) act on the time - independent domain @xmath85 where @xmath86 is given by @xmath87 we have thus achieved our goal , that is mapping the initial family of hamiltonians with time - dependent domains into a family with a common fixed domain of self - adjointness .
this has been possible thanks to the unitary operator ( [ eq : diltr ] ) and , most importantly , to the choice of dilation - invariant boundary conditions ( [ eq : myham ] ) as discussed in the previous section .
we have taken into account those boundary conditions ( [ eq : mybc ] ) which do not mix derivatives and functions at the boundary : these are the only ones which leave the transformed domain @xmath86 in ( [ eq : dom ] ) time - independent .
the main objective of this section will be to exhibit a non - trivial geometric phase associated to a cyclic adiabatic evolution of the physical system described in ( [ eq : myham ] ) .
let @xmath88 be a closed path in the parameter space @xmath89 .
let the @xmath90-th energy level be non degenerate ; then , in the adiabatic approximation , the berry phase associated to the cyclical adiabatic evolution is given by @xmath91 where @xmath92 is the eigenfunction associated to the @xmath90-th eigenvalue , @xmath93 is the external differential defined over the parameter manifold @xmath94 , and @xmath95 in our case @xmath96 reads @xmath97 a technical difficulty arises from equations ( [ eq : berry])-([eq : diff ] ) . in this section we are going to show that , for fixed @xmath48 , the eigenfunctions @xmath98 determine an orthonormal basis in @xmath57 .
however , in general the derivatives in ( [ eq : diff ] ) do not belong to @xmath57 so that the integral in ( [ eq : integ ] ) is ill - posed and needs a prescription of calculation .
no doubt that the ill - posedness of ( [ eq : berry ] ) is due to the presence of a boundary in our system .
first we need to determine the spectral decomposition of the hamiltonian we started with in ( [ eq : myham ] ) or equivalently in ( [ eq : embham ] ) .
of course this would be a difficult problem to handle , but thanks to the unitary operator in ( [ eq : diltr ] ) we can move on to the hamiltonians with fixed domain , compute the spectral decomposition and then make our way unitarily back to the problem with time - dependent domain .
therefore , we need to solve the eigenvalue problem @xmath99 where @xmath100 in ( [ eq : dom ] ) and @xmath101 . the spectral decomposition will heavily rely on the choice of the parameter @xmath48 , which , as already stressed , represents a particular choice of boundary condition .
if @xmath102 the spectrum is non - degenerate , and the normalized eigenfunctions have the form @xmath103 where @xmath104 so that the dispersion relation ( @xmath105 ) reads @xmath106 let @xmath107 be an arbitrary function belonging to @xmath108 ; @xmath107 is clearly an eigenfunction of the * 0 * operator on @xmath108 with zero eigenvalue .
therefore we can extend @xmath109 to an eigenfunction of ( [ eq : embham ] ) , @xmath110 , which can be conveniently chosen to be a test function : @xmath111 , the space of smooth functions with compact support . ) . ] for the sake of the reader , let us exhibit an explicit construction of @xmath112 .
let @xmath113 be a smooth extension of @xmath114 to the whole real line . roughly speaking our eigenfunction
can be written as the restriction of this extension , namely @xmath115 , where @xmath116 is the characteristic function of the set @xmath79 [ @xmath117 if @xmath118 , and @xmath119 otherwise ] , showing why divergent contributions arise from the boundary when taking derivatives .
so the idea which underlies the following discussion is to regularize the contribution of the characteristic function @xmath120 .
let @xmath121 be a nonnegative monotone decreasing function which belongs to @xmath122 , moreover we require that @xmath123 , @xmath124 and @xmath125 for @xmath126 .
we are going to paste two contracted copies of the latter to @xmath120 , such that the final result would be as in figure [ molly2 ] .
given @xmath127 we define the regularized characteristic function of @xmath50 as follows : @xmath128 which is a test function , @xmath129 . in light of the previous discussion we choose the following functions and show that they are eigenfunctions for the hamiltonian ( [ eq : embham ] ) : @xmath130 where @xmath131 see figure [ molly ] .
even if @xmath132 , this will not alter the desired regularity property and the integrability condition of ( [ eq : emb ] ) .
clearly ( [ eq : emb ] ) is still an eigenfunction of ( [ eq : embham ] ) because @xmath133 is an eigenfunction of the hamiltonian defined in ( [ eq : myham ] ) and @xmath134 is trivially an eigenfunction of the * 0 * operator with null eigenvalue .
moreover , from the explicit expression in ( [ eq : normchar ] ) this eigenfunction is normalized . in this renormalization scheme , which is needed for the definiteness of ( [ eq : diff ] ) , we are first embedding @xmath135 into @xmath136 and then regularizing the boundary contribution through the introduction of the regularizer @xmath137 . ) . ]
now it is essential to observe that @xmath138 that is the eigenfunction of a particle confined in @xmath50 . here , the convergence of the limit is pointwise and , by dominated convergence , in @xmath57 .
we are now in the right position to compute ( [ eq : integ ] ) for @xmath127 , which is well posed , and then take the limit @xmath139 .
we start by considering separately both the terms in @xmath140 which , after an integration by parts , become @xmath141 by plugging the explicit expressions of the eigenfunctions we find , by dominated convergence , that for @xmath139 @xmath142 moreover , since @xmath92 has inherited from @xmath143 the right regularity properties , for any @xmath127 , one gets @xmath144
.,scaledwidth=40.0% ] summing up , we finally get the expression of the berry one - form : @xmath145 which is manifestly not closed yielding a nontrivial abelian phase .
notice that the one - form derived in ( [ eq : form ] ) is purely imaginary , consistently with the general theory of berry phases @xcite .
moreover it heavily depends on the energy level through @xmath146 in ( [ eq : eigenvalues ] ) and on the boundary conditions through @xmath147 . as a simple example , we choose a rectangular path @xmath88 in the @xmath148 half - plane , as shown in figure [ fig : path ] , and compute @xmath149 whose
only non - trivial contributions are given by the vertical components of the circuit .
the final result is @xmath150 which , as expected , depends on the particular path chosen . in the spirit of the physical implementation of our system in terms of a ring with a junction ( see section [ boundcond ] ) , our cyclic adiabatic evolution could be illustrated as in figure [ fig : circles ] .
another interesting aspect provided by this problem is linked to a nontrivial berry curvature : @xmath151 the above formula brings to mind the curvature of a hyperbolic riemannian manifold .
indeed , consider the _
poincar half - plane _ , which by definition is the upper - half plane together with the poincar metric : @xmath152 the half - plane is a model of hyperbolic geometry and if we consider the area form on it we have @xmath153 which has the same structure as the berry curvature ( [ eq : curv ] ) of our quantum mechanical model .
.[fig : circles],scaledwidth=50.0% ] we remark that the relevant group in hyperbolic geometry is @xmath154 the group of real mbius transformations .
the lie algebra of this group is the space of real @xmath155 traceless matrices which are spanned by @xmath156 and @xmath157 , where the @xmath158s are the usual pauli matrices .
the two generators @xmath159 and @xmath160 form a closed subalgebra .
the structure of this lie subalgebra is exactly the same as ours : the commutator of the virial operator and the momentum operator is the momentum operator , namely , @xmath161=
one may object to the regularization scheme introduced in the previous section for its artificiality .
in fact , in order to have a well - posed problem we embedded our original problem into a larger space , @xmath162 , and needed to make sense of the differential in ( [ eq : diff ] ) . in this section
we are going to understand better what may be the problem in the definition of the derivative with respect to our parameters , and , moreover , we are going to show an alternative , intrinsic , approach to renormalization which does not make use of any embedding .
let us consider the following map : @xmath163 where @xmath164 is a suitable unit vector independent of @xmath148 , and @xmath165 is defined in ( [ eq : diltr ] ) .
we would like to understand better the following differential : @xmath166 fix @xmath46 and consider the restriction of ( [ eq : unit ] ) to its second argument @xmath167 @xmath168 in ( [ eq : vtransl ] ) form a one - parameter group , whose generator is the momentum @xmath169 defined in ( [ eq : momentum ] ) thus , @xmath170 which is well posed if and only if @xmath171 .
for this reason we can interpret @xmath172 as the distributional derivative of @xmath173 which is forced to belong to @xmath57 . in our case
the extension of the eigenfunction to the real line is smooth , @xmath174 , so that the derivatives can be computed classically .
clearly @xmath175 is only locally summable over the real line .
since the restriction of smooth functions to open subsets is still smooth and since from a physical perspective we can have information only on what happens on the inside of the one dimensional box , @xmath50 , we give the following prescription : @xmath176 being an element of @xmath177 and locally summable .
an analogous prescription works for the derivative with respect to _ l_. let us return to our problem settled in @xmath178 .
this time the one - form is given by @xmath179 where the derivatives in ( [ eq : distr ] ) are to be considered in the sense stated above , that is as locally integrable functions in @xmath180 .
once more , @xmath181 due to normalization the first factor in the second member vanishes so that @xmath182 while as before we get @xmath183 and an analogous expression for the partial derivative with respect to @xmath64 holds .
with this in mind we are able to get the same result ( [ eq : form ] ) as before , by reaching the boundary from the `` inside '' , rather than from the outside " , so that our new prescription , though equivalent to the one discussed above , may appear more natural .
this is coherent from a physical perspective since we can have information only on what happens on the inside of the one dimensional box @xmath50 .
for completeness , we are going to investigate the exceptional cases @xmath36 , which , as mentioned before , correspond to degenerate spectra . for @xmath184
we have that for any @xmath126 the two eigenvalues @xmath185 and @xmath186 in ( [ eq : eigenvalues ] ) coalesce , and an orthonormal basis in the @xmath90-th eigenspace is given by @xmath187 for @xmath188 , we have instead that @xmath189 , and a possible choice of an orthonormal basis is @xmath190 from the general theory of geometric phases @xcite it is well known that a degenerate spectral decomposition gives rise to a one - form connection in terms of a hermitian matrix and from a geometrical perspective this corresponds to a connection on a principal bundle , whose typical fiber is identified with a non - abelian group .
let us consider the case @xmath184 , which physically corresponds to periodic boundary conditions .
we need to compute the following matrix one - form : @xmath191 where the coefficients of the differentials are to be considered in the distributional sense .
the former equation yields the following result : @xmath192 where @xmath193 is the second pauli matrix . for a non - abelian principal fiber bundle , the curvature two - form , according to the cartan structure equation ,
is provided by @xmath194.\ ] ] plugging in the explicit expression of the above one - form ( [ eq : onef ] ) we find that @xmath195 the latter equation shows explicitly that , although every fiber is two dimensional , the overall bundle is trivial .
the one - form connection in ( [ eq : onef ] ) can be _ globally _ diagonalized making use of the basis of plane waves .
indeed , if we had started from a rotated " basis , instead of ( [ eq : basis ] ) : @xmath196 due to euler s identity , and computed ( [ eq : onef ] ) in this new basis , we would have obtained a diagonal matrix . in the most general case , instead
, one is able to determine only a _ local _
basis where the above one - form ( [ eq : onef ] ) is diagonal . on the other hand , in our case the bundle can be globally trivialized .
we have considered the problem of a particle in a box with moving walls with a class of boundary conditions .
unlike the example studied by berry and wilkinson ( two dimensional region with dirichlet boundary conditions ) , our box is one dimensional and we impose more general boundary conditions .
we consider situations in which the location and the size of the box are slowly varied .
our problem is complicated by the fact that different points in the parameter space correspond to different hilbert spaces . in order to deal with this
we need to invoke a larger hilbert space and exercise care while varying our two parameters . within this two parameter space
we conclude that there is a non - trivial geometric phase .
the functional form of this phase two - form is suggestive of the area two - form in hyperbolic geometry .
our boundary conditions in general violate time reversal symmetry , i.e , the complex conjugate of a wave function which satisfies the boundary condition described by @xmath48 may not satisfy the same boundary condition .
in fact , the only boundary conditions that respect time reversal are those where @xmath48 is real . in this case
, we would expect the geometric phase to reduce to the topological phase ( which only takes values @xmath197 ) .
thus the two form describing the phase must vanish .
in fact when @xmath48 is real ( but not equal to @xmath197 , which is a degenerate case ) , @xmath198 in ( [ eq : eigenvalues ] ) is zero or @xmath199 and the corresponding geometric phase two - form ( [ eq : curv ] ) vanishes , as it should .
the case of @xmath200 is exceptional since it has degeneracies in the spectrum . in this case
one may expect to find a @xmath43 non - abelian geometric phase of the type discussed by wilczek and zee @xcite .
however , we find that the phase is a diagonal subgroup of @xmath43 and is essentially abelian .
this is easy to understand from time reversal symmetry .
since translations and dilations are real operations , they commute with time reversal and so the allowed @xmath43 must also be real .
this reduces @xmath43 to @xmath201 , which is abelian . by a suitable choice of basis
one can render the connection diagonal as in ( 86 ) .
the `` non - abelian ''
@xmath43 wilczek - zee phase is in fact in an abelian subgroup .
it is also worth noting that the approach to @xmath200 is a singular limit because of the degeneracy there .
it is also interesting to note that the adiabatic transformations we consider act quite trivially on the spectrum of the hamiltonian .
indeed , the translations are isospectral and the dilations only cause an overall change in the scale of the energy spectrum @xmath202 . in particular , there are no level crossings and no degeneracies ( away from @xmath203 ) .
this illustrates a remark made by berry in the conclusion of @xcite : although degeneracies play an important role in berry s phase , they are not a necessary condition for the existence of geometric phase factors .
indeed , our example reiterates this point .
the berry phases are nonzero even though one of the deformations is isospectral and the other a simple scaling .
it is the twisting of the eigenvectors over the parameter space that determines the berry connection and phase , not the energy spectrum .
this work was partially supported by prin 2010llkjbx on `` collective quantum phenomena : from strongly correlated systems to quantum simulators , '' and by the italian national group of mathematical physics ( gnfm - indam ) .
in this appendix we briefly recall the technique of boundary triples and their main applications to the search of self - adjoint extensions of densely defined symmetric operators . for a review on the subject see @xcite .
von neumann s theory of self - adjoint extensions does not provide an explicit way to construct them .
the theorem , in fact , guarantees their existence once the dimensions of the deficiency subspaces are found to be equal .
however , self - adjoint extensions can be constructed as restrictions of the adjoint operator over suitable domains where a sesquilinear form identically vanishes .
given @xmath204 hermitian , we define the following sesquilinear form : @xmath205 the essential ingredient in the analysis of self - adjoint extensions is given by the deficiency subspaces , where the boundary form usually does not vanish .
every element @xmath206 can be uniquely split into three components @xcite @xmath207 where @xmath208 are the deficiency subspaces , that is the null spaces of @xmath209 . from this decomposition
it is easy to prove that @xmath210 showing how the boundary form can be used as a measure of lack of self - adjointness " .
moreover von neumann s theorem tells us that every self - adjoint extension is in a one - to - one correspondence with a unitary operator @xmath26 between the deficiency subspaces .
it follows that each self - adjoint extension of t is given by @xmath211 following @xcite we now introduce a more general tool useful for unveiling all the self - adjoint extensions of a symmetric operator .
we will show how this naturally arises from von neumann s theory and extends it .
moreover , von neumann s theory and the use of boundary forms are helpful when studying differential operators , but what could one state about self - adjoint extensions of hermitian operators , which are not in general differential operators ?
a possible answer could be given by boundary triples , which are a natural generalization of the notion of boundary values in functional spaces .
let @xmath204 be a hermitian operator with equal deficiency indices .
h _ be an auxiliary hilbert space and take @xmath212 which are supposed linear and with ranges dense in _ h _ , @xmath213 suppose that they satisfy the following condition : @xmath214 where @xmath215 , @xmath216 , and @xmath217 is the boundary form defined in ( [ eq : boundf ] ) .
a triple @xmath218 that satisfies the above conditions is called a _
boundary triple_. recall that from ( [ eq : nonvan ] ) the non - vanishing of the boundary form is due to non - trivial deficiency subspaces , so that one may choose either @xmath219 or @xmath220 , and once more by von neumann s theorem all self - adjoint extensions are in a one - to - one correspondence with unitary operators @xmath221 .
moreover , it could be useful to consider @xmath222 with the same dimension of either one of the two deficiency subspaces .
the latter statement is enforced by the fact that two hilbert spaces are unitarily equivalent if and only if they have the same dimension .
in general , it can be proved that given a boundary triple @xmath223 for a hermitian operator with equal deficiency indices , all the self - adjoint extensions @xmath224 of @xmath204 are given by @xmath225 for every unitary operator @xmath226 @xcite .
in section [ boundcond ] we apply this theorem , by choosing as auxiliary space @xmath222 the space of boundary data , and a suitable pair of maps @xmath227 , in order to get the parametrization of the self - adjoint extensions of the laplacian exhibited in @xcite . | we unveil the existence of a non - trivial berry phase associated to the dynamics of a quantum particle in a one dimensional box with moving walls .
it is shown that a suitable choice of boundary conditions has to be made in order to preserve unitarity .
for these boundary conditions we compute explicitly the geometric phase two - form on the parameter space .
the unboundedness of the hamiltonian describing the system leads to a natural prescription of renormalization for divergent contributions arising from the boundary .
_ keywords _ : geometric phases , quantum boundary conditions , time - dependent systems |
the manner in which granular materials respond to external loads belongs to a class of problems related to mechanical stability .
questions surrounding mechanical stability of particulate media have broad economical , environmental , and societal applications . these include the common practical problem of a collapsing grain silo , numerous natural events such as avalanches and mud slides , along with other unexpected destructive occurrences like dam breakage .
many such events are reported annually and readily convey the dramatic , and often catastrophic , effects that are primarily due to the failure of various granular systems .
the general physical theme that connects such a disparate range of problems is that of determining the underlying stress state of the system . given information on the distribution of stresses within the system , can we predict whether such systems will remain mechanically stable , and what are the conditions that might promote or inhibit failure ?
one possibility is to employ established theories such as continuum , linear , elasticity theory @xcite .
although elasticity theory applies over a wide range of situations , deviations from classical elasticity theory show up on a regular basis as some of the assumptions that underpin the theory break down .
these problems have largely been emphasized over the past 15 years or so in studies of packings of granular materials .
the essential point to note is that granularity , or the discreteness of the particles that constitute the packing , play a vital role in the resulting stress characteristics of the system due to a lack of a separation of length scales .
thus , it is becoming increasingly necessary to probe the relevant limits of classical elasticity theory and its range of valid application .
for instance , classical elasticity is often assumed to remain valid not only at macroscopic scales to infer stress properties from displacement fields during loading events of the lithosphere @xcite , but also at the microscopic level of interrogation in studies of nanoscopic indentation in relation to physical metallurgy @xcite .
a convenient method to test the range of validity of classical elasticity theory in granular systems that has been utilized over the past decade is the localized force perturbation technique .
this procedure has the particular benefit that for an isotropic elastic material the stress profile in response to localized forcing - green function response - can be determined without the need for any free parameters .
we review this result for the case of three dimensions - boussinesq equations - in the following section ( and see the appendix ) . to summarize efforts to date , studies on two- and three - dimensional granular packings indicate that disordered and frictional systems tend to follow the predictions of isotropic elasticity theory . whereas , ordered packings composed of frictionless particles result in strongly anisotropic stress profiles that are not consistent with a linear , isotropic , elastic model .
such anisotropic profiles ( see below for a more qualitative definition ) have been debated in relation to anisotropic elasticity theory based on differential equations of the elliptic class @xcite , and alternative descriptions of the hyperbolic wave variety @xcite . to distinguish between the traditional isotropic result and anisotropic stress profiles we refer to these as _ one - peak _ and _ two - peak _ response profiles respectively , in reference to two dimensional experiments and simulations that have reported these observations @xcite . in recent computer simulations
the role of structure for frictionless granular packings on the stress response in a pseudo - infinite medium was investigated @xcite .
these studies showed that for ordered and quasi - ordered arrangements of frictionless grains , the response function was two - peak but consistent with the framework of an anisotropic elastic theory .
a phenomenological model was employed to fit these anisotropic profiles that necessarily captured the two - peak character of the response profiles . as the structural disorder of the packing
was increased , the stress response crossed over to a one - peak response , more in line with isotropic behavior . the problem addressed here is to determine how confinement affects the stress response of granular packings .
in most previous studies samples were used of sufficient size that the perturbed packings could be considered as semi - infinite in extent . in some cases
the roughness of the supporting base in packings prepared under gravity was seen to influence the magnitude of the stress response , although the response profiles retained isotropic , elastic character @xcite .
we also highlight two particularly relevant efforts that have been proposed to describe stress response specific to two dimensional systems .
extensions to three dimensions and especially for fully confined , shallow packings remains an ongoing effort .
a non - linear elasticity formalism was developed @xcite that captures the main features of experimental results for two dimensional granular packings @xcite .
the isotropic result is recovered through a multipole expansion of the normal stress response component @xmath0 of classical elasticity .
this theory was also able to describe anisotropic profiles by including generalized nonlinear strain components through the introduction of phenomenologically defined elastic constants .
a future extension of this particular formalism may be useful in describing the stress response for a wide range of granular systems in both two and three dimensions .
additionally , force perturbation simulations of confined lennard - jones glasses @xcite found a range of stress response profiles that appear , at least superficially , similar to some of the isotropic results we present here .
moreover , in some of the above examples a stress response crossover was reported either with system size and/or packing arrangement .
this suggests that stress response investigations on granular packings can provide a route to understanding stress states in a range of systems where the discreteness of the constituent particles come into play .
however , there currently remains a distinct lack of a systematic study on the role that system size plays in modifying the stress response in granular materials .
we propose that such a study is needed with the ever - growing emphasis on the design of small devices where the particulate nature of the material eventually dominate the system properties .
moreover , from this study we can estimate relevant length scales over which continuum elasticity theory remains valid , and below which alternative descriptions may be needed .
thus , our goal here is to provide a qualitative guide map to the typical system parameters for which classical elasticity theory is likely to apply and those for which more advanced techniques may be needed to predict their properties .
this paper is organized as follows . in the section [ sec : boussinesqequations ]
we summarize the boussinesq equations of classical , isotropic , linear elasticity theory for an infinite half - plane , and discuss other possibilities that predict anisotropic stress response profiles .
in section [ simulations ] we provide an overview of the computational technique and the packing preparation protocol implemented in this work . section [ results ] presents our main findings for the stress response behavior in frictional and frictionless particle packings .
we end with conclusions and discussion .
the equations of classical elasticity theory are derived from differential equations of the elliptic class that rely on the identification of well defined displacement fields of elements constituting the continuous elastic medium @xcite .
the analytical solutions for the problem of a localized point force applied vertically at the surface of an elastic , isotropic , infinite half - space are called the boussinesq equations @xcite . in cartesian
coordinates , where the origin of the coordinate system coincides with the application point of the vertical force , @xmath1 , the components of the displacement fields are given by , @xmath2 $ \\ \\
$ u_{y}$ & $ = $ & $ \frac{f_{\rm app}}{4\pi g}\left[\frac{yz}{\rho^{3 } } - ( 1 - 2\nu)\frac{y}{\rho(\rho+z)}\right]$ \\ \\
$ u_{z}$ & $ = $ & $ \frac{f_{\rm app}}{4\pi g}\left[\frac{z^{2}}{\rho^{3 } } + \frac{2(1-\nu)}{\rho}\right]$
\end{tabular } \label{eq1}\ ] ] where @xmath3 .
both @xmath4 , the modulus of rigidity , and @xmath5 , the poisson ratio , are material parameters characterizing the bulk properties of the elastic medium .
based on thermodynamic grounds @xcite , the poisson ratio is constrained by , @xmath6 , in three dimensions . through implementation of hooke s law between stress and strain components ,
one then obtains expressions for components of the stress tensor . for the specific case when the force is applied in the ` downwards ' or negative @xmath7-direction ( see fig .
[ fig1 ] ) the normal stress in the direction of the applied force , @xmath0 , characterizes the response of the system .
we focus primarily on this component of the stress although we also provide information on the normal stress components not in the direction of the applied force , @xmath8 and @xmath9 .
we give expressions for all three normal stress components @xmath10 $ \\ \\
$ \sigma_{yy}$&$=$&$\frac{f_{\rm app}}{2\pi}[\frac{1 - 2\nu}{r^2}((1-\frac{z}{\rho})\frac{y^2-x^2}{r^2}+\frac{z x^2}{\rho^{3}})-\frac{3zy^{2}}{\rho^{5}}]$ \\ \\
$ \sigma_{zz}$&$=$&$\frac{3f_{\rm app}}{2\pi}\frac{z^3}{\rho^{5}}$ \label{eq2 } \end{tabular}\ ] ] where @xmath11 .
the stress response scales linearly with the applied force .
deviations from this indicate non - linear behavior , which , in the case of a particulate medium , corresponds to non - affine , or irreversible , particle displacements in response to the applied load .
we reiterate that that there are no fitting parameters in @xmath0 for the case of a semi - infinite half space .
therefore , we can quantify the influence of boundaries , or confinement , through a comparison of the boussinesq result of eq .
[ eq2 ] , with our computed stress response for confined systems .
furthermore , we can also quantify deviations from isotropy from the stress profiles .
we find that although the application of boussinesq theory might at first seem to be inappropriate for the systems investigated here , we find that for packings of sufficient linear size the classical theory captures the essential features of the stress properties and we classify the conditions for which isotropic elasticity theory are suitable .
we implement a granular dynamics ( gd ) variant on molecular dynamics simulations @xcite that has been designed to simulate granular systems @xcite . here
we provide a summary of the technique @xcite .
for this particular study we focus on noncohesive , monodisperse sphere packings composed of elastic spheres of diameter @xmath12 and mass @xmath13 , that interact only on contact , through forces that act in directions that are normal @xmath14 , and tangential @xmath15 , to the contact plane , defined via @xmath16 where @xmath17 , is the separation between particles @xmath18 and @xmath19 , located at positions @xmath20 and @xmath21 respectively , and @xmath22 .
@xmath23 represents the unit vector .
the normal and tangential contact forces @xmath24 are based on a linear spring - dashpot model characterized by normal and tangential linear spring constants @xmath25 , and damping factors @xmath26 , that account for elastic deformation of the contact point and inelasticity respectively , @xmath27 here , @xmath28 is the surface compression of two particles undergoing a collision and @xmath29 the relative normal / tangential velocity . the effective mass , @xmath30 , for the systems studied here .
the quantity @xmath31 represents the integrated displacement of the contact point while two particles remain in contact .
friction is implemented through a local coulomb yield criterion : @xmath32 , for a given friction coefficient @xmath33 . in this work , @xmath33 takes on the values zero and one .
thence , the total contact force is given by @xmath34 .
interactions between the particles and the walls are given by similar expressions where the wall is characterized by an effective mass of unity and the wall friction coefficient @xmath35 takes on the same values as that of the particles .
the results are presented in simulation units .
length scales are given in units of @xmath12 , timescales in units of @xmath36 , where @xmath37 is the acceleration due to gravity , forces in units of @xmath38 , stresses in units of @xmath39 , and energies are given in units of @xmath40 .
table [ simulpar ] shows the simulation parameters used in this study .
we point out that the values of the spring constants are fixed throughout this study resulting in a particle scale poisson ratio of zero . .
[ simulpar ] physical parameters used in this simulation study . @xmath41 and @xmath42 are normal and tangential spring constants in units of @xmath43 .
@xmath44 and @xmath45 are normal and tangential damping factors in units of @xmath46 . for values used here
the particle poisson ratio is zero .
the particle friction coefficient is @xmath33 and the coefficient of inelasticity @xmath47 . [
cols="<,<,<",options="header " , ] stresses are computed from the contact forces . to enable an accurate determination of stresses from microscopic to macroscopic length scales , we follow the protocol developed by goldhirsch and colleagues @xcite .
we implement a coarse graining procedure into the microscopic stress expression @xcite .
the coarse graining function chosen is a 3d gaussian , and the resulting expression for the stress components at some location @xmath48 is , @xmath49 where @xmath50 is the coarse graining length scale .
@xmath51 represents the force experienced by particle @xmath18 in contact with particle @xmath19 , obtained from eq .
the numerical prefactor @xmath52 is the normalization factor . to implement eq .
[ eq5 ] into our calculations we adopted a spatial grid over which to evaluate the coarse grained stress .
to achieve sufficient spatial resolution we fixed the grid size @xmath53 and @xmath54 .
( we computed stresses for a range of @xmath50 and grid sizes for our 3d systems and found that @xmath55 provides a suitable plateau window @xcite , at an acceptable computational cost . ) in this work we investigate how confinement influences the nature of stress response in granular packings . because of the potentially large parameter space we fixed many features of the packings and focus our attention on a subset of variables .
more specifically , we studied face centered cubic ( fcc ) ordered arrays of monodisperse spheres confined within a walled box . to prepare the initial configuration
, we placed non overlapping particles at the lattice points of an fcc structure at the hard sphere packing fraction , @xmath56 .
the packings are confined within open - top boxes , consisting of flat side walls and a bottom base .
we defined a coordinate system as shown in fig .
the bottom plane of the box defines the @xmath57-plane , while gravity points in the @xmath58 direction .
the size of the simulation boxes were chosen to be commensurate with the crystal structure .
for the fcc structures the lattice unit is defined as @xmath59 .
most of the results presented here are for boxes with square bases of side lengths , @xmath60 , ranging from , @xmath61 to @xmath62 .
we also studied rectangular shaped boxes .
the total number of particles @xmath63 , was chosen by fixing the height , @xmath64 , of all packings to be , @xmath65 . and @xmath66 define the center of box in the bottom of box .
gravity points in the @xmath58 direction.,width=264,height=264 ] after the particles were positioned at the lattice sites , we switched on gravity and allowed the particles to settle until the energy per particle , @xmath67 , reached numerical precision .
this then defines our initial configuration .
we recorded the positions of the particles , the contact forces between interacting particles , and consequently the stress , @xmath68 , of this initial state .
we then identified the top middle region of the packing , locating the particles in this region .
the second part of the simulation involves applying a downwards - pointing localized perturbation force , @xmath1 , to the top middle particle(s ) , and allowing the system to relax back to a mechanically stable state .
we varied the magnitude of this applied force over several orders of magnitude , @xmath69 , to check the limits of the linear regime .
again , we tracked particle positions and contact forces , and the final stress state @xmath70 @xcite .
the response function is calculated as the difference between the final and initial stress states , which we denote simply as , @xmath71 to make a clearer connection with previous experimental studies and visualizations @xcite .
we present our results in a convenient format that can be readily compared with previous studies and theoretical predictions .
in particular , we make use of three dimensional contour stress maps and two dimensional stress profiles which view the stress response in a given plane of the box or angularly - averaged slice across this plane , respectively . given the box geometry that we employ , it is convenient to visualize @xmath72 as a stress map in the horizontal plane . using these plots it is relatively straightforward to distinguish between isotropic and anisotropic response profiles .
these typically show up more clearly in stress profiles as single- and double - peak response profiles respectively . in all our stress maps lighter
shading indicates larger stresses .
we also present results on particle displacement fields taken in the plane .
our first results show that ordered arrays of _ frictional _ granular packings display isotropic elastic - like stress response behavior as seen in fig .
these stress maps convey the `` single - peak '' nature of the stress response measured at the bottom of the packing for different system sizes for one value of the applied force , @xmath73 .
thus , we immediately note that the boussinesq equation appears to describe confined systems .
however , we find that system size and force magnitude influence this picture .
we quantify the regime over which the classical expression remains suitably valid in two ways : firstly , we compare our stress profiles to the boussinesq expression and extract the full width at half maximum .
secondly , we identify the linear regime as a function of the applied force . .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath73 , for different system sizes , @xmath74 indicated by scale in panels.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath73 , for different system sizes , @xmath74 indicated by scale in panels.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath73 , for different system sizes , @xmath74 indicated by scale in panels.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath73 , for different system sizes , @xmath74 indicated by scale in panels.,title="fig:",width=151 ] to cast this data into a more familiar form we circularly average the stress map data of fig . [ fig2 ] to construct stress profiles and directly compare with the predictions of eq .
[ eq2 ] , from which we also obtain the full width of the response peaks . in fig .
[ fig3](a ) we show our averaged response profiles for systems of different height , clearly indicating a single - peak response consistent with the classical theory shown by the thick solid line .
\(a ) , for one value of the applied force @xmath73 for systems of fixed height @xmath75 and varying side lengths @xmath76 ( long dash line ) , @xmath77 ( thin solid line ) , @xmath78 ( dashed ) , @xmath79 ( dotted ) , and b ( solid line ) , the boussinesq result . ( b ) @xmath80 full width at half height of the stress profiles for the simulation data ( symbols ) relative to the theoretical result @xmath81 .
perfect agreement indicated by dashed line.,title="fig:",width=132 ] ( b ) , for one value of the applied force @xmath73 for systems of fixed height @xmath75 and varying side lengths @xmath76 ( long dash line ) , @xmath77 ( thin solid line ) , @xmath78 ( dashed ) , @xmath79 ( dotted ) , and b ( solid line ) , the boussinesq result .
( b ) @xmath80 full width at half height of the stress profiles for the simulation data ( symbols ) relative to the theoretical result @xmath81 .
perfect agreement indicated by dashed line.,title="fig:",width=132 ] we note that for the smallest system ( @xmath82 ) the stress profiles do not accurately conform to the boussinesq form . to better quantify deviations from eq .
[ eq2 ] we extract the full width at half height , @xmath80 , and compare with the theoretical result @xmath83 , shown in fig . [ fig3](b ) .
we find that for @xmath84 , the single - peak response profiles are suitably described by eq .
[ eq2 ] , with better agreement for larger system size as expected .
thus , we have identified a minimum system size for which the classical expression describes the data accurately , @xmath85 . to further test our results with eq .
[ eq2 ] , we again fit the data and now allow the applied force parameter in eq .
[ eq2 ] to vary as a fitting parameter , denoted here as @xmath86 , which we then compared to the actual force applied during the simulation , @xmath87 .
our results shown in fig . [ fig4 ] clearly indicate that confined systems of box sizes greater than @xmath85 conform to the semi - infinite system size result . versus @xmath86 , the fitted force using the boussinesq equation , eq .
[ eq2 ] , to fit the stress profile data at the bottom of a frictional granular packing for boxes of different sizes , with side lengths @xmath88 .
perfect agreement indicated by dashed line.,width=264 ] the magnitude of the applied force also strongly influences the response function . for small applied forces
, the stress response remains isotropic consistent with the classical result as seen in the previous figures . beyond some force threshold
the stress response becomes anisotropic due to the fcc arrangement of the grains and eq .
[ eq2 ] is no longer valid as indicated by the stress maps in fig .
[ fig5 ] . to quantify this effect and identify the linear regime as a function of applied force
, we performed a similar analysis as described above .
we varied the applied force over several orders of magnitude , @xmath89 , applied to a system size of side length @xmath90 .
figure [ fig6 ] shows that the fitted force increases linearly with applied force with unit slope when the magnitude of the perturbation force remains below @xmath91 . in ongoing work to be reported elsewhere @xcite we find that this threshold also depends on the friction coefficient .
, for the bottom of the box of a frictional system of size @xmath92 , left panel : @xmath93 .
right panel : @xmath94.,title="fig:",width=151,height=94 ] , for the bottom of the box of a frictional system of size @xmath92 , left panel : @xmath93 .
right panel : @xmath94.,title="fig:",width=151,height=94 ] versus @xmath86 , the fitted force using the boussinesq equation to fit the stress profile data at the bottom of a frictional granular packing of side length @xmath92.,width=264 ] thus , in summary , we find that for ordered frictional packings the stress state is suitably described by the semi - infinite result even for relatively small systems over a wide range in applied forcing and we have quantified the linear regime over which this agreement exists . in an effort to investigate the underlying mechanisms responsible for the changing character of the stress response in our different systems
we have computed the normal stress response at different distances from the source of the force perturbation which are shown in fig .
the left hand panels in fig .
[ fig7 ] show data for our smallest system , @xmath82 , which deviates strongly from the classical predictions despite appearing to retain a single broad peak response profile . while the right hand columns are for a larger system , @xmath90 , that is consistent with the boussinesq expression .
as expected at the top of packing ( top panels ) the stress is localized at the point of perturbation .
deeper into the packings the stress response broadens . for the smallest system
the stress quickly spreads across the entire region of the packing while for larger systems the stress peak remains relatively localized within the central region of the packing .
, in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] + , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] + , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] the reason behind this changing stress response is a strong reflective component of stress from the side walls as the stress is transmitted away from the source of the perturbation . to qualify this effect , in fig .
[ fig8 ] we plot the averaged normal stresses , @xmath95 , exerted at the side walls for our different systems , and data for the classical expression using a value of the poisson ratio of @xmath96 . indeed , we find that for the smaller systems the stress at the side walls is much larger than that expected from the classical behavior of side wall stress from boussinesq equation and the corresponding larger system size .
, at the side wall for frictional packings of height @xmath97 and side length , @xmath98 and the classical expression for a system of size @xmath90 ( bottom right panel ) , using @xmath99.,title="fig:",width=151 ] , at the side wall for frictional packings of height @xmath97 and side length , @xmath98 and the classical expression for a system of size @xmath90 ( bottom right panel ) , using @xmath99.,title="fig:",width=151 ] , at the side wall for frictional packings of height @xmath97 and side length , @xmath98 and the classical expression for a system of size @xmath90 ( bottom right panel ) , using @xmath99.,title="fig:",width=151 ] , at the side wall for frictional packings of height @xmath97 and side length , @xmath98 and the classical expression for a system of size @xmath90 ( bottom right panel ) , using @xmath99.,title="fig:",width=151 ] to generate the data presented in fig . [ fig8 ]
it is necessary to introduce the material poisson ratio @xmath5 .
our final choice for @xmath5 was made on the best fit between the stress profiles obtained via the displacement fields , eq .
[ eq1 ] , and the displacement fields constructed from the simulation data .
we matched our simulation displacement fields for our largest system to the predictions of eq .
[ eq1 ] using @xmath5 as a fitting parameter .
our confined packings therefore behave as a continuum material with an effective poisson ratio , @xmath100 .
this procedure actually provides a convenient method to determine material properties of composite materials such as granular packings @xcite .
the benefit of analyzing the displacement fields is that we can view the direct microscopic response of the system at the particle scale due to the imposed localized force perturbation .
the displacement fields of the middle vertical slice for frictional packings of different size are plotted in fig .
we also include a comparison to eq .
[ eq1 ] for a system of size @xmath101 , using @xmath99 .
these changes in the displacement fields of the perturbed system result in a larger reflection of stress at the side walls hence causing deviations in the overall stress response of smaller systems compared with the semi - infinite size result .
however , it is worth pointing out that these differences occur only for relatively small packings of size @xmath102 . above this size
, these frictional packings appear to be suitably described by the boussinesq formalism . , and the result of eq . [ eq1 ] on a system of size @xmath103 , using @xmath99.,title="fig:",width=151 ] , and the result of eq . [ eq1 ] on a system of size @xmath103 , using @xmath99.,title="fig:",width=151 ] , and the result of eq . [ eq1 ] on a system of size @xmath103 , using @xmath99.,title="fig:",width=4 ] , and the result of eq . [ eq1 ] on a system of size @xmath103 , using @xmath99.,title="fig:",width=4 ] in contrast to frictional packings , _ frictionless _ ordered arrays display primarily strongly anisotropic stress behavior in response to localized force perturbations . in fig .
[ fig10 ] the stress component @xmath0 for frictionless arrays of different sizes are compared .
apart from the smallest system size , the response profiles are ringed or multi - peaked with a minimum in the stress response characteristic of an anisotropic response function in clear contrast to the frictional packings of fig .
surprisingly , however , the smallest system of size @xmath82 exhibits a stress response that initially appears isotropic in nature - a maximum in the stress in the middle of the packing - but with some unusual features not seen in the frictional case .
in fact this smaller system does not conform to the boussinesq stress profile even though an isotropic - like single peak response is observed . .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath104 , for different systems sizes , @xmath74.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath104 , for different systems sizes , @xmath74.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath104 , for different systems sizes , @xmath74.,title="fig:",width=151 ] .
stress response maps of @xmath0 at the bottom boundary due to an applied force @xmath104 , for different systems sizes , @xmath74.,title="fig:",width=151 ] the influence of confinement plays a significant role in modifying the stress properties for frictionless packings even within the same packing . to investigate this effect further
, stress response maps at different distances from the source of the force perturbation are shown in fig .
[ fig11 ] for two system sizes , @xmath105 . in the vicinity of the localized force perturbation
, the stress response also takes on localized features where the stress is concentrated in a region directly below the point of application . at intermediate distances between the top and bottom of the packing
both systems now display strongly anisotropic stress profiles with a stress minimum focused in the region beneath the point of force application .
however , as noted above , the response measured at the bottom of the packing dramatically changes character between the two system sizes .
these features suggest that the stress state of small frictionless packings can not only vary substantially within the packing , but exhibit highly unusual stress properties compared to both larger frictionless systems and frictional packings .
, in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] + , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] + , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] , in response to a localized force of magnitude @xmath73 applied at the top of the packing .
the panels show data at different distances from the perturbation source .
rows show data for layer 1 ( top ) , 3 , and 8 ( near bottom ) , for @xmath82 ( left column panels ) and @xmath90 ( right column panels).,title="fig:",width=151 ] again we find that these unusual stress properties are a direct result of the manner in which stresses are transmitted and interact with the side wall boundaries due to the strongly correlated displacement fields of the particles during the response process .
visualization of the displacement fields for these frictionless packings are particularly illuminating as shown in fig .
[ fig12 ] .
these displacement vectors are taken from a middle slice inside the different packings .
for the smallest packing ( top left panel in fig .
[ fig12 ] ) , the particle displacements in response to the applied perturbation show a directional structure that is responsible for the increased stress reflection at the side walls .
interestingly , this reflection shields the bottom of the packing from the growing stress minimum apparent at intermediate depths . whereas , for larger systems the stress response is seen to be dominated by particle displacements along directed rays through the system responsible for the anisotropic profiles observed at the bottom of the packings .
.,title="fig:",width=151 ] .,title="fig:",width=151 ] .,title="fig:",width=151 ] .,title="fig:",width=151 ] the results presented above suggest that there exists some crossover response for confined , frictionless packings that depends on the size of the system .
we have investigated the changing character of the stress response and its dependence on box size .
this crossover behavior is illustrated in stress response maps of fig .
[ fig13 ] , where the box side lengths have been incrementally increased from @xmath106 . , with increasing box size for systems of side length , @xmath107.,title="fig:",width=151 ] , with increasing box size for systems of side length , @xmath107.,title="fig:",width=151 ] , with increasing box size for systems of side length , @xmath107.,title="fig:",width=151 ] , with increasing box size for systems of side length , @xmath107.,title="fig:",width=151 ] it is apparent that changing box size results in large - scale stress fluctuations that span the system as the stress response crosses over from a more isotropic - like response ( smaller systems ) to strongly anisotropic behavior ( larger system ) .
there is a region of box sizes for these frictionless packings where the stress state of the system exhibits peculiar behavior that does not conform to either a pseudo - isotropic description nor can it be considered anisotropic in character .
these features can have implications in the design of mechanical systems at small scales where confinement can have a large effect on the stress state of the system .
we have also investigated the role of box geometry on the nature of the stress response . as shown earlier we find that for the smaller frictionless systems confined within a square - base box with @xmath108 , the stress response exhibits a broad peak in the central region of the base .
we find that highly unusual and complex features emerge if we now increase the size of only one side of the box into a rectangular base keeping @xmath109 fixed .
stress maps for this scenario are shown in fig .
[ fig14 ] , where we have increased the side ratio from 1 to 10 .
these result emphasize the importance of box size and geometry on the nature of stress response in shallow packings .
, in frictionless packings as one side of the confining box is increased from a square base of side lengths @xmath108 , to a rectangular base of side lengths @xmath109 and @xmath110 , @xmath109 and @xmath111 , @xmath109 and @xmath112.,title="fig:",width=151 ] , in frictionless packings as one side of the confining box is increased from a square base of side lengths @xmath108 , to a rectangular base of side lengths @xmath109 and @xmath110 , @xmath109 and @xmath111 , @xmath109 and @xmath112.,title="fig:",width=151 ] , in frictionless packings as one side of the confining box is increased from a square base of side lengths @xmath108 , to a rectangular base of side lengths @xmath109 and @xmath110 , @xmath109 and @xmath111 , @xmath109 and @xmath112.,title="fig:",width=151 ] , in frictionless packings as one side of the confining box is increased from a square base of side lengths @xmath108 , to a rectangular base of side lengths @xmath109 and @xmath110 , @xmath109 and @xmath111 , @xmath109 and @xmath112.,title="fig:",width=151 ]
using computer simulations of a model granular material , we have studied the nature of the stress response in shallow and confined granular crystalline arrays .
packings composed of highly frictional particles generally exhibit a stress response function measured at the bottom of the packing that resembles that of an isotropic , elastic material that is well described by the boussinesq equations of classical elasticity .
thus , the semi - infinite half - space result remains suitably valid to describe the stress state of the system provided the size of the base of the confining walls is larger than approximately 40 particle diameters . in this regime
the stress response appears to behave as a linear elastic material . for smaller frictional systems , the stress response exhibits confinement effects that are not accurately captured by the classical result .
packing arrangement and the influence of the boundaries have significant effects on the stress state of the system .
furthermore , due to the fcc crystalline arrangements studied here , anisotropic stress response behavior is observed at large applied forces whereby the influence of structure starts to dominate the stress properties .
frictionless packings exhibit stress response properties that are generally anisotropic in nature due to the fcc arrays
. however , smaller packings can be tuned to observe a range of behaviors depending on the location within the packing for square boxes , or on the ratio of the side lengths for rectangular systems .
complex response patterns emerge for frictionless rectangular arrays .
we also point out that for the anisotropic profiles - either with zero friction or for large forcing - the resulting stress response indicates a decrease in the stress relative to the initial state prior to perturbation , i.e. in the measures defined here the stress becomes negative .
this suggests that such force perturbations actually weaken the material relative to the unperturbed state .
our results further suggest that there exists a transition in the underlying character of the response as the friction coefficient is varied . to highlight this possibility ,
we show in fig .
[ fig15 ] , the averaged stress profiles for a square - base , @xmath103 , for different particle friction coefficients in response to an applied force , @xmath73 .
indeed we find that increasing the friction coefficient suppresses the stress dip in the central region of the packing until a single - peak response is recovered for moderate to large friction coefficients .
we expect that the precise location of this transition will also depend on the magnitude of the applied force @xcite . , averaged stress profiles for different friction coefficients , @xmath33 , for one system size , @xmath113 , in response to an applied force @xmath73 .
each stress profile is scaled by its maximum value , @xmath114,width=264 ] our studies indicate that for confined granular arrays , particle properties , magnitude of forcing , and geometry of the confining container can all be used to design materials with specific stress states . through this accessible parameter space ,
confined granular packings can be made to mimic the properties of linear , isotropic , elastic materials , where the stress is concentrated beneath the point of perturbation - single - peak response .
or they can be designed to exhibit strongly anisotropic properties where stresses are transmitted along preferred directions . moreover , for particular box geometries , the stress response can be transmitted approximately uniformly through the system .
thus , it might be necessary to take into account the possible stress states indicated here during the design of microscopic granular devices and components .
here we provide a short derivation of the normal stress response component @xmath0 ( see eq .
[ eq2 ] ) , within the framework of linear elasticity , indicating how the material parameters , @xmath4 and @xmath5 , cancel from the expression given in eq .
this derivation follows directly from landau and lifshitz @xcite .
the basic ingredient is the implementation of a constitutive relation between stress and strain that corresponds to hooke s law : the normal stress @xmath0 is linearly proportional to the strain components @xmath115 , @xmath116 , and @xmath117 , @xmath118 \label{hookeszz}\ ] ] where the strain components are obtained from derivatives of the components of the displacement fields leading to @xmath119 $ , \\ \\ $ u_{yy}$&$=$&$\frac{1}{4\pi g}\left[\frac{z}{\rho^{3}}-\frac{3 y^2 z } { \rho^{5}}-\frac{1 - 2\nu}{\rho^{2}+ \rho z}-\frac{(1 - 2 \nu)y^2(2+\frac{z}{\rho})}{(\rho^{2}+\rho z)^2}\right]$ , \\
\\ $ u_{zz}$&$=$&$\frac{1}{4\pi g}\left[\frac{2 \nu z}{\rho^{3}}-\frac{3 z^3}{\rho^{5}}\right]$ , \end{tabular } \label{strains}\ ] ] and we have set the value of the applied , perturbing force to unity for notational convenience . combining the above equations we find , @xmath120\nonumber\end{aligned}\ ] ] explicitly emphasizing the cancellation of @xmath4 from the equation .
the resulting expression can be written as @xmath121 \nonumber \label{hookeszzstep1}\end{aligned}\ ] ] and with further algebra ,
@xmath122 \nonumber \label{hookeszzstep2}\ ] ] the terms containing @xmath5 cancel out and one obtains eq .
[ eq2 ] for an arbitrary applied force , @xmath1 , directed in the @xmath58 direction .
this work is dedicated to isaac goldhirsch .
we acknowledge support from the national science foundation cbet-0828359 . | stress transmission inside three dimensional granular packings is investigated using computer simulations .
localized force perturbation techniques are implemented for frictionless and frictional shallow , ordered , granular arrays confined by solid boundaries for a range of system sizes .
stress response profiles for frictional packings agree well with the predictions for the semi - infinite half plane of classical isotropic elasticity theory down to boxes of linear dimensions of approximately forty particle diameters and over several orders of magnitude in the applied force .
the response profiles for frictionless packings exhibit a transitional regime to strongly anisotropic features with increasing box size .
the differences between the nature of the stress response are shown to be characterized by very different particle displacement fields . |
the recent work of e.lieb and j.yngvason @xcite on the foundations of thermostatics has given a new interest to an old but still controversial problem of thermodynamics .
this is the so called `` adiabatic piston problem '' described as follows @xcite ( see also recent references in @xcite ) .
the system is composed of a finite volume cylinder containing two fluids separated by an adiabatic movable piston , as shown in fig.1 .
a brake maintains initially the piston at rest , and the two fluids are in equilibrium with temperature , pressure and volume given by @xmath0 , and @xmath1 , respectively .
@xmath2 denotes here the cross section area of the cylinder , and @xmath3 the initial position of the piston . at a certain time
the brake is released , and the problem is to find the final equilibrium state .
+ figure 1 : adiabatic piston problem the laws of thermostatics give immediately the condition for the mechanical equilibrium , namely that in the final equilibrium state the pressures must be equal , i.e. @xmath4 ( with @xmath5 ) .
however , they give no information concerning the final temperatures , except for the condition that they must be such that for both fluids the final entropy can not be smaller than its initial value , i.e. @xmath6 , and @xmath7 . on the other hand , from the laws of thermodynamics one can derive ordinary differential equations for the time evolution of the system . solving these equations
would then give the unique solution of the problem , and in particular the final temperatures .
this idea has been recently followed in @xcite , where a system of thermodynamic evolution equations has been derived within a very simple model .
one had to distinguish then between the case where the two fluids have vanishing viscosities and the case where at least one of the fluids has a non - zero viscosity .
for non - viscous fluids the thermodynamic equations predicts a periodic motion subject to the energy conservation law , with constant subsystem entropies : @xmath8 , @xmath9 .
when at least one of the fluids is viscous , the system will attain the final equilibrium state with equal pressures @xmath10 .
however , to find the final entropies ( or temperatures ) one has to solve the evolution equations which so far could not be achieved explicitly .
these same evolution equations were also obtained in the framework of the kinetic theory of gases , assuming the fluctuations of the piston to be negligibly small and considering perfect gases , and several numerical solutions have been obtained @xcite .
but even having obtained the macroscopic description of the evolution of the system the controversial question remains whether the stochastic frictionless motion of the piston induced by collisions with the fluid particles could eventually lead ( on a much longer time scale ) to a final state where the temperatures of the fluids are equal , and whether such a microscopic fluctuation mechanism would not violate the second law of thermodynamics .
clearly , the analysis of such problem requires the microscopic description of the dynamical evolution . in this article
we thus present a microscopic study of the motion of the adiabatic piston in contact with two infinite volume reservoirs .
our analysis , based on the boltzmann equation , has been inspired by the above fundamental adiabatic piston question . in section 2
we define the microscopic model and we discuss the applicability of the boltzmann equation to the description of the motion of the piston .
the rigorous determination of the stationary state in one dimension for the case where the mass of the piston is equal to the fluid particle mass is presented in section 3 .
the derived solution at equal temperatures reduces simply to the maxwell equilibrium state .
but when the temperatures are different a non - equilibrium stationary state appears characterized by an average velocity of the piston oriented toward the higher temperature region , although no macroscopic force is present ( the pressures on both sides of the piston are supposed to be equal ) .
this situation is thus precisely that of the `` molecular motor problem '' . the rigorous derivation ( within boltzmann s theory ) of the fluctuation induced macroscopic motion is the main result of our work .
the relevance of this prediction to a much more realistic problem of a massive piston is discussed in the closing section 4 .
the case where the mass of the piston is much larger than that of a fluid particle will be discussed in the forthcoming paper @xcite .
in order to simplify the analysis as much as possible we consider in the following the one - dimensional system occupying the whole infinite line . the two fluids are assumed to be perfect gases composed of identical point particles of mass @xmath11 .
they are separated by a structureless particle of mass @xmath12 which plays the role of the piston .
the perfectly elastic collisions between the fluid particles result in instantaneous exchanges of velocities .
so , from the point of view of the dynamics of the piston the fluid particles can be considered as non - interacting , penetrating each other in their unperturbed free motion .
the only perturbation of the fluid states is due to collisions with the piston .
the similar model has been already investigated almost forty years ago by lebowitz @xcite and few years later by alkemade , van kampen and mac donald @xcite .
when a collision occurs , the pre - collisional velocities @xmath13 and @xmath14 of the piston and the fluid particle , respectively , suffer a change according to the elastic collision law @xmath15 let us suppose that at the initial moment the fluids to the left and to the right of the piston are at equilibrium with temperatures @xmath16 and @xmath17 , and number densities @xmath18 and @xmath19 , respectively . moreover , the macroscopic mechanical equilibrium will be assumed : the pressures on both sides of the piston are equal @xmath20 ( @xmath21 is boltzmann s constant ) .
when @xmath22 , the probability that the piston interacts back with the fluid particle it already collided with becomes negligible . in the absence of recollisions
the piston always `` sees '' the unperturbed maxwell distributions of the arriving fluid particles , and no correlations can occur between its velocity and the precollisional velocity of the fluid particles .
the perturbations caused by collisions with the piston go away to infinity .
so , in the large mass limit @xmath23 , we can expect the linear boltzmann equation to give a correct description of the dynamical evolution of the piston . in order to write down this equation
we denote by @xmath24 and @xmath25 the maxwell velocity distributions of the two surrounding fluids , where @xmath26 let @xmath27 denote the probability density for finding the piston at point @xmath28 with velocity @xmath13 at time @xmath29 .
when @xmath30 , @xmath27 satisfies the linear boltzmann equation of the form @xmath31 \nonumber\end{aligned}\ ] ] where @xmath32 is the unit heaviside step function .
our aim is to determine the velocity distribution of the piston in the asymptotic stationary state attained for @xmath33 .
it turns out that this problem can be given a rigorous answer in the special case of @xmath34 .
but when the mass @xmath12 of the piston is equal to the fluid particle mass @xmath11 , the recollisions can take place , and the validity of the boltzmann equation becomes questionable .
however , what we are really interested in is the case of @xmath35 .
we thus expect that neglecting the effect of recollisions in the equal mass case will lead to the evolution qualitatively similar to that occuring for @xmath35 , simply because the same class of dynamical events will be taken into account .
this should apply in particular to the nature of the stationary state .
it should be noticed here that even when compared with the exact results the linear boltzmann equation for @xmath34 represented a very reasonable approximation to the velocity correlation function and to the van hove self - correlation function of the one - dimensional hard rod fluid @xcite .
having adopted the above point of view we are going to study in the next section the equation satisfied by the stationary probability density @xmath36 for finding the piston with velocity @xmath13 in the equal pressure case ( see eq.([2 ] ) ) . denoting by @xmath37 the stationary solution of equation ( [ 4 ] ) , we have @xmath38
according to the linear boltzmann equation ( [ 4 ] ) the stationary velocity distribution @xmath36 in the equal pressure case ( [ 2 ] ) satisfies the equation @xmath39\phi ( v ) \label{6}\ ] ] @xmath40\phi ( v)\ ] ] in the following we shall assume @xmath41 .
it is convenient to express eq.([6 ] ) in terms of the following two functions @xmath42 using the identities @xmath43\ ] ] @xmath44\ ] ] we rewrite eq.([6 ] ) in the form @xmath45g(v)= \frac{1}{2 } ( v-<v>)[{\frac{1}{t_{1}}\phi^{m}_{t_{1}}(v)}- { \frac{1}{t_{2}}\phi^{m}_{t_{2}}(v ) } ] \label{9}\ ] ] where @xmath46 denotes the average velocity of the piston @xmath47 we now notice that the first- and second - order derivatives of the functions @xmath48 and @xmath49 are given by @xmath50 therefore ( [ 9 ] ) takes the form of a simple differential equation @xmath51 ' \label{15}\ ] ] @xmath52\ ] ] using the asymptotic formulae @xmath53 @xmath54 we can readily evaluate the limit @xmath55 = \left(\frac{1}{t_{1 } } + \frac{1}{t_{2}}\right)<v > \label{18}\ ] ] integrating equation ( [ 15 ] ) with the boundary condition ( [ 18 ] ) yields the relation @xmath56 \label{19 } \\ { } & + & < v>[\int_{v}^{\infty } dv{\frac{1}{t_{1}}\phi^{m}_{t_{1}}(v)}+ \int_{-\infty}^{v } dv{\frac{1}{t_{2}}\phi^{m}_{t_{2}}(v ) } ] \nonumber\end{aligned}\ ] ] we thus arrived at a first order differential equation for the unknown function @xmath57 .
the solution of eq.([19 ] ) can be now constructed by standard methods .
introducing the auxiliary function @xmath58 @xmath59 and using the relations @xmath60 one readily determines the function @xmath61 .
it reads @xmath62 \label{22 } \\ { } & + & \int \frac{du}{i^{2}(u)}{\rm sgn}(v - u)\left [ \int_{-\infty}^{u}dv{\frac{1}{t_{2}}\phi^{m}_{t_{2}}(v)}+ \int_{u}^{\infty}dv{\frac{1}{t_{1}}\phi^{m}_{t_{1}}(v)}\right]<v > \nonumber\end{aligned}\ ] ] here @xmath63 .
the relations ( [ 14 ] ) , ( [ 20 ] ) permit to determine the shape of the stationary probability density @xmath36 .
we find @xmath64 \label{23}\ ] ] @xmath65 \left\ { t_{1}+t_{2 } + \phantom{\frac{1}{t}}\right.\ ] ] @xmath66 \right\}\ ] ] in order to complete the calculation of @xmath67 we have still to evaluate the average piston velocity @xmath46 of the piston which shows in equation ( [ 23 ] ) .
this can be achieved from the normalization condition @xmath68 to perform the calculation we group together the terms multiplying @xmath46 , and rewrite equation ( [ 23 ] ) in the form @xmath69 it turns out that the function @xmath70 can be written as a derivative @xmath71 \nonumber\end{aligned}\ ] ] we then deduce from ( [ 26 ] ) the relation @xmath72 which follows directly from the asymptotic formulae ( [ 16 ] ) .
the right hand side of equation ( [ 25 ] ) equals @xmath73 - \frac{k_{b}}{m}\frac{\partial}{\partial v}a(v ) \label{28}\ ] ] where @xmath74\ + \ ] ] @xmath75\ ] ] taking again into account the asymptotics ( [ 16 ] ) and the normalization ( [ 24 ] ) we get @xmath76 \label{29}\ ] ] integrating equation ( [ 25 ] ) over the whole velocity space we thus arrive at the following formula for the stationary mean velocity of the piston @xmath77 \right\}^{-1 } \nonumber\end{aligned}\ ] ] the formula ( [ 30 ] ) is the main result of our analysis . whereas the mean velocity vanishes for @xmath78 ,
the piston acquires a macroscopic systematic motion when @xmath79 , and this under the equal pressures condition ( [ 2 ] ) . in order to further investigate this phenomenon
let us analyze the case of a small temperature difference .
we thus put @xmath80 when @xmath81 , the denominator in equation ( [ 30 ] ) takes the asymptotic form @xmath82 where @xmath83^{-2 } \label{32}\ ] ] and @xmath84 is the dimensionless maxwell distribution @xmath85 on the other hand the numerator equals @xmath86 + { \rm o}((\delta t)^{2 } ) \label{33}\ ] ] so , when @xmath87 , the mean velocity of the piston is given by @xmath88\left [ \int dw h(w)\right]^{-1 } + { \rm o}(\delta t ) \label{34}\ ] ] integration by parts yields the relation @xmath89 the inequality @xmath90^{3/2}w\int dw ' { \rm sign}(w - w')\phi ( w')\leq 0 \label{35}\ ] ] permits to determine the sign of the mean velocity of the piston .
we find @xmath91 the inequality ( [ 37 ] ) shows that in the stationary state the piston performs an uniform motion in the direction of the higher temperature region .
in fact , we checked by numerical analysis that the inequality ( [ 37 ] ) remained true for any @xmath41 .
we have studied the effect of fluctuations on the motion of the piston separating two semi - infinite regions filled with perfect gases at equal pressures .
the explicit stationary solution of the boltzmann equation ( [ 23 ] ) , ( [ 30 ] ) obtained for the special case where the mass of the piston equals that of a fluid particle reduces to the equilibrium maxwell distribution when the temperatures on both sides of the piston are equal .
the mean velocity then vanishes .
in fact , it has been rigorously shown that the piston ( tagged particle ) performs in this case a clasical diffusive motion @xcite , @xcite .
when the temperatures are different , the space asymmetry induces asymmetry of fluctuations which turns out to have a macroscopic effect .
the piston attains a stationary state with a non - zero average velocity oriented toward the higher temperature region . in view of the absence of macroscopic forces
we discover here a genuine molecular motor .
there is some analogy between our prediction and the `` stochastic ratchet '' problem where one observes that fluctuations conspire with the spatial anisotropy to induce a macroscopic directed motion in the absence of macroscopic forces @xcite , @xcite .
we expect the same qualitative behavior in the case of a massive piston whose mass is much larger than that of a gas molecule . indeed , when @xmath35 the boltzmann equation , which was the basis of our analysis , gives an adequate description of the dynamics because the creation of dynamical correlations between the piston and the gas molecules via the recollision mechanism becomes negligible
. the persistence of the non - vanishing average velocity in the @xmath35 regime at equal pressures and different temperatures will be the subject of the forthcoming article@xcite .
one can thus expect that the stochastic motion of the piston as a whole can lead to an effective energy transfer between the separated fluids in the absence of pressure difference .
hence , although the piston has no internal degrees of freedom calling it adiabatic becomes a misuse of the term .
it remains to be seen what is the relevance of our analysis to the original adiabatic piston problem where the fluids occupy finite volumes .
this is a delicate question indeed , because in one dimension even for a very massive piston the collisions with the boundaries will inevitably induce recollisions : the piston will interact back with the particles it had already met before .
so , the boltzmann description can not be reliable in the long time limit .
indeed , at the time scale characterized by the time necessary to cover the total length of the cylinder @xmath92 with the thermal velocity of the fluids the effects of dynamical correlations will start to play an important role .
however , for sufficiently large volumes , after the mechanical equilibrium has been established , one can expect ( on a larger time scale ) the occurence of a net displacement of the piston , reducing the volume of the the higher temperature fluid , as it is the case for the conducting piston .
it seems likely that for finite volumes the boltzmann description can work much better in two and three dimensions because then the recollision effects due to boundaries become weaker .
anyway , for finite volumes one will have to consider the system of kinetic equations coupling the evolution of the state of the piston to that of the fluids .
the authors are greatful to j.l lebowitz and n.g .
van kampen for stimulating information about their own work on a similar model .
j.piasecki greatly acknowledges the hospitality at the institue of theoretical physics of the cole polytechnique fdrale de lausanne where a part of this research has been performed .
e. lieb and j. yngvason , _ the physics and mathematics of the second law of thermodynamics _ , physics reports ( in press ) , austin , math.phys .
thermodynamics _ , john wiley and sons , new york , 1963 .
appendix c. + see also h.b .
callen , _ thermodynamics and thermostatics _ , john wiley and sons , 2nd edition , new york , 1985 , pp .
51 and 53 .
gruber , _ thermodynamics of systems with internal adiabatic constrains : time evolution of the adiabatic piston _ , preprint epfl ( 1998 ) .
b. crosignani , p. di porto and m. segev , _ approach to thermal equilibrium in a system with adiabatic constraint _ , am.j.phys .
* 64 * ( 1996 ) , 610 .
ch.gruber and j.piasecki ( in preparation ) .
lebowitz , _ stationary non - equilibrium gibbsian ensembles _ ,
* 114 * ( 1959 ) 1192 .
p. rsibois , _ solution de lquation de boltzmann pour des btonnets durs unidimensionnels _ , physica * 90a * ( 1978 ) 273 .
jepsen , _ dynamics of a simple many - body system of hard rods _ , j.math.phys .
* 6 * ( 1965 ) 405 .
j.l . lebowitz and j.k .
percus , _ kinetic equations and density expansions : exactly solvable one - dimensional system _ , phys.rev .
* 155 * ( 1967 ) 122 .
alkemade , n.g .
van kampen and d.k.c .
mac donald , f.r.s .
, _ non - linear brownian movement of a generalized rayleigh model
_ , proc.roy.soc . *
a271 * ( 1963 ) 449 .
e. abad and a. mielke , _ brownian motion in fluctuating periodic potentials _
ann.physik * 7 * ( 1998 ) 9 . h. kohler and a. mielke , _ noise - induced transport at zero temperature _ , j.phys.a * 31 * ( 1998 ) 1929 . | 0.75 cm the controversial problem of the evolution of an isolated system with an internal adiabatic wall is investigated with the use of a simple microscopic model and the boltzmann equation . in the case of two infinite volume one - dimensional ideal fluids separated by a piston
whose mass is equal to the mass of the fluid particles we obtain a rigorous explicit stationary non - equilibrium solution of the boltzmann equation .
it is shown that at equal pressures on both sides of the piston , the temperature difference induces a non - zero average velocity , oriented toward the region of higher temperature .
it thus turns out that despite the absence of macroscopic forces the asymmetry of fluctuations results in a systematic macroscopic motion .
this remarkable effect is analogous to the dynamics of stochastic ratchets , where fluctuations conspire with spatial anisotropy to generate directed motion .
however , a different mechanism is involved here .
the relevance of the discovered motion to the adiabatic piston problem is discussed . |
in 1961 , james and stein @xcite considering the problem of estimating the mean vector @xmath4 of a @xmath5-dimensional normal distributed random vector @xmath2 with a covariance matrix @xmath6 introduced an estimator @xmath7 which outperforms the maximum likelihood estimate ( mle ) @xmath8 for dimension @xmath9 , under the common quadratic risk @xmath10 in the sense that for all parameter values @xmath4
@xmath11 this unexpected result draw a great interest of mathematical statisticians and stimulated a number of authors to contribute to the theory of improved estimation by extending the problem of james and stein in different directions to more general models with unknown covariance matrix and considering other types of estimates ( see @xcite for more details and other references ) .
a considerable effort has been directed towards the problems of improved estimation in non - gaussian models with the spherically symmetric distributions ( see @xcite ) and in the non - parametric regression models @xcite .
now the james stein estimator and other improved shrinkage estimators are widely used in econometrics and the problems associated with the signal processing . in this paper
we will consider the problem of estimating the mean in a conditionally gaussian distribution .
suppose that the observation @xmath2 is a @xmath5-dimensional random vector which obeys the equation @xmath12 where @xmath4 is a constant vector parameter , @xmath13 is a conditionally gaussian random vector with a zero mean and the covariance matrix @xmath14 , i.e. @xmath15 , where @xmath16 is some fixed @xmath17-algebra .
we propose to consider a shrinkage estimator of the form @xmath18 where @xmath19 is a positive constant which will be specified below .
it will be shown that such an estimator allows one to obtain an explicit upper bound for the quadratic risk in case of the regression model with a conditionally gaussian noise .
theorem [ le.sec:gas.1 ] in section [ sec : gas ] claims that the estimator outperforms the maximum likelihood estimate @xmath20 uniformly in @xmath4 from any compact set @xmath21 for any dimension @xmath5 starting from two . in section [ sec : per ] , we apply the estimator to solve the problem of improved parametric estimation in the regression model in continuous time with a non - gaussian noise . the rest of the paper is organized as follows . in section [ sec : gas ] , we impose some conditions on the random covariance matrix @xmath14 and derive to upper bound for the difference of risks @xmath22 corresponding to @xmath23 and @xmath20 respectively . in section [ sec : aut ] , the estimate is used for the parameter estimation in a discrete time regression with a gaussian noise depending on some nuisance parameters .
appendix contains some technical results .
in this section we will derive an upper bound for the risk of estimate under some conditions on the random covariance matrix @xmath14 .
assume that @xmath24 there exists a positive constant @xmath25 , such that the minimal eigenvalue of matrix @xmath14 satisfies the inequality @xmath26
@xmath27 the maximal eigenvalue of the matrix @xmath14 is bounded on some compact set @xmath21 from above , i.e. @xmath28 where @xmath29 is some known positive constant .
let denote the difference of the risks of estimate and that of as @xmath30 we will need also the following constant @xmath31 where @xmath32,@xmath33 [ le.sec:gas.1 ] let the noise @xmath13 in have a conditionally gaussian distribution @xmath34 and its covariance matrix @xmath14 satisfy conditions @xmath35 with some compact set @xmath21 .
then the estimator with @xmath36 dominates the mle @xmath20 for any @xmath37 , i.e. @xmath38 ^ 2.\ ] ] first we will establish the lower bound for the random variable @xmath39 .
[ le.sec:gas.2 ] under the conditions of theorem 2.1 @xmath40 the proof of lemma is given in the appendix . in order to obtain the upper bound for @xmath41 we will adjust the argument in the proof of stein s lemma @xcite to the model with a random covariance matrix .
we consider the risks of mle and of @xmath42 r(\theta^{*},\theta)=r(\hat{\theta}_{ml},\theta)+\mathbf{e}_{\theta}[\mathbf{e}((g(y)-1)^{2}\|y\|^{2}|\mathcal{g})]\\[2 mm ] + 2\sum_{j=1}^{p}\mathbf{e}_{\theta}[\mathbf{e}((g(y)-1)y_{j}(y_{j}-\theta_j)|\mathcal{g})],\end{gathered}\ ] ] where @xmath43 . denoting @xmath44 and applying the conditional density of distribution of a vector @xmath2 with respect to @xmath17-algebra
@xmath16 @xmath45 one gets @xmath46 making the change of variable @xmath47 and assuming @xmath48 , one finds that @xmath49 where @xmath50 denotes the @xmath51-th element of matrix @xmath52 .
these quantities can be written as @xmath53 thus , the risk for an estimator takes the form @xmath54|_{u = y}\right).\end{gathered}\ ] ] therefore one has @xmath55 where @xmath56 this implies that @xmath57 since @xmath58 , one comes to the inequality @xmath59 from here it follows that @xmath60 taking into account the condition @xmath24 and the lemma [ le.sec:gas.2 ] , one has @xmath61 minimizing the function @xmath62 with respect to @xmath19 , we come to the desired result @xmath63 ^ 2.\ ] ] hence theorem [ le.sec:gas.1 ] .
[ le.sec:gas.3 ] let in the noise @xmath64 with the positive definite non random covariance matrix @xmath65 and @xmath66
. then the estimator with @xmath67 dominates the mle for any @xmath37 and compact set @xmath21 , i.e. @xmath38 ^ 2.\ ] ] note that if @xmath68 then @xmath69 ^ 2.\ ] ] [ le.sec:gas.4 ] if @xmath70 and @xmath71 in model then the risk of estimate is given by the formula @xmath72 ^ 2=:r_p.\ ] ] by applying the stirling s formula for the gamma function @xmath73 one can check that @xmath74 as @xmath75 .
the behavior of the risk for small values of @xmath5 is shown in fig.1 .
it will be observed that in this case the risk of the james stein estimate remains constant for all @xmath9 , i.e. @xmath76 and the risk of the mle @xmath20 is equal to @xmath5 and tends to infinity as @xmath75 . at @xmath71 .
, scaledwidth=90.0% ]
in this section we apply the proposed estimate to a non - gaussian continuous time regression model .
let observations @xmath77 obey the equation @xmath78 here a vector @xmath79 of unknown parameters from some compact set @xmath21.assume that @xmath80 is a one - periodic @xmath81 functions , bounded and orthonormal in @xmath82 $ ] .
the noise @xmath83 in is a non - gaussian ornstein uhlenbeck process given by stochastic differential equation @xmath84 where @xmath85 is unknown parameter , @xmath86 is a levy process satisfying the equation @xmath87 here @xmath88 are unknown constants , @xmath89 is a standard brownian motion , @xmath90 is a compound poisson process defined as @xmath91 where @xmath92 is a poisson process with unknown intensity @xmath93 and @xmath94 is a sequence of i.i.d .
gaussian random variables with parameters ( 0,1 ) .
the problem is to estimate the unknown vector parameter @xmath4 on the basis of observations @xmath77 .
let @xmath95 denote the @xmath17-algebra generated by the poisson process .
it will be noted that the model is conditionally gaussian given the @xmath17-algebra @xmath16 . therefore one can use estimate to obtain an improved estimate of the unknown vector parameter @xmath4 . to this end
we have to reduce the initial continuous time regression model to a discrete time model of the form .
the quality of an estimator @xmath96 will be measured by the quadratic risk @xmath97 a commonly used estimator of an unknown vector @xmath4 in model is the least squares estimate ( lse ) @xmath98 with @xmath99 from here taking into account , one has @xmath100 where @xmath101 and @xmath102 is the random vector with coordinates @xmath103 note that the vector @xmath102 has a conditionally gaussian distribution with a zero mean and conditional covariance matrix @xmath104 with the elements @xmath105 thus the initial problem of estimating parameter @xmath4 in can be reduced to the that of estimating parameter @xmath4 in conditionally gaussian regression model .
[ le.sec:per.1 ] let the regression model be given by the equations , @xmath106 . then , for any @xmath107 and @xmath108 , the estimator of @xmath4@xmath109 dominates the lse @xmath110:@xmath111 ^ 2.\ ] ] to proved this theorem one can apply theorem [ le.sec:gas.1 ] and it suffices to check conditions @xmath24 , @xmath27 on the matrix @xmath112 .
the proof of conditions @xmath24 and @xmath27 is given in the appendix .
in this section we consider the problem of improved estimating the unknown mean of a multivariate normal distribution when the dispersion matrix is unknown and depends on some nuisance parameters .
let in the noise @xmath113 , be described by a gaussian autoregression process @xmath114 where @xmath115 , @xmath116 and @xmath117 are independent gaussian ( 0,1 ) random variables .
assume that the parameter @xmath118 in is unknown and belongs to interval @xmath119 $ ] , where @xmath120 is known number .
it is easy to check that the covariance of the noise @xmath13 has the form @xmath121 a & 1 & ... & a^{p-2 } \\[2 mm ] & \ddots & & \\[2 mm ] a^{p-1 } & a^{p-2 } & ... & 1 \end{array } \right ) \ ] ] let @xmath13 in be specified by with @xmath122 $ ] . then for any @xmath123
the mle is dominated by the estimator @xmath124 and @xmath125 one has that @xmath126
. now we find the estimation of the maximal eigenvalue of matrix @xmath127 . by definition @xmath128 one
has @xmath129 by applying the cauchy bunyakovskii inequality
we obtain that @xmath130 thus , @xmath131 hence , taking into account the theorem [ le.sec:gas.1 ] we come to assertion of proposition .
in this paper we propose a new type improved estimation procedure . the main difference from the well - known james stein estimate is that in the dominator in the corrected term we take the first power of the observation norm @xmath132 .
this allow us to improve estimation with respect to mle begining with any dimension @xmath0 .
moreover , we apply this procedure to the estimation problem for the non - gaussian ornstein uhlenbeck levy regression model .
6.1 . proof of the lemma [ le.sec:gas.2 ] . from one
has @xmath133 using a repeated conditional expectation and since the random vector @xmath13 is distributed conditionally normal with a zero mean , then @xmath134 making the change of variable @xmath135 and applying the estimation @xmath136 we find @xmath137 6.2 .
the proof of conditions @xmath24 and @xmath27 on the matrix @xmath112 .
the elements of matrix @xmath112 can be written as @xcite @xmath140 where @xmath141 and @xmath142 are the jump times of the poisson process @xmath92 , i.e. @xmath143 notice that by one can the matrix @xmath112 present as @xmath146 where @xmath147 is non random matrix with elements @xmath148 and @xmath149 is a random matrix with elements @xmath150 .
\end{gathered}\ ] ] this implies that @xmath151 therefore @xmath152 and we come to the assertion of lemma [ lem.sec:app.1 ] .
[ lem.sec:app.2 ] let @xmath83 be defined by with @xmath85 .
then a maximal eigenvalue of the matrix @xmath153 with elements defined by , satisfy the following inequality @xmath154 where @xmath155 and @xmath156 .
one has @xmath157 where @xmath158 since the @xmath159 is a one - periodic orthonormal functions therefore the first integral is equal to @xmath160 and in view of the inequality @xmath161 for any @xmath85 @xmath162 from here denoting @xmath163 we obtain that @xmath164 hence lemma [ lem.sec:app.2 ] .
w. james , c. stein , estimation with quadratic loss , in : proceedings of the fourth berkeley symposium on mathematics statistics and probability , vol . 1 , university of california press , berkeley , 1961 , pp .
361 - 380 . | the paper considers the problem of estimating a @xmath0 dimensional mean vector of a multivariate conditionally normal distribution under quadratic loss .
the problem of this type arises when estimating the parameters in a continuous time regression model with a non - gaussian ornstein uhlenbeck process driven by the mixture of a brownian motion and a compound poisson process .
we propose a modification of the james stein procedure of the form @xmath1 where @xmath2 is an observation and @xmath3 is a special constant .
this estimate allows one to derive an explicit upper bound for the quadratic risk and has a significantly smaller risk than the usual maximum likelihood estimator for the dimensions @xmath0 .
this procedure is applied to the problem of parametric estimation in a continuous time conditionally gaussian regression model and to that of estimating the mean vector of a multivariate normal distribution when the covariance matrix is unknown and depends on some nuisance parameters .
_ keywords _ : conditionally gaussian regression model ; improved estimation ; james stein procedure ; non - gaussian ornstein uhlenbeck process . _ ams 1991 subject classifications _ : primary : 62c20 ; secondary : 62c15 |
ensembles of ordinary random walks in one dimension @xcite are used to model physical , chemical , and biological processes ranging from wetting @xcite to the motion of colloidal particles in narrow channels @xcite and reaction - diffusion processes @xcite . in particular , first - passage properties @xcite of multiple random walks explain dynamics of interacting spins @xcite , lead changes in a voting process @xcite , and lifetime of knots in polymer chains @xcite . in these examples , first - passage properties
are intertwined with the ordering of the walkers @xcite .
previous studies focused on first - passage statistics of extreme particles ; for example , the probability that a single prey particle survives the predators to its left @xcite .
these problems involve a single first - passage probability and hence , a single first - passage exponent @xcite . in this paper ,
we ask first - passage questions that concern the bulk particles , not necessarily the extreme ones .
we find that a family of first - passage exponents characterizes the first - passage kinetics .
these exponents depend on two parameters : the particle order and the total number of particles . yet ,
when the number of particles is very large , the exponents depend on a single scaling variable .
this scaling behavior is unusual . in equilibrium as well as in non - equilibrium settings ,
one or two scaling exponents quantify a scaling behavior @xcite . in the present case , however , the exponents themselves obey scaling , and remarkably , there are scaling laws for the scaling exponents . , the position of the rightmost particle is @xmath10 and that of the leftmost particle is @xmath11.,scaledwidth=42.0% ] we consider an array of @xmath0 identical particles that undergo simple random walk on a one - dimensional line , and investigate two different first - passage problems where the particle order plays an essential role .
the first problem , discussed in sections ii - v , concerns the probability that the @xmath2th rightmost particle hasnt crossed the origin , if all particles start to the right of the origin ( figure [ fig - initial ] ) .
this quantity equals the likelihood that @xmath12 out of all random walks have yet to simultaneously reside in the negative half - line . in the second problem , described in section vi ,
we identify the particle order with rank , such that the rightmost particle is viewed as the leader , and similarly , the leftmost particle as the laggard .
we ask : what is the probability that the rank of the initial leader does not fall below some specified threshold ? in both problems , we find a family of nontrivial first - passage exponents . in both cases ,
the exponents depend on two variables : the particle order and the total number of particles , @xmath0 .
interestingly , in the large-@xmath0 limit , the exponents become a function of a single scaling variable . however , the similarities between the two problems end here as the two scaling variables are fundamentally different and moreover , the two scaling functions are dissimilar .
our analysis relies on mapping the @xmath0 noninteracting random walks onto a single compound random walk in @xmath0 dimensions . by combining this mapping with exact and asymptotic properties for kinetics of first - passage inside a cone @xcite
, we obtain approximate values for the first - passage exponents .
the cone approximation is straightforward to implement , yet it yields useful estimates for exponents and in particular , this framework faithfully captures typical and extremal properties of the first - passage exponents .
our system consists of @xmath0 identical particles .
each particle undergoes a random walk on a one - dimensional lattice . at every time step ,
one particle is selected at random , and it moves to the left , @xmath13 , or to the right @xmath14 , with equal probabilities .
time is augmented by the inverse number of particles after each such step , @xmath15 .
the particles always undergo independent random walks and hence , they are noninteracting .
let @xmath16 be the position of the @xmath2th rightmost particle at time @xmath3 ( figure [ fig - initial ] ) .
we consider the initial configuration where all particles are located in the positive half - line ( figure [ fig - initial ] ) , @xmath17 for all @xmath2 .
we stress that it is not the initial order , but instead , the order at time @xmath3 , that sets the index @xmath2 .
we are interested in @xmath1 , the probability that the @xmath2th rightmost particle remains in the positive half - line until time @xmath3 .
hence , @xmath1 is the likelihood that @xmath18 for all @xmath19 .
in particular , @xmath20 is the probability that the rightmost particle has yet to cross the origin , while @xmath21 is the probability that the leftmost particle has not crossed the origin .
clearly , the probabilities @xmath22 decrease monotonically with @xmath2 , @xmath23 the `` survival '' probabilities @xmath1 generalize the classic survival probability of a single one - dimensional random walk in the presence of a trap .
indeed , when @xmath24 , we have @xmath25 @xcite . as usual , the survival probability @xmath22 immediately gives the first - passage probability as @xmath26\times dt$ ] is the probability that the @xmath2th rightmost particle crosses the origin for the first time during the infinitesimal time interval @xmath27 .
the analytically solvable case of two particles yields valuable insights into the general behavior .
when @xmath28 , we map the two random walks onto a single random walk in two dimensions . the position of the two - dimensional walk is specified by the positions of the two independent walks . at @xmath29 , the two - dimensional walk is always inside the first quadrant ( figure [ fig - rw2 ] ) .
, the random walk is confined to the exterior of the third ( negative ) quadrant .
( b ) for @xmath30 , the random walk is confined to the interior of the first ( positive ) quadrant.,scaledwidth=45.0% ] the quantity @xmath20 is the probability that the coordinates of the two random walks have not become negative _ simultaneously _ up to time @xmath3 , or equivalently , the probability that the two - dimensional walk remains in the exterior of the third quadrant ( figure [ fig - rw2]a ) . to find the survival probability , we impose an absorbing boundary condition along the edge of the third quadrant ( figure [ fig - rw2]a ) .
then @xmath20 is the probability that the compound particle avoids this absorbing boundary up to time @xmath3 .
the region in which the random walk can move is a two - dimensional cone , or equivalently , a wedge with opening angle @xmath31 .
( the opening angle @xmath32 is defined as the angle between the cone axis and the cone surface , so it lies within the bounds @xmath33 . )
thus , @xmath20 equals the survival probability of a particle diffusing inside a wedge with an absorbing surface .
this survival probability decays algebraically , @xmath34 in the long - time limit @xcite . by substituting @xmath31 into the general expression
, we find the intriguing behavior @xmath35 as @xmath36 . as in
, this asymptotic behavior holds regardless of the initial position , although the prefactor does depend on the initial conditions . along the same lines ,
@xmath37 is the probability that the positions of the two random walks remain positive up to time @xmath3 , or alternatively , the probability that the two - dimensional walk remains in the _ interior _ of the first quadrant ( figure [ fig - rw2]b ) .
now , the boundary of the first quadrant is absorbing ( figure [ fig - rw2]b ) , and the random walk is confined to a wedge with opening angle @xmath38 . using , we again find power - law behavior albeit with a larger exponent , @xmath39 as shown below , this particular behavior follows from an elementary argument .
the analytic results for a two - particle system suggest that generally , the survival probabilities decay algebraically , @xmath40 in the long - time limit .
moreover , we expect that the decay exponents are distinct , and that @xmath41 increases monotonically with @xmath2 , @xmath42 consistent with
. and @xmath37 for three particles .
these results are from numerical simulations.,scaledwidth=40.0% ] .the exponent @xmath43 , obtained from numerical simulations for @xmath44 . [ cols="^,<,<,<,<,<,<",options="header " , ] using the asymptotic behaviors @xmath45 we deduce that the variable @xmath46 defined in has the asymptotic behaviors @xmath47 substituting these expressions into the asymptotic behavior of the scaling function in , we find the limiting behaviors ( see also @xcite ) @xmath48 the exponent @xmath49 decreases monotonically with @xmath50 : it weakly diverges when @xmath51 and it vanishes as @xmath52 . by substituting @xmath53 and @xmath54 , respectively , into the appropriate expressions in , we find the leading large-@xmath0 behaviors @xmath55 both expressions match estimates based on heuristic arguments @xcite .
the smallest and the largest exponents , listed respectively in tables vi - vii , show that the cone yields an excellent approximation . yet , the quality of the approximation declines ever so slightly as @xmath0 increases .
monte carlo simulations with a large number of particles confirm the scaling behavior .
moreover , we numerically verified that the entire survival probability becomes a universal function of the scaling variable @xmath50 , in analogy with , that is @xmath56 with @xmath57 , as @xmath7 .
interestingly , the numerical results strongly suggest that the scaling function specified in is asymptotically _ exact _ ( figure [ fig - gamma ] ) .
finally , the exponent @xmath58 , that characterizes the probability that the original leader always ranks higher than median has a simple limiting value , @xmath59 . versus the scaling variable @xmath50 .
the line shows the prediction of the cone approximation given in equation , as well as results of numerical simulations with @xmath60 ( bullets ) and @xmath61 particles ( squares).,scaledwidth=45.0% ]
in summary , we investigated two first - passage problems involving ordered random walks in one dimension . in both cases , there is a series of survival probabilities and a nontrivial family of first - passage exponents .
a universal function describes the exponents when the number of particles is very large .
remarkably , there are scaling laws for the scaling exponents . in general
, a first - passage process with @xmath0 random walks is equivalent to diffusion in a high - dimensional space with a complicated absorbing boundary .
this boundary is typically formed by multiple intersecting planes .
solving the first - passage problem is equivalent to solving the electrostatic problem given these boundary conditions @xcite .
for example , to find @xmath62 for a three - particle system , we must solve for the potential in the exterior of an insulating three - dimensional corner .
yet , the solution , which likely requires an ingenious implementation of the image method , remains unknown even in this physically relevant geometry @xcite .
to circumvent this difficulty , we introduced the cone approximation in which the absorbing boundary is replaced with the surface of a suitably chosen cone .
this approximation is straightforward to implement and utilizes exact and asymptotic properties for first passage in a cone .
the cone approximation provides useful estimates for the exponents and moreover , it faithfully captures both typical and extremal features . in particular , the cone approximation yields the correct scaling variable for the first - passage exponents .
we used the term `` cone approximation '' , yet this framework produces the lower bound whenever the allowed space is a generalized cone with the crucial property of invariance under dilation , @xmath63 @xcite .
the computation of the first - passage exponent in a generalized @xmath64-dimensional cone requires computation of the lowest eigenvalue @xmath65 of the angular portion @xmath66 of the laplacian @xmath67 subject to the dirichlet boundary condition , @xmath68 , on the surface of the cone .
the choice of the smallest eigenvalue ensures that @xmath69 inside the allowed region .
the first - passage exponent @xmath70 is related to the lowest eigenvalue via @xmath71 , see @xcite .
we now invoke a theorem that , in its simplest form , states that amongst all domains with the same volume , the lowest eigenvalue of the laplace - dirichlet operator occurs when the domain is a ball .
rayleigh conjectured this result for two dimensions @xcite , and faber and krahn eventually proved it @xcite .
the rayleigh - faber - krahn theorem generalizes to higher dimensions and under mild conditions , to riemannian manifolds @xcite . in the context of first - passage processes
, this theorem implies that of all generalized cones of fixed solid angle , the first - passage exponent is minimal for the circular cone .
since the regions discussed in our study qualify as generalized cones , results of the cone approximation constitute rigorous lower bounds for the exponents .
cones can also provide upper bounds .
to establish an upper bound , we choose a cone that is inscribed by the absorbing boundary .
clearly , the first - passage process is faster inside the inscribed cone and therefore , the corresponding exponent must be an upper bound .
the cone approximation appears to be asymptotically exact for the leadership problem as it predicts the bulk of the exponents with excellent accuracy .
a first step toward proving this exactness is the observation that all opening angles in approach a right angle as the number of particles diverges .
the other major challenge is finding an exact , or at least asymptotically exact , framework for calculating the family of exponents in the origin - crossing problem .
finally , we mention that in the absence of a theoretical method for obtaining statistics of first - passage exactly , monte carlo simulations play vital role .
we utilized a straightforward simulation technique .
it is interesting to find out how accelerated monte carlo simulations fare in producing accurate estimates for the exponents @xcite .
a generalized cone is an infinite convex domain that contains a special point , called an apex , and has the property that each ray emanating from the apex and going through any point inside a cone lies inside the cone . | we investigate first - passage statistics of an ensemble of @xmath0 noninteracting random walks on a line . starting from a configuration in which all particles are located in the positive half - line ,
we study @xmath1 , the probability that the @xmath2th rightmost particle remains in the positive half - line up to time @xmath3 .
this quantity decays algebraically , @xmath4 , in the long - time limit .
interestingly , there is a family of nontrivial first - passage exponents , @xmath5 ; the only exception is the two - particle case where @xmath6 . in the @xmath7 limit , however , the exponents attain a scaling form , @xmath8 with @xmath9 .
we also demonstrate that the smallest exponent decays exponentially with @xmath0 .
we deduce these results from first - passage kinetics of a random walk in an @xmath0-dimensional cone and confirm them using numerical simulations .
additionally , we investigate the family of exponents that characterizes leadership statistics of multiple random walks and find that in this case , the cone provides an excellent approximation . |
in this paper we develop the connection between triangulations of type @xmath15 root polytopes and a commutative algebra @xmath18 , the subdivision algebra of type @xmath15 root polytopes .
a type @xmath15 root polytope is a convex hull of the origin and some of the points @xmath3 for @xmath4 $ ] , where @xmath19 denotes the @xmath20 standard basis vector in @xmath21 .
a polytope @xmath22 corresponds to each monomial @xmath23 , and each relation of the algebra equating a monomial with three others , @xmath24 , can be interpreted as cutting the polytope @xmath25 into two polytopes @xmath26 and @xmath27 with interiors disjoint such that @xmath28 ; thus the name subdivision algebra for @xmath18 .
a subdivision algebra @xmath29 for type @xmath30 root polytopes was studied in @xcite yielding an exciting interplay between polytopes and algebras .
the algebra @xmath29 is related to the algebras studied by fomin and kirillov in @xcite and by kirillov in @xcite , which have tight connections to schubert calculus . using techniques for polytopes ,
the algebra @xmath29 can be understood better , and using the properties of @xmath29 results for root polytopes can be deduced .
the subdivision algebra @xmath18 is a type @xmath15 generalization of @xmath29 and its intimate connection to type @xmath15 root polytopes is displayed by a variety of results obtained by using this connection .
root polytopes were first defined by postnikov in @xcite , although the full root polytope of type @xmath30 already appeared in the work of gelfand , graev and postnikov @xcite , where they gave a canonical triangulation of it into simplices corresponding to noncrossing alternating trees .
properties of this triangulation are studied in ( * ? ? ?
* exercise 6.31 ) .
canonical triangulations for a family of type @xmath30 root polytopes were constructured in @xcite extending the result of @xcite . in this paper
we define type @xmath15 analogs for noncrossing and alternating graphs , and show that a family of type @xmath15 root polytopes , containing the full root polytope , has canonical triangulations into simplices corresponding to noncrossing alternating graphs . using the canonical triangulations we compute the volumes for these root polytopes .
the subdivision algebra @xmath18 is closely related to the noncommutative bracket algebra @xmath31 of type @xmath15 defined by a. n. kirillov @xcite .
kirillov conjectured the uniqueness of the reduced form of a coxeter type element in @xmath31 .
as the algebras @xmath18 and @xmath31 have over ten not - so - simple - looking relations , we postpone their definitions and the precise statement of kirillov s conjecture till section [ sec : alg ] . while at the first sight the relations of @xmath31 might appear rather mysterious , we interpret them similarly to the relations of @xmath18 , as certain subdivisions of root polytopes
this connection ultimately yields a proof of kirillov s conjecture along with more general theorems on reduced forms , of which there are two types . in the noncommutative algebra @xmath31 we show that for a family of monomials @xmath32 , including the coxeter type element defined by kirillov , the reduced form is unique . in the commutative algebra @xmath18 and the commutative counterpart
@xmath33 $ ] of @xmath31 , the reduced forms are not unique ; however , we show that the number of monomials in a reduced form of @xmath34 is independent of the order of reductions performed .
we also study the noncommutative bracket algebra @xmath35 of type @xmath17 defined by a. n. kirillov @xcite .
using noncommutative grbner bases techniques we prove that a family of monomials has unique reduced forms in it .
a special case of our results proves a conjecture of a. n. kirillov about the uniqueness of the reduced form of a coxeter type element in the bracket algebra of type @xmath17 .
this paper is organized as follows . in section [ sec : alg ] we give the definition of @xmath31 , as well as two related commutative algebras @xmath36 and @xmath18 .
we also state kirillov s conjecture pertaining to @xmath31 in section [ sec : alg ] . in section
[ sec : red ] we introduce signed graphs , define the type @xmath16 analogue of alternating graphs , and show how to reformulate the relations of the algebras @xmath37 into reductions on graphs . in section [ sec : root ] we introduce coned root polytopes of type @xmath15 and state the reduction lemma which connects root polytopes and the algebras @xmath38 . in section [ sec : play ] we prove a characterization of the vertices of coned type @xmath15 root polytopes , while in section [ sec : reduction_lemma ] we prove the reduction lemma . in section
[ sec : forest ] we establish the relation between volumes of root polytopes and reduced forms of monomials in the algebras @xmath37 using the reduction lemma . in section [ reductionsb ]
we reformulate the noncommutative relations of @xmath31 in terms of egde - labeled graphs and define well - structured and well - labeled graphs , key for our further considerations . in section [ reductionsb1 ] we prove a simplified version of kirillov s conjecture , construct a canonical triangulation for the full type @xmath15 root polytope @xmath39 and calculate its volume . in section [ sec : gen ] we generalize kirillov s conjecture to all monomials arising from well - structured and well - labeled graphs and give the triangulations and volumes of the corresponding root polytopes . in section [ 11 ] we prove the general form of kirillov s conjecture in a weighted bracket algebra @xmath40 , and show a way to calculate ehrhart polynomials of certain type @xmath15 root polytopes . in section [ sec : d ] the definition of @xmath35 is given along with kirillov s conjecture pertaining to it . in section [ sec:13 ]
combinatorial results regarding a family of monomials are proved .
finally , in section [ sec : grobi ] we prove a general result on the reduced forms of monomials implying kirillov s type @xmath17 conjecture .
in this section the definition of the bracket algebra @xmath31 is given , along with a conjecture of kirillov pertaining to it .
we introduce the subdivision algebra @xmath18 , which , as its name suggests , will be shown to govern subdivisions of type @xmath15 root polytopes .
kirillov @xcite defined the algebra we are denoting @xmath31 as a type @xmath41 bracket algebra @xmath42 , but since we can interpret its generating variables as corresponding to either the type @xmath41 and type @xmath15 roots , we refer to it as a type @xmath15 bracket algebra @xmath31 .
the reason for our desire to designate @xmath31 as a type @xmath15 algebra is its essential link to type @xmath15 root polytopes , which we develop in this paper .
here we define a simplified form of the bracket algebra @xmath31 ; for a more general definition , see section [ 11 ] .
let the * bracket algebra * @xmath43 * of type @xmath15 * be an associative algebra over @xmath44 with a set of generators @xmath45 subject to the following relations : \(1 ) @xmath46 @xmath47 , for @xmath48 , \(2 ) @xmath49 ( @xmath50 ) @xmath51 , @xmath52 , @xmath53 , for @xmath54 distinct .
( @xmath55 ) @xmath56 , @xmath57 , for all @xmath58 \(5 ) @xmath59 , for @xmath60 , ( @xmath61 ) @xmath62 , for @xmath60 , \(6 ) @xmath63 , for @xmath60 , ( @xmath64 ) @xmath65 , for @xmath60 , \(7 ) @xmath66 , for @xmath60 , ( @xmath67 ) @xmath68 , for @xmath60 , \(8 ) @xmath69 , for @xmath60 , ( @xmath70 ) @xmath71 , for @xmath60 , \(9 ) @xmath72 , for @xmath10 ( @xmath73 ) @xmath74 , for @xmath10 let @xmath75 be a coxeter type element in @xmath43 and let @xmath76 be the polynomial in variables @xmath77 obtained from @xmath78 by successively applying the defining relations @xmath79 in any order until unable to do so .
we call @xmath76 a * reduced form * of @xmath78 and consider the process of successively applying the defining relations @xmath80 as a reduction process , with possible commutations ( 2)-(4 ) between reductions , as we show in the following example
. @xmath81 in the example above the pair of variables on which one of reductions @xmath80 is performed is in boldface , and the variables which we commute according to one of ( 2)-(4 ) are underlined . * conjecture 1 .
( kirillov @xcite ) * _ apart from applying the relations ( 1)-(4 ) , the reduced form @xmath76 of @xmath78 does not depend on the order in which the reductions are performed .
_ note that the above statement does not hold true for any monomial .
we show one simple example of how it fails .
@xmath82 @xmath83 note that we reduced the monomial @xmath84 in two different ways yielding two different polynomials .
the reader can also check another example of this phenomenon by reducing the monomial @xmath85 in two different ways to obtain two different reduced forms .
we prove conjecture 1 in section [ reductionsb1 ] , as well as its generalizations in sections [ sec : gen ] and [ 11 ] .
we first define and study a commutative algebra @xmath18 closely related to @xmath31 , though more complicated than its commutative counterpart , @xmath86 $ ] , which is simply the commutative associative algebra over @xmath44 with a set of generators @xmath45 subject to relations ( 1 ) and @xmath80 from above .
our motivation for defining @xmath18 is a natural correspondence between the relations of @xmath18 and ways to subdivide type @xmath15 root polytopes , which correspondence is made precise in the reduction lemma ( lemma [ reduction_lemma ] ) . in order to emphasize this connection , we call @xmath18 the subdivision algebra of type @xmath15 . the subalgebra @xmath87 of @xmath18 generated by @xmath88 has been studied in @xcite , and an analogous correspondence between the relations of @xmath87 and ways to subdivide type @xmath89 root polytopes has been established . moreover , results in the spirit of conjecture 1 for type @xmath89 can also be found in @xcite .
let the * subdivision algebra * @xmath18 be the commutative algebra over @xmath90 $ ] , where @xmath91 is a variable ( and a central element ) , with a set of generators @xmath45 subject to the following relations : \(1 ) @xmath46 @xmath47 , for @xmath48 , \(2 ) @xmath92 , for @xmath60 , \(3 ) @xmath93 , for @xmath60 , \(4 ) @xmath94 , for @xmath60 , \(5 ) @xmath95 , for @xmath60 , \(6 ) @xmath96 , for @xmath10 \(7 ) @xmath97 , for @xmath98 notice that when we set @xmath99 relations ( 2)-(5 ) of @xmath18 become relations ( 5)-(8 ) of @xmath31 , and if we combine relations ( 6 ) and ( 7 ) of @xmath18 we obtain relation @xmath100 of @xmath31 . in some cases we will in fact simply work with the commutative counterpart of @xmath31 , @xmath36 .
we treat relations ( 2)-(7 ) of @xmath18 as * reduction rules * : @xmath101 @xmath102 @xmath103 @xmath104 @xmath105 @xmath106 0.1 in a * reduced form * of the monomial @xmath107 in variables @xmath108,$ ] in the algebra @xmath18 is a polynomial @xmath109 obtained by successive applications of reductions ( [ red1])-([red6 ] ) until no further reduction is possible , where we allow commuting any two variables . requiring that @xmath107 is in variables @xmath108,$ ]
is without loss of generality , since otherwise we can simply replace @xmath110 with @xmath111 and @xmath112 with @xmath113 .
note that the reduced forms are not necessarily unique .
however we show in section [ sec : forest ] that the number of monomials in a reduced form of a suitable monomial @xmath107 is independent of the order of the reductions performed .
in this section we rephrase the reduction process described in section [ sec : alg ] in terms of graphs
. this view will be useful throughout the paper .
we use the language of signed graphs .
signed graphs have appeared in the literature before , for example in zaslavsky s and reiner s work @xcite .
their notation is not the same , and we use a notation closer to reiner s . in particular , positive and negative edges in our notation mean something different than in zaslavsky s language .
we request the reader to read the definitions with full attention for this reason .
a * signed graph * @xmath5 on the vertex set @xmath114 $ ] is a multigraph with each edge labeled by @xmath115 or @xmath116 .
all graphs in this paper are signed and in each of them the loops are labeled positive .
we denote an edge with endpoints @xmath117 and sign @xmath118 by @xmath119 .
note that @xmath120 . as a result ,
we drop the signs from the loops in figures . a positive edge , that is an edge labeled by @xmath115 ,
is said to be * positively incident * , or , * incident with a positive sign * , to both of its endpoints .
a negative edge is positively incident to its smaller vertex and * negatively incident * to its greater endpoint .
we say that a graph is * alternating * if for any vertex @xmath121 the edges of @xmath5 incident to @xmath122 are incident to @xmath122 with the same sign .
think of a monomial @xmath123 in variables @xmath124,$ ] as a signed graph @xmath5 on the vertex set @xmath114 $ ] with a negative edge @xmath125 for each appearance of @xmath110 in @xmath107 and with a positive edge @xmath126 for each appearance of @xmath112 in @xmath107 and with a loop @xmath127 for each appearance of @xmath128 in @xmath107 .
let @xmath129 $ ] denote this graph .
it is straighforward to reformulate the reduction rules ( [ red1])-([red6 ] ) in terms of reductions on graphs .
if @xmath123 , then we replace each monomial @xmath107 in the reductions by corresponding graphs @xmath129 $ ] .
* reduction rules for graphs : * given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath131 for some @xmath132 , let @xmath133 be graphs on the vertex set @xmath114 $ ] with edge sets @xmath134 given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath135 for some @xmath132 , let @xmath133 be graphs on the vertex set @xmath114 $ ] with edge sets @xmath136 given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath137 for some @xmath132 , let @xmath133 be graphs on the vertex set @xmath114 $ ] with edge sets @xmath138 given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath139 for some @xmath132 , let @xmath133 be graphs on the vertex set @xmath114 $ ] with edge sets @xmath140 given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath141 for some @xmath142 , let @xmath133 be graphs on the vertex set @xmath114 $ ] with edge sets @xmath143 given a graph @xmath130 on the vertex set @xmath114 $ ] and @xmath144 for some @xmath142 , let @xmath133 be graphs on the vertex set @xmath145 $ ] with edge sets @xmath146 we say that @xmath130 * reduces * to @xmath133 under the reduction rules defined by equations ( [ graphs1])-([graphs6 ] ) .
an * @xmath147-reduction tree *
@xmath148 for a monomial @xmath149 , or equivalently , the graph @xmath150 $ ] , is constructured as follows .
the root of @xmath148 is labeled by @xmath151 $ ] .
each node @xmath129 $ ] in @xmath148 has three children , which depend on the choice of the edges of @xmath129 $ ] on which we perform the reduction .
e.g. , if the reduction is performed on edges @xmath152)$ ] , @xmath132 , then the three children of the node @xmath153 $ ] are labeled by the graphs @xmath133 as described by equation ( [ graphs1 ] ) . for an example of an @xmath147-reduction tree ,
see figure [ fig : tau ] .
-reduction tree with root corresponding to the monomial @xmath154 . summing the monomials corresponding to the graphs labeling the leaves of the reduction tree multiplied by suitable powers of @xmath91
, we obtain a reduced form @xmath155 of @xmath154 , @xmath156 . , scaledwidth=110.0% ] of course , given a graph we can also easily recover the corresponding monomial .
namely , given a graph @xmath5 on the vertex set @xmath114 $ ] we associate to it the monomial @xmath157=m^{\mathcal{b}^c}[g]=\prod_{(i , j , \e ) \in e(g)}\w(i , j,\e)$ ] , where @xmath158 for @xmath10 , @xmath159 for @xmath160 , @xmath161 and @xmath162 .
summing the monomials corresponding to the graphs labeling the leaves of the reduction tree @xmath148 multiplied by suitable powers of @xmath91 , we obtain a reduced form of @xmath149 .
generalizing the terminology of ( * ? ? ?
* definition 12.1 ) , a root polytope of type @xmath0 is the convex hull of the origin and some of the points @xmath8 , @xmath11 and @xmath163 for @xmath164 , @xmath165 $ ] , where @xmath19 denotes the @xmath20 coordinate vector in @xmath2 .
a very special root polytope is the full type @xmath15 root polytope @xmath166 ) \\ & = \textrm{convhull}(0 , e_{ij}^- , 2e_k \mid 1\leq i < j \leq
n , k \in [ n]),\end{aligned}\ ] ] where @xmath167 and @xmath168
. we study a class of root polytopes including @xmath1 , which we now discuss .
let @xmath5 be a graph on the vertex set @xmath114 $ ] .
let @xmath169 define @xmath170 @xmath171 @xmath172 where @xmath173\}$ ] is the set of positive roots of type @xmath15 .
the idea to consider the positive roots of a root system inside a cone appeared earlier in reiner s work @xcite , @xcite on signed posets .
coned type @xmath174 root polytopes were studied in @xcite .
define the * transitive closure * of a graph @xmath5 as @xmath175 the * root polytope * @xmath12 associated to graph @xmath5 is @xmath176 the root polytope @xmath12 associated to graph @xmath5 can also be defined as @xmath177 the equivalence of these two definition is proved in lemma [ equivalent ] in section [ sec : reduction_lemma ] .
note that @xmath178 for the graph @xmath179 , \{(n , n , + ) , ( i , i+1 , - ) \mid i \in [ n-1]\}).$ ] while the choice of @xmath5 such that @xmath180 is not unique , it becomes unique if we require that @xmath5 is * minimal * , that is for no edge @xmath181 can the corresponding vector @xmath182 be written as a nonnegative linear combination of the vectors corresponding to the edges @xmath183 .
graph @xmath184 is minimal .
we can describe the vertices in @xmath185 in terms of paths in @xmath5 .
a * playable route * @xmath186 of a graph @xmath5 is an ordered sequence of edges @xmath187 @xmath188 , @xmath189 for @xmath190 $ ] , such that @xmath191 and @xmath192 , @xmath190 $ ] , are incident to @xmath189 with opposite signs . for a playable route of @xmath5 , @xmath193 .
a * playable pair * @xmath194 in a graph @xmath5 is a pair of playable routes @xmath195 @xmath196 @xmath197 and @xmath198 @xmath199 such that @xmath200 and @xmath201 .
it follows that @xmath202 .
define a map @xmath203 from the playable routes and playable pairs to @xmath204 as follows .
@xmath205 [ playable_route ] let @xmath5 be a graph on the vertex set @xmath114 $ ] .
any @xmath206 is @xmath207 or @xmath208 for some playable route @xmath186 or playable pair @xmath194 of @xmath5 .
if the set of vectors @xmath209 is linearly independent , then the correspondence between playable routes and pairs of @xmath5 and vertices in @xmath210 is a bijection .
the proof of proposition [ playable_route ] appears in section [ sec : play ] . define @xmath211 , e(g ) )
\mid \mbox { $ \mathcal{v}_g$ is a linearly independent set}\},\ ] ] and @xmath212 with linearly independent generators .
since all polytopes in this paper are coned root polytopes with linearly independent generators , we simply refer to them as coned root polytopes .
the next lemma characterizes graphs @xmath5 which belong to @xmath213 ; a version of it appears in @xcite .
[ fong ] ( @xcite ) a graph @xmath5 on the vertex set @xmath114 $ ] belongs to @xmath213 if and only if each connected component of @xmath5 is a tree or a graph whose unique simple cycle has an odd number of positively labeled edges .
the full root polytope @xmath214 , since the graph @xmath215 by lemma [ fong ] .
we show below how to obtain central triangulations for all polytopes @xmath216 .
a * central triangulation * of a @xmath217-dimensional root polytope @xmath218 is a collection of @xmath217-dimensional simplices with disjoint interiors whose union is @xmath218 , the vertices of which are vertices of @xmath218 and the origin is a vertex of all of them .
depending on the context we at times take the intersections of these maximal simplices to be part of the triangulation .
we now state the crucial lemma which relates root polytopes and the algebras @xmath219 and @xmath18 defined in section [ sec : alg ] .
[ reduction_lemma ] * ( reduction lemma ) * given a graph @xmath220 with @xmath217 edges let @xmath133 be as described by any one of the equations ( [ graphs1])-([graphs6 ] ) .
then @xmath221 , @xmath222 where all polytopes @xmath223 are @xmath217-dimensional and @xmath224 what the reduction lemma really says is that performing a reduction on graph @xmath220 is the same as cutting " the @xmath217-dimensional polytope @xmath225 into two @xmath217-dimensional polytopes @xmath226 and @xmath227 , whose vertex sets are subsets of the vertex set of @xmath225 , whose interiors are disjoint , whose union is @xmath225 , and whose intersection is a facet of both .
we prove the reduction lemma in section [ sec : reduction_lemma ] .
in this section we prove proposition [ playable_route ] , which characterizes the vertices of any root polytope @xmath12 .
we start by proving the statement for connected @xmath228 .
[ conn ] let @xmath229 be a connected graph .
the correspondence between playable routes of @xmath5 and vertices in @xmath210 given by @xmath230 is a bijection .
denote by @xmath231w$ ] the coefficient of @xmath19 when @xmath232 is expressed in terms of the standard basis @xmath233 of @xmath21 .
_ proof of proposition [ conn ] . _ given a playable route @xmath186 of @xmath5 , @xmath234 by definition .
it remains to show that for each vertex @xmath235 there exists a playable route @xmath186 in @xmath5 such that @xmath207 .
the uniqueness of such a route follows from the linear independence of the set of vectors @xmath209 for @xmath229 .
consider @xmath235 .
then @xmath236 , for some @xmath237 , or @xmath238 , for @xmath165 $ ] , and @xmath239 let @xmath240 , \ { e \in e(g ) \mid c_e \neq 0\})$ ] .
observe that @xmath241 has at most one connected component containing edges .
this follows since a connected @xmath229 contains at most one simple cycle , and if there were two connected components of @xmath241 , one would be a tree contributing at least two nonzero coordinates to the right hand side of ( [ h ] ) and each connected component containing edges contributes at least one nonzero coordinate to the right hand side of ( [ h ] ) .
but , the left hand side of ( [ h ] ) has one or two nonzero coordinates .
if @xmath242 is a leaf of @xmath241 then @xmath243v\neq 0 $ ] . therefore
, @xmath241 can have at most two leaves .
we consider three cases depending on the number of leaves @xmath241 has : @xmath244 . in all cases
we show that there exists a playable route @xmath186 of @xmath5 with all its edges among the edges of @xmath241 , such that @xmath245 , yielding the desired conclusion . * case 1 .
* @xmath241 has @xmath246 leaves . since @xmath247
, it follows that @xmath241 is a simple cycle .
relabel the vertices of the cycle so that @xmath241 is now a graph on @xmath248 $ ] .
then @xmath249 since @xmath250 only has edges positively incident to it .
regardless of which vertex of @xmath241 is @xmath251 , there is a playable route @xmath186 starting at vertex @xmath252 and ending at @xmath253 such that @xmath245 .
* @xmath241 has @xmath250 leaf .
then @xmath241 is a union of a simple cycle @xmath16 and a simple path @xmath254 . relabel the vertices of @xmath241 so that it is a graph on the vertex set @xmath248 $ ] .
let @xmath255 be the leftmost vertex of the cycle @xmath16 of @xmath241 and let @xmath256 be the vertex in common to @xmath16 and @xmath254 .
let @xmath242 be the unique leaf .
if @xmath257 , then @xmath258 .
thus , at least one of the edges of @xmath16 incident to @xmath256 are incident with an opposite sign to @xmath256 than the edge of @xmath254 incident to @xmath256 .
therefore , the edges on the path from @xmath255 to @xmath256 through the edge that is incident to @xmath256 in @xmath16 with the opposite sign to that of the edge of @xmath254 , and then the edges of path @xmath254 form a playable route @xmath186 such that @xmath245 .
if @xmath259 then we consider two possibilities , depending on whether @xmath260 or @xmath261 . if @xmath260 then @xmath262 and @xmath263 .
if @xmath264 , then the edges of @xmath254 ( from @xmath250 to @xmath255 ) and the edges on the path from @xmath255 to @xmath253 through the edge that is incident to @xmath253 in @xmath16 with the sign of @xmath265 in @xmath122 make up a playable route @xmath186 with @xmath245 . if @xmath266 however , then , either the edges on the path from @xmath252 to @xmath253 along @xmath254 make up a playable route @xmath186 with @xmath245 , or the the edges of @xmath254 ( from @xmath250 to @xmath255 ) and the edges of @xmath16 and then the edges on the path from @xmath255 to @xmath253 make up a playable route @xmath186 with @xmath245 .
if @xmath259 and @xmath261 then either @xmath267 or @xmath268 .
if @xmath267 then the edges on the path @xmath254 from @xmath269 to @xmath270 make up a playable route @xmath186 with @xmath245 . on the other hand if @xmath268 then @xmath249 and if the edge of @xmath254 is incident to @xmath255 with the same sign as that of the sign of @xmath265 in @xmath122 , than the edges of @xmath254 make up a playable route @xmath186 with @xmath245 .
if , however , that sign is different , then it must be that @xmath271v=1 $ ] in which case all edges of @xmath241 ( suitably ordered ) make up a playable route @xmath186 with @xmath245 .
* case 3 . *
@xmath241 has @xmath272 leaves .
then @xmath241 could be a path , or a union of a simple cycle @xmath16 and two disjoint paths @xmath273 attached to @xmath16 at vertices @xmath274 , or a union of a cycle @xmath16 and a tree @xmath275 with two leaves attached to @xmath16 at @xmath276 . as in cases @xmath250 and @xmath272 , in each case we can identify a playable route by inspection .
we omit the details here .
proposition [ conn ] yields a characterization of the vertices of @xmath12 for a connected @xmath277 .
[ phi ] let @xmath229 .
the map @xmath203 defined by ( [ phieq ] ) is a one - to - one correspondence between playable routes and playable pairs of @xmath5 and the vertices in @xmath210 .
the proof is almost identical to that of proposition [ conn ] .
the only difference is that the graph @xmath241 defined in the proof of proposition [ conn ] could have two connected components containing edges .
the case of @xmath241 with one connected component containing edges is the same as in the proof of proposition [ conn ] .
let the two connected components of @xmath241 containing edges be @xmath278 and @xmath279 .
then , @xmath278 and @xmath279 each contributes exactly one coordinate with a nonzero coefficient , and thus each of them is a union of a simple cycle ( since @xmath229 ) and a possibly empty simple path .
the edges of @xmath278 and @xmath279 , in a suitable order , constitute playable pairs .
[ vertices ] for any graph @xmath5 the set of vertices @xmath280 is the image of playable routes and pairs of @xmath5 under the map @xmath203 defined by ( [ phieq ] ) .
let @xmath281 , and let @xmath282 be a central triangulation of @xmath283 .
for each @xmath284 we define @xmath285 , where the vertex set of @xmath286 is @xmath287 , @xmath288 and @xmath289 .
then , @xmath290 thus , any @xmath291 belongs to some @xmath292 .
therefore , @xmath293 , for @xmath289 , @xmath288 . by proposition [ phi ] , there is a playable route @xmath186 or pair @xmath194 in @xmath294 , such that @xmath207 or @xmath208 .
but all playable routes and pairs of @xmath294 are also playable routes and pairs of @xmath5 . propositions [ conn ] , [ phi ] and [ vertices ] imply proposition [ playable_route ] .
this section is devoted to proving the reduction lemma ( lemma [ reduction_lemma ] ) . as we shall see in section [ sec : forest ] , the reduction lemma is the secret force " that makes everything fall into its place for coned root polytopes .
we start by characterizing the root polytopes which are simplices , then in lemma [ equivalent ] we prove that equations ( [ eq1 ] ) and ( [ eq2 ] ) are equivalent definitions for the root polytope @xmath12 , and finally we prove the cone reduction lemma ( lemma [ cone_reduction_lemma ] ) , which , together with lemma [ equivalent ] implies the reduction lemma . [ simplex ] for a graph @xmath5 on the vertex set @xmath114 $ ] with @xmath217 edges , the polytope @xmath12 as defined by ( [ eq1 ] ) is a simplex if and only if @xmath5 is alternating and @xmath228 .
it follows from equation ( [ eq1 ] ) that for a minimal graph @xmath5 the polytope @xmath12 is a simplex if and only if the vectors corresponding to the edges of @xmath5 are linearly independent and @xmath295 .
the vectors corresponding to the edges of @xmath5 are linearly independent if and only if @xmath228 . by proposition [ playable_route ] ,
@xmath295 if and only if @xmath5 contains no edges incident to a vertex @xmath296 with opposite signs , i.e. @xmath5 is alternating .
[ equivalent ] for any graph @xmath5 on the vertex set @xmath114 $ ] , @xmath297 for a graph @xmath241 on the vertex set @xmath114 $ ] , let @xmath298 .
then , @xmath299 . let @xmath300 be a @xmath217-dimensional polytope for some @xmath301 and consider any central triangulation of it : @xmath302 , where @xmath303 is a set of @xmath217-dimensional simplices with disjoint interiors , @xmath304 , @xmath305 .
since @xmath302 is a central triangulation , it follows that @xmath306 , for @xmath305 , and @xmath307 . since @xmath308 , @xmath305 ,
is a @xmath217-dimensional simplex , it follows that @xmath309 and has @xmath217 edges .
furthermore , @xmath305 is alternating , as otherwise there are edges @xmath310 incident to @xmath253 with opposite signs , and while @xmath311 , @xmath312 , contradicting that @xmath313 is a central triangulation of @xmath300 .
thus , @xmath314 , and @xmath315 .
it is clear that @xmath316 @xmath305 . since if @xmath317 is in the facet of @xmath318 opposite the origin , then @xmath319 and for any point @xmath320 , @xmath321 it follows that @xmath322
thus , @xmath323 .
finally , @xmath324 as desired . [ cone_reduction_lemma ] * ( cone reduction lemma ) * given a graph @xmath220 with @xmath217 edges , let @xmath133 be the graphs described by any one of the equations ( [ graphs1])-([graphs6 ] ) .
then @xmath221 , @xmath325 where all cones @xmath326 are @xmath217-dimensional and @xmath327 the proof of lemma [ cone_reduction_lemma ] is the same as that of the cone reduction lemma in the type @xmath174 case ; see ( * ? ? ?
* lemma 7 ) .
_ proof of the reduction lemma ( lemma [ reduction_lemma ] ) .
_ straightforward corollary of lemmas [ equivalent ] and [ cone_reduction_lemma ] .
in this section we use the reduction lemma to establish the link between the volumes of root polytopes and the number of monomials in reduced forms .
in fact we shall see that if we know either of these quantities , we also know the other . [ 2/n ! ] let @xmath220 be a connected graph on the vertex set @xmath114 $ ] with @xmath328 edges , and let @xmath148 be an @xmath147-reduction tree with root labeled @xmath130 .
then , @xmath329 where @xmath330 denotes the number of leaves of @xmath148 labeled by graphs with @xmath328 edges . by the reduction lemma ( lemma [ reduction_lemma ] )
@xmath331 , where @xmath5 runs over the leaves of @xmath148 labeled by graphs with @xmath328 edges .
we now prove that for each @xmath5 with @xmath328 edges labeling a leaf of @xmath148 with root labeled @xmath130 , @xmath332 . since @xmath220 is a connected graph on the vertex set @xmath114 $ ] with @xmath328 edges
, so are all its successors with @xmath328 edges .
if @xmath5 labels a leaf of @xmath148 , then @xmath5 satisfies the conditions of lemma [ simplex ] .
thus , @xmath12 is a simplex .
the volume of @xmath12 can be calculated by calculating the determinant @xmath333 of the matrix @xmath334 whose rows are the vectors @xmath7 , @xmath335 , written in the standard basis . if @xmath336 $ ] is a vertex of degree @xmath250 in @xmath5 , the @xmath337 column contains a single @xmath250 or @xmath338 in the row corresponding to the edge incident to @xmath122 .
let this row be the @xmath339 .
delete the @xmath337 column and @xmath340 row from @xmath334 and delete the edge incident to @xmath122 in @xmath5 obtaining a new graph .
successively identify the leaves in the new graphs and delete the corresponding columns and rows from their matrices until we obtain a graph @xmath16 that is a simple cycle and the corresponding matrix @xmath341 .
the rows of @xmath341 are the vectors @xmath7 , @xmath342 . by laplace expansion , @xmath343 .
since @xmath228 , so is @xmath344 .
thus , @xmath345 .
expand @xmath341 by any of its rows obtaining matrices @xmath346 and @xmath347 .
then we get @xmath348 , since both @xmath346 and @xmath347 are such that their entries are all @xmath349 or @xmath338 , each row ( column ) except one has exactly two nonzero entries , and the remaining one exactly one nonzero entry .
thus , @xmath350 . a general version of proposition [ 2/n ! ] can be proved for any connected @xmath220 using the following lemma .
[ 2^k / d ! ] let @xmath229 be an alternating graph on the vertex set @xmath114 $ ] with @xmath217 edges , with @xmath351 connected components of which @xmath352 contain simple cycles .
then , @xmath353 let @xmath354 be the matrix whose rows are the vectors @xmath182 , @xmath181 , written in the standard basis .
matrix @xmath354 is a @xmath355 matrix .
the rows and columns of @xmath354 can be rearraged so that it has a block form in which the blocks @xmath356 on the diagonal correspond to the connected components of @xmath5 , while all other blocks are @xmath246 . since @xmath229 satisfies the conditions of lemma [ simplex ]
, @xmath12 is a simplex , @xmath357 and @xmath358 can be calculated by dropping some @xmath359 columns of @xmath354 such that the resulting matrix @xmath334 has nonzero determinant .
then , @xmath360 .
drop a column @xmath361 from the block matrix @xmath362 if the block @xmath362 corresponds to a tree on @xmath107 vertices , obtaining matrix @xmath363 with nonzero determinant . then , @xmath364 . if @xmath362 corresponds to a connected component of @xmath130 with @xmath107 vertices and @xmath107 edges , then @xmath365 and @xmath366 . since there are @xmath359 connected components which are trees , if we drop the columns @xmath361 from @xmath354 for all blocks @xmath362 corresponding to a tree obtaining a matrix @xmath334 , then @xmath367 .
since @xmath334 has a special block form with blocks @xmath363 along diagonal and zeros otherwise , we have that @xmath368 . [ 2^k ]
let @xmath220 be a graph on the vertex set @xmath114 $ ] with @xmath217 edges , with @xmath351 connected components of which @xmath352 contain cycles .
let @xmath148 be an @xmath147-reduction tree with root labeled @xmath130 .
then , @xmath369 where @xmath330 denotes the number of leaves of @xmath148 labeled by graphs with @xmath217 edges .
the proof of proposition [ 2^k ] proceeds analogously to proposition [ 2/n ! ] , in view of lemma [ 2^k / d ! ] .
let @xmath370 and let @xmath371 $ ] be the monomial corresponding to it .
then for any reduced form @xmath109 of @xmath371 $ ] , the value of @xmath372 is independent of the order of reductions performed .
note that @xmath373 as defined in proposition [ 2^k ] .
since @xmath374 is only dependent on @xmath130 , the value of @xmath372 is independent of the particular reductions performed . with analogous methods
the following proposition about reduced forms in @xmath36 can also be proved .
[ ce ] let @xmath370 and let @xmath371=m^{\mathcal{b}^c}[g_0]$ ] be the monomial corresponding to it .
then for any reduced form @xmath375 of @xmath371 $ ] in @xmath36 , the value of @xmath376 is independent of the order of reductions performed .
in this section we turn our attention to the noncommutative algebra @xmath31 .
we consider reduced forms of monomials in @xmath31 and the reduction rules correspond to the relations @xmath80 of @xmath31 : \(5 ) @xmath377 , for @xmath60 , ( @xmath61 ) @xmath378 , for @xmath60 , \(6 ) @xmath379 , for @xmath60 , ( @xmath64 ) @xmath380 , for @xmath60 , \(7 ) @xmath381 , for @xmath60 , ( @xmath67 ) @xmath382 , for @xmath60 , \(8 ) @xmath383 , for @xmath60 , ( @xmath70 ) @xmath384 , for @xmath60 , \(9 ) @xmath385 , for @xmath10 ( @xmath73 ) @xmath386 , for @xmath10 as observed in proposition [ ce ] , in the commutative counterpart of @xmath31 , @xmath36 , the number of monomials in a reduced form of @xmath78 is the same , regardless of the order of the reductions performed . in this section
we develop the tools necessary for proving the uniqueness of the reduced form in @xmath31 for @xmath78 and other monomials .
the key concept is that of a good " graph , which property is preserved under the reductions . as in the commutative case before , we can to phrase the reduction process in terms of graphs .
let @xmath387 be a monomial in variables @xmath388 $ ] , where @xmath158 for @xmath10 , @xmath159 for @xmath160 , @xmath161 and @xmath162 .
we can think of @xmath107 as a graph @xmath5 on the vertex set @xmath114 $ ] with @xmath256 edges labeled @xmath389 , such that the edge labeled @xmath255 is @xmath197 .
let @xmath390 $ ] denote the edge - labeled graph just described .
let @xmath391 denote an edge @xmath392 labeled @xmath393 .
recall that in our edge notation @xmath120 , i.e. , vertex - label @xmath252 might be smaller or greater than @xmath253 .
we can reverse the process and obtain a monomial from an edge labeled graph @xmath5 .
namely , if @xmath5 is edge - labeled with labels @xmath389 , we can also associate to it the noncommutative monomial @xmath394=\prod_{a=1}^p \w(i_a , j_a,\e_a)$ ] , where @xmath395\}$ ] . in terms of graphs
the partial commutativity of @xmath31 , as described by relations ( 2)-(4 ) , means that if @xmath5 contains two edges @xmath396 and @xmath397 with @xmath398 distinct , then we can replace these edges by @xmath399 and @xmath400 , and vice versa . for illustrative purposes we write out
the graph reduction for relation ( 5 ) of @xmath31 .
if there are two edges @xmath401 and @xmath402 in @xmath130 , @xmath132 , then we replace @xmath130 with two graphs @xmath403 on the vertex set @xmath114 $ ] and edge sets @xmath404 relations @xmath405 of @xmath31 can be translated into graph language analogously .
we say that @xmath130 reduces to @xmath406 and @xmath407 under reductions @xmath80 .
while in the commutative case reductions on @xmath129 $ ] could result in crossing graphs , we prove that in @xmath31 all reductions preserve the noncrossing nature of graphs , provided that we started with a suitable noncrossing graph @xmath5 .
a graph @xmath5 is * noncrossing * if there are no vertices @xmath408 such that @xmath409 and @xmath410 are edges of @xmath5 .
we also show that under reasonable circumstances , if in @xmath36 a reduction could be applied to edges @xmath411 and @xmath412 , then after suitably many allowed commutations in @xmath31 it is possible to perform a reduction on @xmath411 and @xmath412 in @xmath31 .
we now define two central notions of the noncommutative case , that of a well - structured graph and that of a well - labeled graph .
a graph @xmath241 on the vertex set @xmath114 $ ] is * well - structured * if it satisfies the following conditions : @xmath413 @xmath241 is noncrossing .
@xmath414 for any two edges @xmath415 , @xmath416 , it must be that @xmath417 and @xmath418 .
@xmath419 for any two edges @xmath420 , @xmath421 , it must be that @xmath422 .
@xmath423 there are no edges @xmath424 with @xmath425 .
@xmath426 there are no edges @xmath427 with @xmath428 .
@xmath429 graph @xmath241 is connected , contains exactly one loop , and contains no nonloop cycles .
condition @xmath429 implies that any well - structured graph on the vertex set @xmath114 $ ] contains @xmath328 edges .
a graph @xmath241 on the vertex set @xmath114 $ ] and @xmath256 edges labeled @xmath389 is * well - labeled * if it satisfies the following conditions : @xmath413 if edges @xmath396 and @xmath430 are in @xmath241 , @xmath132 , @xmath431 , then @xmath432 .
@xmath414 if edges @xmath396 and @xmath433 in @xmath241 are such that @xmath434 , @xmath431 , then @xmath435 .
@xmath419 if edges @xmath396 and @xmath436 in @xmath241 are such that @xmath437 , @xmath431 , then @xmath435 .
@xmath423 if edges @xmath438 and @xmath439 in @xmath241 are such that @xmath440 , then @xmath432 .
@xmath426 if edges @xmath441 and @xmath439 in @xmath241 are such that @xmath440 , then @xmath435 .
@xmath429 if edges @xmath438 and @xmath442 in @xmath241 are such that @xmath440 , then @xmath435 .
@xmath443 if edges @xmath441 and @xmath442 in @xmath241 are such that @xmath440 , then @xmath432 . note that no graph @xmath241 with a nonloop cycle can be well - labeled
. however , every well - structured graph can be well - labeled .
we call graphs that are both well - structured and well - labeled * good * graphs . a * @xmath444-reduction tree * @xmath445 is defined analogously to an @xmath147-reduction tree , except we use the noncommutative reductions to describe the children
. see figure [ fig : eps ] for an example .
a graph @xmath446 is called a * @xmath444-successor * of @xmath5 if it is obtained by a series of reductions from @xmath5 .
-reduction tree with root corresponding to the monomial @xmath447 .
note that in order to perform a reduction on this monomial we commute variables @xmath448 and @xmath449 . in the @xmath444-reduction tree
we only record the reductions , not the commutations .
summing the monomials corresponding to the graphs labeling the leaves of the reduction tree we obtain a reduced form @xmath76 of @xmath447 , @xmath450 . , scaledwidth=120.0% ]
[ huh ] if the root of a @xmath444-reduction tree is labeled by a good graph , then all nodes of it are also labeled by good graphs .
the proof of lemma [ huh ] is an analysis of the local changes that happen during the noncommutative reduction process .
an analogous lemma for type @xmath174 is proved in ( * ? ? ?
* lemma 12 ) . a reduction applied to a noncrossing graph
@xmath5 is * noncrossing * if the graphs resulting from the reduction are also noncrossing .
the following is then an immediate corollary of lemma [ huh ] .
if @xmath5 is a good graph , then all reductions that can be applied to @xmath5 and its @xmath444-successors are noncrossing .
let @xmath451 be edges of the graph @xmath241 such that in the commutative algebra @xmath36 a reduction could be performed on @xmath411 and @xmath412 as well as on @xmath411 and @xmath452 .
suppose that @xmath453
. then we say , in the noncommutative case @xmath31 , that performing reduction on edges @xmath411 and @xmath412 is a * priority * over performing reduction on edges @xmath411 and @xmath452 .
we give a few concrete examples of this priority below .
performing reduction ( 6 ) on edges @xmath454 , @xmath132 , is a * priority * over performing reduction ( 9 ) on edges @xmath455 .
performing reduction ( 9 ) on edges @xmath455 , is a * priority * over performing reduction ( 5 ) on edges @xmath456 , @xmath132 . performing reduction ( 9 ) on edges @xmath455 ,
is a * priority * over performing reduction ( 9 ) on edges @xmath457 , @xmath458 . performing reduction ( 9 ) on edges
@xmath455 is a * priority * over performing reduction ( 8) on edges @xmath459 , @xmath425 .
[ cross ] let @xmath5 be a good graph .
let @xmath411 and @xmath412 be edges of @xmath5 such that one of the reductions @xmath80 could be applied to them in the commutative case , and such that the reduction would be noncrossing .
then after finitely many applications of allowed commutations in @xmath31 we can perform a reduction on edges @xmath411 and @xmath412 , provided there is no edge @xmath452 in the graph such that reducing @xmath411 and @xmath452 or @xmath412 and @xmath452 is a priority over reducing @xmath411 and @xmath412 .
the proof of lemma [ cross ] proceeds by inspection .
an analogous lemma for type @xmath174 is proved in ( * ? ? ?
* lemma 14 ) .
in this section we prove conjecture 1 , construct a triangulation of @xmath39 and compute its volume . in order to do this we study alternating well - structured graphs .
recall that an alternating well - structured graph @xmath460 is the union of a noncrossing alternating tree @xmath275 on the vertex set @xmath114 $ ] and a loop , that is , @xmath461 , e(t ) \cup \{(k , k , + ) \}),$ ] for some @xmath165 $ ] for which @xmath460 is alternating .
a well - labeling that will play a special role in this section is the lexicographic labeling , defined below .
the * lexicographic order * on the edges of a graph @xmath5 with @xmath107 edges is as follows .
edge @xmath462 is less than edge @xmath463 , @xmath118 , in the lexicographic order if @xmath464 , or @xmath465 and @xmath466 .
furthermore , any positive edges is less than any negative edges in the lexicographic ordering .
graph @xmath5 is said to have * lexicographic edge - labels * if its edges are labeled by integers @xmath467 such that if edge @xmath468 is less than edge @xmath469 in lexicographic order , then the label of @xmath468 is less than the label of @xmath469 in the usual order on the integers . given any graph @xmath5
there is a unique edge - labeling of it which is lexicographic .
note that our definition of lexicographic is closely related to the conventional definition , but it is not the same .
for an example of lexicographic edge - labels , see the graphs labeling the leaves of the @xmath444-reduction tree in figure [ fig : eps ] .
[ labeling ] if @xmath460 is an alternating good graph , then upon some number of commutations performed on @xmath460 , it is possible to obtain @xmath470 with lexicographic edge - labels .
if edges @xmath411 and @xmath412 of @xmath460 share a vertex and if @xmath411 is less than @xmath412 in the lexicographic order , then the label of @xmath411 is less than the label of @xmath412 in the usual order on integers by the definition of well - labeling on alternating well - structured graphs . since commutation swaps the labels of two vertex disjoint edges labeled by consecutive integers in a graph
, these swaps do not affect the relative order of the labels on edges sharing vertices .
continue these swaps until the lexicographic order is obtained .
[ non ] by choosing the series of reductions suitably , the set of leaves of a @xmath444-reduction tree with root labeled by @xmath471 $ ] can be all alternating well - structured graphs @xmath460 on the vertex set @xmath114 $ ] with lexicographic edge - labels .
the number of such graphs is @xmath472 . by the correspondence between the leaves of a @xmath444-reduction tree and simplices in a subdivision of @xmath473)$ ] obtained from the reduction lemma ( lemma [ reduction_lemma ] )
, it follows that no graph with edge labels disregarded appears more than once among the leaves of a @xmath444-reduction tree .
thus , it suffices to prove that any alternating well - structured graph @xmath460 on the vertex set @xmath114 $ ] appears among the leaves of a @xmath444-reduction tree and that all these graphs have lexicographic edge - labels .
first perform all possible reductions on the graph and its successors not involving the loop @xmath474 . according to (
* theorem 18 ) the outcome is all noncrossing alternating spanning trees with lexicographic ordering on the vertex set @xmath114 $ ] and edge @xmath475 present .
let @xmath476 be the trees just described and @xmath477 , e(t_i ) \cup \{(n , n , + ) \})$ ] , @xmath478 $ ] .
it is clear from the definition of reductions that the only edges involved in further reducing @xmath479 , @xmath478 $ ] are the ones incident to vertex @xmath328 .
thus , in order to understand what the leaves of a reduction tree with root labeled @xmath480 , @xmath478 $ ] , are , it suffices to understand the leaves of a reduction tree with root labeled @xmath481 , \{(k+1 , k+1 , + ) , ( i , k+1 , - ) \mid i \in [ k])$ ] , @xmath482 .
it follows by inspection that the leaves of a reduction tree with root labeled @xmath5 are of the form @xmath483 , e(g_1 ) \cup e(g_2))$ ] , where @xmath406 is a connected well - structured graph with only positive edges ( having exactly one loop ) on @xmath484 $ ] , @xmath485 $ ] , of which there are @xmath486 and @xmath487 , \ { ( i , k+1 ) \mid i \in \{l , l+1 , \ldots , k\})$ ] .
it follows that all alternating well - structured graphs @xmath460 are among the leaves of the particular @xmath444-reduction tree described . since all these graphs are well - labeled , having started with a good graph , by lemma [ labeling ] we can assume they have lexicographic edge - labels . from the description of the reductions above it is clear that the number of leaves of this particular reduction tree is @xmath488 where @xmath489 is the number of noncrossing alternating trees on the vertex set @xmath114 $ ] with exactly @xmath242 edges incident to @xmath328 , and @xmath490 is the number of leaves of the reduction tree with root labeled @xmath491 , \{(k+1 , k+1 , + ) , ( i , k+1 , - ) \mid i \in [ k])$ ] as above .
the formula for @xmath492 follows by a simple bijection between noncrossing alternating trees on the vertex set @xmath114 $ ] with exactly @xmath242 edges incident to @xmath328 and ordered trees on the vertex set @xmath114 $ ] with the root having degree @xmath242 . by equations ( 6.21 ) , ( 6.22 ) , ( 6.28 ) and the bijection presented in appendix e.1 . in @xcite , ordered trees on the vertex set @xmath114 $ ] with the root having degree @xmath242 are enumerated by @xmath492 . since @xmath493 the proof is complete .
[ ajaj ] the set of leaves of a @xmath444-reduction tree with root labeled by @xmath471 $ ] is , up to commutations , the set of all alternating well - structured graphs on the vertex set @xmath114 $ ] with lexicographic edge - labels . by proposition [ non ]
there exists a @xmath444-reduction tree which satisfies the conditions above . by proposition [ 2^k ]
the number of graphs with @xmath328 of edges among the leaves of an @xmath147-reduction tree is independent of the particular @xmath147-reduction tree , and , thus , the same is true for a @xmath444-reduction tree . since all graphs labeling the leaves of a @xmath444-reduction tree with root labeled by @xmath471 $ ] have to be good by lemma [ huh ] , and no graph , with edge - labels disregarded , can appear twice among the leaves of a @xmath444-reduction tree , imply , together with lemma [ labeling ] , the statement of theorem [ ajaj ] . as corollaries of theorem [ ajaj ] we obtain the characterziation of reduced forms of the noncommutative monomial @xmath78 , a triangulation of @xmath39 and a way to compute its volume .
[ main ] if the polynomial @xmath494 is a reduced form of @xmath78 , then up to commutations @xmath495,\ ] ] where the sum runs over all alternating well - structured graphs @xmath460 on the vertex set @xmath114 $ ] with lexicographic edge - labels .
[ main2 ] if the polynomial @xmath496 is a reduced form of @xmath78 in @xmath36 , then @xmath497 proposition [ ce ] and theorem [ main ] imply @xmath498 [ main3 ] let @xmath499 be all alternating well - structured graphs on the vertex set @xmath114 $ ] . then @xmath500 are @xmath328-dimensional simplices forming a triangulation of @xmath39 .
furthermore , @xmath501 the reduction lemma implies the first claim , and proposition [ 2/n ! ] implies @xmath502 the value of the volume of @xmath39 has previously been observed by fong @xcite .
in this section we find analogues of theorems [ ajaj ] , [ main ] , [ main2 ] and [ main3 ] for any well - structured graph @xmath460 on the vertex set @xmath114 $ ] .
[ gen : non ] let @xmath460 be a well - structured graph on the vertex set @xmath114 $ ] . by choosing the series of reductions suitably , the set of leaves of a @xmath444-reduction tree with root labeled by @xmath460 can be all alternating well - structured spanning graphs @xmath5 of @xmath503 on the vertex set @xmath114 $ ] with lexicographic edge - labels .
all graphs labeling the leaves of a @xmath444-reduction tree must be alternating well - structured spanning graphs @xmath5 of @xmath503 . also , it is possible to obtain any well - structured graph @xmath460 on the vertex set @xmath114 $ ] as an @xmath444-successor of @xmath184 .
furthermore , if @xmath460 and @xmath504 are two @xmath444-successor of @xmath184 in the same @xmath444-reduction tree , and neither is the @xmath444-successor of the other , then the intersection of @xmath503 and @xmath505 does not contain a well - structured graph @xmath5 , as the existence of such a graph would imply that @xmath506 and @xmath507 have a common interior point , contrary to the reduction lemma .
since the set of leaves of a @xmath444-reduction tree with root labeled by @xmath184 is , up to commutations , the set of all alternating well - structured graphs on the vertex set @xmath114 $ ] with lexicographic edge - labels according to theorem [ ajaj ] , proposition [ gen : non ] follows .
[ gen : ajaj ] let @xmath460 be a well - structured graph on the vertex set @xmath114 $ ] .
the set of leaves of a @xmath444-reduction tree with root labeled @xmath460 is , up commutations , the set of all alternating well - structured spanning graphs @xmath5 of @xmath503 on the vertex set @xmath114 $ ] with lexicographic edge - labels .
the proof is analogous to that of theorem [ ajaj ] using proposition [ gen : non ] instead of proposition [ non ] .
as corollaries of theorem [ gen : ajaj ] we obtain the characterziation of reduced forms of the noncommutative monomial @xmath508 $ ] , a triangulation of @xmath506 and a way to compute its volume , for a well - structured graph @xmath460 on the vertex set @xmath114 $ ] . *
( noncommutative part . ) * if the polynomial @xmath509 is a reduced form of @xmath508 $ ] for a well - structured graph @xmath460 on the vertex set @xmath114 $ ] , then up to commutations @xmath510,\ ] ] where the sum runs over all alternating well - structured spanning graphs @xmath5 of @xmath503 on the vertex set @xmath114 $ ] with lexicographic edge - labels . *
( commutative part . ) * if the polynomial @xmath511 is a reduced form of @xmath512 $ ] for a well - structured graph @xmath460 on the vertex set @xmath114 $ ] , then @xmath513 where @xmath514 is the number of alternating well - structured spanning graphs @xmath5 of @xmath503 . *
( triangulation and volume . )
* let @xmath499 be all alternating well - structured spanning graphs of @xmath503 for a well - structured graph @xmath460 on the vertex set @xmath114 $ ]
. then @xmath500 are @xmath328-dimensional simplices forming a triangulation of @xmath506 . furthermore , @xmath515 where @xmath514 is the number of alternating well - structured spanning graphs @xmath5 of @xmath503 .
in this section we define the noncommutative algebra @xmath40 , which specializes to @xmath31 when we set @xmath99 . we prove analogs of the results presented so far for this more general algebra .
we also provide a way for calculating ehrhart polynomials for certain type @xmath15 root polytopes .
let the * @xmath91-bracket algebra * @xmath40 * of type @xmath15 * be an associative algebra over @xmath90 $ ] , where @xmath91 is a variable ( and a central element ) , with a set of generators @xmath45 subject to the following relations : \(1 ) @xmath46 @xmath47 , for @xmath48 , \(2 ) @xmath49 ( @xmath50 ) @xmath51 , @xmath52 , @xmath53 , for @xmath54 distinct .
( @xmath55 ) @xmath56 , @xmath57 , for all @xmath58 \(5 ) @xmath92 , for @xmath60 , ( @xmath61 ) @xmath516 , for @xmath60 , \(6 ) @xmath93 , for @xmath60 , ( @xmath64 ) @xmath517 , for @xmath60 , \(7 ) @xmath94 , for @xmath60 , ( @xmath67 ) @xmath518 , for @xmath60 , \(8 ) @xmath95 , for @xmath60 , ( @xmath70 ) @xmath519 , for @xmath60 , \(9 ) @xmath520 , for @xmath521 ( @xmath73 ) @xmath522 , for @xmath523 kirillov @xcite made conjecture 1 not just for @xmath31 , but for a more general @xmath91-bracket algebra of type @xmath15 , which is almost identical to @xmath40 ; it differs in a term in relations ( 9 ) and @xmath524 .
we prove the analogue of conjecture 1 for @xmath40 .
notice that the commutativization of @xmath40 yields the relations of @xmath18 , except for relations ( 9 ) and @xmath524 of @xmath40 , which can be obtained by combining relations ( 6 ) and ( 7 ) of @xmath18 .
since the reduction lemma ( lemma [ reduction_lemma ] ) hold for @xmath18 , so does it for @xmath40 , keeping in mind that relations ( 9 ) and @xmath524 of @xmath40 are obtained by combining relations ( 6 ) and ( 7 ) of @xmath18 . as a result
, we can think of relations @xmath80 of @xmath40 as operations subdividing root polytopes into smaller polytopes and keeping track of their lower dimensional intersections .
a * @xmath525-reduction tree * is analogous to an @xmath147-reduction tree , just that the children of the nodes are obtained by the relations @xmath80 of @xmath40 , and now some nodes have five , and some nodes have three children
. see figure [ fig : beta ] for an example .
if @xmath526 is a @xmath525-reduction tree with root labeled @xmath5 and leaves labeled by graphs @xmath527 , then @xmath528 by an analogue of the reduction lemma .
-reduction tree with root corresponding to the monomial @xmath529 .
summing the monomials corresponding to the graphs labeling the leaves of the reduction tree multiplied by suitable powers of @xmath91 , we obtain a reduced form @xmath530 of @xmath529 , @xmath531 .
, scaledwidth=120.0% ] in order to prove an analogue of proposition [ non ] for the algebra @xmath40 , we need a definition more general than well - structured .
thus we now define weakly - well - structured graphs .
a graph @xmath241 on the vertex set @xmath114 $ ] and @xmath532 edges is * weakly - well - structured * if it satisfies the following conditions : @xmath413 @xmath241 is noncrossing .
@xmath414 for any two edges @xmath415 , @xmath416 , it must be that @xmath417 and @xmath418 .
@xmath419 for any two edges @xmath420 , @xmath421 , it must be that @xmath422 .
@xmath423 there are no edges @xmath424 with @xmath425 .
@xmath426 there are no edges @xmath427 with @xmath428 .
@xmath429 graph @xmath241 contains at most one loop , and @xmath241 contains no nonloop cycles . @xmath443 graph @xmath241 contains a positive edge incident to vertex @xmath250 .
note that well - structured graphs are also weakly - well - structured .
[ prop : p^l ] by choosing the set of reductions suitably , the set of leaves of a @xmath525-reduction tree @xmath526 with root labeled by @xmath179 , \{(n , n , + ) , ( i , i+1 , - ) \mid
i \in [ n-1]\})$ ] can be the set of all alternating weakly - well - structured subgraphs @xmath5 of @xmath533 with lexicographic edge - labels .
the proof of proposition [ prop : p^l ] proceeds analogously as that of proposition [ non ] , using equation ( [ red ] ) , instead of the original statement of the reduction lemma , and using the full statement of ( * ? ? ?
* theorem 18 ) which says that the leaves of a reduction tree with root labeled by @xmath534 , \{(i , i+1 , - ) \mid i \in [ n-1]\})$ ] are all noncrossing alternating forests with negative edges on the vertex set @xmath114 $ ] containing edge @xmath475 with lexicographic edge - labels .
[ p^l ] the set of leaves of a @xmath525-reduction tree @xmath526 with root labeled @xmath184 is , up commutations , the set of all alternating weakly - well - structured subgraphs @xmath5 of @xmath533 with lexicographic edge - labels . proposition [ prop : p^l ] proves the existence of one such @xmath525-reduction tree .
an analogue of lemma [ huh ] states that if the root of a @xmath525-reduction tree is a weakly - well - structured well - labeled graph , then so are all its nodes . together with equation ( [ red ] ) these imply theorem [ p^l ] . as corollaries of theorem [ p^l ] we obtain the characterziation of reduced forms of the noncommutative monomial @xmath78 in @xmath40 as well as a canonical triangulation of @xmath535 and an expression for its ehrhart polynomial .
[ beta - main ] if the polynomial @xmath536 is a reduced form of @xmath78 in @xmath40 , then @xmath537,\ ] ] where the sum runs over all alternating weakly - well - structured graphs @xmath5 on the vertex set @xmath114 $ ] with lexicographic edge - labels .
[ canonical ] * ( canonical triangulation . )
* let @xmath538 be all the alternating well - structured graphs on the vertex set @xmath114 $ ] .
then the root polytopes @xmath539 @xmath540 are @xmath328-dimensional simplices forming a triangulation of @xmath535 .
furthermore , the intersections of the top dimensional simplices @xmath541 are simplices @xmath542 , where @xmath241 runs over all alternating weakly - well - structured graphs on the vertex set @xmath114 $ ] .
given a polytope @xmath543 , the * @xmath544 dilate * of @xmath218 is @xmath545 the * ehrhart polynomial of an integer polytope * @xmath543 is @xmath546 for background on the theory of ehrhart polynomials see @xcite .
[ ehrhart ] * ( ehrhart polynomial . )
* @xmath547 where @xmath548 is the number of alternating weakly - well - structured graphs on the vertex set @xmath114 $ ] with @xmath217 edges one of which is a loop and @xmath549 is the number of alternating weakly - well - structured graphs on the vertex set @xmath114 $ ] with @xmath217 edges and no loops . by theorem [ canonical ] ,
@xmath550 where @xmath551 is the set of all alternating weakly - well - structured graphs on the vertex set @xmath114 $ ] with no loops and @xmath552 is the set of all alternating weakly - well - structured graphs on the vertex set @xmath114 $ ] with a loop
. then @xmath553 by ( * ? ?
? * theorem 1.3 ) the ehrhart series of @xmath554 , @xmath555 , @xmath556 , and @xmath557 , @xmath558 , @xmath559 , respectively , are @xmath560 and @xmath561 .
equivalently , @xmath562 , @xmath563 thus , @xmath564 where @xmath565 , @xmath566 . using the ehrhart - macdonald reciprocity ( * ? ? ?
* theorem 4.1 ) @xmath567 theorems [ p^l ] , [ beta - main ] , [ canonical ] and [ ehrhart ] can be generalized to any well - structured graph @xmath5 by adding further technical requirements on the weakly - well - structured graphs that can appear among the leaves of a @xmath525-reduction tree with root labeled by @xmath5 . due to the technical nature of these results
, we omit them here .
in the rest of the paper we study the reduced forms of elements in the type @xmath17 bracket algebra with combinatorial methods fused with noncommutative grbner basis theory . while the connection with subdivisions of type @xmath15 root polytopes is present in this case as well , for brevity we choose to suppress this aspect .
let the * @xmath91-bracket algebra * @xmath568 * of type @xmath17 * be an associative algebra over @xmath90 $ ] , where @xmath91 is a variable ( and a central element ) , with a set of generators @xmath569 subject to the following relations : \(1 ) @xmath46 @xmath47 , for @xmath48 , \(2 ) @xmath49 ( @xmath50 ) @xmath51 , @xmath52 , @xmath53 , for @xmath54 distinct .
( @xmath55 ) @xmath56 , @xmath57 , for all @xmath58 \(5 ) @xmath92 , for @xmath60 , ( @xmath61 ) @xmath516 , for @xmath60 , \(6 ) @xmath93 , for @xmath60 , ( @xmath64 ) @xmath517 , for @xmath60 , \(7 ) @xmath94 , for @xmath60 , ( @xmath67 ) @xmath518 , for @xmath60 , \(8 ) @xmath95 , for @xmath60 , ( @xmath70 ) @xmath519 , for @xmath60 , note that @xmath40 is the quotient of @xmath568 , since @xmath40 has all the above relation and in addition relations ( 9 ) , ( @xmath73 ) ; see section [ 11 ] .
let @xmath570 be a coxeter type element in @xmath568 and let @xmath76 be the polynomial in variables @xmath571 obtained from @xmath572 by successively applying the defining relations @xmath573 in any order until unable to do so , in the algebra @xmath574\langle x_{ij } , y_{ij } \mid 1\leq i < j\leq n \rangle / i,$ ] where @xmath575 is the ( two - sided ) ideal generated by the relations @xmath576 .
we call @xmath76 a * reduced form * of @xmath572 and consider the process of successively applying the defining relations @xmath573 as a reduction process in @xmath574\langle x_{ij } , y_{ij } \mid 1\leq i < j\leq n \rangle / i,$ ] with the * reduction rules * : \(5 ) @xmath577 , for @xmath60 , ( @xmath61 ) @xmath578 , for @xmath60 , \(6 ) @xmath579 , for @xmath60 , ( @xmath64 ) @xmath580 , for @xmath60 , \(7 ) @xmath581 , for @xmath60 , ( @xmath67 ) @xmath582 , for @xmath60 , \(8 ) @xmath583 , for @xmath60 , ( @xmath70 ) @xmath584 , for @xmath60
. the reduced form of any other element of @xmath568 is defined analogously . as in the type @xmath15 case
, the relations of @xmath568 can be interpreted as subdividing type @xmath15 root polytopes and the reduced form of an element as a subdivision , though not a triangulation , of a type @xmath15 polytope .
we pursue a different approach to studying reduced forms here .
we can think of the reduction process in @xmath585 where the generators of the ( two - sided ) ideal @xmath586 are those of @xmath575 and in addition the commutators of @xmath91 with all the other variables @xmath587 , @xmath588 . * conjecture 2 .
( kirillov @xcite ) * _ apart from applying the relations ( 1)-(4 ) , the reduced form @xmath76 of @xmath572 does not depend on the order in which the reductions are performed .
_ note that the above statement does not hold true for any monomial ; some examples illustrating this were already explained in the comments after conjecture 1 in section [ sec : alg ] .
it is straighforward to reformulate the reduction rules ( 5)-(@xmath70 ) in terms of reductions on graphs .
if @xmath589 , then we replace each monomial @xmath107 in the reductions by corresponding graphs @xmath390 $ ] . the analogous procedure for type @xmath15
is explained in detail in section [ reductionsb ] .
we now define a central notion for those signed graphs whose corresponding monomials turn out to have a unique reduced form in @xmath568 .
we reuse the expression good graph " from the type @xmath15 case , though the meaning in type @xmath17 is different .
previously we used good in the type @xmath15 sense ; in the following we use good in the type @xmath17 sense .
a graph @xmath241 on the vertex set @xmath114 $ ] and @xmath242 edges labeled @xmath590 is * good * if it satisfies the following conditions : @xmath413 the negative edges of @xmath241 form a noncrossing graph .
@xmath414 if edges @xmath401 and @xmath430 are in @xmath241 , @xmath132 , @xmath591 , then @xmath432 .
@xmath419 if edges @xmath401 and @xmath433 are in @xmath241 , @xmath132 , @xmath591 , then @xmath435 .
@xmath423 if edges @xmath592 and @xmath433 are in @xmath241 , @xmath132 , @xmath591 , then @xmath432 .
@xmath426 if edges @xmath593 and @xmath594 are in @xmath241 , @xmath132 , then @xmath435 .
@xmath429 if edges @xmath595 and @xmath596 are in @xmath241 , @xmath597 , then @xmath435 .
[ good ] if @xmath241 is a good graph , then reduction rules @xmath598 can not be performed on it .
if we perform any of the reduction rules @xmath599 on @xmath241 , then we obtain a graph @xmath600 , which is also a good graph .
note that there is no way of commuting the labels of good graphs as to obtain an order on the edges which would allow rules @xmath598 to be performed .
that the following properties carry over from @xmath241 to @xmath600 follows from ( * ? ? ?
* lemma 12 ) , noting that only reduction rule @xmath601 creates new negative edges : * the negative edges of @xmath241 form a noncrossing graph . * if edges @xmath401 and @xmath602 are in @xmath241 , @xmath132 , then @xmath432 . * if edges @xmath401 and @xmath594 are in @xmath241 , @xmath132 , then @xmath435 . * if edges @xmath592 and @xmath594 are in @xmath241 , @xmath132 , then @xmath432 .
inspection shows that the following properties carry over from @xmath241 to @xmath600 , keeping in mind that the above properties carry over for negative edges . * if edges @xmath401 and @xmath603 are in @xmath241 , @xmath132 , then @xmath432 . * if edges @xmath592 and @xmath604 are in @xmath241 , @xmath132 , then @xmath432 . * if edges @xmath593 and @xmath594 are in @xmath241 , @xmath132 , then @xmath435 . * if edges @xmath595 and @xmath596 are in @xmath241 , @xmath597 , then @xmath435 .
finally , given that all the above properties carry over from @xmath241 to @xmath600 , it follows that the property * if edges @xmath401 and @xmath604 are in @xmath241 , @xmath132 , then @xmath435
. also carries over .
why are good graphs so good ?
well , if the relations @xmath599 were a noncommutative grbner basis for the ideal they generate in @xmath585 with tips @xmath605 , @xmath606 , @xmath607 , @xmath608 , respectively , then it would follow immediately that the reduced form of monomials corresponding to good graphs are unique by results in noncommutative grbner bases theory . as it turns out
the previous is not the case , however , we can still use grbner bases to prove the uniqueness of the reduced forms of the monomials corresponding to good graphs , which we call * good monomials * , with a little bit more work .
we show how to do this in the next section .
in this section we briefly review some facts about noncommutative grbner bases and use them to show that the reduced forms of good monomials are unique .
we use the terminology and notation of @xcite , but state the results only for our special algebra . for the more general statements , see @xcite . throughout this section
we consider the noncommutative case only .
let @xmath609 with multiplicative basis @xmath610 , the set of noncommutative monomials up to equivalence under the commutativity relations described by @xmath611 .
the * tip * of an element @xmath612 is the largest basis element appearing in its expansion , denoted by tip@xmath613 .
let ctip@xmath613 denote the coefficient of tip@xmath613 in this expansion .
a set of elements @xmath614 is * tip reduced * if for distinct elements @xmath615 , tip@xmath616 does not divide tip@xmath617 .
a well - order @xmath618 on @xmath610 is * admissible * if for @xmath619 : \1 . if @xmath620 then @xmath621 if both @xmath622 and @xmath623 ; \2 . if @xmath620 then @xmath624 if both @xmath625 and @xmath626 ; \3 . if @xmath627 , then @xmath628 and @xmath629 .
let @xmath630 and suppose that there are monomials @xmath631 such that \1 .
tip@xmath632=@xmath633tip@xmath634 .
tip@xmath613 does not divide @xmath633 and tip@xmath634 does not divide @xmath635 then the * overlap relation of @xmath636 and @xmath637 by @xmath633 and @xmath351 * is @xmath638 [ suf ] ( ( * ? ? ?
* theorem 2.3 ) ) a tip reduced generating set of elements @xmath639 of the ideal @xmath640 of @xmath641 is a grbner basis , where the ordering on the monomials is admissible , if for every overlap relation @xmath642 where @xmath643 and the above notation means that dividing @xmath644 by @xmath639 yields a remainder of @xmath246 .
see ( * ? ? ?
* theorem 2.3 ) for the more general formulation of proposition [ suf ] and ( * ? ? ?
* section 2.3.2 ) for the formulation of the division algorithm .
[ grobi ] let @xmath640 be the ideal generated by the elements * @xmath645 , for @xmath60 , * @xmath646 , for @xmath60 , * @xmath647 , for @xmath60 , * @xmath648 , for @xmath60 , in @xmath649 where @xmath650 is the ideal in @xmath641 generated by the elements * @xmath651 , for @xmath60 .
then there is a monomial order in which the above generators of @xmath640 form a grbner basis @xmath639 of @xmath640 in @xmath652 and the tips of the generators are , respectively , * @xmath605 , * @xmath606 , * @xmath607 , * @xmath608 .
let @xmath653 for any @xmath10 , @xmath421 , and let @xmath654 and @xmath655 if @xmath656 is less than @xmath657 lexicographically .
the degree of a monomial is determined by setting the degrees of @xmath587 to be @xmath250 and the degrees of @xmath91 and scalars to be @xmath246 .
a monomial with higher degree is bigger in the order @xmath618 , and the lexicographically bigger monomial of the same degree is greater than the lexicographically smaller one . since in @xmath641
two equal monomials can be written in two different ways due to commutations , we can pick a representative to work with , say the one which is the largest " lexicographically among all possible ways of writing the monomial , to resolve any ambiguities .
the order @xmath618 just defined is admissible , in it the tips of * @xmath645 , for @xmath60 , * @xmath646 , for @xmath60 , * @xmath647 , for @xmath60 , * @xmath648 , for @xmath60 , are * @xmath605 , * @xmath606 , * @xmath607 , * @xmath608 . in particular the generators of @xmath640 are tip reduced .
a calculation of the overlap relations shows that @xmath658 in @xmath659 , where @xmath643 .
proposition [ suf ] then implies proposition [ grobi ] .
[ /y ] the reduced form of a good monomial @xmath107 is unique in @xmath660 .
since the tips of elements of the grbner basis @xmath639 of @xmath640 are exactly the monomials which we replace in the prescribed reduction rules @xmath661 , the reduced form of a good monomial @xmath107 is the remainder @xmath662 upon division by the elements of @xmath639 with the order @xmath618 described in the proof of proposition [ grobi ] .
since we proved that in @xmath663 the basis @xmath639 is a grbner basis f @xmath640 , it follows by ( * ? ? ?
* proposition 2.7 ) that the remainder @xmath662 of the division of @xmath107 by @xmath639 is unique in @xmath663 .
that is , the reduced form of a good monomial @xmath107 is unique in @xmath660 .
we would , however , like to prove uniqueness of the reduced form of a good monomial @xmath107 in @xmath641 .
this is what the next series of statements accomplish .
[ y ] there is a monomial order in which the elements * @xmath651 , for @xmath60 , are a grbner basis of @xmath650 in @xmath641 , and the tip of @xmath651 is @xmath664 .
let @xmath665 for any @xmath10 , @xmath421 , and let @xmath654 and @xmath655 if @xmath656 is less than @xmath657 lexicographically .
the degree of a monomial is determined by setting the degrees of @xmath587 to be @xmath250 and the degrees of @xmath91 and scalars to be @xmath246 .
a monomial with higher degree is bigger in the order @xmath618 , and the lexicographically bigger monomial of the same degree , the variables being read from left to right , is greater than the lexicographically smaller one .
since in @xmath641 two equal monomials can be written in two different ways due to commutations , we can pick a representative to work with , say the one which is the largest " lexicographically among all possible ways of writing the monomial , to resolve any ambiguities .
the order @xmath618 just defined is admissible , the tip of @xmath651 is @xmath664 , and thus the generators of @xmath650 are tip reduced .
since there are no overlap relations at all , by proposition [ suf ] lemma [ y ] follows .
[ div ] if @xmath666 then there is a term of @xmath636 which can be written as @xmath667 for some @xmath60 , where @xmath668 are some monomials in @xmath641 .
lemma [ y ] implies that @xmath669 from here the statement follows .
[ unique ] the reduced form of a good monomial @xmath107 is unique in @xmath641 . by corollary [ /y ]
the reduced form of a good monomial @xmath107 is unique in @xmath670 .
since by corollary [ div ] every @xmath666 contains a term divisible by @xmath664 for some @xmath60 , it follows that the reduced form of a good monomial @xmath107 is unique in @xmath585 since a good monomial can not contain any term divisible by @xmath664 because of property @xmath419 , with @xmath671 .
a special case of theorem [ unique ] is the statement of conjecture 2 , since @xmath572 is a good monomial .
i am grateful to my advisor richard stanley for suggesting this problem and for many helpful suggestions .
i would like to thank alex postnikov for sharing his insight into root polytopes .
i would also like to thank anatol kirillov for his intriguing conjectures . | the type @xmath0 root polytope @xmath1 is the convex hull in @xmath2 of the origin and the points @xmath3 for @xmath4 $ ] . given a graph @xmath5 , with edges labeled positive or negative , associate to each edge @xmath6 of @xmath5 a vector @xmath7 which is @xmath8 if @xmath9 , @xmath10 , is labeled negative and @xmath11 if it is labeled positive .
for such a signed graph @xmath5 , the associated root polytope @xmath12 is the intersection of @xmath1 with the cone generated by the vectors @xmath7 , for edges @xmath6 in @xmath5 .
the reduced forms of a certain monomial @xmath13 $ ] in commuting variables @xmath14 under reductions derived from the relations of a bracket algebra of type @xmath15 , can be interpreted as triangulations of @xmath12 . using these triangulations ,
the volume of @xmath12 can be calculated .
if we allow variables to commute only when all their indices are distinct , then we prove that the reduced form of @xmath13 $ ] , for good " graphs @xmath5 , is unique and yields a canonical triangulation of @xmath12 in which each simplex corresponds to a noncrossing alternating graph in a type @xmath16 sense .
a special case of our results proves a conjecture of a. n. kirillov about the uniqueness of the reduced form of a coxeter type element in the bracket algebra of type @xmath15 .
we also study the bracket algebra of type @xmath17 and show that a family of monomials has unique reduced forms in it .
a special case of our results proves a conjecture of a. n. kirillov about the uniqueness of the reduced form of a coxeter type element in the bracket algebra of type @xmath17 . |
in the last decades , surface effects in solid state physics have been widely investigated from the applied and fundamental point of view .
an appropriate thermodynamic framework has been developed to describe the interplay between elasticity and surface effects in crystalline solids @xcite .
an essential result is that the concept of surface tension can not be applied without any caution to a solid condensed phase .
this description is usually , however , not considered in the soft matter community , which has focused on problems of adhesion , elasto - capillarity of slender bodies , deformation of an elastomer by a liquid drop , etc . @xcite .
as already pointed out in @xcite a correct treatment of elastocapillarity requires to distinguish between two excess interfacial quantities : the surface energy and the surface stress , cf .
[ fig : intro ] . a thermodynamical approach to derive the relation between surface energy and surface stress yields the well - known shuttleworth - relation @xcite : @xmath0 with @xmath1 the elastic strain parallel to the interface . from this relation
it is immediately clear that for incompressible liquids , and _ only _ for such systems , the surface energy and surface stress are equal and are usually called `` surface tension '' .
the aim of this paper is to adapt this framework previously proposed for crystalline solids , to the case of elastomers and gels : we will extend the concept of surface stress to the case of two soft condensed phases in contact .
specifically , we will propose a microscopic model for these macroscopic excess quantities , in the case of soft elastic amorphous solids , based on the density functional theory in the sharp interface approximation @xcite .
we will show that the compressibility of the superficial layer , quantified by the poisson ratio @xmath2 , is the key characteristic of elastomers for elasto - capillary effects .
the central result will be that the solid - liquid surface stress @xmath3 relates to the interfacial energies in the following way : @xmath4 if the surface layer is perfectly incompressible , i.e. @xmath5 , this gives a strict equality between surface energy and surface stress : @xmath6 , as for liquid interfaces .
the other extreme limit is that the stress directions perfectly decouple , i.e. @xmath7 , for which @xmath8 .
schematic of a planar interface between a liquid and solid phase .
the surface free energy associated with this interface is @xmath9 . in the case of a solid - liquid interface
this surface free energy is generally _ not _ equal to the surface stress ( force per unit length ) at the interface @xmath10 indicated by the arrows at the dashed control volume . ] in sect .
[ sect : micmac ] we derive a microscopic model to describe fluid and solid behaviour , which we use to derive @xmath3 in sect .
[ sect : forces ] . the consequence of this result [ eq .
[ eq : upsilon_sl ] ] will be discussed in detail in the discussion of sect .
[ sect : disc ] , where we compare our findings to recent experiments and numerical results .
we wish to define a model to compute the mechanical stresses that arise in the vicinity of a fluid - fluid or fluid - solid interface .
these surface or capillary forces originate from molecular interactions , and predicting their strength thus calls for a description at that scale . before developing this in detail ,
let us first consider the purely macroscopic viewpoint and define a macroscopic ( or thermodynamic ) stress tensor @xmath11 . in the absence of bulk forces ,
the mechanical equilibrium condition is @xmath12 for a planar interface whose normal points along the @xmath13-direction , this implies that the normal stress @xmath14 is constant and is identical in both phases ( assuming homogeneity in the @xmath15-directions ) .
when the interface is curved , the boundary condition at the interface involves a normal stress discontinuity , to account macroscopically for the microscopic interactions in the surface layers . for a fluid - fluid interface , this can be expressed as the difference in thermodynamic pressure , @xmath16 , where @xmath17 is the surface tension and @xmath18 is the curvature @xcite .
this macroscopic result can be contrasted with a purely microscopic approach that explicitly includes the molecular interactions .
microscopically , the interfacial region will be continuous , but one should still recover _ as a result _ that the bulk stress far away from a curved interface is different in the two phases .
for fluid - fluid interfaces , such a microscopic framework has been developed based on density functional theory using the so - called sharp - kink approximation @xcite . as will be further outlined in the next subsection ,
the key assumption of this theory is that the molecular interactions can be split in a long - range attractive part and a short - ranged repulsive interaction @xcite .
the attraction is described by a ( mean - field ) potential @xmath19 , while the repulsion is a contact interaction described by a microscopic stress tensor @xmath20 .
the mechanical equilibrium equation then reads @xmath21 the sharp - kink approximation assumes that the interface between the two phases is perfectly sharp , but that the interaction potential is long - ranged and extends across the interface most notably , the potential varies only near the interface and @xmath22 directly accounts for interfacial stresses . since capillary effects are already included in this description , the boundary condition is now the _ continuity _ of the normal stress @xmath23 across the interface .
this model is very close to the original work by laplace @xcite , who postulated an internal pressure to balance the strong attractive forces between molecules . for incompressible liquids near a straight interface
, we will derive the following connection between macroscopic and microscopic stress : @xmath24 where @xmath25 is a constant .
one should bear in mind that even under atmospheric conditions , the repulsive stress @xmath26 is orders of magnitude larger than @xmath27 , as it balances the strong attractive interaction @xmath19 .
an estimate for the repulsive pressure is obtained by @xmath28 , where @xmath29 is the scale of the repulsion ( a few angstroms ) and thus @xmath30 n / m@xmath31 , as already estimated by laplace @xcite .
the description above is valid for fluids .
below , we will generalize this formulation to account for elastic solids whose elastic properties arise from entropic effects .
in other words , the description focuses on solids such as gels and elastomers , which are polymer solutions in which there is no connected structure based on enthalpic ( chemical ) elastic effects .
the goal is to derive an expression for the surface stress [ eq . ] from the microscopic interactions .
we consider a simple model of liquid based on a pair potential @xmath32 .
the key idea of the density functional theory is to express the grand potential @xmath33 as a functional of the particle density @xmath34 and to perform a functional minimisation for given values of the chemical potential @xmath35 and of the temperature @xmath36 .
the free energy @xmath37 is not known explicitly and needs to be built by integration , starting from the free energy of a known state . for a perturbation theory
, one may choose hard spheres or the same system with only a repulsive potential , and then treat the van der waals attraction as a corrective effect . in the sharp interface approximation ,
one uses a standard van der waals expansion of the free energy around the hard - sphere reference system ( whose pair potential is denoted @xmath38 ) , computed in the local density approximation .
@xmath39&=&\int f_{hs}(\rho(\vec r ) ) d\vec r\\ & + & \frac{1}{2 } \int_0 ^ 1 d\lambda \int d\vec r_1\int d\vec r_2 \rho(\vec r_1)\rho(\vec r_2 ) g_\lambda(\vec r_1,\vec r_2)\varphi(|\vec r_2-\vec r_1|)\nonumber\end{aligned}\ ] ] where @xmath40 is the pair correlation function in a system of same geometry and same volume , for which the interaction is @xmath41 $ ] above the sphere radius .
one can use , for instance , a low density approximation for @xmath40 : @xmath42 where @xmath43 is boltzmann s constant and @xmath36 the temperature .
we now apply the gibbs interface idealization and assume that the density is homogeneous on both sides of the interface .
we furthermore extend the description to several phases and introduce @xmath44 the effective potential associated to the influence of phase @xmath45 on phase @xmath46 , at @xmath47 : @xmath48 note that in general @xmath49 , unless the domains of @xmath45 and @xmath46 have an identical shape .
altogether , we get contribution to the free energy associated to the phase @xmath45 as : @xmath50 d\vec r\ ] ] the long - range attractive interaction between liquid and solid molecules gives rise to a potential @xmath51 i.e. an energy per unit volume .
the short - range repulsive interactions are assumed to lead to an isotropic stress i.e. to a contact pressure @xmath52 @xcite .
minimising the free energy , we obtain the ( microscopic ) equilibrium condition inside the liquid , @xmath53 so that the total potential @xmath54 must be homogeneous @xcite .
this is the same result as obtained in eq .
, as for liquids @xmath55 and the macroscopic stress @xmath11 is homogeneous in the liquid phase . in this work
, we work with planar interfaces only and therefore , for future reference , we define the potential @xmath19 in phase @xmath45 due to the presence of a semi - infinite phase @xmath46 at distance @xmath56 from an interface as : @xmath57 note that , altough in general @xmath58 the equivalence @xmath59 does hold , since in that case the phases @xmath45 and @xmath46 are identically shaped ( both are infinite half - spaces ) . within this formalism
we can readily compute the adhesion energies and hence the surface free energies . as the surface tensions are defined for flat interfaces between two semi - infinite phases ,
these energies can be expressed in terms of integrals over @xmath60 and one finds @xcite : @xmath61 we now extend the density theory description to describe soft solids such as elastomers or gels .
the elastic properties in such solids are of entropic origin and their structure is close to that of a liquid .
therefore , we hypothesize that these solids can be described using the density functional theory in the sharp interface approximation , by replacing the repulsive pressure @xmath62 by an elastic stress tensor @xmath20 . just like the liquid in the previous section , this solid
is assumed to be submitted to the long - range attractive potentials @xmath63 and @xmath64 , which are not included in the stress tensor @xmath65 .
we therefore get a very similar expression as in the liquid case : @xmath66 as anticipated in eq . .
the next step is to find expressions for the components of @xmath11 and @xmath20 , which are generally not isotropic in the elastic case .
the description of the solid differs from the liquid discussed in the previous section by the existence of a base state with respect to which deformations are measured .
one standardly considers , for bulk elasticity , a base state which is stress free .
however , this choice is problematic for capillary effects as these induce a pre - stress , even when no external force is exerted on the solid . to follow experimental constraints
, we will therefore consider that the base state is the state in which the system was prepared .
if a gel is prepared in a solid mold as in @xcite , the base state will therefore be sensitive to the molecular interactions between this mold and the gel .
for the sake of simplicity , we will reduce the discussion to the case were the solid is prepared in void . in order to express
the constitutive relation of the material , we split both stresses into a reference part that is associated with the base state , and the elastic contribution relative to this base state : @xmath67 @xmath68 and @xmath69 are the stresses in a reference state , taken as a semi - infinite solid in contact with vacuum ( i.e. a solid - vacuum interface ) .
for @xmath70 and @xmath71 we apply hooke s law .
we will write the expressions using the thermodynamic stress @xmath72 only , but the same relations hold for the microscopic stress @xmath73 .
hooke s law relates the components of the stress @xmath74 to the elastic strain @xmath75 : @xmath76\ ] ] or equivalently : @xmath77 where @xmath78 with @xmath79 is the displacement vector , and @xmath80 the young s modulus .
we consider plane strain conditions ( @xmath81 , with the @xmath13-direction normal to the interface ) , and therefore find : @xmath82 as explained above , the same expression holds for @xmath73 , so : @xmath83 this relation between the @xmath84 and @xmath85 components of the stress - tensor is at the core of the inequality between @xmath86 and @xmath3 .
note that the stress - tensor for an incompressible surface layer ( @xmath5 ) is isotropic ( @xmath87 ) and we would recover liquid behaviour : @xmath88 schematic of a liquid and solid interface . ( a ) a liquid - vapour interface located at @xmath89 .
the vapour above the liquid is at pressure @xmath90 . due to the low density of the vapour phase ,
this pressure @xmath90 consists entirely of short - ranged repulsive ( kinetic ) stress such that @xmath91 .
since @xmath52 is continuous at the interface this can be used as a boundary condition for eq . .
( b ) a solid - vacuum interface , which serves as the reference system without elastic stresses ( @xmath92 ) .
similarly as in the liquid , this system fixes the value of @xmath93 at @xmath89 , providing the boundary condition for eq . .
] we now consider the equilibrium near a planar interface , as sketched in fig .
[ fig : eqs ] .
first , we note that the macroscopic equilibrium condition for the fluid reads : @xmath94 where @xmath90 is the thermodynamic pressure . in the liquid case
we find that , cf .
: @xmath95 the unknown constant is set by both equilibrium conditions eqs . and at the liquid - vapour ( or equivalently for a liquid - liquid ) interface ,
[ fig : eqs](a ) , with the vapour phase at pressure @xmath90 . at the liquid - vapour interface [ @xmath89 , cf .
fig [ fig : eqs](a ) ] the local potential due to liquid - liquid interactions reads @xmath96 and due to the absence of any solid @xmath97 .
the vapour phase above the liquid is of such low density that only the repulsive contribution applies , and therefore we find in the vapour @xmath91 .
furthermore , microscopic the stress tensor must be continuous at @xmath89 , which implies @xmath91 at @xmath89 .
this allows us to compute the integration constant in eq . , and we obtain @xmath98 everywhere within the liquid .
we now have an explicit relation between the thermodynamic pressure @xmath90 ( or stress @xmath11 ) and the microscopic contributions to the pressure @xmath52 and @xmath19 . in planar geometries
, we recover that in the liquid phase @xmath99 [ fig .
[ fig : intdomains](a ) ] where the @xmath100-term represents the attractive interactions due to the surrounding liquid in the bulk , and the @xmath101-term the missing interactions ( compared to the bulk ) due to the presence of an interface at distance @xmath13 .
we then find , for the liquid , using eq .
: @xmath102 in the bulk of the liquid , where @xmath103 and @xmath97 , @xmath52 tends to : @xmath104 finally , we derive the microscopic equilibrium equation for the elastic solid .
we define the base state as the state where the solid is in contact with vacuum . by definition , the elastic stress
@xmath105 then vanishes .
we follow the same path as in the liquid case , with the distinction that in the solid case the stress tensor is not isotropic . at the solid - vacuum interface [ @xmath89 , cf . fig .
[ fig : eqs](b ) ] we find @xmath106 , because a vacuum represents @xmath107 .
we use this boundary condition to determine the integration constant in eq . , and find that : @xmath108 where we used that @xmath64 anywhere in the solid phase at distance @xmath13 from the interface is given by @xmath109 , cf . fig .
[ fig : intdomains](a ) . furthermore , since @xmath110 and @xmath111 we find : @xmath112 in order to calculate the excess stress in sect .
[ sect : forces ] , a relation between @xmath113 and @xmath114 is required .
posing that the excess stress in the bulk is zero ( which is the definition of an excess quantity ) , we find that in the bulk : @xmath115 this is the solid analogue of for the liquid .
although the bulk value of @xmath116 is now known , we did not yet account for the presence of the vacuum , which is felt at small distances from an interface .
analogously to the liquid in eq .
this will give : @xmath117 note that the term @xmath118 does not change the value of the stress in the bulk , eq .
, as it vanishes at large @xmath13 .
we also anticipate that we do not require an explicit expression for @xmath113 , as it will cancel out when computing the excess stress as described in the next section .
in this section , we derive the surface stress by calculating the total excess stress across the interface . introducing the coordinate @xmath13 normal to the interface the interfacial excess quantity of an extensive measure @xmath119 is defined as @xmath120 dz\;,\ ] ] with @xmath121 the corresponding intensive quantity and
@xmath122 the heaviside step - function .
the total excess @xmath123 is therefore the ( integrated ) difference between the bulk values of @xmath124 and @xmath125 and the true value of @xmath121 that varies continuously in the vicinity of the interface . throughout the analysis
we will use superscripts to denote the bulk value of a given phase .
thermodynamic quantities such as mass density , entropy or energy exhibit an interfacial excess .
for example , the surface free energy @xmath126 can be evaluated from the microscopic interactions using the classical relations ( [ eq : ab]-[eq : bb ] ) . in this section
we calculate the surface stress @xmath127 , which is depicted schematically in fig . [
fig : sketch ] .
the surface stress is defined as the integrated excess stress , and has the dimension of a force per unit length , or equivalently of an energy per unit area . while for liquid - vapor or liquid - liquid interfaces the surface stress @xmath127 is identical to the surface free energy @xmath126 ( following from the virtual work principle ) , this is not the case when one of the phases is elastic . in this section
we provide explicit expressions for @xmath127 , starting from the definition @xmath128 where @xmath129 and @xmath130 are thermodynamic bulk stresses ; note that these thermodynamic stresses implicitly contain all microscopic interactions ( short- and long - ranged ) , so that for the liquid one simply recovers the thermodynamic pressure @xmath131 .
@xmath132 is the total force per unit length acting on the control volume @xmath133 , left of the dashed line in fig .
[ fig : sketch ] , projected parallel to the interface .
this force is exerted by the volume indicated @xmath134 , right of the dashed line in fig .
[ fig : sketch ] . by computing this force @xmath132 explicitly from the microscopic models described in the previous section
, we will obtain explicit expressions for @xmath127 for all combinations of liquid / vapour , liquid / liquid , and solid / liquid interfaces .
schematic of a planar interface between two semi - infinite phases @xmath135 and @xmath136 . near the interface , an excess stress develops and the total ( integrated ) excess stress is defined as the surface stress @xmath127 .
note that , by definition , the excess stress is zero in the bulk ( far away from the interface ) . to evaluate the surface stress associated with this interface @xmath127
we define a control surface perpendicular to the @xmath133-interface , across which the excess stress on the control volume ( left ) due to contact forces and long - range forces originating from outside the control volume is integrated . ]
the interactions in the dft models consist of a short - ranged contact stress @xmath26 , and long - ranged potential interactions @xmath137 .
the total mechanical force on the control volume therefore reads @xmath138 where we integrated the the long - ranged forces per unit volume , @xmath139 , up to the dashed line in fig .
[ fig : sketch ] .
the terms @xmath140 represents the potential induced by volume @xmath141 ( outside the control volume ) on the part of the same phase that lies within the control volume .
note the subtle difference with @xmath142 appearing in the equilibrium for @xmath52 which is the potential energy due to the entire volume .
combined with the definition ( [ eq : def ] ) , we find the surface stress : @xmath143 as the long - ranged potentials @xmath144 follow from volume integrals [ see eq .
( [ eq : phi ] ) ] , their contribution is completely independent of the rheology of the materials and can be readily evaluated .
in addition , all the energies can be expressed directly in terms of @xmath145 , for which phase @xmath45 is a flat semi - infinite space , owing to the symmetry of the domains in fig .
[ fig : sketch ] .
let us first exploit this connection for the `` cross terms '' @xmath146 and @xmath147 . from the geometry
one can see that @xmath148 , due to a missing quadrant of phase @xmath136 . since by symmetry @xmath149 , the last two terms in ( [ eq : poep ] ) combine to an integral over @xmath150 and according to ( [ eq : ab ] ) this yields the adhesion energy @xmath151 .
the terms @xmath152 and @xmath153 are slightly more difficult to interpret , as can be seen from fig .
[ fig : intdomains ] .
as explained in detail in the caption of the figure , one finds @xmath154 . using the above expressions for the potentials , we can split the surface stress into three contributions : @xmath155 where [ eq : poep3 ] @xmath156 the last contribution of ( [ eq : poep2 ] ) is the work of adhesion
; it is completely independent of the elastic properties of phases @xmath135 and @xmath136 .
by contrast , the direct interactions @xmath157 and @xmath158 involve the stress @xmath159 , which has to be determined from the constitutive relations . from shuttleworth s thermodynamic relation , it follows that for incompressible liquids there is a strict equality between surface energy and surface stress @xcite .
hence , one usually does not distinguish between the two concepts , and simply use the nomenclature of surface tension . as a consistency check for the formalism that is developed in this paper ,
we first verify explicitly that this equality is recovered for a liquid / vapor interface , i.e. @xmath160 and a liquid / liquid interface , i.e. @xmath161 with @xmath135 , @xmath136 liquids .
then , we will study the particular case of a liquid in contact with an elastic solid , for which we will see that in general @xmath162 .
let us therefore consider phase @xmath135 to be vacuum and phase @xmath136 to be a liquid .
the thermodynamic stress inside the liquid is isotropic , i.e. @xmath163 . for a true vacuum the thermodynamic pressure is strictly zero , but in real physical systems the liquid will equilibrate at a small nonzero vapor pressure .
for such a liquid / vapor interface , the surface energies become @xmath164 , @xmath165 and @xmath166 . as a consequence of the vacuum , the work of adhesion @xmath167 , and also @xmath168 . hence
, the total surface stress @xmath169 only receives a contribution from liquid - liquid interactions , @xmath170 , as defined in ( [ eq : poep3 ] ) .
the integral ( [ eq : poep3b ] ) involves @xmath159 , which from eqs .
( [ eq : liqhom ] , [ eq : liqpr2 ] ) can be written @xmath171 inserting this in ( [ eq : poep3b ] ) , one finds a simplified expression @xmath172 according to ( [ eq : aa ] ) , the last integral can indeed be identified with the surface energy @xmath173 .
this completes the demonstration that the mechanical excess stress at the liquid / vapor interface , defined in fig .
[ fig : sketch ] and computed from the dft model , is strictly equal to the surface energy : @xmath160 .
we now consider an interface between two immiscible liquids @xmath135 and @xmath136 , so again the stress tensor will be isotropic @xmath174 . retracing the steps of the analysis for the liquid / vapor interface
, the main difference is that the equilibrium equation for the repulsive pressure @xmath52 now has a contribution @xmath175 .
hence , one finds in phase @xmath135 : @xmath176 while in phase @xmath136 @xmath177 note that this result is equivalent to eq . , with the solid phase replaced by the other liquid
this gives for the integrals ( [ eq : poep3 ] ) @xmath178 adding the work of adhesion , one indeed obtains that the surface stress is equal to the surface energy : @xmath179 finally , we discuss the solid - liquid interface .
as before , we evaluate the terms in the integral seperately .
the elastic properties of the solid enter through @xmath180 , whereas @xmath181 is the same as that in the liquid / liquid case , @xmath182 to obtain @xmath159 in the equation for @xmath180 [ eq . ] we use once more that the normal stress @xmath23 is continuous across the liquid - solid inteface . at the liquid side
we know that @xmath183 , which allows us to determine the integration constant in eq .
and we obtain , for the solid : @xmath184 to obtain the elastic response due to the presence of the liquid , we subtract the reference stress ( solid in contact with vacuum , eq . ) and obtain : @xmath185 compared to the base state ( sect . [
sect : boundaryconds ] ) , the liquid interactions @xmath186 are added to the pressure @xmath90 .
then , via hooke s law under plane strain conditions in the @xmath187- and @xmath188-directions we get ( eq . ):
@xmath189 and thus , following eqs . and : @xmath190 the final step before we can evaluate @xmath191 is to determine @xmath192 by applying hooke s law on @xmath193 : @xmath194 we can now evaluate @xmath191 , eq . , and find : @xmath195 and we obtain for the total surface stress : @xmath196 which is the central result of this paper , eq . .
dependence of the transmitted tangential stress on the superficial poisson ratio @xmath2 : eq . .
the dashed circle indicate a control volume around the contact line , which is larger than the molecular interaction range . for @xmath7 , the full stress on the solid induced by the presence of the contact line
is transmitted to the solid bulk , where it is balanced by elastic effects . in the incompressible limit @xmath5 ,
no tangential stress is transmitted .
this must necessarily hold for liquids , which can offer no elastic resistance to any transmitted tangential stress . ]
the shuttleworth equation is a thermodynamic relation showing that one needs to distinguish between the surface stress @xmath197 , i.e. the excess force per unit length in the interface , and the surface free energy @xmath17 . by computing
the elastic and enthalpic interactions from a microscopic model , we have for the first time derived an explicit relation between @xmath197 and @xmath17 .
we find that , in general , the surface stress of a solid interface is indeed different from the surface energy of the same interface , as is quantified by eq . .
it turns out that the difference between surface stress and surface energy depends on the compressibility in the superficial zones of the solid - liquid interface , characterized by the poisson s ratio @xmath2 .
the relation shows that for solids that are incompressible in the superficial zone ( @xmath5 ) the surface stress is equal to the surface energy , in line with the fact that for incompressible liquids @xmath198 .
the model predicts an equality @xmath199 for a solid - vapor interface of arbitrary value of @xmath2 [ consider with @xmath200 and @xmath201 . in our previous work @xcite we have shown that the difference between surface stress and surface energy is crucial for understanding the elastic deformation of partially wetted deformable surfaces . in particular , it was demonstrated that the contact line region transmits a residual tangential force to the substrate , which has to be balanced by bulk elasticity .
the magnitude of this tangential force can be expressed as ( see inset fig . [
fig : nurftel ] ) : @xmath202 with this becomes : @xmath203 the key feature of this result is that a finite tangential force is transmitted whenever the poisson ratio @xmath204 , see fig .
[ fig : nurftel ] .
we can directly compare this result with recent force measurements performed in molecular dynamics simulations @xcite .
for such a cubic lattice with springs of equal spring constant between nearest and next - nearest neighbors , it can be shown that @xmath205 . in these simulations ,
parameters were chosen such that @xmath206 , so @xmath207 .
the numerically measured value of the transmitted tangential stress ( @xmath208 ) agrees quantitatively with eq . , which predicts @xmath209 for such a system .
the influence of the tangential force recovered in eq .
[ eq : pies ] is usually not taken into account when describing elasto - capillary deformations .
one might argue that for most experiments this is a valid assumption : soft rubbers and gels are essentially incompressible , in which case our model recovers @xmath210 .
intriguingly , however , recent experiments on a soft elastomeric wire clearly demonstrate that a tangential force is transmitted across the surface , and give rise to a significant tangential bulk strain @xcite .
this strongly suggests that , while the bulk material clearly has @xmath211 , the interfacial layers of soft elastomers _ do _ exhibit a substantial degree of compressibility .
this would explain why a tangential stress is transmitted in the vicinity of the contact line , even for incompressible bulk materials .
a final consequence of such a tangential force is that the contact angles on very soft surfaces do not follow neumann s law @xcite .
this provides a strong limitation on the recently proposed method to determine surface stresses from contact angle measurements @xcite .
our work clearly shows that a more detailed microscopic description of the interfacial chemistry , combined with further experimental characterizations , are necessary to fully resolve the physics of elastocapillarity .
j. w. van honschoten , j. w. berenschot , t. ondarcuhu , r. g. p. sanders , j. sundaram , m. elwenspoek , and n. r. tas , _ elastocapillary fabrication of three - dimensional microstructures _ , appl .
phys . lett . *
97 * , 014103 ( 2010 ) .
r. pericet - camara , a. best , h .- j
. butt , and e. bonaccurso , _ effect of capillary pressure and surface tension on the deformation of elastic surfaces by sessile liquid microdrops : an experimental investigation _
, langmuir * 24 * , 10565 ( 2008 ) .
r. w. style , r. boltyanskiy , y. che , j. s. wettlaufer , l. a. wilen , and e. r. dufresne , _ universal deformation of soft substrates near a contact line and the direct measurement of solid surface stresses _ , phys .
* 110 * , 066103 ( 2013 ) .
r. w. style , r. boltyanskiy , y. che , j. s. wettlaufer , l. a. wilen , and e. r. dufresne , _ universal deformation of soft substrates near a contact line and the direct measurement of solid surface stresses _ , phys .
lett . * 110 * , 066103 ( 2013 ) . | the elastic deformation of a soft solid induced by capillary forces crucially relies on the excess stress inside the solid - liquid interface . while for a liquid - liquid interface
this surface stress " is strictly identical to the surface free energy " , the thermodynamic shuttleworth equation implies that this is no longer the case when one of the phases is elastic . here we develop a microscopic model that incorporates enthalpic interactions and entropic elasticity , based on which we explicitly compute the surface stress and surface free energy .
it is found that the compressibility of the interfacial region , through the poisson ratio near the interface , determines the difference between surface stress and surface energy .
we highlight the consequence of this finding by comparing with recent experiments and simulations on partially wetted soft substrates . |
the monte carlo calculations in the lattice qcd at finite temperature @xmath1 ( see , e.g. , refs . @xcite and references therein ) reveal two physical phases of strongly interacting matter : hadron phase at small @xmath1 and deconfined quark - gluon phase at high @xmath1 . in fig . [ fig - latt ]
the lattice results of ref .
@xcite for the pressure and energy density ( @xmath2 and @xmath3 ) obtained at zero baryonic chemical potential in qcd with 2 + 1 quark flavors and extrapolated to the thermodynamical and continuum limits are shown as functions of temperature @xmath1 .
( circles ) and @xmath3 ( squares ) at zero baryonic chemical potential .
, scaledwidth=69.0% ] from fig .
[ fig - latt ] one observes a steep increase of thermodynamical quantities near the crossover temperature @xmath4 .
this temperature is estimated in the range of 150 - 160 mev .
the values of @xmath2 and @xmath3 in the deconfined quark - gluon phase approach slowly from below the stefan - boltzmann limit @xmath5 , which equals to @xmath6 in the 3-flavor qcd . at @xmath7
the confined hadron phase emerges . in the present paper
the lattice data @xcite will be used to constrain an equation of state of the hadronic matter .
a description of hadron multiplicities in high - energy nucleus - nucleus collisions shows a surprisingly good agreement between the results of the hadron resonance gas ( hrg ) model ( see , e.g. , refs .
@xcite ) and the experimental data . in most statistical model formulations
the ideal hrg ( id - hrg ) is used .
it is argued that a presence of all known resonance states in the thermal system takes into account attractive interactions between hadrons @xcite .
two extensions of the id - hrg model have been widely discussed .
the first one is the excluded volume hrg ( ev - hrg ) model which introduces the effects of hadron repulsions at short distances .
one usually uses the van der waals procedure @xcite and substitutes a system volume @xmath8 by the available volume @xmath9 , where @xmath10 is the volume parameter for @xmath11th hadron species , @xmath12 is the number of particles of @xmath11th type , and the sum is taken over all types @xmath11 of hadrons and resonances .
another example of attractive and repulsive interactions between hadrons is the relativistic mean field theory @xcite ( see also the recent paper @xcite and references therein ) .
note that the ev - hrg model can be equivalently formulated in terms of the mean - field ( see refs .
this makes it possible to incorporate other hadron interactions within unified mean - field approach .
the second extension of the hrg model is an inclusion of the exponentially increasing mass spectrum @xmath13 proposed by hagedorn about 50 years ago @xcite .
these excited colorless states ( named fireballs or strings ) are considered as a continuation of the resonance spectrum at masses @xmath14 higher than @xmath15 gev . in the present paper
we use the lattice data @xcite at small temperature @xmath16 mev to confirm a presence of the excluded volume effects and effects of the hagedorn mass spectrum .
the hrg model had been used for comparison with the lattice data in the hadronic sector @xcite .
there were as well several publications , where the ev - hrg model ( see , e.g. , refs .
@xcite ) or the hagedorn mass spectrum ( see , e.g. , refs .
@xcite ) were confronted with the lattice data .
our analysis extends these previous attempts to the case when both these physical effects are treated simultaneously .
included together they improve essentially an agreement of the hrg model with the lattice results , while treated separately none of them can be clearly established .
the paper is organized as follows . in sec .
[ sec - hrg ] the grand canonical ensemble formulation of the id - hrg and ev - hrg are considered . in sec .
[ sec - hag ] the ev - hrg model is extended by inclusion of the hagedorn mass spectrum . a summary in sec .
[ sec - sum ] closes the article .
in the grand canonical ensemble the pressure and energy density of the id - hrg are given by @xmath17^{-1}~ , \\
\varepsilon^{\rm id}(t,\mu ) & = & \sum_i \varepsilon_i^{\rm id}(t,\mu_i ) \nonumber \\ & = & \sum_i\frac{d_i}{2\pi^2}\int d m\ , f_i(m)\int_0^{\infty}k^2dk\,\sqrt{k^2+m^2 } \left [ \exp\left(\frac{\sqrt{k^2+m^2}-\mu_i}{t}\right)~+~\eta_i\right]^{-1}~ , \label{e - id}\end{aligned}\ ] ] where @xmath18 is the spin degeneracy of @xmath11th particle and the normalized function @xmath19 takes into account the breit - wigner shape of resonance with finite width @xmath20 around their average mass @xmath21 , for the stable hadrons , @xmath22 .
the sum over @xmath11 in eqs . and
is taken over all non - strange and strange hadrons that are listed in particle data tables @xcite .
this includes mesons up to @xmath23 and ( anti-)baryons up to @xmath24 .
we also note that in these equations @xmath25 and @xmath26 for bosons and fermions , respectively , while @xmath27 corresponds to the boltzmann approximation .
the chemical potential for @xmath11th hadron is given by @xmath28 with @xmath29 , @xmath30 , and @xmath31 being the corresponding baryonic number , strangeness , and electric charge of @xmath11th hadron .
the notation @xmath32 will be used to denote all chemical potentials , @xmath33 . in this subsection a role of the repulsive interactions is considered within the ev - hrg model .
the van der waals excluded volume procedure corresponds to a substitution of the system volume @xmath8 by the available volume @xmath34 , @xmath35 where @xmath12 is the particle number , @xmath36 is the excluded volume parameter with @xmath37 being the corresponding hard sphere radius of particle @xmath11 , and the sum is taken over all hadrons and resonances .
this result , in particular , the presence of a factor of 4 in the expression for @xmath10 , can be rigorously obtained for a low density gas of particles of a single type ( see , e.g. , ref .
@xcite ) . in the grand canonical ensemble
, the substitution ( [ vdw ] ) leads to a transcendental equation for the ev - hrg pressure @xcite : @xmath38 and the energy density is calculated as @xmath39 where @xmath40 is the ideal - gas particle number density of @xmath11th hadron species , @xmath41^{-1}~. \end{aligned}\ ] ] in what follows we restrict our consideration to the case of equal volume parameters @xmath10 for all hadrons and resonances , @xmath42 .
the boltzmann approximation @xmath43 in eqs .
( [ p - id],[e - id],[n - id ] ) simplifies eqs .
( [ p - ev ] ) and ( [ e - ev ] ) : @xmath44 where the excluded volume suppression factor @xmath45 and the total particle number density @xmath46 in the id - hrg are introduced as @xmath47 expressions ( [ p - ev-1]-[kappa ] ) can be also obtained in the framework of thermodynamically self - consistent mean - field theory ( see sec .
@xcite ) . this approach gives a sequential treating of the problem when one can examine various different mean fields that mimic the repulsive and attractive interactions ( for details see refs .
@xcite ) .
the ev - hrg was used to fit the data on hadron multiplicities in ref .
@xcite with values of @xmath48 in the region of 0.2 - 0.8 fm .
a numerical value of the hard - core radius was estimated as @xmath49 fm in ref .
note that if radii of all hadrons are assumed to be the same , the chemical freeze - out parameters , temperature and baryon chemical potential , fitted to data on hadron multiplicities , are identical to those obtained within the id - hrg model .
indeed , the particle number ratios are not sensitive to the numerical value of @xmath48 .
hence , in order to establish a presence of non - zero hard - core hadron radii , the independent measurements of a total system volume are needed . on the other hand
, it was shown that the particle number fluctuations depend straightforwardly on the hard - core hadron radius @xcite .
thus , an interpretation of these data within the ev - hrg opens the way to estimate the value of @xmath48 from the data .
this is , however , not an easy task as there are many other effects which influence the particle number fluctuations . in fig .
[ fig - id ] ( a ) the id - hrg pressure ( [ p - id ] ) divided by @xmath50 is shown as a function of temperature for the case of zero chemical potentials , @xmath51 equation ( [ mu0 ] ) corresponds to zero values of all conserved charges , baryonic number , strangeness , and electric charge , of the strongly interacting matter .
this is approximately valid for the matter created at the large hadron collider ( lhc ) of the european organization for nuclear research ( cern ) . as it is seen from fig .
[ fig - id ] ( a ) the boltzmann approximation @xmath43 ( solid line ) gives a very accurate evaluation of @xmath52 calculated with the quantum statistics ( dashed line ) .
in fact , a difference between the solid and dashed lines is hardly seen in fig .
[ fig - id ] ( a ) .
the boltzmann approximation will be thus adopted for our further analysis .
note that a shift of the chemical potential according to eq .
( [ p - ev ] ) makes the boltzmann approximation in the ev - hrg even more accurate than in the case of the id - hrg model . ,
is shown as a function of temperature at @xmath53 by the dotted line .
the boltzmann approximation @xmath43 is shown by the solid line .
( b ) : the id - hrg pressure and ev - hrg pressure functions for several different values of hard - core radius @xmath48 are presented .
the stefan - boltzmann limit for the deconfined quark - gluon phase , @xmath54 is indicated by the horizontal dotted line .
, title="fig:",scaledwidth=49.0% ] , is shown as a function of temperature at @xmath53 by the dotted line .
the boltzmann approximation @xmath43 is shown by the solid line .
( b ) : the id - hrg pressure and ev - hrg pressure functions for several different values of hard - core radius @xmath48 are presented .
the stefan - boltzmann limit for the deconfined quark - gluon phase , @xmath54 is indicated by the horizontal dotted line .
, title="fig:",scaledwidth=49.0% ] are compared to the lattice data for @xmath55 ( a ) and @xmath3 ( c ) .
the values of @xmath56 and @xmath57 are shown as functions of @xmath48 in ( b ) and ( d ) , respectively , and the shaded grey area corresponds to @xmath58 fm .
, title="fig:",scaledwidth=49.0% ] are compared to the lattice data for @xmath55 ( a ) and @xmath3 ( c ) .
the values of @xmath56 and @xmath57 are shown as functions of @xmath48 in ( b ) and ( d ) , respectively , and the shaded grey area corresponds to @xmath58 fm . , title="fig:",scaledwidth=49.0% ] are compared to the lattice data for @xmath55 ( a ) and @xmath3 ( c ) .
the values of @xmath56 and @xmath57 are shown as functions of @xmath48 in ( b ) and ( d ) , respectively , and the shaded grey area corresponds to @xmath58 fm .
, title="fig:",scaledwidth=49.0% ] are compared to the lattice data for @xmath55 ( a ) and @xmath3 ( c ) .
the values of @xmath56 and @xmath57 are shown as functions of @xmath48 in ( b ) and ( d ) , respectively , and the shaded grey area corresponds to @xmath58 fm .
, title="fig:",scaledwidth=49.0% ] in fig .
[ fig - id ] ( b ) the id - hrg pressure ( [ p - id ] ) is compared with ev - hrg pressures ( [ p - ev-1 ] ) at several values of the hard - core radius @xmath48 .
note that the id - hrg model shows a strong increase of @xmath59 at high @xmath1 which exceeds the stefan - boltzmann pressure of the deconfined quarks and gluons , @xmath60 .
therefore , according to the gibbs criterium the point - like hadrons would always be the dominant phase at high temperatures @xcite due to the large number of different types of mesons and baryons .
this feature of the id - hrg equation of state contradicts to the lattice qcd results , hence it shows a shortcoming of the model based on the concept of the point - like particles .
just the excluded volume effects ensure a transition from a gas of hadrons and resonances to the quark - gluon plasma .
one needs therefore the ev - hrg equation of state for the hydrodynamic model calculations of nucleus - nucleus collisions ( see , e.g. , refs .
for any finite particle volume , i.e. @xmath61 , the behavior of pressure is found as @xmath62 at @xmath63 .
however , as seen from fig .
[ fig - id ] ( b ) , a more rigid restriction , @xmath64 fm , is needed to guarantee that @xmath65 at all @xmath66 . in fig .
[ fig - ev ] ( a ) , the ev - hrg results for @xmath67 are compared to the lattice data @xmath68 @xcite at @xmath16 mev for several different values of hard - core radius @xmath48 . in fig .
[ fig - ev ] ( b ) , the value of @xmath56 at different @xmath48 is shown .
this quantity is calculated as @xmath69 ^ 2}{[\delta(p^{\rm lat}/t^4)_i]^2}~ , \end{aligned}\ ] ] where @xmath70 is number of points @xmath71 ( equal 10 in our case ) minus the number of fitting parameters ( one parameter @xmath48 in our fit ) .
a most essential part of uncertainties in the lattice data is not statistical .
the systematical uncertainties dominate , and these uncertainties are significantly correlated . in this case ,
the usage of @xmath72 criterion is not perfectly reasonable .
however , we still use this quantity as a way to quantify the deviations of hrg calculations from the lattice data . from fig .
[ fig - ev ] ( b ) one observes that the lattice data for @xmath68 are fitted well in a rather wide range of hard - core radius of @xmath73 fm .
therefore , the lattice data for the hadron pressure are consistent with a presence of rather significant excluded volume effects and suggest reasonable numerical values for the hard - core radius @xmath48 . however , a comparison of the ev - hrg model with the lattice results @xmath74 shown in fig .
[ fig - ev ] ( c ) does not indicate a presence of the excluded volume effects .
this is clear from fig .
[ fig - ev ] ( d ) , where we show the value of @xmath57 calculated as in eq .
( [ chi - p ] ) but with @xmath3 instead of @xmath55 .
the value of @xmath75 corresponds to the best fit of @xmath74 .
let us , however , remind fig .
[ fig - id ] ( b ) which shows that too small values of the hard - core radius , @xmath58 fm , look doubtful . for these small values of @xmath48 the ev hrg pressure becomes larger than the stefan - boltzmann limit for quarks and gluons .
the ` forbidden ' region of the hard - core radius , @xmath76 fm , is shown as the grey area in figs .
[ fig - ev ] ( b ) , ( d ) and [ fig - hag2 ] ( a ) , ( b ) . therefore , while a reasonable @xmath48-value , e.g. , @xmath49 fm , gives a good agreement , @xmath77 , with the @xmath68 lattice data , it also leads to rather large value of @xmath78 and , thus , looks unreasonable for the fit of @xmath74 .
the analysis presented in the previous section gives no conclusive answer about a presence of the excluded volume effects .
the value of @xmath49 fm in the ev - hrg model leads to sizeable suppression effects of the id - hrg pressure and to a good agreement with the lattice data @xmath68 , whereas the @xmath74 data prefer the value of @xmath79 and is thus consistent with the id - hrg model . from our point of view
, this observation may indicate a presence of additional contributions to @xmath80 and @xmath81 in the ev - hrg model .
these contributions should be small enough for the pressure and much larger for the energy density .
we argue that massive hagedorn states are the ideal candidates for this role .
indeed , each heavy particle with @xmath82 gives its contribution @xmath1 to the pressure , and much larger contribution , @xmath83 , to the energy density .
for a further analysis we use the following parametrization for the hagedorn mass spectrum @xcite : @xmath84 the spectrum ( [ rho ] ) with the following parameters , @xmath85 gev , @xmath86 mev , @xmath87=0.5 gev , and @xmath88 , will be used .
the parameter @xmath89 in ( [ rho ] ) will be the only free parameter in the following analysis .
we have checked that another set of parameters , e.g. , the same set with @xmath90 gev , and also a set with @xmath91 , @xmath92 , and @xmath93 gev , leads to very similar results .
our final assumption concerns the proper volume for the hagedorn states . to avoid additional free parameters we adopt the same value of @xmath94 for all known hadrons and resonances as well as for the hagedorn states .
the ev - hrg model with the hagedorn mass spectrum will be denoted as ev - hrg - h .
the pressure @xmath95 in ev - hrg - h model is given by the following equation @xmath96 ~. \label{phag}\ ] ] the expression ( [ phag ] ) can be equivalently rewritten similar to eq .
( [ kappa ] ) , @xmath97 where @xmath98 , and @xmath99 and @xmath100 are , respectively , the expressions for the pressure and total number density of all particles ( hadrons , resonances , and hagedorn excited states ) in the ideal hrg with the haggedorn mass spectrum ( id - hrg - h ) , i.e. , with the hagedorn mass spectrum but without excluded volume effects .
one can easily calculate the energy density as @xmath101 where @xmath102 ~ \label{eq : hag - energy - density}\ ] ] denotes the energy density of the id - hrg - h model .
are compared to the lattice data for @xmath55 and @xmath3 in ( a ) and ( b ) , respectively .
the value of @xmath89 is fixed as @xmath103 gev@xmath104 .
, title="fig:",scaledwidth=49.0% ] are compared to the lattice data for @xmath55 and @xmath3 in ( a ) and ( b ) , respectively .
the value of @xmath89 is fixed as @xmath103 gev@xmath104 .
, title="fig:",scaledwidth=49.0% ] which minimizes @xmath72 at each value of @xmath48 is shown as a function of @xmath48 .
( b ) : the quantity @xmath72 as a function of @xmath48 . for each value of @xmath48 parameter @xmath89
is fitted in order to minimize @xmath72 .
the shaded grey area corresponds to @xmath58 .
, title="fig:",scaledwidth=49.0% ] which minimizes @xmath72 at each value of @xmath48 is shown as a function of @xmath48 .
( b ) : the quantity @xmath72 as a function of @xmath48 . for each value of @xmath48 parameter @xmath89
is fitted in order to minimize @xmath72 .
the shaded grey area corresponds to @xmath58 .
, title="fig:",scaledwidth=49.0% ] in fig .
[ fig - hag ] , a comparison of the ev - hrg - h with the lattice data @xcite is presented .
the results for @xmath105 and @xmath106 are presented by the solid lines in fig .
[ fig - hag ] ( a ) and ( b ) , respectively .
these lines correspond to different values of @xmath48 but fixed @xmath107 gev@xmath104 .
the dashed lines show the ev - hrg results without hagedorn mass spectrum , i.e. , at @xmath108 .
an inclusion of the hagedorn mass spectrum become clearly visible at @xmath109 mev , and its contribution to the energy density is essentially larger than that to the pressure .
a simultaneous fit of the @xmath68 and @xmath74 lattice data is done .
a quality of the fit is now controlled by @xmath72 with @xmath110 number of points and 2 fitting parameters ( @xmath48 and @xmath89 ) . at each value of @xmath48
one can find the @xmath89 parameter which minimize @xmath72 at fixed @xmath48 .
this introduces the correlation between parameters @xmath89 and @xmath48 which is shown in fig .
[ fig - hag2 ] ( a ) .
the data are well fitted for a rather wide range of values for hard - core radius : @xmath111 for @xmath112 fm . in fig .
[ fig - hag2 ] ( b ) a dependence of @xmath72 on @xmath48 is shown for @xmath108 and @xmath113 , where @xmath114 is depicted in fig [ fig - hag2 ] ( a ) .
a simultaneous fit of @xmath68 and @xmath74 within ev - hrg model ( i.e. , for @xmath108 ) does not show a necessity of @xmath61 .
similarly , the id - hrg - h model ( i.e. , at @xmath75 ) admits only very small contributions from hagedorn states to the thermodynamical functions ( a small value of @xmath89 at @xmath75 seen in fig .
[ fig - hag2 ] ( a ) ) .
therefore , no clear evidence for @xmath61 or @xmath115 can be found if these two effects are considered separately , i.e. , within ev - hrg or id - hrg - h model . on the other hand , taking them simultaneously within ev - hrg - h model we indicate a presence of these two effects and improvement of the id - hrg model .
let us also note that an inclusion of the hagedorn mass spectrum led to a successful description of lattice data for the confined glueball phase in the pure su(3 ) case ( without quarks ) in ref .
this analysis is based on the assumption that glueball decay widths and interactions ( i.e. , excluded volume effects too ) are rather small and can be neglected . in the full su(3 )
theory ( with quarks ) this assumption may not be valid .
in summary , the lattice data of ref . @xcite for @xmath68 and @xmath74 are considered with the hrg model .
two extensions of this model are analyzed : the excluded volume effects ( with the same hard - core radius @xmath48 for all particles ) and the exponential hagedorn mass spectrum . a condition that the pressure of the hrg should not exceed the stefan - boltzmann limit for quarks and gluons indicates that hadrons should have a non - zero hard - core radius of at least @xmath116 fm .
however , a comparison of the excluded volume hrg model with the lattice data at @xmath16 mev yields no conclusive evidences in favor of a presence of the excluded volume effects .
namely , the fit of @xmath68 prefers values of @xmath117 fm , while the best fit of @xmath74 corresponds to @xmath79 .
if @xmath75 , there is also not much room for the contribution from the hagedorn states , the best fit in this case corresponds to @xmath118 , i.e. it suggests an absence of the contributions from the hagedorn states .
it means that neither excluded volume hrg nor ideal hrg with additional hagedorn states being considered separately demonstrates any advantages for fitting the lattice data in a comparison to the ideal hrg model ( with no excluded volume effects and no hagedron states ) . on the other hand ,
if both these physical effects are considered simultaneously the situation is changed : the data are well fitted for @xmath112 fm and @xmath119 gev@xmath104 with @xmath111 , i.e. , there is a clear indication that both the hard core repulsion and the hagedorn mass spectrum should be taken into account simultaneously in the framework of the hadron resonance gas model .
they lead to rather essential contributions : suppression effects for @xmath105 and @xmath106 due to the excluded volume effects and enhancement due to the hagedorn mass spectrum .
these _ simultaneous _ contributions ensure a better agreement with the lattice data and lead , therefore , to improvement of the ideal hrg model .
we would like to thank z. fodor , m. gadzicki , f. karsch , k. redlich , and e. shuryak for fruitful comments .
publication is based on the research provided by the grant support of the state fund for fundamental research ( project no . f58/175 - 2014 ) .
the work was partially supported by hic for fair within the loewe program of the state of hesse and by the program of fundamental research of the department of physics and astronomy of nas of ukraine ( project no . 0112u000056 ) .
j. cleymans and h. satz , z. phys .
c * 57 * , 135 ( 1993 ) .
f. becattini , j. cleymans , a. keranen , e. suhonen and k. redlich , phys .
c * 64 * , 024901 ( 2001 ) .
braun - munzinger , d. magestro , k. redlich , and j. stachel , phys .
b * 518 * , 41 ( 2001 ) .
j. rafelski and j. letessier , nucl .
a * 715 * , 98c ( 2003 ) .
a. andronic , p. braun - munzinger , and j. stachel , nucl .
phys . a * 772 * , 167 ( 2006 ) .
f. becattini , j. manninen and m. gazdzicki , phys .
c * 73 * , 044905 ( 2006 ) .
j. noronha - hostler , m. beitel , c. greiner , and i. shovkovy , phys .
c * 81 * 054909 ( 2010 ) .
a. majumder and b. muller , phys .
lett . * 105 * , 252002 ( 2010 ) .
j. noronha - hostler , j. noronha , and c. greiner , phys .
c * 86 * , 024913 ( 2012 ) .
y. hama , t. kodama and o. socolowski , jr .
* 35 * , 24 ( 2005 ) .
k. werner , iu .
karpenko , t. pierog , m. bleicher and k. mikhailov , phys .
c * 82 * , 044904 ( 2010 ) .
l. m. satarov , m. n. dmitriev , and i. n. mishustin , phys .
. nucl . * 72 * , 1390 ( 2009 ) ; a. v. merdeev , l. m. satarov , and i. n. mishustin , phys .
c * 84 * , 014907 ( 2011 ) . | the monte carlo results in lattice qcd for the pressure and energy density at small temperature @xmath0 mev and zero baryonic chemical potential are analyzed within the hadron resonance gas model .
two extensions of the ideal hadron resonance gas are considered : the excluded volume model which describes a repulsion of hadrons at short distances and hagedorn model with the exponential mass spectrum . considering both of these models _
one by one _ we do not find the conclusive evidences in favor of _ any of them_. the controversial results appear because of rather different sensitivities of the pressure and energy density to both excluded volume and hagedorn mass spectrum effects . on the other hand , we have found a clear evidence for a _ simultaneous _ presence of _ both of them_. they lead to rather essential contributions : suppression effects for thermodynamical functions of the hadron resonance gas due to the excluded volume effects and enhancement due to the hagedorn mass spectrum . |
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this is a late parrot !
it s a stiff !
bereft of life , it rests in peace ! if you had nt nailed him to the perch he would be pushing up the daisies !
its metabolical processes are of interest only to historians !
it s hopped the twig !
it s shuffled off this mortal coil !
it s rung down the curtain and joined the choir invisible !
this is an ex - parrot ! _ monty python , `` pet shop '' _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a mechanism for automatically generating multiple paraphrases of a given sentence would be of significant practical import for text - to - text generation systems .
applications include summarization and rewriting @xcite : both could employ such a mechanism to produce candidate sentence paraphrases that other system components would filter for length , sophistication level , and so forth . not surprisingly , therefore , paraphrasing has been a focus of generation research for quite some time @xcite .
one might initially suppose that sentence - level paraphrasing is simply the result of word - for - word or phrase - by - phrase substitution applied in a domain- and context - independent fashion .
however , in studies of paraphrases across several domains , this was generally not the case .
for instance , consider the following two sentences ( similar to examples found in ): 0.9 after the latest fed rate cut , stocks rose across the board .
+ observe that `` fed '' ( federal reserve ) and `` greenspan '' are interchangeable only in the domain of us financial matters .
also , note that one can not draw one - to - one correspondences between single words or phrases .
for instance , nothing in the second sentence is really equivalent to `` across the board '' ; we can only say that the entire clauses `` stocks rose across the board '' and `` winners strongly outpaced losers '' are paraphrases .
this evidence suggests two consequences : ( 1 ) we can not rely solely on generic domain - independent lexical resources for the task of paraphrasing , and ( 2 ) _ sentence - level _ paraphrasing is an important problem extending beyond that of paraphrasing smaller lexical units .
_ our work presents a novel knowledge - lean algorithm that uses _ multiple - sequence alignment _ ( msa ) to _ learn _ to generate sentence - level paraphrases essentially from unannotated corpus data alone .
_ in contrast to previous work using msa for generation , we need neither parallel data nor explicit information about sentence semantics .
rather , we use two _ comparable corpora _ , in our case , collections of articles produced by two different newswire agencies about the same events .
the use of related corpora is key : we can capture paraphrases that on the surface bear little resemblance but that , by the nature of the data , must be descriptions of the same information . note that we also acquire paraphrases from each of the individual corpora ; but the lack of clues as to sentence equivalence in single corpora means that we must be more conservative , only selecting as paraphrases items that are structurally very similar .
our approach has three main steps .
first , working on each of the comparable corpora separately , we compute _ lattices _ compact graph - based representations to find commonalities within ( automatically derived ) groups of structurally similar sentences .
next , we identify pairs of lattices from the two different corpora that are paraphrases of each other ; the identification process checks whether the lattices take similar arguments . finally , given an input sentence to be paraphrased , we match it to a lattice and use a paraphrase from the matched lattice s mate to generate an output sentence .
the key features of this approach are : * focus on paraphrase generation . * in contrast to earlier work , we not only extract paraphrasing rules , but also automatically determine which of the potentially relevant rules to apply to an input sentence and produce a revised form using them
. * flexible paraphrase types . * previous approaches to paraphrase acquisition focused on certain rigid types of paraphrases , for instance , limiting the number of arguments .
in contrast , our method is not limited to a set of _ a priori_-specified paraphrase types . * use of comparable corpora and minimal use of knowledge resources . *
in addition to the advantages mentioned above , comparable corpora can be easily obtained for many domains , whereas previous approaches to paraphrase acquisition ( and the related problem of phrase - based machine translation ) required parallel corpora .
we point out that one such approach , recently proposed by , also represents paraphrases by lattices , similarly to our method , although their lattices are derived using parse information .
moreover , our algorithm does not employ knowledge resources such as parsers or lexical databases , which may not be available or appropriate for all domains a key issue since paraphrasing is typically domain - dependent .
nonetheless , our algorithm achieves good performance .
previous work on automated paraphrasing has considered different levels of paraphrase granularity . learning synonyms via distributional similarity
has been well - studied . and identify phrase - level paraphrases , while and acquire structural paraphrases encoded as templates .
these latter are the most closely related to the sentence - level paraphrases we desire , and so we focus in this section on template - induction approaches .
extract inference rules , which are related to paraphrases ( for example , x wrote y implies x is the author of y ) , to improve question answering .
they assume that _ paths _ in dependency trees that take similar arguments ( leaves ) are close in meaning .
however , only two - argument templates are considered .
also use dependency - tree information to extract templates of a limited form ( in their case , determined by the underlying information extraction application ) .
like us ( and unlike lin and pantel , who employ a single large corpus ) , they use articles written about the same event in different newspapers as data .
our approach shares two characteristics with the two methods just described : pattern comparison by analysis of the patterns respective arguments , and use of non - parallel corpora as a data source .
however , _ extraction _ methods are not easily extended to _ generation _ methods .
one problem is that their templates often only match small fragments of a sentence .
while this is appropriate for other applications , deciding whether to use a given template to generate a paraphrase requires information about the surrounding context provided by the entire sentence .
[ [ overview ] ] overview + + + + + + + + we first sketch the algorithm s broad outlines .
the subsequent subsections provide more detailed descriptions of the individual steps .
the major goals of our algorithm are to learn : * recurring patterns in the data , such as x ( injured / wounded ) y people , z seriously , where the capital letters represent variables ; * pairings between such patternsthat represent paraphrases , for example , between the patternx ( injured / wounded ) y people , z of them seriously and the patterny were ( wounded / hurt ) by x , among them z were in serious condition .
figure [ fig : arch ] illustrates the main stages of our approach . during training ,
patterninduction is first applied independently to the two datasets making up a pair of comparable corpora .
individual patternsare learned by applying _ multiple - sequence alignment _ to clustersof sentences describing approximately similar events ; these patternsare represented compactly by _ lattices _ ( see figure [ fig : lattice ] ) .
we then check for latticesfrom the two different corpora that tend to take the same arguments ; these latticepairs are taken to be paraphrase patterns .
once training is done , we can generate paraphrases as follows : given the sentence `` the surprisebombing injured twenty people , five of them seriously '' , we match it to the lattice x ( injured / wounded ) y people , z of them seriously which can be rewritten as y were ( wounded / hurt ) by x , among them z were in serious condition , and so by substituting arguments we can generate `` twenty were wounded by the surprisebombing , among them five were in serious condition '' or `` twenty were hurt by the surprisebombing , among them five were in serious condition '' .
our first step is to cluster sentences into groups from which to learn useful patterns ; for the multiple - sequence techniques we will use , this means that the sentences within clustersshould describe similar events and have similar structure , as in the sentences of figure [ fig : cluster ] .
this is accomplished by applying hierarchical complete - link clustering to the sentences using a similarity metric based on word n - gram overlap ( @xmath0 ) .
the only subtlety is that we do not want mismatches on sentence details ( e.g. , the location of a raid ) causing sentences describing the same type of occurrence ( e.g. , a raid ) from being separated , as this might yield clusterstoo fragmented for effective learning to take place .
( moreover , variability in the _ arguments _ of the sentences in a cluster is needed for our learning algorithm to succeed ; see below . )
we therefore first replace all appearances of dates , numbers , and proper names with generic tokens .
clusterswith fewer than ten sentences are discarded . in order to learn patterns ,
we first compute a _ multiple - sequence alignment _ ( msa ) of the sentences in a given cluster .
pairwise msa takes two sentences and a scoring function giving the similarity between words ; it determines the highest - scoring way to perform insertions , deletions , and changes to transform one of the sentences into the other .
pairwise msa can be extended efficiently to multiple sequences via the iterative pairwise alignment , a polynomial - time method commonly used in computational biology @xcite .
the results can be represented in an intuitive form via a word _ lattice _ ( see figure [ fig : lattice ] ) , which compactly represents ( n - gram ) structural similarities between the cluster s sentences . to transform latticesinto generation - suitable patternsrequires some understanding of the possible varieties of latticestructures .
the most important part of the transformation is to determine which words are actually instances of arguments , and so should be replaced by _ slots _
( representing variables ) .
the key intuition is that because the sentences in the clusterrepresent the same _ type _ of event , such as a bombing , but generally refer to different _ instances _ of said event ( e.g. a bombing in jerusalem versus in gaza ) , areas of large variability in the latticeshould correspond to arguments . to quantify this notion of variability ,
we first formalize its opposite : commonality .
we define _ backbone _ nodes as those shared by more than 50% of the cluster s sentences .
the choice of 50% is not arbitrary it can be proved using the pigeonhole principle that our strict - majority criterion imposes a unique linear ordering of the backbone nodes that respects the word ordering within the sentences , thus guaranteeing at least a degree of well - formedness and avoiding the problem of how to order backbone nodes occurring on parallel `` branches '' of the lattice .
once we have identified the backbonenodes as points of strong commonality , the next step is to identify the regions of variability ( or , in latticeterms , many parallel disjoint paths ) between them as ( probably ) corresponding to the arguments of the propositions that the sentences represent .
for example , in the top of figure [ fig : lattice ] , the words southern city , `` settlement of name'',``coastal resort of name '' , etc . all correspond to the location of an event and could be replaced by a single slot .
figure [ fig : lattice ] shows an example of a latticeand the derived slotted lattice ; we give the details of the slot - induction process in the appendix .
now , if we were using a parallel corpus , we could employ sentence - alignment information to determine which lattices correspond to paraphrases . since we do not have this information , we essentially approximate the parallel - corpus situation by correlating information from descriptions of ( what we hope are ) the same event occurring in the two different corpora .
our method works as follows .
once latticesfor each corpus in our comparable - corpus pair are computed , we identify latticeparaphrase pairs , using the idea that paraphrases will tend to take the same values as arguments .
more specifically , we take a pair of latticesfrom different corpora , look back at the sentence clusters from which the two lattices were derived , and compare the slot values of those cross - corpus sentence pairs that appear in articles written on the _ same day _ on the same topic ; we pair the latticesif the degree of matching is over a threshold tuned on held - out data .
for example , suppose we have two ( linearized ) lattices bombed slot2 and slot3 was bombed by slot4 drawn from different corpora .
if in the first lattice s sentence cluster we have the sentence `` the plane bombed the town '' , and in the second lattice s sentence cluster we have a sentence written on the same day reading `` the town was bombed by the plane '' , then the corresponding lattices may well be paraphrases , where slot1 is identified with slot4 and slot2 with slot3 . to compare the set of argument values of two lattices ,
we simply count their word overlap , giving double weight to proper names and numbers and discarding auxiliaries ( we purposely ignore order because paraphrases can consist of word re - orderings ) .
given a sentence to paraphrase , we first need to identify which , if any , of our previously - computed sentence clustersthe new sentence belongs most strongly to .
we do this by finding the best alignment of the sentence to the existing lattices .
if a matching latticeis found , we choose one of its comparable - corpus paraphrase latticesto rewrite the sentence , substituting in the argument values of the original sentence .
this yields as many paraphrases as there are lattice paths .
all evaluations involved judgments by native speakers of english who were not familiar with the paraphrasing systems under consideration .
we implemented our system on a pair of comparable corpora consisting of articles produced between september 2000 and august 2002 by the agence france - presse ( afp ) and reuters news agencies . given our interest in domain - dependent paraphrasing , we limited attention to 9 mb of articles , collected using a tdt - style document clustering system , concerning individual acts of violence in israel and army raids on the palestinian territories . from this data
( after removing 120 articles as a held - out parameter - training set ) , we extracted 43 slotted latticesfrom the afp corpus and 32 slotted latticesfrom the reuters corpus , and found 25 cross - corpus matching pairs ; since latticescontain multiple paths , these yielded 6,534 template pairs . before evaluating the quality of the rewritings produced by our templates and lattices , we first tested the quality of a random sample of just the template pairs . in our instructions to the judges , we defined two text units ( such as sentences or snippets ) to be paraphrases if one of them can generally be substituted for the other without great loss of information ( but not necessarily vice versa ) . given a pair of _ templates _ produced by a system , the judges marked them as paraphrases if for many instantiations of the templates variables , the resulting text units were paraphrases .
( several labelled examples were provided to supply further guidance ) .
to put the evaluation results into context , we wanted to compare against another system , but we are not aware of any previous work creating templates precisely for the task of generating paraphrases .
instead , we made a good - faith effort to adapt the dirt system to the problem , selecting the 6,534 highest - scoring templates it produced when run on our datasets .
( the system of was unsuitable for evaluation purposes because their paraphrase extraction component is too tightly coupled to the underlying information extraction system . )
it is important to note some important caveats in making this comparison , the most prominent being that dirt was not designed with sentence - paraphrase generation in mind its templates are much shorter than ours , which may have affected the evaluators judgments and was originally implemented on much larger data sets.tide@xmath1 n : nn : n '' , which we transformed into `` y tide of x '' so that its output format would be the same as ours . ]
the point of this evaluation is simply to determine whether another corpus - based paraphrase - focused approach could easily achieve the same performance level . in brief
, the dirt system works as follows .
dependency trees are constructed from parsing a large corpus .
leaf - to - leaf paths are extracted from these dependency trees , with the leaves serving as slots .
then , pairs of paths in which the slots tend to be filled by similar values , where the similarity measure is based on the mutual information between the value and the slot , are deemed to be paraphrases .
we randomly extracted 500 pairs from the two algorithms output sets .
of these , 100 paraphrases ( 50 per system ) made up a `` common '' set evaluated by all four judges , allowing us to compute agreement rates ; in addition , each judge also evaluated another `` individual '' set , seen only by him- or herself , consisting of another 100 pairs ( 50 per system ) .
the `` individual '' sets allowed us to broaden our sample s coverage of the corpus .
the pairs were presented in random order , and the judges were not told which system produced a given pair .
as figure [ msa - dirt - accuracy ] shows , our system outperforms the dirt system , with a consistent performance gap for all the judges of about 38% , although the absolute scores vary ( for example , judge 4 seems lenient ) .
the judges assessment of correctness was fairly constant between the full 100-instance set and just the 50-instance common set alone . in terms of agreement , the kappa value ( measuring pairwise agreement discounting chance occurrences ) on the common set was 0.54 , which corresponds to moderate agreement .
multiway agreement is depicted in figure [ msa - dirt - accuracy ] there , we see that in 86 of 100 cases , at least three of the judges gave the same correctness assessment , and in 60 cases all four judges concurred .
finally , we evaluated the quality of the paraphrase sentences generated by our system , thus ( indirectly ) testing all the system components : pattern selection , paraphrase acquisition , and generation .
we are not aware of another system generating sentence - level paraphrases .
therefore , we used as a baseline a simple paraphrasing system that just replaces words with one of their randomly - chosen wordnet synonyms ( using the most frequent sense of the word that wordnet listed synonyms for ) .
the number of substitutions was set proportional to the number of words our method replaced in the same sentence .
the point of this comparison is to check whether simple synonym substitution yields results comparable to those of our algorithm . [ cols="<,<",options="header " , ] for this experiment , we randomly selected 20 afp articles about violence in the middle east published later than the articles in our training corpus .
out of 484 sentences in this set , our system was able to paraphrase 59 ( 12.2% ) .
( we chose parameters that optimized precision rather than recall on our small held - out set . )
we found that after proper name substitution , only seven sentences in the test set appeared in the training set , which implies that latticesboost the generalization power of our method significantly : from seven to 59 sentences .
interestingly , the coverage of the system varied significantly with article length .
for the eight articles of ten or fewer sentences , we paraphrased 60.8% of the sentences per article on average , but for longer articles only 9.3% of the sentences per article on average were paraphrased .
our analysis revealed that long articles tend to include large portions that are unique to the article , such as personal stories of the event participants , which explains why our algorithm had a lower paraphrasing rate for such articles .
all 118 instances ( 59 per system ) were presented in random order to two judges , who were asked to indicate whether the meaning had been preserved .
of the paraphrases generated by our system , the two evaluators deemed 81.4% and 78% , respectively , to be valid , whereas for the baseline system , the correctness results were 69.5% and 66.1% , respectively .
agreement according to the kappa statistic was 0.6 .
note that judging full sentences is inherently easier than judging templates , because template comparison requires considering a variety of possible slot values , while sentences are self - contained units .
figure [ fig : wordnet ] shows two example sentences , one where our msa - based paraphrase was deemed correct by both judges , and one where both judges deemed the msa - generated paraphrase incorrect .
examination of the results indicates that the two systems make essentially orthogonal types of errors .
the baseline system s relatively poor performance supports our claim that whole - sentence paraphrasing is a hard task even when accurate word - level paraphrases are given .
we presented an approach for generating sentence level paraphrases , a task not addressed previously .
our method learns structurally similar patterns of expression from data and identifies paraphrasing pairs among them using a comparable corpus .
a flexible pattern - matching procedure allows us to paraphrase an unseen sentence by matching it to one of the induced patterns .
our approach generates both lexical and structural paraphrases .
another contribution is the induction of msa lattices from non - parallel data .
lattices have proven advantageous in a number of nlp contexts , but were usually produced from arallel data , which may not be readily available for many applications .
we showed that word lattices can be induced from a type of corpus that can be easily obtained for many domains , broadening the applicability of this useful representation .
in this appendix , we describe how we insert slots into latticesto form slotted lattices .
recall that the backbone nodes in our latticesrepresent words appearing in many of the sentences from which the lattice was built . as mentioned above
, the intuition is that areas of high variability between backbone nodes may correspond to arguments , or slots .
but the key thing to note is that there are actually two different phenomena giving rise to multiple parallel paths : _ argument variability _
, described above , and _
synonym variability_. for example , figure [ fig : variability](b ) contains parallel paths corresponding to the synonyms `` injured '' and `` wounded '' .
note that we want to _ remove _ argument variability so that we can generate paraphrases of sentences with arbitrary arguments ; but we want to _ preserve _ synonym variability in order to generate a variety of sentence rewritings . to distinguish these two situations , we analyze the _ split level _ of backbonenodes that begin regions with multiple paths .
the basic intuition is that there is probably more variability associated with arguments than with synonymy : for example , as datasets increase , the number of locations mentioned rises faster than the number of synonyms appearing .
we make use of a _ synonymy threshold _
@xmath2 ( set by held - out parameter - tuning to 30 ) , as follows .
* if no more than @xmath2% of all the edges out of a backbonenode lead to the same next node , we have high enough variability to warrant inserting a slot node .
* otherwise , we incorporate reliable synonyms , identified only single - word synonyms , phrase - level synonyms can similarly be acquired by considering chains of nodes connecting backbone nodes . ] into the backbone structure by preserving all nodes that are reached by at least @xmath2% of the sentences passing through the two neighboring backbonenodes .
furthermore , all backbonenodes labelled with our special generic tokens are also replaced with slotnodes , since they , too , probably represent arguments ( we condense adjacent slotsinto one ) .
nodes with in - degree lower than the synonymy threshold are removed under the assumption that they probably represent idiosyncrasies of individual sentences .
see figure [ fig : variability ] for examples . | we address the text - to - text generation problem of sentence - level paraphrasing a phenomenon distinct from and more difficult than word- or phrase - level paraphrasing .
our approach applies _ multiple - sequence alignment _ to sentences gathered from unannotated comparable corpora : it learns a set of paraphrasing patterns represented by _ word lattice _
pairs and automatically determines how to apply these patterns to rewrite new sentences .
the results of our evaluation experiments show that the system derives accurate paraphrases , outperforming baseline systems . |
[ sec : intro ] understanding of certain fine - tunings in the parameters of the standard model is an important frontier in modern hadron and nuclear physics . in connection with the anthropic considerations
, a question has been raised whether the light quark masses have to take very specific values in order to maintain conditions essential for the development of life , see ref .
@xcite for discussion .
in particular , the proximity of the hoyle state , the first @xmath3 excited state of @xmath4c , to the triple alpha - particle threshold is known to be crucial for the enhanced resonance formation of the life - important elements @xmath4c and @xmath5o in red giant stars @xcite .
the dependence of the excitation energy of the hoyle state on the light - quark masses was analyzed recently within _ ab initio _ nuclear lattice simulations @xcite .
it was found that the variation of the light quark masses by a few percent is likely to be not detrimental for the development of life .
more conclusive statements would require a better knowledge of the quark mass dependence of the nuclear force or , more precisely , of the nucleon - nucleon ( nn ) @xmath6-wave scattering lengths which is by far the dominant source of theoretical uncertainty in this calculation .
the quark mass dependence of the nuclear force also plays an important role for constraining a time variation of the standard model parameters as predicted by various extensions of the standard model at the time of big bang nucleosynthesis by comparing the observed and calculated primordial deuterium and helium abundances @xcite . in recent years
, there has been significant progress in lattice - qcd calculations of nuclear systems which constitute the primary source of information about the light - quark or , equivalently , pion mass dependence of nuclear observables . in particular , the fully dynamical calculations at unphysical pion masses as low as @xmath7-@xmath8 mev
have been performed , see e.g. refs .
@xcite . to connect these results with experimentally observed quantities corresponding to the physical value of the quark masses one can employ chiral effective field theory ( eft ) which is still expected to be applicable at such pion masses
notice that not only the binding energies but also the nn scattering observables such as the phase shifts and effective range parameters have been calculated on a lattice , see refs .
@xcite for recent analyses of the nn @xmath2 and @xmath1 channels by the nplqcd collaboration at @xmath9 and @xmath10 mev and ref .
@xcite for the analysis of the higher partial waves at @xmath11 mev by the callat collaboration .
interestingly , the general trend of lattice calculations by different groups suggests a stronger attraction both in the @xmath2 and @xmath1 channels when going away from the physical point towards heavier quark masses @xcite .
both the deuteron and dineutron systems in these calculations are found to be bound at unphysically heavy pion masses with the deuteron binding energy being significantly larger than the experimentally observed one .
these results are , however , not supported by the hal qcd collaboration which finds no bound states in these channels for pion masses ranging from @xmath12 to @xmath13 mev @xcite .
it should be noted that unlike the lattice calculations mentioned above , a different approach is employed by the hal qcd collaboration , which makes use of a two - nucleon potential at the intermediate step .
the puzzle is getting even more intriguing given that the chiral eft calculations tend to indicate less attraction at @xmath14 larger than the physical value @xcite , see also ref .
@xcite for a related work .
these calculations , however , rely on the naturalness assumption and/or make use of resonance saturation estimates for @xmath14-dependent four nucleon contact interactions . in our recent paper
@xcite , we have argued that the low - energy theorems ( lets ) in nn scattering can provide important consistency checks of both lattice - qcd results and their chiral extrapolations .
specifically , the knowledge of the analytic properties of the scattering amplitude allows one to predict its energy dependence and , under certain circumstances , extract the parameters of the effective range expansion . in ref .
@xcite , we have tested this approach by predicting the effective range parameters for @xmath6-wave nn scattering in the spin - triplet and the spin - singlet channels , see also refs .
@xcite for earlier studies along this line .
further , we generalized the lets to unphysical pion masses and applied the resulting approach to selected lattice - qcd results .
in particular , using the linear behavior of the quantity @xmath15 , where @xmath16 refers to the effective range , conjectured in ref .
@xcite , and visualized in fig .
[ fig:3s1eff_rad_mpi ] , we employed the lets to predict the @xmath14 dependence of the deuteron binding energy and of the other parameters in the effective range expansion of the nn scattering amplitude in the @xmath1 channel . remarkably , the resulting @xmath14-dependence of the deuteron binding energy turned out to be in good agreement with the general trend of different lattice calculations @xcite except for the results of ref .
@xcite which do not support the existence of bound states for large pion masses .
( color online ) linear with @xmath14 behavior of @xmath15 in the @xmath17s@xmath18 partial wave conjectured in ref .
red solid triangles correspond to the nplqcd results at @xmath11 mev @xcite while the shaded area shows the uncertainty of the suggested linear interpolation .
the black squares show the empirical values of the effective range at the physical pion mass @xcite .
, scaledwidth=35.0% ] recently , the nplqcd collaboration has reported their new results for the @xmath6-wave nn observables both in the spin - triplet and the spin - singlet channels at @xmath19 mev @xcite .
in particular , they have calculated the values of the deuteron and the dineutron binding energies and the phase shifts in the @xmath1 and @xmath2 channels at several values of the center of mass ( cms ) momenta above threshold . in this paper
we confront these results with the lets . in particular , we demonstrate that the lattice phase shifts at the two lowest energies in the @xmath1 channel are inconsistent ( within the quoted uncertainties ) with the deuteron binding energy obtained in the same lattice calculation .
the situation is less conclusive in the @xmath2 channel due to larger uncertainties of the lets in this channel .
however , the results of our analysis in this channel also indicate an inconsistency between the phase shifts and the large value of the dineutron binding energy reported by the nplqcd collaboration . taking the nplqcd results for the deuteron and dineutron binding energies as input
, we use the lets to infer the corresponding phase shifts and extract the values of the scattering lengths and effective ranges .
we perform a detailed comparison of our results for these quantities with the ones of ref .
@xcite and argue that their determination by means of the effective range approximation is not self - consistent .
our paper is organized as follows . in section [ sec : letphys ] , we discuss in detail our formalism , explain the meaning of the lets and discuss their generalization to unphysical values of the pion mass .
implications of the lets for the recent lattice - qcd results at @xmath20 mev are considered in section [ sec : latticeqcd ] .
section [ sec : rangempi ] addresses the dependence of the effective range on the pion mass .
finally , our main findings are summarized in section [ sec : summary ] .
[ sec : letphys ] the concept of the low - energy theorems for nn scattering and their generalization to unphysical pion masses have been discussed in ref .
@xcite , see also refs .
@xcite for related earlier studies . in this section
we formulate the main idea of the lets from a somewhat different perspective as compared to ref .
@xcite , where a quantum mechanical framework of the modified effective range expansion @xcite was employed .
we assume that the nn interaction is characterized by two distinct scales @xmath21 and @xmath22 , @xmath23 , so that the potential can be written as @xmath24 with the interaction ranges of the order of @xmath25 and @xmath26 , respectively .
the analytic structure of the scattering amplitude near threshold is governed by the long - range interactions .
taking into account the discontinuity across the left - hand cut from the long - range potential @xmath27 , the energy dependence of the scattering amplitude can be predicted in a model independent way up to the energies corresponding to the branch point of a more distant left - hand cut associated with the potential @xmath28 .
this prediction can be regarded as a low - energy theorem .
alternatively , one can view the lets as correlations between the parameters in the effective - range expansion of the inverse scattering amplitude induced by long - range interactions .
notice that the inverse scattering amplitude may possess poles in the near - threshold region , whose appearance does not affect the validity range of the lets if the scattering amplitude is kept unexpanded .
the longest - range part of the nn force is due to the one - pion exchange potential ( opep ) .
thus , the let for nn scattering are expected to be governed by the left - hand cut generated by the opep .
the opep is , however , singular at the origin and requires regularization and renormalization .
therefore , instead of using the quantum mechanical approach , we formulate the let within the modified weinberg approach of chiral eft @xcite . since the correlations between the effective range expansion parameters are inherently long - range phenomena , the results after renormalization and removing the ultraviolet cutoff should be model and regularization - scheme independent .
to be specific , we calculate the scattering amplitude @xmath29 by solving the lippmann - schwinger - type integral equation introduced originally by kadyshevsky @xcite which , for the case of the fully off - shell kinematics , has the form @xmath30 where @xmath31 is the free green function @xmath32 further , @xmath33 ( @xmath34 ) is the incoming ( outgoing ) three - momentum of the nucleon in the cms and @xmath35 with @xmath36 denoting the nucleon mass and @xmath37 being the corresponding ( on - mass - shell ) three - momentum .
the @xmath6-wave potential at leading order ( lo ) consists of the opep and two derivative - less contact interactions ( here denoted @xmath38 for each given partial wave ) @xmath39 where @xmath40 ( @xmath41 ) denote the spin ( isospin ) pauli matrices of the nucleon @xmath42 , while @xmath43 and @xmath44 refer to the axial vector coupling of the nucleon and the pion decay constant , respectively . as discussed in ref .
@xcite , the lo integral equation in this framework is exactly renormalizable scattering see ref . @xcite .
this paper also addresses chiral extrapolations of the @xmath45 binding energy .
] , that is all ultraviolet divergencies appearing from iterations of the lo potential can be removed via an appropriate redefinition of the contact interaction @xmath38 . as a consequence , the cutoff in the integral equation can be put to infinity , so that no finite - cutoff artifacts can affect the let .
the integral equation ( [ meqlok0integrated ] ) with the potential ( [ vlo ] ) can , in general , be only solved numerically .
however , the correlations between the parameters of the effective range expansion implied by the lets can be demonstrated analytically . for simplicity reasons , we consider the partial wave projected @xmath29-matrix in an uncoupled channel with the zero orbital angular momentum , whereas generalizations to the coupled - channel case and to non - zero angular momenta are straightforward . since the lo contact interaction is separable , it is possible to write the solution for the @xmath29-matrix in the semi - analytic ( operator ) form @xmath46 where @xmath47 denotes the off - shell solution of the projected integral equation ( [ g ] ) with the opep alone ( @xmath48 ) , whereas the functions @xmath49 and @xmath50 , shown graphically in fig .
[ fig : blocks ] , involve convolutions of this pionic amplitude with the point - like vertex , namely @xmath51 1 true cm graphical illustration of the functions @xmath49 and @xmath50 in eq . , which correspond to convolutions of the pionic amplitude @xmath52 with the point - like vertices as explained in the text.,title="fig:",scaledwidth=90.0% ] to complete the renormalization program at lo , one has to express the contact interaction @xmath38 in terms of the scattering amplitude at zero momentum , that is the scattering length @xmath53 , which is assumed to be an input quantity .
thus , the on - shell nn amplitude reads @xmath54 where we used the shorthand notation @xmath55 .
for an uncoupled channel with the zero orbital angular momentum , the scattering amplitude @xmath56 can be expressed in terms of the so - called effective range function @xmath57 via @xmath58 we are now in the position to formulate the lets at lo : using the scattering length as the only input quantity to fix the unknown low - energy constant , we employ eqs .
( [ tmaton])-([fk ] ) to predict the momentum dependence of the scattering amplitude and to calculate the phase shifts @xmath59 .
such a prediction is possible because all information about the longest - range ope potential and , in particular , about the discontinuity across the corresponding left - hand cut starting from the momentum @xmath60 , is explicitly taken into account in the calculation . unlike the scattering amplitude
, the effective range function does not possess the kinematic unitarity cut and is a real meromorphic function of @xmath61 near the origin @xmath62 @xcite .
it can , therefore , be taylor - expanded about the origin leading to the effective range expansion @xmath63 where @xmath16 is the effective range , while @xmath64 are the so - called shape parameters .
thus , a single piece of information in the form of the scattering length ( or the energy of a bound / virtual state ) allows one to predict all the coefficients in the effective range expansion @xmath65 where @xmath66 and @xmath67 are polynomials in the inverse scattering length , namely @xmath68 with the coefficients @xmath69 and @xmath70 being calculable from the various quantities appearing in eq .
( [ tmaton ] ) and their derivatives evaluated at @xmath62 .
their explicit form can be easily obtained by performing taylor expansion of the inverse amplitude @xmath71 around @xmath72 .
given that the left - hand cut from the opep is explicitly included , the convergence radius of the lets is restricted by the next - to - lowest - lying left - hand singularity associated with the two - pion exchange potential ( tpep ) . in the present analysis ,
we take into account the contributions of the tpep implicitly by including the relevant momentum - dependent short - range interactions at next - to - leading order ( nlo ) , whose strengths have to be adjusted to reproduce the empirical value of the effective range in the corresponding channel . then , the shape parameters can be predicted up - to - and - including the nlo corrections .
after renormalization , the predictions for the lets become insensitive to details of the short - range interaction once its strength is adjusted to reproduce the physical observable . as argued in ref .
@xcite , it is convenient to employ resonance saturation via a heavy - meson exchange in order to model higher - order contact interaction without destroying explicit renormalizability of the integral equation or having to rely on perturbation theory . specifically , the nlo correction to the potential is taken in the form @xmath73 where the heavy - meson mass @xmath74 is set to be @xmath75 mev and the strength @xmath76 is adjusted to reproduce the empirical value of the effective range in the @xmath2 and the @xmath1 channels .
our results are not sensitive to the functional form of the term parameterizing the subleading short - range interaction in eq .
( [ satur ] ) , see ref .
@xcite for more details .
we now summarize the main findings of ref .
@xcite for the physical value of the pion mass : * in the @xmath1 channel , the lets yield very accurate results already at lo .
for example , the effective range is predicted with an accuracy better than @xmath77 .
the accuracy at lo even appears to be better than one could naively expect from the ratio of scales corresponding to the explicitly included lowest left - hand cut from the opep and the next - to - lowest one from the tpep , which is not considered explicitly at this order .
this observation can be understood by noticing that iterations of the opep do actually generate the dominant contributions to the left - hand cuts due to the two- and multiple - pion exchange . *
the accuracy of the lets in the @xmath2 channel is much worse than in the spin - triplet case .
this has to be expected due to the weakness of the opep in that channel .
indeed , the opep contributes less than @xmath78 to the magnitude of the empirical @xmath2 phase shift at its maximum value . * as expected , the predicted values of the shape parameters at nlo show in both channels a clear improvement as compared with the lo results .
in particular , the nlo lets appear to be accurate at the level of a few percent for the @xmath1 channel ( except for @xmath79 which is unnaturally small ) , while the accuracy of the predictions in the @xmath2 channel is improved to @xmath80 .
a generalization of the lets to the case of unphysical pion masses can be carried out straightforwardly @xcite .
the main dynamical effect of changing the pion mass in the opep corresponds to shifts of the branch points of all left - hand cuts .
the discontinuity across the left - hand cuts also changes due to the dependence of the ratio @xmath81 , which determines the strength of the opep , on the pion mass . ]
finally , the discontinuity across the left - hand cuts is also affected by the @xmath14-dependence of the nucleon mass through its appearance in the integral equation ( [ meqlok0integrated ] ) . to account for the last two effects in a way that minimizes the theoretical uncertainty , we use lattice - qcd results for the @xmath14 dependence of @xmath82 and @xmath36 .
in particular , we performed quadratic polynomial regression fits ( as functions of @xmath83 ) of the lattice - qcd data for pion masses up to @xmath84mev as shown in fig . 3 of ref .
we refer the reader to this paper for more details on the fits and references to the included lattice - qcd calculations of these quantities . using the above results , we can generalize the lo lets by calculating the effective range function @xmath85 , see eq .
( [ fk ] ) , as a function of the ( inverse ) scattering length at _ arbitrary _ values of the pion mass .
using a single input quantity such as the binding energy or the scattering length , we can then predict the phase shifts and extract the effective range @xmath16 and the shape parameters @xmath64 ( provided the effective range function does not have poles near the origin ) .
some exemplary results for such predictions are discussed in ref .
@xcite for different values of the pion mass .
notice that contrary to the chiral extrapolations performed in the framework of chiral eft , no assumptions about the short - range interaction @xmath38 as a function of the pion mass is made when calculating the lets .
instead , we perform fully independent calculations at each given value of the pion mass as if we lived in different worlds characterized by a specific value of @xmath14 . thus , for each considered value of the pion mass , the lec @xmath38 has to be adjusted to reproduce the given value of the scattering length used as input . by providing relations between various low - energy observables at unphysical values of the pion mass
, the lets may serve as consistency checks of the lattice qcd calculations . as already emphasized above , the extension of the lets to nlo
is achieved by including subleading contact interactions parametrized via resonance saturation , see eq .
( [ satur ] ) . retaining the light - quark mass variation in the subleading contact interaction
is formally suppressed according to the chiral eft estimates . on the other hand , allowing for a variation of this term with @xmath14 can be used to estimate the theoretical uncertainty of our analysis . following ref .
@xcite , this is achieved by adjusting the strength @xmath76 of the short - range interaction to reproduce the effective range at the physical point and by assuming that the @xmath83-dependence of @xmath76 is within the envelope built by the straight lines which go through the physical point and describe a @xmath86 change in the value of @xmath76 for @xmath87 mev , i.e. : @xmath88 with @xmath89 and @xmath90 .
such a choice of @xmath91 is motivated by the fact that it would cover the known @xmath14-dependence of @xmath43 , @xmath44 and @xmath36 if the same procedure is applied to these quantities . in the next section
, we will also give results corresponding to the more conservative choice of @xmath92 .
as explained in the previous section , the lets allow one to perform consistency checks of lattice - qcd results for nn scattering provided more than a single observable is extracted .
unfortunately , most of the lattice calculations in the nn sector have so far focused on the determination of the binding energies .
one exception is the work by the nplqcd collaboration at the pion mass of @xmath94 mev @xcite , which provides , in addition to the binding energies , also the values of the scattering length , effective range and even the first shape parameter .
it is , furthermore , conjectured in that paper that the effective range , expressed in units of the pion mass , may be approximated by a linear function of @xmath14 . while the lets are certainly beyond their range of applicability at such heavy pion masses , this conjecture was tested using the lets in our previous work @xcite , where the resulting @xmath95-dependence of the deuteron binding energy was indeed found to be in good agreement with the general trend of lattice data @xcite .
recently , new results for nn scattering in the @xmath1 and @xmath2 channels have been reported by the nplqcd collaboration at @xmath96 mev @xcite .
the calculations were performed for @xmath97 flavors of light quarks at three lattice volumes of @xmath98 fm , @xmath99 fm , and @xmath100 fm using the lattice spacing of @xmath101 fm .
in analogy to their previous work , the scattering phase shifts for the @xmath1 and @xmath2 partial waves were extracted for several values of the cms nn momenta using the extended lscher approach @xcite as shown by the black filled regions in fig .
[ fig:3s1let ] for the case of the @xmath1 channel .
channel calculated on the lattice at @xmath102 mev @xcite ( filled black regions ) in comparison with the predictions based on the lets at lo ( orange light - shaded bands ) and nlo ( blue dark - shaded and hatched blue light - shaded bands ) using the nplqcd result for the deuteron binding energy @xmath103 as input . the uncertainty at lo shown by the orange bands is entirely given by the uncertainty of @xmath103 in eq . .
the nlo dark - shaded ( hatched light - shaded ) bands correspond to the uncertainty in @xmath103 and the theoretical uncertainty of the lets estimated via the variation of @xmath76 with @xmath89 ( @xmath92 ) combined in quadrature .
the grey light- and dark - shaded bands in the right panel depict the fit results of the lattice points of ref .
@xcite based on the effective range approximation .
the energy of the bound ( virtual ) states corresponds to the intersection points of the effective - range function @xmath104 and the unitarity term @xmath105 , shown by the dotted line in the right panel , in the lower ( upper ) half - plane .
the phase shift corresponds to the blatt - biedenharn parametrization of the s - matrix @xcite .
[ fig:3s1let ] , title="fig:",scaledwidth=49.0% ] channel calculated on the lattice at @xmath102 mev @xcite ( filled black regions ) in comparison with the predictions based on the lets at lo ( orange light - shaded bands ) and nlo ( blue dark - shaded and hatched blue light - shaded bands ) using the nplqcd result for the deuteron binding energy @xmath103 as input . the uncertainty at lo shown by the orange bands is entirely given by the uncertainty of @xmath103 in eq . .
the nlo dark - shaded ( hatched light - shaded ) bands correspond to the uncertainty in @xmath103 and the theoretical uncertainty of the lets estimated via the variation of @xmath76 with @xmath89 ( @xmath92 ) combined in quadrature .
the grey light- and dark - shaded bands in the right panel depict the fit results of the lattice points of ref .
@xcite based on the effective range approximation .
the energy of the bound ( virtual ) states corresponds to the intersection points of the effective - range function @xmath104 and the unitarity term @xmath105 , shown by the dotted line in the right panel , in the lower ( upper ) half - plane .
the phase shift corresponds to the blatt - biedenharn parametrization of the s - matrix @xcite .
[ fig:3s1let ] , title="fig:",scaledwidth=49.0% ] in addition to the phase shifts , the binding energies of the deuteron and the dineutron were extracted .
thus , it is interesting to test whether these results fulfill the lets introduced above . the deuteron binding energy calculated in ref .
@xcite at @xmath106 mev at three lattice volumes and extrapolated to the infinite volume is @xmath107 where the errors include statistical and systematic uncertainties as well as the extrapolation uncertainty combined in quadrature .
further , the first two coefficients in the effective range expansion , namely the scattering length and the effective range , were determined in ref .
@xcite by fitting the effective range approximation of the effective range function , @xmath108 to the two lowest - energy scattering data points and the deuteron binding energy , see the grey bands in the right panel of fig .
[ fig:3s1let ] .
notice that all three lattice data correspond to nucleon momenta below the branch point @xmath109 of the left - hand cut from the opep .
the resulting values for the inverse scattering length and the effective range in units of the pion mass reported in ref .
@xcite are @xmath110 where the uncertainties in the first and second brackets are statistical and systematic , respectively . in fig .
[ fig:3s1let ] , we confront the lattice - qcd phase shifts of ref .
@xcite with the predictions of the lets at lo and nlo .
we use the nplqcd result for the deuteron binding energy given in eq . as input to adjust the leading - order contact term @xmath38 .
this is sufficient to predict the phase shift at lo . as explained in the previous section , there are no additional parameters at nlo . as shown in the left panel of fig .
[ fig:3s1let ] , the change in the phase shifts when going from lo to nlo is reasonably small which confirms a good convergence of the lets in this channel .
the expected accuracy of the nlo prediction can be roughly estimated by the width of the blue band generated by the variation of the parameter @xmath76 as described above and appears to be consistent with the shift from lo to nlo .
notice that the lo ( orange ) band reflects the uncertainty in the nplqcd prediction of the binding energy and does not include the theoretical uncertainty of the lets . as required by the levinson theorem for the case of a bound deuteron , the phase shifts generated by the lets go through @xmath111 at the origin . comparing the nplqcd results for phase shifts and the effective range function in the @xmath1 channel with those based on the lets as visualized in fig .
[ fig:3s1let ] , we end up with the following conclusions : * first , as shown in the right panel of fig .
[ fig:3s1let ] , only _ positive _ values of the scattering length appear to be consistent with the nplqcd result for the deuteron binding energy quoted in eq .
( [ bd ] ) as opposed to the negative central value for @xmath112 reported in ref .
our results for the inverse scattering length extracted from @xmath103 by means of the lets disagree with the nplqcd ones given in eq .
( [ ere3s1 ] ) as can be inferred from the right panel of fig .
[ fig:3s1let ] .
* while the lattice phase shifts at higher momenta are in reasonable agreement with the ones predicted by the lets , their low - momentum behavior is incompatible ( within the quoted errors ) with that predicted by the lets as demonstrated in both panels in fig .
[ fig:3s1let ] .
in particular , the phase shift calculated on the lattice at the lowest considered momentum of @xmath113 mev , @xmath114 degrees , is a factor of three smaller than the corresponding value of @xmath115 degrees extracted from the lets . *
an extrapolation of the lattice data to zero momenta in the left panel of fig .
[ fig:3s1let ] seems to indicate that the phase shift goes to zero . this would , however , contradict the existence of a bound state in this partial wave as a consequence of the levinson theorem ( or require shifting @xmath116 by @xmath117 degrees in the entire plotted energy range which would be inconsistent with the lets )
. one may raise a question whether the observed inconsistencies between the lattice - qcd results for phase shifts and the lets predictions could originate from underestimating the quark mass dependence of the nlo contact interaction by constraining the function @xmath118 as described in the previous section . to clarify this issue ,
we have increased the allowed variation of @xmath76 by a factor of two , i.e. we set @xmath119 instead of @xmath89 .
this corresponds to the allowed variation of the strength of the short - range term at @xmath87 mev by @xmath120 as compared to its value at the physical point .
the resulting predictions for the phase shifts and the effective range function are shown by the hatched blue light - shaded bands in fig .
[ fig:3s1let ] . with the resulting uncertainty nearly covering the shift from our lo to nlo results
, we expect such an error estimation to be too conservative .
still , none of our conclusions appear to be affected by employing this very conservative uncertainty estimation .
we are now in the position to employ the lets in order to extract the scattering length and the effective range from the deuteron binding energy calculated by the nplqcd collaboration .
such an extraction is possible because the effective range function does not possess poles at low momenta and , therefore , can be taylor expanded around the origin mev where the phase shift crosses zero but these momenta are already beyond the region of the validity of the effective range expansion . ] . in fig .
[ fig : lets_3s1 ] , we plot the deuteron binding energy and the effective range as functions of the inverse scattering length in units of the pion mass predicted by the lets at lo ( shown by the lines ) and nlo ( shown by the bands ) . , effective range @xmath16 and the binding energy in the @xmath1 partial wave induced by the one - pion exchange potential .
the red solid and dashed magenta lines show the predictions of the lo lets for @xmath121 and @xmath103 .
the light - shaded bands between the red solid and dashed magenta lines visualize the predictions of the nlo lets for @xmath121 and @xmath103 , respectively , and reflect the theoretical uncertainty estimated via the variation of @xmath76 with @xmath89 as described in the text .
the horizontal dotted lines specify the range of values for @xmath103 consistent with the lattice - qcd results of ref .
@xcite for this observable .
the solid dark - red circle ( blue rectangle ) shows the lo ( nlo ) let predictions for the effective range .
the open black circle gives the result for the inverse scattering length and effective range reported by the nplqcd collaboration @xcite while the grey area around it shows the estimated uncertainty from that paper .
all results correspond to the blatt - biedenharn parametrization of the s - matrix @xcite .
[ fig : lets_3s1 ] , scaledwidth=70.0% ] specifically , the red band between two solid lines represents the nlo let calculation for the effective range as a function of the inverse scattering length .
similarly , the magenta band between two dashed lines shows the deuteron binding energy versus the inverse scattering length at nlo .
further , the two horizontal dotted lines separate the region of the binding energies consistent with the nplqcd result of ref .
@xcite , eq . , for the binding energy . projecting this area onto the @xmath122-axis , as shown by the vertical lines ,
one obtains the corresponding values of the scattering length and the effective range from the lets .
in particular , we find @xmath123 which correspond to the following values in units of fm @xmath124 here , the errors in the first brackets reflect the uncertainty in the value of the deuteron binding energy in eq .
( [ bd ] ) used as input . for the nlo results , we also give in the second brackets an estimation of the theoretical uncertainty corresponding to the choice of @xmath89 . clearly , the above values are at variance with those extracted by the nplqcd collaboration and given in eq .
( [ ere3s1 ] ) . in particular , our value for the effective range is about a factor of @xmath125 smaller than the one found in ref .
interestingly , the nlo let prediction for the effective range is in excellent agreement with the assumed linear in @xmath126 behavior of the quantity @xmath127 conjectured in ref .
@xcite , cf .
[ fig:3s1eff_rad_mpi ] and the right panel of fig .
[ fig : eff_rad_mpi ] . for the sake of completeness
, we also give the nlo let results based on a more conservative uncertainty estimation resulting by employing a weaker constraint on the allowed @xmath14-dependence of the subleading contact interaction corresponding to the choice of @xmath128 : @xmath129 or @xmath130 in units of fm .
to understand the origin of the disagreement between our results for the scattering length and effective range with those of ref .
@xcite , it is instructive to take a closer look at the procedure for their determination employed by the nplqcd collaboration . to this aim ,
a fit of the lattice phase - shift data at the two lowest energies and the deuteron pole was performed using the effective - range approximation ( [ eq : ere ] ) .
note that the considered phase - shifts and the deuteron pole correspond to momenta below the branch point of the @xmath131-channel cut due to the opep . for non - singular potentials of a finite range , the applicability region of the effective range expansion is given by the inverse range of the interaction which determines the position of the first left - hand singularity .
consequently , the effective range and shape parameters may be expected to scale with the corresponding powers of the pion mass .
for example , for the physical value of the pion mass , one has @xmath132 and @xmath133 .
the very large value of the effective range reported by the nplqcd collaboration , @xmath134 , either indicates that the range of the nuclear force is considerably larger than that of the opep or signals the appearance of a pole in the effective - range function in the near - threshold region .- wave channel for neutron - deuteron scattering . ] in both cases , the applicability range of the effective range expansion of @xmath135 would be significantly smaller than one may expect based on the position of the left - hand cut due to the opep . as a consequence , the solution for @xmath136 and @xmath137 reported in ref .
@xcite and listed in eq .
( [ ere3s1 ] ) is not self - consistent in the sense that it is obtained by fitting the effective range approximation to the data points outside of its validity region which can be roughly estimated as @xmath138 .
specifically , the deuteron binding momentum at @xmath139 mev is of the order of @xmath140 , whereas the phase - shifts data employed in the analysis correspond to @xmath141 and @xmath142 . to get further insights into this issue ,
consider the two roots of the quadratic equation @xmath143 which determines the pole positions of the scattering amplitude within the effective - range approximation , @xmath144 where we have expanded the square root in powers of @xmath145 and neglected terms of order @xmath146 .
this is justified both for the physical value of the pion mass and for the solution given in eq .
( [ ere3s1 ] ) , since in both cases one has @xmath147 . at the physical pion mass , the second root yields the deuteron binding momentum @xmath148 mev while the first root , @xmath149 mev , lies outside of the applicability region of the effective range expansion and is an artifact of the effective range approximation .
in particular , it disappears or changes the position upon including higher - order terms in the effective range expansion . on the contrary , for the solution in eq .
( [ ere3s1 ] ) at @xmath150 mev , the deuteron pole corresponds to the first root , @xmath151 mev , where the dominant contribution comes from the effective range .
meanwhile , because the scattering length in eq . is negative , the second root corresponds to the momentum , @xmath152 mev , lying on the imaginary axis in the lower half plane .
therefore , the results of ref .
@xcite imply the existence of a shallow virtual state with the excitation energy less than @xmath153 mev in addition to the deuteron , which is not supported by our analysis based on the lets . finally , it is interesting to compare our results based on the lets with the ones obtained using an alternative approach proposed in ref .
@xcite , which will be referred to as bbsvk , where the expansion of the nuclear force around the chiral limit was employed , see , however , ref .
@xcite for a criticism .
this approach was used in ref .
@xcite to calculate the phase shifts and the mixing angle in the @xmath154 channel .
a comparison of results from the two approaches is presented in fig .
[ fig : comp_3s1 ] .
( color online ) neutron - proton phase shifts in the @xmath1 channel ( left panel ) and the mixing angle @xmath155 ( right panel ) at @xmath93 mev based on the lets at lo and nlo in comparison with the results obtained in ref .
@xcite using the eft formulation of ref .
@xcite , labelled as bbsvk , at lo ( light - shaded band between pink dotted lines ) and nlo ( light - shaded band between blue dashed lines ) .
the results for the @xmath17s@xmath18 phase shift correspond to the blatt - biedenharn parametrization of the s - matrix @xcite while the mixing parameter is shown for the stapp parameterization to allow for the comparison with the results of ref .
@xcite . for remaining notation
see fig .
[ fig:3s1let ] .
, title="fig:",scaledwidth=49.0% ] ( color online ) neutron - proton phase shifts in the @xmath1 channel ( left panel ) and the mixing angle @xmath155 ( right panel ) at @xmath93 mev based on the lets at lo and nlo in comparison with the results obtained in ref .
@xcite using the eft formulation of ref .
@xcite , labelled as bbsvk , at lo ( light - shaded band between pink dotted lines ) and nlo ( light - shaded band between blue dashed lines ) .
the results for the @xmath17s@xmath18 phase shift correspond to the blatt - biedenharn parametrization of the s - matrix @xcite while the mixing parameter is shown for the stapp parameterization to allow for the comparison with the results of ref .
@xcite . for remaining notation
see fig .
[ fig:3s1let ] .
, title="fig:",scaledwidth=49.0% ] while the @xmath1 phase shift and the mixing angle show a very similar behavior at lo , there are more sizable differences at nlo .
notice that apart from the different treatment of pions , the two approaches also differ in the way the nlo short - range interaction is taken into account . in particular , in ref .
@xcite , the strength of this subleading short - range term was adjusted to fit the lattice phase shifts .
in contrast , in our approach , the strength of the subleading contact interaction @xmath76 is determined by the value of the effective range at the physical point while its allowed @xmath14 dependence at unphysical pion masses is used to estimate the theoretical uncertainty as explained in sec .
[ general ] .
this procedure ensures that both the lo and nlo let results depend on a single unknown parameter .
we further emphasize that the low - energy behavior of the mixing angle found in ref .
@xcite and shown in the right panel of fig . [ fig : comp_3s1 ] seems to be at variance with the expected threshold behavior for this quantity , @xmath156 ( for details see , e.g. @xcite ) .
irregardless of these differences , the two approaches yield similar numerical results for the @xmath1 phase shift and the mixing angle @xmath155 in the considered range of momenta .
the values of the scattering length and effective range extracted in ref .
@xcite from the lattice data using the framework of ref .
@xcite read @xmath157 where the uncertainties in the first and second brackets correspond to the statistical and systematic uncertainties of the lattice results . as already pointed out , the lo values are in agreement with our lo predictions given in eq .
( [ erenlofm ] ) , while the deviations at nlo and , in particular , the large value of the effective range are presumably caused by an attempt to reproduce the lattice - qcd result for the @xmath1 phase shift at @xmath158 gev within the bbsvk approach .
we now turn to the spin - singlet channel . in fig .
[ fig:1s0_letseb ] , we confront the phase shifts extracted based on the lets with the lattice - qcd results for the @xmath2 partial wave . here
we apply the same procedure as in the @xmath1 channel and use the nplqcd result for the dineutron binding energy , @xcite @xmath159 as input to fix the short - range interaction at lo .
the nlo short - range interaction is again taken into account by means of resonance saturation , see eq .
( [ satur ] ) , with the strength @xmath76 being determined by the effective range at the physical point .
the allowed @xmath14-dependence of @xmath76 is specified by eq .
( [ beta_range ] ) , and the blue dark - shaded bands in fig .
[ fig:1s0_letseb ] correspond to the choice @xmath160 .
notice that the shift in the predictions when going from lo to nlo is now much larger than in the spin - triplet channel which is in line with the lower predictive power of the lets in the @xmath2 partial wave .
consequently , we believe that a variation of the strength @xmath76 with @xmath89 does not provide a realistic estimation of the theoretical uncertainty at nlo in this channel . to have a more conservative estimation , we will allow for a larger @xmath14-dependence in this channel and set @xmath119 as visualized by the hatched blue light - shaded bands in fig .
[ fig:1s0_letseb ] . as shown in fig .
[ fig:1s0_letseb ] , we arrive at similar conclusions as in the case of the spin - triplet channel . while our nlo let predictions for @xmath161 mev are in very good agreement with the phase shifts calculated by the nplqcd collaboration , there is a clear discrepancy for the two lowest values of the momentum @xmath162 .
in particular , for the lowest momentum of @xmath163 mev , the phase shift from the nlo lets is roughly a factor of two larger than that from the lattice - qcd analysis .
similarly to the @xmath1 channel , the predictions of the lets based on the dineutron binding energy are only compatible with positive values of the scattering length , see the right panel of fig .
[ fig:1s0_letseb ] .
specifically , we obtain @xmath164 which correspond to the following values in units of fm @xmath165 here , the errors at lo and in the first brackets at nlo correspond to the uncertainty in the dineutron binding energy while the ones in the second brackets at nlo reflect the unknown @xmath14-dependence of @xmath76 subject to the constraint @xmath119 .
these results are in conflict with the nplqcd determination based on the effective range expansion , namely @xcite : @xmath166 channel calculated on the lattice at @xmath102 mev @xcite in comparison with the predictions based on the lets at lo and nlo using the nplqcd result for the dineutron binding energy @xmath167 as input . for notation see fig .
[ fig:3s1let ] .
[ fig:1s0_letseb ] , title="fig:",scaledwidth=49.0% ] channel calculated on the lattice at @xmath102 mev @xcite in comparison with the predictions based on the lets at lo and nlo using the nplqcd result for the dineutron binding energy @xmath167 as input . for notation
see fig .
[ fig:3s1let ] .
[ fig:1s0_letseb ] , title="fig:",scaledwidth=49.0% ] again , we believe that the analysis performed by the nplqcd collaboration and based on the effective range approximation is not self - consistent .
all arguments given in the previous section apply to the @xmath2 channel too , even though our conclusions in this case are somewhat less stringent due to the lower accuracy of the lets . to further elaborate on this point and to provide an assessment of the robustness of our conclusions ,
we have re - done the calculations by using the lattice phase shifts instead of the dineutron binding energy as input .
specifically , we vary the scattering length , which is now used as input for the lets at nlo , in the range consistent with the lattice - qcd phase shifts at the two lowest energies .
the resulting phase shifts , corresponding to the inverse scattering length in the range of @xmath168 are shown in the left panel of fig .
[ fig:1s0_letsa ] . here , we set @xmath169 , and the width of the band reflects the uncertainty of the lattice - qcd phase shifts used as input .
notice that while the nplqcd value of the inverse scattering length given in eq .
( [ ere1s0nplqcd ] ) is indeed consistent with the range of values in eq .
( [ ainvphases ] ) , the obtained solutions correspond to the bound ( virtual ) state binding energy of @xmath170 mev ( @xmath171 mev ) which is in conflict with the lattice - qcd prediction .
the apparent bound state corresponding to the leftmost intersection point of the gray bands with the unitarity term @xmath172 in the right panel of fig .
[ fig:1s0_letseb ] is an artifact of the effective range approximation . finally , it is interesting to compare our results for the scattering length and effective range with the values obtained in ref .
@xcite within the ksw approach to chiral eft @xcite , namely @xmath173 notice that the effective range vanishes at lo in the ksw approach , and the number of independent parameters fitted to lattice data is equal to one , two and three at lo , nlo and next - to - next - to - leading order ( nnlo ) , respectively .
our nlo let results are in excellent agreement with the nlo ksw values and also nearly consistent with the nnlo ksw results .
channel calculated on the lattice at @xmath102 mev @xcite in comparison with the predictions based on the lets at nlo ( blue shaded bands ) using the scattering length in eq .
( [ ainvphases ] ) as input . for remaining notation
see fig .
[ fig:3s1let ] .
[ fig:1s0_letsa ] , title="fig:",scaledwidth=49.0% ] channel calculated on the lattice at @xmath102 mev @xcite in comparison with the predictions based on the lets at nlo ( blue shaded bands ) using the scattering length in eq .
( [ ainvphases ] ) as input . for remaining notation
see fig .
[ fig:3s1let ] .
[ fig:1s0_letsa ] , title="fig:",scaledwidth=49.0% ]
as already discussed , ref .
@xcite conjectured that the effective range calculated on lattice at @xmath11 mev and expressed in units of the pion mass may be extrapolated to the physical point by a linear function of @xmath14 .
we are now in the position to test this hypothesis by an explicit calculation based on the lets . using the nn bound state energies calculated on the lattice at @xmath174 mev @xcite , @xmath175 mev @xcite , @xmath139 mev @xcite and @xmath176 mev @xcite
, we employ the lets at nlo to predict the values of the effective range in the @xmath177s@xmath178 and @xmath17s@xmath18 partial waves .
the results are visualized in fig .
[ fig : eff_rad_mpi ] . at @xmath11
mev given in ref .
@xcite , @xmath179 , is somewhat different from the one plotted in their fig .
11 and corresponding to the linear extrapolation specified in eq .
( 8) of that work . the lattice - qcd result at @xmath11 mev shown in the left panel of fig .
[ fig : eff_rad_mpi ] is based on the linear extrapolation specified in eq .
( 8) of ref . @xcite . ]
note that the last point in both panels at @xmath11 mev represents the result of lattice calculations by nplqcd @xcite , while the lets are already beyond their range of validity at such heavy pion masses . as seen from the right panel of fig .
[ fig : eff_rad_mpi ] , the nlo lets predictions for the effective range in the @xmath17s@xmath18 partial wave are in very good agreement with the linear in @xmath126 behavior of the quantity @xmath127 .
interestingly , the lattice data point at @xmath11 mev is consistent with the nlo let results linearly extrapolated to higher pion masses . the results from the lets for the @xmath177s@xmath178 partial wave ,
although less conclusive due to larger uncertainties , are also generally consistent with the linear in @xmath14 behavior of @xmath180 .
we note at this point that the lattice data at @xmath11 mev were obtained in ref .
@xcite by using the effective range approximation .
the same procedure was employed by the nplqcd to extract the scattering length and effective range at @xmath139 mev and criticized in this work .
it is conceivable that the data at @xmath11 mev might also suffer from underestimated systematic uncertainties .
( color online ) nucleon - nucleon effective range ( @xmath15 ) in the @xmath177s@xmath178 ( left panel ) and @xmath17s@xmath18 ( right panel ) partial waves predicted based on the next - to - leading order lets using the bound state energies calculated on the lattice as input .
blue open triangles , orange crosses , green open circles and purple solid circles show @xmath15 for the binding energies at @xmath174 mev @xcite , @xmath175 mev @xcite , @xmath139 mev @xcite and @xmath176 mev @xcite , respectively .
the uncertainty of our results for @xmath15 is twofold : the smaller error bars reflect the uncertainty of the lattice results for the binding energies used as input while larger ones correspond to the theoretical uncertainty of the lets estimated by setting @xmath119 and the uncertainty of the lattice results added in quadrature .
red solid triangles correspond to the nplqcd results for @xmath15 at @xmath11 mev @xcite .
the black squares show the empirical values of the effective range at the physical pion mass @xcite .
, scaledwidth=95.0% ]
in this paper , we have employed the low - energy theorems for nn scattering , which have been generalized in ref .
@xcite to the case of unphysical pion masses , to analyze the recent lattice - qcd results at @xmath9 mev reported by the nplqcd collaboration @xcite .
the pertinent results of our work can be summarized as follows .
* we have used the lets along with the lattice - qcd results for the deuteron and dineutron binding energies in order to extract the energy behavior of the nn phase shifts in the @xmath1 and @xmath2 partial waves and the mixing angle @xmath181 at @xmath139 mev .
our lo and nlo calculations suggest a good ( fair ) convergence of our theoretical approach in the spin - triplet ( spin - singlet ) channel . in both channels ,
the resulting phase shifts are in good agreement with the lattice - qcd results of ref .
@xcite for momenta of @xmath182 mev , but are inconsistent with the lattice - qcd predictions at lower energies .
* we have used the lets to extract the values of the scattering length and effective range in the @xmath1 and @xmath2 partial waves from the bound state energies obtained on the lattice .
the extracted value of @xmath183 is in excellent agreement with the linear in @xmath14 behavior of this quantity conjectured in ref .
@xcite . on the other hand ,
our results are in strong disagreement with the values obtained by the nplqcd collaboration from fits to the lattice - qcd data based on the effective range approximation .
we have argued that the very large values for the effective range found in ref .
@xcite make the effective range approximation invalid in the energy region corresponding to the lattice data .
* our results for phase shifts , scattering lengths and effective ranges agree reasonably well with those obtained in ref .
@xcite by analyzing lattice - qcd data within the various eft approaches .
given considerable evidence of a bound dineutron and a stronger bound deuteron at heavy pion masses @xcite , our findings indicate that the lattice - qcd calculations of the nn phase shifts of ref .
@xcite using the extended lscher approach may possibly suffer from underestimated systematic errors at the lowest considered energies .
in addition , using lattice results for the binding energies of the deuteron and dineutron at various pion masses as input @xcite , we demonstrate that the effective range expressed in units of the pion mass behaves as a linear function of @xmath126 .
our work demonstrates that the lets provide a useful tool to analyze lattice qcd results for the nn system by allowing one to extract the scattering phase shifts from the calculated bound state energies and/or test consistency of lattice calculations if several observables are computed .
we would like to thank jambul gegelia for sharing his insights into the topic discussed here and for a careful reading of the manuscript .
we are also grateful to the organizers of the yipqs long - term and nishinomiya - yukawa memorial international workshop on `` computational advances in nuclear and hadron physics '' at yukawa institute for theoretical physics , kyoto university and to the organizers of the int-16 - 1 program `` nuclear physics from lattice qcd '' at int seattle , where a part of this work was carried out .
work supported in part by the erc ( project 259218 nucleareft ) , the bmbf ( verbundprojekt 05p2015 - nustar r&d ) and the dfg ( grant gz : ba 5443/1 - 1 aobj : 616443 ) .
meiner , sci . bull . *
60*(1 ) , 43 ( 2015 ) , [ arxiv:1409.2959 [ hep - th ] ] .
h. oberhummer , a. csoto and h. schlattl , nucl .
a * 689 * , 269 ( 2001 ) , [ nucl - th/0009046 ] .
e. epelbaum _ et al .
_ , phys .
* 110 * , no .
11 , 112502 ( 2013 ) , [ arxiv:1212.4181 [ nucl - th ] ] .
e. epelbaum , h. krebs , t. a. lhde , d. lee and u .- g .
meiner , eur .
j. a * 49 * , 82 ( 2013 ) , [ arxiv:1303.4856 [ nucl - th ] ] .
f. bedaque , t. luu and l. platter , phys .
c * 83 * , 045803 ( 2011 ) , [ arxiv:1012.3840 [ nucl - th ] ] .
j. c. berengut _ et al .
_ , phys .
d * 87 * , no .
8 , 085018 ( 2013 ) , [ arxiv:1301.1738 [ nucl - th ] ]
. s. r. beane _ et al . _
[ nplqcd collaboration ] , phys .
d * 85 * , 054511 ( 2012 ) , [ arxiv:1109.2889 [ hep - lat ] ]
. t. yamazaki , k. i. ishikawa , y. kuramashi and a. ukawa , phys .
d * 92 * , no . 1 , 014501 ( 2015 ) , [ arxiv:1502.04182 [ hep - lat ] ] .
k. orginos , a. parreno , m. j. savage , s. r. beane , e. chang and w. detmold , phys . rev .
d * 92 * , no .
11 , 114512 ( 2015 ) [ arxiv:1508.07583 [ hep - lat ] ] .
s. r. beane , p. f. bedaque , m. j. savage and u. van kolck , nucl .
a * 700 * , 377 ( 2002 ) , [ nucl - th/0104030 ] .
s. r. beane and m. j. savage , nucl .
a * 713 * , 148 ( 2003 ) , [ hep - ph/0206113 ] .
e. epelbaum , u .- g .
meiner and w. glckle , nucl .
a * 714 * , 535 ( 2003 ) , [ nucl - th/0207089 ] .
s. r. beane and m. j. savage , nucl .
a * 717 * , 91 ( 2003 ) , [ nucl - th/0208021 ] .
e. epelbaum , u .- g .
meiner and w. glckle , nucl - th/0208040 .
j. w. chen , t. k. lee , c .- p . liu and y. s. liu , phys .
c * 86 * , 054001 ( 2012 ) , [ arxiv:1012.0453 [ nucl - th ] ] . j. soto and j. tarrus , phys .
c * 85 * , 044001 ( 2012 ) , [ arxiv:1112.4426 [ nucl - th ] ] .
e. epelbaum and j. gegelia , pos cd * 12 * , 090 ( 2013 ) , [ arxiv:1301.6134 [ nucl - th ] ]
. s. r. beane _ et al . _
[ nplqcd collaboration ] , phys .
d * 87 * , no . 3 , 034506 ( 2013 ) , [ arxiv:1206.5219 [ hep - lat ] ]
. s. r. beane _ et al .
_ [ nplqcd collaboration ] , phys . rev .
c * 88 * , no . 2 , 024003 ( 2013 ) , [ arxiv:1301.5790 [ hep - lat ] ] .
e. berkowitz _ et al .
_ , arxiv:1508.00886 [ hep - lat ]
. t. yamazaki , k. i. ishikawa , y. kuramashi and a. ukawa , phys .
d * 86 * , 074514 ( 2012 ) , [ arxiv:1207.4277 [ hep - lat ] ]
. t. inoue _ et al .
[ hal qcd collaboration ] , nucl .
a * 881 * , 28 ( 2012 ) , [ arxiv:1112.5926 [ hep - lat ] ] . v. v. flambaum and r. b. wiringa , phys .
c * 76 * , 054002 ( 2007 ) , [ arxiv:0709.0077 [ nucl - th ] ] .
v. baru , e. epelbaum , a. a. filin and j. gegelia , phys .
c * 92 * , no . 1 , 014001 ( 2015 ) , [ arxiv:1504.07852 [ nucl - th ] ] . t. d. cohen and j. m. hansen , phys .
c * 59 * , 13 ( 1999 ) , [ nucl - th/9808038 ] .
t. d. cohen and j. m. hansen , phys .
c * 59 * , 3047 ( 1999 ) , [ nucl - th/9901065 ] . j. v. steele and r. j. furnstahl , nucl .
a * 645 * , 439 ( 1999 ) , [ nucl - th/9808022 ] .
e. epelbaum and j. gegelia , eur .
j. a * 41 * , 341 ( 2009 ) , [ arxiv:0906.3822 [ nucl - th ] ] .
e. epelbaum and j. gegelia , pos cd * 09 * , 077 ( 2009 ) .
e. epelbaum , arxiv:1001.3229 [ nucl - th ] .
j. j. de swart , c. p. f. terheggen and v. g. j. stoks , nucl - th/9509032 . h. van haeringen and l. p. kok , phys .
* a26 * , 1218 ( 1982 ) .
e. epelbaum and j. gegelia , phys .
b * 716 * , 338 ( 2012 ) , [ arxiv:1207.2420 [ nucl - th ] ] .
e. epelbaum , a. m. gasparyan , j. gegelia and h. krebs , eur .
phys . j. a * 51 * , no . 6 , 71 ( 2015 ) , [ arxiv:1501.01191 [ nucl - th ] ] . v. g. kadyshevsky , nucl .
b * 6 * , 125 ( 1968 ) .
v. baru , e. epelbaum , a. a. filin , j. gegelia and a. v. nefediev , phys .
d * 92 * , no .
11 , 114016 ( 2015 ) , [ arxiv:1509.01789 [ hep - ph ] ] . j. m. blatt and j. d. jackson , phys . rev . * 76 * , 18 ( 1949 ) .
b. midya _ et al .
_ , arxiv:1501.04011 [ quant - ph ] .
m. lscher , commun .
phys . *
105 * , 153 ( 1986 ) .
m. lscher , nucl .
b * 354 * , 531 ( 1991 ) .
r. a. brice@xmath184o , z. davoudi , t. luu and m. j. savage , phys .
d * 88 * , no .
11 , 114507 ( 2013 ) , [ arxiv:1309.3556 [ hep - lat ] ] . j. m. blatt and l. c. biedenharn , rev .
* 24 * , 258 ( 1952 ) .
d. b. kaplan , m. j. savage and m. b. wise , nucl .
b * 534 * , 329 ( 1998 ) , [ nucl - th/9802075 ] .
d. b. kaplan , m. j. savage and m. b. wise , phys .
b * 424 * , 390 ( 1998 ) [ nucl - th/9801034 ] . m. pavon valderrama and e. ruiz arriola , phys . rev .
c * 72 * , 044007 ( 2005 ) . | we apply the low - energy theorems to analyze the recent lattice qcd results for the two - nucleon system at a pion mass of @xmath0 mev obtained by the nplqcd collaboration .
we find that the binding energies of the deuteron and dineutron are inconsistent with the low - energy behavior of the corresponding phase shifts within the quoted uncertainties and vice versa .
using the binding energies of the deuteron and dineutron as input , we employ the low - energy theorems to predict the phase shifts and extract the scattering length and the effective range in the @xmath1 and @xmath2 channels . our results for these quantities are consistent with those obtained by the nplqcd collaboration from effective field theory analyses but are in conflict with their determination based on the effective - range approximation . |
white dwarf ( wd ) stars are stellar remnants composed almost entirely of the inert byproducts of previous nuclear reactions ; they are the burnt - out cores of stars with initial masses below about @xmath7 @xmath4 @xcite .
the majority of wds have an overall mass near @xmath80.6 @xmath4 @xcite .
wds near this canonical mass are expected to harbor remnant carbon - oxygen ( co ) cores after core hydrogen burning and subsequent helium burning .
however , an isolated progenitor star with an initial mass larger than about @xmath9 @xmath4 will reach sufficiently high temperature to achieve stable carbon burning , and may possibly end up as an ultramassive wd with an oxygen - neon ( one ) or oxygen - neon - magnesium ( onemg ) core , if the progenitor had insufficient conditions to start further nuclear burning and detonate as a type ii supernova @xcite .
@xcite found that a 9 @xmath4 progenitor model undergoes repeated carbon - burning shell flashes when its core exceeds @xmath81.05 @xmath4 , ultimately ending up as a wd with an one core , although rotation may also play a role in the outcome of an intermediate - mass progenitor @xcite .
we note that there are other possible formation channels for ultramassive wds , most importantly binary evolution , specifically the merger of double - degenerate systems @xcite .
a handful of ultramassive wds ( @xmath10 @xmath4 ) have been found in nature . @xcite
have reviewed the evidence for ultramassive wds and find much of it compelling .
however , since we can not see below the photosphere of these wds , our understanding of their interiors is essentially superficial . direct evidence that ultramassive wds harbor one cores comes from heavy isotope anomalies found in classical novae , which match predicted abundances from explosive nucleosynthesis on massive wds with onemg cores @xcite .
additionally , two oxygen - rich wds were recently discovered in the sloan digital sky survey ( sdss ; @xcite ) .
these wds are likely exposed one cores , as the observed o / c abundance ratio indicates a very low overall carbon mass fraction , a prediction for some of the most massive progenitors avoiding core collapse @xcite .
a more direct test of ultramassive wd core composition would be to find a massive wd undergoing pulsations .
asteroseismology offers the unique opportunity to use these pulsations to probe below the photosphere and into the interior of stars , and has had numerous successful applications with wds ( see reviews by @xcite ) .
we have thus engaged in a search for pulsations in massive wds in or near the dav ( or zz ceti ) instability strip , a region for which wds with hydrogen - dominated atmospheres have the appropriate temperature to develop a hydrogen partial ionization zone , which in turn drives global pulsations . that search has already yielded multiple new massive davs @xcite after the 1.1 @xmath4 bpm 37093 discovered by @xcite . here
we report a new success in that search : the discovery of the most massive pulsating wd known , gd 518 .
model fits to the optical spectrum first reported by @xcite show this is a @xmath11 @xmath12 k wd with @xmath1 @xmath2 , which would correspond to a mass of @xmath3 @xmath4 using the one wd models of @xcite or a mass of @xmath13 @xmath4 using the co wd models of @xcite . in this letter
we present our discovery of pulsations in gd 518 . in sections [ sec : observations ] and
[ sec : analysis ] we outline our observations and analysis .
we conclude with a discussion of the impact of this finding in section [ sec : discussion ] .
we targeted gd 518 ( wd j165915.11 + 661033.3 ) as a candidate ultramassive pulsating wd based on model atmosphere fits to its optical spectrum . the object was first classified in @xcite .
evolutionary models by @xcite suggest that such a wd has a cooling age of roughly 1.7 gyr and an absolute @xmath14-band magnitude of 13.6 mag , which indicates that gd 518 ( @xmath6=17.2 mag ) is roughly 53 pc from earth .
we display the optical spectrum analyzed by @xcite in figure [ fig : gd518spec ] .
the spectrum was obtained in 2009 march using the 2.3 m telescope at steward observatory , equipped with the boller & chivens spectrograph at a resolution of @xmath86 full width at half - maximum ( fwhm ) , covering a wavelength range from roughly @xmath15 .
a more detailed explanation of the observations , models , and fitting can be found in sections @xmath16 of @xcite .
we obtained additional spectroscopy from the flwo 1.5 m telescope in 2013 april using the fast spectrograph @xcite .
these observations , four 20 minute exposures at a resolution of 1.7 fwhm , cover a wavelength range from @xmath17 .
using the same models and fitting method as in @xcite , we confirm that this wd has a very high surface gravity . although this new summed spectrum is much lower signal - to - noise ratio ( s / n@xmath815 ) , our fits formally yield @xmath11 @xmath18 k and @xmath1 @xmath19 , which agree with the previous determination within the stated uncertainties .
we thus adopt the primary parameters derived from the higher quality ( s / n@xmath855 ) spectrum analyzed in @xcite , displayed in figure [ fig : gd518spec ] .
a @xmath20 k temperature puts gd 518 inside an extrapolated empirical instability strip for davs of high mass @xcite .
in fact , @xcite noted that gd 518 was a `` most intriguing '' candidate to target for possible pulsations .
we obtained time - series photometric observations of gd 518 at the mcdonald observatory over nine nights in 2013 march , eight of them consecutive , and five nights in 2013 april , for a total of more than 42.9 hr of coverage over 33 nights .
a full journal of observations can be found in table [ tab : jour ] .
we used the argos instrument , a frame - transfer ccd mounted at the prime focus of the 2.1 m otto struve telescope @xcite , to obtain @xmath21 s exposures on this @xmath22 mag wd .
the seeing averaged 2.0 and conditions were generally fair .
observations were obtained through a 3 mm bg40 filter to reduce sky noise .
lcccc 2013 mar 10 & 3.1 & 3.9 & 10 & 437.6[4.9 ] + 2013 mar 12 & 3.0 & 2.2 & 5 & 437.6[2.0 ] + 2013 mar 13 & 3.0 & 1.7 & 5 & 418.4[1.8 ] + 2013 mar 14 & 2.3 & 1.4 & 5 & 438.0[4.3 ] + 2013 mar 15 & 2.5 & 1.7 & 5 & 441.0[3.8 ] + 2013 mar 16 & 3.0 & 1.8 & 5 & 441.4[6.5 ] + 2013 mar 17 & 3.4 & 1.9 & 5 & 439.9[6.2 ] + 2013 mar 18 & 3.5 & 2.2 & 5 & 438.5[4.6 ] + 2013 mar 19 & 2.9 & 1.5 & 5 & 437.0[4.1 ] + 2013 apr 4 & 3.6 & 1.8 & 5 & 440.0[5.4 ] + 2013 apr 6 & 4.4 & 2.0 & 5 & 441.1[3.7 ] + 2013 apr 7 & 3.1 & 1.4 & 5 & 524.2[1.6 ] + 2013 apr 9 & 2.6 & 2.5 & 5 & 519.1[3.6 ] + 2013 apr 12 & 2.9 & 1.7 & 5 & 514.2[4.3 ] the raw science frames were calibrated by dark subtraction and flat - fielding .
we performed weighted aperture photometry on the calibrated frames using the external iraf package @xmath23 written by antonio kanaan ( the reduction method is outlined in @xcite ) .
we divided the sky - subtracted light curves by the sum of the three nearest brighter comparison stars in the field to correct for transparency variations , and applied a timing correction to each observation to account for the motion of the earth around the barycenter of the solar system @xcite .
the top panel of figure [ fig : gd518lc ] shows the light curve of gd 518 from 2013 march 16 .
a two - frequency solution to this 3.0-hr run finds variability at @xmath24 s ( @xmath25 mma ) and @xmath26 s ( @xmath27 mma ) .
the bottom panel of figure [ fig : gd518lc ] shows a fourier transform ( ft ) for our entire data set , some @xmath28 points from more than 42.9 hr of observations in 2013 march and april .
we display the 4@xmath29 reference line , calculated from the average amplitude , @xmath29 , of the ft of the entire data set from 0 to @xmath30 @xmath31hz .
the optical light curve of gd 518 shows low - amplitude but statistically significant variability at multiple periods , ranging from roughly @xmath5 s , with amplitudes that can reach up to 0.7% over a single night of observations .
this can be seen by eye in the top panel of figure [ fig : gd518lc ] , as well as in the ft of our entire data set in the bottom panel of that same figure .
we have attempted to identify the periodicities present in the star , which will form the basis for future asteroseismic modeling . complicating our analysis , however , is the fact that the amplitudes ( and perhaps frequencies ) of the observed variability are not consistent from night - to - night .
in fact , the ft for a few nights had no significant peaks above 1.5 mma .
we have included the period ( @xmath32 ) and amplitude ( @xmath33 ) of the highest peaks for each night in table [ tab : jour ] .
there is thus some strong frequency and/or amplitude modulation occurring in gd 518 acting on the timescale of days , perhaps caused by beating of closely spaced periodicities or perhaps due to a physical mechanism in the star .
we have therefore broken up the data into different subsets of minimum length allowed by the frequency splitting in the overall frequency solution : the first five nights ( 2013 mar @xmath34 ) , the second five nights ( 2013 mar @xmath35 ) , and the final nine nights ( 2013 apr @xmath36 ) .
we present a frequency solution for each subset in table [ tab : gd518freq ] .
it was determined by computing an ft , then a nonlinear least - squares fit on the frequency with the highest amplitude , then prewhitening by that frequency until there are no peaks above a 4@xmath29 significance line , which came from the average amplitude of an ft from 0 to @xmath30 @xmath31hz of the unprewhitened data .
we have included the 508 s periodicity in the solution for our first subset even though it is not above 4@xmath29 , based on its presence in other subsets .
for more realistic estimates , the quoted uncertainties in table [ tab : gd518freq ] are not formal least - squares uncertainties to the data but rather the product of 1000 monte carlo simulations of perturbed data using the software package period04 @xcite .
lccc + @xmath37 & 440.2 @xmath38 1.5 & 2271.7 @xmath38 7.6 & + @xmath39 & 513.2 @xmath38 2.4 & 1948.6 @xmath38 9.2 & + @xmath40 & 583.7 @xmath38 1.5 & 1713.3 @xmath38 4.5 & + + @xmath41 & 438.47 @xmath38 0.64 & 2280.7 @xmath38 3.3 & 2.92 @xmath38 0.44 + @xmath42 & 438.098 @xmath38 0.057 & 2282.59 @xmath38 0.30 & 2.24 @xmath38 0.48 + @xmath43 & 508.2 @xmath38 1.5 & 1967.9 @xmath38 5.9 & 1.30 @xmath38 0.29 + + @xmath41 & 439.6 @xmath38 4.5 & 2275 @xmath38 24 & 4.05 @xmath38 0.57 + @xmath42 & 438.89 @xmath38 0.16 & 2278.45 @xmath38 0.82 & 2.57 @xmath38 0.24 + @xmath44 & 440.26 @xmath38 0.25 & 2271.4 @xmath38 1.3 & 2.42 @xmath38 0.39 + @xmath45 & 511.3 @xmath38 2.9 & 1956 @xmath38 11 & 2.0 @xmath38 1.3 + @xmath46 & 509.405 @xmath38 0.099 & 1963.08 @xmath38 0.38 & 1.8 @xmath38 1.6 + @xmath47 & 518.99 @xmath38 0.14 & 1926.82 @xmath38 0.52 & 1.49 @xmath38 0.37 + @xmath48 & 592 @xmath38 33 &
1690 @xmath38 95 & 1.24 @xmath38 0.49 + + @xmath49 & 519.238 @xmath38 0.043 & 1925.90 @xmath38 0.16 & 2.51 @xmath38 0.48 + @xmath45 & 441.244 @xmath38 0.046 & 2266.32 @xmath38 0.23 & 2.38 @xmath38 0.41 + @xmath46 & 440.156 @xmath38 0.062 & 2271.92 @xmath38 0.32 & 2.12 @xmath38 0.40 + @xmath47 & 512.6 @xmath38 5.3 & 1951 @xmath38 20 & 1.63 @xmath38 0.47 + + @xmath41 & 442.12 @xmath38 0.42 & 2261.8 @xmath38 2.1 & 2.38 @xmath38 0.73 + @xmath42 & 441.15 @xmath38 0.17 & 2266.81 @xmath38 0.88 & 2.36 @xmath38 0.72 + @xmath44 & 439.5 @xmath38 1.4 & 2275.5 @xmath38 7.3 & 1.94 @xmath38 0.53 + @xmath43 & 519.2 @xmath38 1.8 & 1925.9 @xmath38 6.7 & 1.55 @xmath38 0.59 + @xmath50 & 440.59 @xmath38 0.47 & 2269.7
@xmath38 2.4 & 1.21 @xmath38 0.52 + @xmath51 & 511.455 @xmath38 0.009 & 1955.207 @xmath38 0.035 & 1.21 @xmath38 0.26 + @xmath52 & 510.824 @xmath38 0.008 & 1957.622 @xmath38 0.032 & 1.13 @xmath38 0.18 + @xmath53 & 437.79 @xmath38 0.13 & 2284.20 @xmath38 0.67 & 1.13 @xmath38 0.46 + @xmath54 & 503.800 @xmath38 0.010 & 1984.914 @xmath38 0.040 & 0.95 @xmath38 0.23 + @xmath55 & 501.44 @xmath38 0.50 & 1994.3 @xmath38 2.0 & 0.93 @xmath38 0.24 + @xmath56 & 426.71 @xmath38 0.86 & 2343.5 @xmath38 4.7 & 0.81 @xmath38
0.34 + @xmath57 & 587.25 @xmath38 0.96 & 1702.8 @xmath38 2.8 & 0.78 @xmath38 0.24 we calculate the chance these detections are real by computing the false alarm probability ( fap ) using the formalism described in @xcite .
we find that all periodicities in each subset have a fap @xmath58% except for @xmath43 in the first subset , which has a fap of @xmath59% .
computing a full frequency solution for our entire dataset , using the same method as we have for each subset , yields 12 formally significant frequencies , many of which are quite closely spaced ( see the bottom panel of table [ tab : gd518freq ] ) . since we observe large - scale amplitude changes over the course of days , we have chosen not to adopt these 12 frequencies as a formal solution , because we can not confirm the coherence of each periodicity .
some frequencies may represent sampling artifacts or frequency drifting rather than truly excited modes in the star .
still , we include these 12 frequencies in table [ tab : gd518freq ] since our full dataset allows us to detect low - amplitude features that may be additional independent periods .
every periodicity in the frequency solution for the entire dataset has a fap @xmath58% except for @xmath57 , which has a 99.8% fap .
we calculate a more conservative overall frequency solution by fitting a lorentzian function to the three significant bands of power for the ft of each of the three subsets , shown in figure [ fig : halves ] .
we adopt the mean of the centroids , weighted by their fwhm , as the overall frequency solution in table [ tab : gd518freq ] , with the uncertainty determined by the standard deviation of the three measurements .
figure [ fig : halves ] shows the fourier analysis for each subset .
we display the original ft for that subset in black , overlaid with the ft prewhitened by the frequencies marked as significant in table [ tab : gd518freq ] .
it is evident that the amplitudes of the variability near 440 s and 513 s ( 2270 and 1949 @xmath31hz , respectively ) change significantly , even over the timescale of a few days .
we have discovered pulsations in gd 518 , which is to date the most massive pulsating wd known .
this star has a mass of roughly 1.2 @xmath4 , derived from model fits to its pressure - broadened balmer lines .
the object offers the best opportunity , to date , to explore the interior of a possible one - core wd using asteroseismology .
since our best current evidence on the high - mass nature of gd 518 rests on its optical spectrum , we have been careful to ensure that this wd truly has high surface gravity .
masses of wds derived from the spectroscopic method , as we have here with gd 518 , show an unphysical upturn in derived surface gravity for effective temperatures below @xmath60 k ( e.g. , @xcite ) .
however , the models used to calculate the surface gravity of gd 518 include improved stark broadening profiles with non - ideal gas effects , which have slightly moderated this upturn @xcite . a full three - dimensional treatment of convection for a wd atmosphere near @xmath61 k and @xmath1 @xmath62 shows that corrections to the one - dimensional models we have used for this spectroscopic analysis do not diverge by more than @xmath63 dex ( p .- e .
tremblay 2013 , private communication ) .
additionally , we do not observe evidence for splitting of the balmer lines caused by a high surface magnetic field , which can sometimes be confused as a high - surface - gravity wd @xcite . fits to follow - up spectroscopy on gd 518 agree with the high surface gravity first reported in @xcite , which confirms a high - mass interpretation for this wd .
additionally , the star is in the footprint of the sdss , and matching @xmath64 colors with synthetic models suggests this is an ultramassive wd @xcite . obtaining a parallax distance to gd 518 will help settle its mass .
there is theoretical support to expect that it will be possible to distinguish the core composition of a massive wd .
@xcite explored the adiabatic pulsational properties of massive wds and found several noticeable differences between co - core models and one - core models of a @xmath65 @xmath4 wd , the only mass they calculated .
their one - core models were characterized by strong deviations in their forward period spacing , and the mean period spacing for their one models was noticeably larger than the mean period spacing for their co - core modes . additionally , the pulsation modes in their one - core models had consistently lower kinetic energies than those in the co - core models . however , the reason they found lower kinetic energies ( and larger period spacings ) for pulsations in their one - core models is that those one - core models were significantly more crystallized ( @xmath66% by mass ) at the same temperature , @xmath67 k , than their co - core models ( @xmath850% by mass ) .
crystallization occurs when the coulomb energy between neighboring ions becomes more than two orders of magnitude larger than the thermal energy of the ions in the wd core , and is a naturally occurring stage of wds as they cool @xcite .
a 1.2 @xmath4 wd should be significantly more crystallized at a similar temperature than a 1.05 @xmath4 wd .
we expect that pulsation energy would be largely excluded from the interior crystallized mass . with less of the stellar material participating in the global pulsations
, it is conceivable that the oscillations have less mode inertia , and can vary on shorter timescales relative to the pulsation periods .
indeed , we observe large amplitude changes in this massive wd ( see figure [ fig : halves ] ) , which may be a consequence of its large crystallized mass fraction .
this relatively short - term amplitude modulation , especially in which pulsation amplitudes fall below detectability , has been seen before in other massive pulsating wds , notably bpm 37093 and sdss j005047.60 - 002316.9 @xcite .
as the highest - mass pulsating wd ever discovered , gd 518 will provide rich insight into the physics of crystallization , as initiated by studies of bpm 37093 @xcite .
however , the degeneracy in parameters caused by crystallization will pose a significant challenge to finding a robust asteroseismic differentiation of the core composition of gd 518 . we expect to need a significant number of observed independent pulsation modes in order to overcome the many free parameters in our asteroseismic fits .
still , we are encouraged by the number of independent periods we have already determined ( at least three ) with a relatively short , single - site campaign . it is possible that the three highest - amplitude periods we observe at @xmath68 s , @xmath69 s , and @xmath70 s are of the same spherical degree ( @xmath71 ) , since they are spaced by roughly 73.0 s and 70.5 s , respectively .
they are unlikely consecutive radial orders .
the 90% crystallized 1.15 @xmath4 co - core models of @xcite find mean period spacings for @xmath72 modes of @xmath73 s , depending on the hydrogen layer mass . likewise , @xcite expect period spacings of roughly 20 s for @xmath72 modes of a 1.05 @xmath4 one - core model .
period spacings for @xmath74 modes should be @xmath75 times longer . aside from the imprint the core chemical profile makes on the pulsation spectrum
, the core composition also affects the rate of cooling for a wd .
this is a consequence of the fact that the cooling time of a wd is inversely proportional to the mean atomic weight of the ions in the core . in some cases we can directly measure this cooling rate by monitoring , long - term , the rate of period change of stable pulsation modes in a dav ( e.g. , @xcite ) . measuring
the rate of period change of any coherent modes will allow another direct test of core composition for this ultramassive wd , albeit a longer - term endeavor .
successfully unraveling both the overall mass and the core composition of gd 518 will constrain intermediate - mass stellar evolution .
it also provides an opportunity to put an upper limit on the primary in a type ia supernovae progenitor system , which theory predicts is a co - core wd , since an onemg - core wd is expected to collapse due to electron capture before detonation as a type ia supernova @xcite .
we thank r. e. falcon and e. l. robinson for useful comments .
this work is supported by the norman hackerman advanced research program , under grant 003658 - 0252 - 2009 , and by the national science foundation , under grant ast-0909107 .
we acknowledge the mcdonald observatory staff for their support , especially john kuehne and dave doss . | we announce the discovery of the most massive pulsating hydrogen - atmosphere white dwarf ( wd ) ever discovered , gd 518 .
model atmosphere fits to the optical spectrum of this star show it is a @xmath0 k wd with a @xmath1 @xmath2 , which corresponds to a mass of @xmath3 @xmath4 .
stellar evolution models indicate that the progenitor of such a high - mass wd endured a stable carbon - burning phase , producing an oxygen - neon - core wd .
the discovery of pulsations in gd 518 thus offers the first opportunity to probe the interior of a wd with a possible oxygen - neon core .
such a massive wd should also be significantly crystallized at this temperature .
the star exhibits multi - periodic luminosity variations at timescales ranging from roughly @xmath5 s and amplitudes up to 0.7% , consistent in period and amplitude with the observed variability of typical zz ceti stars , which exhibit non - radial @xmath6-mode pulsations driven by a hydrogen partial ionization zone .
successfully unraveling both the total mass and core composition of gd 518 provides a unique opportunity to investigate intermediate - mass stellar evolution , and can possibly place an upper limit to the mass of a carbon - oxygen - core wd , which in turn constrains type ia supernovae progenitor systems . |
the observation of a ferromagnetic insulating phase in the low - doped manganites @xmath0 with @xmath10 raised a large number of studies aimed to explain the properties of this phase ( e.g. ) .
several studies have shown that charge and orbital ordering phenomena are crucial to understand the obvious contradiction to the bare double exchange ( de ) model , which predicts metallic behavior in case of ferromagnetic spin order .
recently , resonant x - ray diffraction succeeded to detect the formation of an orbital polaron lattice ( opl ) at low temperatures , which unifies ferromagnetic and insulating properties in a natural way .
it was suggested that local charge hopping processes contribute significantly to the stabilization of the orbital polarons .
further support for this conclusion was found in a recent thermodynamic study which proved the magnetic de energy to be crucial for the stabilization of the opl.@xcite the substitution of la by smaller pr , i.e. the application of chemical pressure , provides a direct route to influence the balance between the various couplings that stabilize the opl in the fmi phase .
more specifically , pr doping causes an increase of the octahedral tilt angles , thereby reducing the mn - o - mn bond angles and the hopping term @xmath9 .
it is well known that this structural change dramatically alters the relative stability of different possible phases .
@xcite in particular , metallic phases with delocalized charges are suppressed upon decreasing @xmath9 .
in contrast to the application of negative chemical pressure , where different samples with different amounts of pr are studied , hydrostatic positive pressure effects can be studied on a single particular compound . here , we present data on the doping dependence of the ordering phenomena in ( la@xmath3pr@xmath4)@xmath5sr@xmath6mno@xmath7 , with @xmath11=0 , 0.1 , 0.25 , 0.5 , 0.75 , and on their hydrostatic pressure dependence in @xmath8 .
our data imply that the opl phase is stabilized by hydrostatic pressure , i.e. upon increasing the charge mobility @xmath9 . whereas , application of negative chemical pressure yields the opposite results .
the verification of the grneisen scaling and the analysis of short range orbital correlations reinforces our conclusions .
our study hence clearly confirms the microscopic opl model .
we mention that recently ferromagnetic correlations have been found in the charge stripe ordered phase of @xmath12 , which might be due to local de , too.@xcite the application of chemical or external pressure might hence also affect the ferromagnetism in the doped nickelates as it is well known in the low - doped manganites .
we report on measurements of the magnetisation , the specific heat , the thermal expansion , and on high energy x - ray diffraction data of @xmath13 single crystals in external magnetic fields up to 16 t .
the experiments were carried out on single crystals , with @xmath11=0 , 0.1 , 0.25 , 0.5 , 0.75 , grown by the floating zone method.@xcite the magnetisation @xmath14 was measured with a squid and a vibrating sample magnetometer .
magnetization measurements under hydrostatic pressure were performed in the temperature range 4.2 - 350k using an mpms squid magnetometer ( quantum design ) and a miniature piston - cylinder pressure cell made of cu - be alloy ( psm 100 , tetcon international ) .
silicon oil was used as the pressure transmitting medium .
constant magnetic fields of 0.05 t in the temperature region 50 - 250k and of 0.5 t were used in the temperature range 250 - 350k for all measurements under pressure .
the pressure at low temperature was determined by the pressure dependence of superconducting transition temperature of high purity pb placed with the sample .
the specific heat @xmath15 was measured using a high resolution calorimeter . here ,
we have applied two different quasi - adiabatic methods , continuous heating and heating pulses .
@xcite for the thermal expansion studies we applied a high resolution capacitive dilatometer.@xcite hard x - ray data have been measured at the beamline bw5 at the hasylab , hamburg . for experimental details
see ref . .
the phase diagram of ( la@xmath3pr@xmath4)@xmath5sr@xmath6mno@xmath7 as it is presented in fig . [ pd ] was established by measurements of the magnetization , the specific heat , the electrical resistivity , and by x - ray diffraction data . for @xmath16 ,
several ordering phenomena are observed upon cooling . at
= 270k , a cooperative jahn - teller effect evolves , and ferromagnetic spin occurs at = 183k .
eventually , the ferromagnetic insulating phase exhibiting the orbital polaron lattice is realized below = 150k . upon pr doping , i.e. by applying chemical pressure , all ordering temperatures significantly change .
( 1 ) the orbital polaron lattice is rapidly suppressed upon pr doping . already for @xmath17 the orbital polaron lattice is considerably destabilized , while it is completely suppressed for @xmath18 .
( 2 ) is suppressed , too , but the effect on the ferromagnetic spin ordering is less than on .
( 3 ) increases upon pr doping , i.e. the cooperative jahn - teller distortions are stabilized . in fig .
[ mag](a ) we show the effect of the spin and the orbital ordering on the magnetization of @xmath8 , under various applied magnetic fields up to 14 t .
the effect of the magnetic field on the opl phase is also illustrated by fig .
[ mag](b ) which displays the field dependence of . at low temperatures
the ferromagnetic opl phase is realized . upon heating
, the orbital polaron lattice melts at @xmath19 110k , at which the transition is associated with a kink in the magnetization @xmath20 .
this kink is visible in fig .
[ mag](a ) e.g. at @xmath21 t .
the measurements in different magnetic fields imply that increases by 22k upon application of @xmath22 t [ see fig .
[ mag](b ) ] .
a similar behavior was observed for @xmath23 ( see fig.6 of ref . ) .
when the temperature is further increased , the ferromagnetic spin order vanishes at @xmath19 168k , and the structural jahn - teller phase transition is observed at @xmath19 296k.@xcite in order to study the effect of ( _ positive _ ) hydrostatic pressure on the orbital and spin ordering phenomena , we studied the temperature dependence of the magnetization of ( la@xmath24pr@xmath25)@xmath26sr@xmath27mno@xmath28 single crystals at various hydrostatic pressures up to 0.82gpa [ see fig .
[ pressure1](a ) ] .
the data show the increase of the magnetization in the ferromagnetic metallic ( fmm ) and in the ferromagnetic opl phase upon application of external hydrostatic pressure , whereas there are only minor pressure effects on the paramagnetic phase at @xmath29 . from the data in fig .
[ pressure1](a ) , we have derived the pressure dependencies of the ordering temperatures , and , respectively , which are displayed in fig .
[ pressure2 ] .
the data show that and increase upon applying hydrostatic pressure , whereas becomes smaller .
thermodynamically , this behavior is in accordance with the fact that the volume of the unit cell shrinks at the jahn - teller transition and increases at and .@xcite the fact that both the magnetization below down to 50k and itself increase as a function of pressure is well understood in the framework of the de model .
enhancement of under hydrostatic pressure is a common trend of perovskite manganites , whereas in the single layered manganites uniaxial pressure is necessary to affect the spin ordering phenomena significantly.@xcite in the case of ( la@xmath24pr@xmath25)@xmath26sr@xmath27mno@xmath28 , the hydrostatic pressure yields an increase of @xmath9 , thereby stabilizing the fmm ordered phase by enhancing the mobility of the charge carriers via the de .
remarkably , the hydrostatic pressure dependence not only of but also of is positive .
this observation proves the important role of the charge hopping @xmath9 and the de for the stabilization of the opl .
the comparison of the magnetic field and the hydrostatic pressure effect on suggests that the application of @xmath30 t is approximately equivalent to the application of 1gpa pressure . at ambient pressure ,
the structural phase transition to the co - operative jahn - teller - distorted phase occurs close to room temperature , i.e. @xmath19 296k , for @xmath17 . under applied hydrostatic pressure , i.e. if the hopping parameter @xmath9 and hence the charge mobility is increased , the cooperative jt distorted phase becomes suppressed and the dynamic jt phase is stabilized .
these effects are just opposite to the chemical pressure effect upon pr doping ( @xmath11=0.1 ) at the la site .
note , that the effect of destabilization of the cooperative jt phase and stabilization of dynamic jt phase is observed in @xmath0 under applied magnetic field , too.@xcite a striking feature of our data is that , at high pressure , increases faster than under pressure .
this means that changes of @xmath9 influence the opl phase more effectively than the fmm phase , i.e. the de promotes the opl even more efficiently than the fmm phase .
faster increase of than was reported under pressure in the low - doped regime around @xmath31.@xcite for small external pressure , however , the effect on is similar to that on , which agrees with the grneisen scaling presented in sec.[secgruen ] .
the experimental data presented in the previous sections show that chemical ( i.e. _ negative _ hydrostatic ) pressure due to the substitution of pr causes the destabilization of the opl phase in @xmath23 , whereas _ positive _ hydrostatic pressure yields the opposite effect .
we have shown experimentally that the hopping @xmath9 significantly contributes to the stabilization of the opl .
this fact provides a natural explanation for the absence of the opl in the fmi - phase of lightly doped la@xmath32ca@xmath33mno@xmath28 and pr@xmath32ca@xmath33mno@xmath28 , where @xmath9 is reduced due to the smaller pr and ca sites . for these systems
there is evidence that a fmi and cooperative jahn - teller distorted phase exists , which corresponds to the low temperature phase of ( la@xmath3pr@xmath4)@xmath5sr@xmath6mno@xmath7 with @xmath34 .
the relevance of the charge hopping @xmath9 for the stabilization of the opl in la@xmath35sr@xmath36mno@xmath7 is further supported by the measurements shown in fig.[fig : dot_vgl_hxs_m ] . in this figure ,
the temperature dependencies of the magnetization measured in an external field of 0.5tesla and the intensities of the superlattice reflections measured by high - energy x - ray diffraction are compared for @xmath37 , 1/8 and 0.14 .
the magnetic entropy in the fmm - phase between and decreases with increasing doping level ; i.e. the magnetic disorder in this phase is largest for @xmath37 and decreases towards @xmath38 .
according to the double exchange ( de ) , the charge hopping in the fmm - phase of la@xmath39sr@xmath40mno@xmath28 is considerably suppressed , as the magnetic disorder is large . at the same time
, the intensity of the superstructure reflection corresponding to the opl vanishes completely in the fmm - phase [ fig.[fig : dot_vgl_hxs_m](a ) ] of this compound .
the increase of the strontium doping to @xmath31 leads to a reduction of the magnetic disorder in the fmm - phase as compared to the @xmath37 sample . in this case
, the superstructure reflection corresponding to the opl is already observed in the fmm - phase [ fig.[fig : dot_vgl_hxs_m](b ) ] , revealing the presence of short range opl order in the fmm - phase of la@xmath5sr@xmath6mno@xmath7 .
similarly , the hxs - measurements on la@xmath41sr@xmath42mno@xmath28 indicate the presence of short range opl correlations above , too [ cf . fig.[fig : dot_vgl_hxs_m](c ) ] . a remarkable observation concerns the temperatures which mark the onset of the short range correlations .
as indicated in fig.[fig : dot_vgl_hxs_m ] , the onset of the short range correlations in the samples with @xmath31 and @xmath38 corresponds in both cases to the temperature where the magnetization reaches about 75% of its saturation value . for @xmath37 ,
the magnetization never reaches the 75% value ( gray horizontal line in fig.[fig : dot_vgl_hxs_m ] ) in the fmm phase of @xmath37 ; i.e. the magnetic disorder is too large to allow for the short range opl order .
the results presented above clearly show that the formation of the opl phase and of the fmm phase is based on a common energy scale which is given by the hopping parameter @xmath9 .
hence , the pressure dependencies of the relevant energy scale for the ordering phenomena must agree , since it is the same energy scale . this can be checked experimentally by the verification of the grneisen scaling , i.e by measuring the anomalous contributions to the specific heat @xmath15 and to the thermal expansion coefficient @xmath43 .
if the grneisen scaling is fulfilled , both quantities are proportional .
quantitatively , the grneisen relation reads @xmath44 with @xmath45 being the relevant energy scale of the respective ordering phenomenon , i.e. in the case at hand @xmath46 .
we hence studied the specific heat and the thermal expansion coefficient of @xmath23 . from both data we subtracted the phonic contribution , which was estimated from the specific heat data as is explained e.g. in ref . .
the resulting anomalous contributions to the specific heat and to the thermal expansion are shown in fig .
[ gruen ] .
as has been shown earlier@xcite , the formation of the fmm phase causes a jump of the specific heat at and a large peak is associated with the onset of the opl phase at .
in addition , there is a regime of anomalous entropy changes at @xmath47 .
all these features are visible in the thermal expansion data in fig .
[ gruen](a ) , too .
there is a peak at , a jump at and a regime of anomalous length changes in the whole fmm phase .
the data hence clearly prove very similar temperature dependencies of @xmath15 and @xmath43 .
this observation implies that both ordering phenomena at and at have the same pressure dependence , which reinforces our conclusions drawn above .
in addition , this is also true for the entropy changes in the whole temperature regime @xmath48 .
this conclusion does not only hold for @xmath23 but also for @xmath37 and 0.14 .
this is illustrated by fig .
[ gruen2 ] , which shows that the grneisen scaling is valid in its integral form for these compositions .
the figure presents the temperature dependencies of the orthorhombicity @xmath49 and of the anomalous entropy changes @xmath50 .
our results show that charge hopping processes are essential to stabilize the opl in la@xmath35sr@xmath36mno@xmath7 .
however , since the opl leads to macroscopically insulating behavior , these hopping processes take place on a local scale , i.e. within the orbital polarons . to be more specific ,
the orbital polarons are ferromagnetic objects which get stabilized by local de processes.@xcite the pressure effects enhance the local de processes by increasing @xmath9 and hence lead to the stabilization of opl at low temperature . the magnetic disorder and the suppression of the orbital polaron lattice created by the pr ( @xmath11=0.1 ) doping is counterbalanced by the external pressure for stabilizing the orbital polaron lattice at low temperatures .
our analysis of the short range orbital correlations and the verification of the grneisen scaling reinforces our conclusions regarding the relevance of @xmath9 and de for the formation of the opl .
+ y. endoh , k. hirota , s. ishihara , s. okamoto , y. murakami , a. nishizawa , t. fukuda , h. kimura , h. nojiri , k. kaneko , s. maekawa , phys .
lett . * 82 * , 4328 ( 1999 ) h. nojiri , k. kaneko , m. motokawa , k. hirota , y. endoh , k. takahashi , phys . rev .
b * 60 * , 4142 ( 1999 ) s. uhlenbruck , r. teipen , r. klingeler , b. bchner , o. friedt , m. hcker , h. kierspel , t. niemller , l. pinsard , a. revcolevschi and r. gross , phys .
lett . * 82 * , 185 ( 1999 ) v.a .
ivanshin , j. deisenhofer , h .- a .
krug von nidda , a. loidl , a.a .
mukhin , a.m. balbashov , m.v .
eremin , phys .
b * 61 * , 6213 ( 2000 ) j. deisenhofer , d. braak , h .- a .
krug von nidda , j. hemberger , r.m .
eremina , v.a .
ivanshin , a.m. balbashov , g. jug , a. loidl , t. kimura , y. tokura , phys .
95 * , 257202 ( 2005 ) r. klingeler , j. geck , r. gross , l. pinsard - gaudart , a. revcolevschi , s. uhlenbruck , and b. bchner , phys .
b * 65 * , 174404 ( 2002 ) t. lorenz , u. ammerahl , t. auweiler , b. bchner , a. revcolevschi , g. dhalenne , phys . rev .
b * 55 * , 5914 ( 1997 ) t. niemller , m. v. zimmermann , j.r . schneider , s. uhlenbruck , o. friedt , b. bchner , p. berthet , l. pinsard , a.m. de leon guevara and a. revcolevschi , eur .
j. b * 8 * , 5 ( 1999 ) j. geck , p. wochner , s. kiele , r. klingeler , p. reutler , a. revcolevschi , m. v. zimmermann , and b. bchner , new . j. phys .
* 6 * , 152 ( 2004 ) j. geck , p. wochner , s. kiele , r. klingeler , p. reutler , a. revcolevschi , and b. bchner , phys .
* 95 * , 236401 ( 2005 ) y. tokura and n. nagaosa , science * 288 * , 462 ( 2000 ) r. klingeler , b. bchner , s - w .
cheong , m. hcker .
b * 72 * , 104424 ( 2005 ) p. reutler , phd thesis , rwth aachen 2003 j. geck , p. wochner , d. bruns , b.bchner , u. gebhardt , s.kiele , p. reutler , and a. revcolevschi , phys . rev .
b * 69 * , 104413 ( 2004 ) .
zhou , j.b .
goodenough , a. asamitsu and y. tokura , phys .
. lett . * 79 * , 3234 ( 1997 ) ; b. martinez , r. senis , ll .
balcells , v. laukin , j. fontcuberta , l. pinsard and a. revcolevschi , phys .
b * 61 * , 8643 ( 2000 ) ; a.e .
petrova , e.s .
itskevich , v.a .
ventcel , v.f .
kraidenov , and a.v .
rudnev , low temp phys .
* 27 * , 831 ( 2001 ) r. klingeler , d. bruns , c. baumann , p. reutler , a. revcolevschi , b. bchner .
magn . mat .
* 290 - 291 * , 944 ( 2005 ) r. kilian and g. khaliullin , phys .
b * 60 * , 13458 ( 1999 ) | we report on the pressure effects on the orbital polaron lattice in the lightly doped manganites @xmath0 , with @xmath1 .
the dependence of the orbital polaron lattice on @xmath2 chemical pressure is studied by substituting pr for la in ( la@xmath3pr@xmath4)@xmath5sr@xmath6mno@xmath7 .
in addition , we have studied its hydrostatic pressure dependence in @xmath8 .
our results strongly indicate that the hopping @xmath9 significantly contributes to the stabilization of the orbital polaron lattice and that the orbital polarons are ferromagnetic objects which get stabilized by local double exchange processes . the analysis of short range orbital correlations and the verification of the grneisen scaling by hard x - ray , specific heat and thermal expansion data reinforces our conclusions . |
the nuclear spins of the host material and the electrons in the dot interact via the fermi contact hyperfine interaction described by @xmath4 as illustrated in figure [ fig : nucspins ] . here
, @xmath5 is the microscopic hyperfine coupling for the different nuclear species @xmath6 , which is weighted by the local electron density @xmath7 , and @xmath8 is the volume of the primitive unit cell containing two nuclei ( in iii - v materials ) .
@xmath9 is the electron spin and @xmath10 the angular momentum operator of the _ _ i__th nuclear spin . .
this polarization influences the electron s larmor frequency.,scaledwidth=40.0% ] in consideration of the large number of nuclear spins interacting significantly with the electron ( typically @xmath11 ) , we describe their effect in terms of a classical effective magnetic field , the overhauser field , @xmath12 where @xmath13 is the effective g - factor , @xmath14 the bohr magneton and @xmath15ev the hyperfine coupling strength . for simplicity , we assume a constant hyperfine coupling strength for all species . the model can be extended straightforwardly to drop this assumption , but this would not change our overall conclusion . due to the high temperature of the nuclear spin bath
, the overhauser field fluctuates around its average .
this field fluctuation can be modelled as a classical , gaussian distributed noise variable with a standard deviation of @xmath16 in gaas quantum dots , where @xmath17 is the maximum overhauser field corresponding to full polarization and @xmath18 the effective number of nuclei in one dot.@xcite for @xmath19 the overhauser field can be expressed as @xmath20 here @xmath21 is a dimensionless local nuclear polarization density and @xmath22 where @xmath23 denotes the total nuclear spin quantum number . in the following only the component of the overhauser field aligned with the external magnetic field is evaluated because this is the principal component seen and influenced by the electron .
the relative polarization density can then be expressed as @xmath24 .
for @xmath25s nuclear dipole - dipole interactions lead to a diffusion - like redistribution of the local nuclear polarization inside and outside the quantum dot.@xcite at magnetic fields exceeding @xmath26mt , hyperfine mediated diffusion can be neglected.@xcite we neglect nuclear spin - lattice relaxation as well because the relaxation time @xmath27 is much longer than the typical correlation time of the overhauser field ( seconds to about 1 minute).@xcite we also neglect the knight shift , which slightly suppresses diffusion in the region occupied by the electron .
the nuclear dipole - dipole interaction can thus be described by the diffusion equation @xcite on distance scales larger than the lattice spacing and time scales longer than the nearest - neighbour dipole - dipole interaction time .
the corresponding dynamics of the relative polarization are given by @xmath28 where @xmath29 is the diffusion constant.@xcite the random nuclear fluctuations can be incorporated via langevin forces .
a spatial fourier transformation of equation gives @xmath30 where @xmath31 is the langevin force acting on the fourier mode with wave vector k and satisfies @xmath32 here , @xmath33 is the magnitude of the frequency independent noise of the langevin forces for each fourier mode.@xcite to obtain the dynamics of the total overhauser field , we rewrite equation in terms of the contributions of the fourier components to the total overhauser field @xmath34 with @xmath35 @xmath36 and @xmath37 denoting the fourier transform of @xmath38 .
this transformation leads to @xmath39 if dnp is applied , a term describing the resulting change in polarization has to be added .
pumping introduces a change of @xmath40 by locally polarizing spins . if feedback is used , the pump rate @xmath41 is a function of instantaneous @xmath40 .
we linearise this relation by introducing the feedback gain @xmath42 and thus obtain @xmath43 as the effective nuclear field is weighted by the fourier transformed electronic charge density @xmath37 , modes with a wave length small in comparison with the spread of the electron wave function are not relevant .
the dominant low-@xmath44 modes on the other hand are slow compared to the feedback time scale since typical decorrelation times of nuclear spin diffusion are seconds , whereas feedback acts in the sub - second regime.@xcite thus the term @xmath45 will be small compared to @xmath46 and can be neglected for all k - modes .
finally the change in the nuclear magnetic field can be described as @xmath47 solving this first - order inhomogeneous linear differential equation in @xmath48 leads to @xmath49 the electronic dephasing time @xmath0 depends on the variance of the overhauser field , @xmath50 substituting @xmath51 from equation and carrying out the integrals , we thus find that the variance in the presence of pumping is given by @xmath52 where @xmath53 is the spectrum of the rate of change of the overhauser field .
consequently the overhauser field variance is directly related to this measure of the strength of the langevin forces and the feedback gain determined by the pump rate .
note that as a result of neglecting the @xmath54-terms , this result is of the same form as previous models not explicitly considering spin diffusion , but with a different expression for @xmath55 .
the feedback gain @xmath46 can be measured ( e.g. see section [ sec : comp ] ) or computed from a microscopic model . to experimentally obtain @xmath55 , we relate it to the fluctuation spectrum of @xmath40 without dnp , @xmath56 , which can be measured independently .
starting from equation and following ref .
, one obtains @xmath57 for a gaussian electron wave function @xmath58\exp\left[-\left(\mathbf{r}^2_\perp/\sigma_\perp^2+z^2/\sigma_z^2\right)/2\right]$ ] with @xmath59 as the vertical and @xmath60 as the lateral spread of the wave function . approximating the @xmath61-dependence of the wave function as gaussion
is not expected to introduce major errors because the exact results for all relevant quantum well potentials ( triangular , rectangular ) also exhibit a quadratic maximum and a fast decay of the wave function , thereby leading to the same qualitative diffusion behaviour.@xcite the resulting noise correlator can be read off as @xmath62 on time scales shorter than the correlation time @xmath63 , the correlator can be approximated to first order as @xmath64 \nonumber \\ & = & \langle b_{\rm{nuc}}^2(0)\rangle- \frac{1}{2}s_{\dot{b}}t \label{eq : nuccor}.\end{aligned}\ ] ] the last form is obtained by again neglecting the @xmath54 terms in equation , which is valid for large @xmath65 corresponding to short times @xmath66 .
the reason for this simple result is that different @xmath44-modes have lorentzian spectra with different corner frequencies @xmath67 , but on the feedback time scale ( shorter than correlation time @xmath68 ) , all relevant modes show an universal @xmath69 behaviour leading to a linear time dependence of the time domain correlator . note that since the different diffusion components have different relaxation times , we can not relate the narrowing of the overhauser distribution to the maximal achievable field by dnp.@xcite in the diffusive case considered here , this maximum field depends on dimension , relaxation rates and diffusion constants .
instead , we employ the directly measurable fluctuation spectrum without dnp .
while we have only considered a single quantum dot so far for simplicity , our model is also applicable to double quantum dots .
for the most widely studied s - t@xmath70 qubits , the nuclear field @xmath40 is replaced by the nuclear field gradient between the two dots , @xmath71 .
as one does not expect correlations between the dots , the total spectrum is the sum of those from each dot .
we now apply our model to the experiment from ref . performed on a gated double quantum dot in a gaas / algaas heterostructure .
dnp was obtained from landau - zener sweeps across the crossing of the singlet state s and the @xmath72 triplet state t@xmath2 , where spin angular momentum can be transferred between electrons and nuclei .
feedback was introduced by preceding each sweep with a free evolution of the qubit driven by the current hyperfine field gradient @xmath73 .
this precession modulates the pump rate as a function of @xmath73 because only one of the two possible qubit states at its end allow spin transfer to the nuclei in the subsequent sweep .
the independently measured noise correlator of the overhauser field fluctuations in the same sample is shown in figure [ fig : nuccor ] . by fitting the noise correlator with equation
, the noise magnitude @xmath55 can be calculated according to equation yielding @xmath74 .
@xmath75 was fixed at 0.32 , and the fitted value of @xmath76 agrees with typical values for @xmath0 . with @xmath77 nuclear spins corresponding to @xmath78 ,
the fit gives a value for the diffusion constant of @xmath79 , which is also reasonable.@xcite via fid measurements with a sampling rate of @xmath80hz .
we plot the difference between the correlator and the variance because this subtraction cancels a relatively large statistical uncertainty in the variance arising from the long correlation time .
the short - time behavior of the correlator is not affected by this uncertainty .
the slope near @xmath81 yields the noise magnitude @xmath74 , which can be used to predict the variance of the overhauser field distribution . ]
the feedback gain @xmath46 can be derived from the dnp - induced rate of change of the nuclear polarization , which was reported as @xmath82 in ref . .
as the feedback pulse was only applied for 30 % of the time , one obtains a mean pump rate of @xmath83 . for the particular feedback implementation
considered , the modulation of this rate with the probability of the free evolution ending in an s state leads to @xmath84 with @xmath85 being the electron larmor frequency and @xmath86 the duration of the precession between singlet and triplet state which leads to the feedback effect .
hence a feedback gain of @xmath87 is obtained with @xmath88 .
equation then leads to an rms width of the narrowed nuclear field gradient distribution of @xmath89 , corresponding to a dephasing time of @xmath90 .
this result is a factor 2 larger than the measured value of @xmath91 .
the discrepancy likely arises partly from the fact that polarization and measurement pulses were each applied for 30 to @xmath92ms , which is too long for the approximation of an average pump rate as used here to be very accurate .
in addition , the actual pump rate during feedback may have been lower than in the measurement used here .
the noise spectrum is mainly determined by dot geometry and independent of the feedback scheme . therefore to increase the effectiveness of the narrowing of the distribution , the feedback gain and accordingly the pump rate should be maximized .
we propose a new polarization scheme , the edsr scheme , based on the work of laird _ et al _ @xcite to improve the feedback efficiency .
hyperfine - mediated edsr can be used to control the nuclear magnetic field in the quantum dot as spin angular momentum conservation requires a simultaneous nuclear and electron spin flip .
an oscillating electric field shifts the electron wave function in the dot , which in turn creates an effective perpendicular ac magnetic field because the electron experiences an oscillating hyperfine field @xmath93.@xcite @xmath93 drives the electron spin with a rabi frequency @xmath94 which depends on the amplitude of the electric field and the hyperfine coupling strength . competing flipping mechanisms such as spin - orbit coupling decrease the nuclear spin flip probability as no nuclear spin is required to flip . however , by aligning the external magnetic field perpendicular to the direction of electron motion , the spin orbit channel can be suppressed.@xcite ( part 1 of the edsr feedback scheme , red ) and the pump curve created by the positively detuned microwave pulse driving @xmath95 ( part 2 of the edsr feedback scheme , blue ) . ( b )
pump curve created by combining the positively and negatively detuned part of the edsr scheme . in green ( black ) the maximum - gain ( most stable ) pump curve yielding a feedback gain @xmath96 ( @xmath97 ) is depicted .
the parameters fulfil the conditions @xmath98 ( @xmath99 ) and @xmath100 ( @xmath101 ) and the exact parameters used are @xmath102 , @xmath103 ( @xmath104 ) , @xmath105 ( @xmath106 ) and @xmath107 .
the ratio of the distance between the pair of void ( filled ) diamonds to the fluctuation width is 12 ( 19 ) so that the feedback is stable . ] to stabilize the overhauser field , two ac electric fields with frequencies each slightly detuned from the desired electron larmor frequency are sequentially applied to drive transitions between the electron up ( @xmath108 ) and the electron down ( @xmath109 ) state in a single quantum dot .
first a @xmath108 state is initialized and a microwave pulse of duration @xmath110 at a frequency negatively detuned by @xmath111 from the targeted electron larmor frequency @xmath112 is applied .
if the nuclear magnetic field decreases and reaches the value corresponding to this detuned frequency , the transition from @xmath108 to @xmath109 is driven .
together with the electron spin a nuclear spin is flipped via the hyperfine interaction and the nuclear magnetic field increases ( part 1 ) .
the rate of change of the nuclear field as a function of the nuclear field ( pump curve ) is illustrated in figure [ fig : pumpmax]a .
next the qubit is initialized in a @xmath109 state and a microwave pulse at a frequency positively detuned by @xmath113 is applied ( part 2 ) .
now the transition from @xmath109 to @xmath108 is driven and the nuclear field decreases if the field strength has reached the value corresponding to this positively detuned frequency .
the pump curve is also shown in figure [ fig : pumpmax]a .
alternating between both parts of the scheme leads to a total pump curve with a stable point at the desired nuclear field @xmath114 as shown in figure [ fig : pumpmax]b .
this dependence of the polarization rate on the nuclear state allows the qubit to act as a feedback loop .
the negative feedback reduces fluctuations by driving the system to the stable point . in order to estimate the effectiveness of the edsr scheme , i.e. the narrowing of the overhauser distribution
, we calculate the pump curve and thus the feedback gain .
firstly the total spin flip probability is evaluated following ref . . because of the parabolic confinement potential , the displacement of the electron wave function is approximately proportional to the electric field . furthermore , since the displacement is much smaller than the spread of the electron wave function , the change in nuclear field is approximately proportional to it .
we assume that for each electron spin also a nuclear spin is flipped .
hence the probability distribution of the rabi frequency arising from the 2d gaussian distribution of the transverse nuclear field is given by @xmath115 , where @xmath116 is the typical rabi frequency .
multiplying the distribution with the edsr spin - flip probability from an initial @xmath109 state , we arrive at a total averaged spin flip probability for hyperfine - mediated edsr of @xmath117 here @xmath118 is the detuning of the microwave from the current larmor frequency , @xmath110 the pulse duration and @xmath119 the rabi frequency relative to the typical rabi frequency .
we see that the pump curve depends on the detuning and pulse duration relative to the typical rabi frequency @xmath120 and @xmath121 , implying that the pump curve and in turn the narrowing is independent of the absolute value of @xmath116 . the probability @xmath122 to flip from @xmath108 to @xmath109
is given by the same expression .
converting this spin - flip probability to a change in nuclear field and adding up the negatively ( part 1 ) and positively ( part 2 ) detuned pulses results in a pump curve of @xmath123\end{gathered}\ ] ] here @xmath124 is the duration of the feedback cycle and @xmath125 the combined detuning of the microwave and the current nuclear field from the desired electron larmor frequency . maximizing the feedback gain with respect to @xmath126 and @xmath110 for rabi frequencies in the order of mhz ( as measured in ref . ) , gives the conditions @xmath127 the resulting pump curve is plotted in figure [ fig : pumpmax ] in green . for @xmath128
this pump curve yields the optimal solution for all sets of parameters @xmath126 , @xmath110 and @xmath116 fulfilling but if @xmath129 due to the initialization time , the effective feedback gain still depends slightly on the individual parameters and is not completely fixed by the conditions . using the largest value from ref . , @xmath102 , for the rabi frequency leads to the parameters @xmath103 and @xmath105 and a feedback gain of @xmath130 .
@xmath124 is typically @xmath131 longer than @xmath110 due to the initialization process . in combination with the noise magnitude from figure [ fig : nuccor ]
, the above feedback gain leads to an rms width of the narrowed nuclear field distribution of @xmath132 ( equation ) and a dephasing time of @xmath3s .
thus , the proposed pump scheme promises an improvement of dephasing time by more than two orders of magnitude in comparison to the intrinsic @xmath133ns .
the nuclear field gradient variation in turn is larger by a factor of @xmath134 .
for the feedback to be stable , i.e. to prevent the mean of the distribution from escaping , the nuclear field fluctuations have to be at least one order of magnitude smaller than the the distance of the global maximum and minimum of the pump curve ( see distance between @xmath135 and @xmath136 in figure [ fig : pumpmax ] ) . although the narrowing is independent of @xmath116 , a smaller @xmath116 deteriorates the feedback stability .
furthermore , it may be advantageous if the second order local maxima and minima are small , i.e. the pump curve has to be smooth .
appropriate parameters for the latter are @xmath99 and @xmath101 , which give a dephasing time of @xmath137 with @xmath102 , @xmath104 and @xmath138 .
the respective curve is shown in figure [ fig : pumpmax ] by the black trace .
the ratio of peak distance to fluctuation width is @xmath139 for the maximum ( green ) curve while it is @xmath140 for the smoother ( black ) pump curve .
thus the stabilization criterion is fulfilled for both curves though the smoother curve is more stable .
so far we have only considered polarization noise arising from the diffusive spin dynamics .
in addition , shot noise generated by the discrete flipping of spins is a noise source intrinsic to the feedback scheme .
the spectral density of this shot noise to be added to the noise from diffusion @xmath55 is given by @xmath141 where @xmath142 is the magnetic field change per nuclear spin flip .
the remaining factor reflects the variance of the binomially distributed number of flips .
the flip probability at the locking point is @xmath143 for the curve leading to the largest pump rate . with an average cycle time of @xmath144s , @xmath145 t and
@xmath146 the noise magnitude is @xmath147 .
this value is comparable to the diffusion noise .
@xmath148 has to be added to @xmath55 in equation .
consequently , @xmath0 is reduced by a factor of order unity .
this shot noise can be reduced by a factor of order unity by choosing a smaller @xmath116 corresponding to a larger @xmath124 and a reduced microwave power without changing the feedback gain @xmath46 .
ultimately , this reduction is limited by the distance between the peaks of the pump curve becoming comparable to the fluctuation width , which will compromise the stability of the feedback scheme .
the edsr pump scheme described for a single dot in section [ sec : polscheme ] can easily be adapted for a double dot system for which the spin states are @xmath149 , where l , r@xmath150 refer to the electron state in the left and right dot , respectively . in this case , the individual magnetic fields in the dots are stabilized so that the field gradient @xmath151 is stabilized as well .
initialization of the qubit in @xmath152 ( @xmath153 ) can be achieved by adiabatic preparation from a singlet @xmath154 , where both electrons are in the same dot , ( and applying a @xmath155-pulse )
. consequently both dots are jointly initialized .
then the microwave pulses from section [ sec : polscheme ] are applied to each individual dot which can be addressed either locally or through frequency selection . for the latter the desired nuclear field gradient @xmath73
has to be large enough so that @xmath152 and @xmath153 are energetically well separated . for the right dot
the transitions driven by the microwave are illustrated in figure [ fig : pumpsequence]a .
the electron is driven into the triplet state @xmath156 ( @xmath157 ) with a positively ( negatively ) detuned microwave pulse if the nuclear field in the right dot has fluctuated downwards ( upwards ) .
consequently the nuclear field in the right dot increases ( decreases ) dependent on the state of the nuclear field in that dot .
the same scheme is applied to the left dot at a different frequency , either simultaneously or sequentially .
a sketch of the pulsing scheme is shown in figure [ fig : pumpsequence]b . charge state for @xmath158 .
the detuned microwave signals are indicated by arrows .
a drift of @xmath73 brings one closer to and the other further from resonance .
( b ) schematic of @xmath159 for the positively detuned ( top ) and negatively detuned ( bottom ) feedback pulses.,scaledwidth=50.0% ] in a double dot system , an alternative possibility to drive transitions is the exchange coupling @xmath160.@xcite at finite @xmath161 the microwave pulse leads to an oscillation of the exchange coupling described by @xmath162 where @xmath163 is the amplitude of the detuning oscillation dependent on the power of the edsr burst , which in turn gives the exchange oscillation amplitude @xmath164 .
this exchange oscillation can be used as an alternative nuclear spin flipping mechanism . without edsr
the transitions @xmath165 are driven by the perpendicular magnetic field difference @xmath166 , but as @xmath167 @xmath168 the transition probability is negligible due to conservation of energy .
this suppression is overcome by the ac - drive , which corresponds to a hamiltonian with a time - dependent variation of the level splitting given by @xmath169 where @xmath170 are the pauli matrices , @xmath171 and @xmath172 .
transforming the eigenbasis for static @xmath173 ( @xmath174 ) and neglecting the time dependence of the level splitting in this basis leads to a standard rabi hamiltonian with a rabi frequency of @xmath175 for a typical external magnetic field @xmath176 , the rabi frequency can be approximated to @xmath177 in order for the exchange flipping mechanism to create a feedback effect , the two eigenstates need to have substantially different components of @xmath153 and @xmath152 , thus requiring @xmath178 . on the other hand
, the exchange energy has to be large enough to drive the transitions as the level splitting becomes @xmath161-independent for @xmath179 and @xmath164 is bounded by @xmath160 . inserting typical experimental parameters for @xmath180 qubits in gaas ( @xmath181 , @xmath182 , @xmath183 yields a rabi frequency of @xmath184 .
the rabi frequency is of the same order of magnitude as the edsr rabi frequency .
however , a spin - orbit contribution to the @xmath185 coupling with a direct @xmath173-dependence @xcite may introduce another strong driving channel in this regime , which would have to be avoided carefully .
in conclusion , we have developed a model that provides a simple relation between the feedback gain of dnp procedures with feedback , the spectrum of nuclear spin fluctuations , and the achievable degree of narrowing of the overhauser field distribution .
the dynamics of the nuclear spins are modelled as a diffusive process .
the results are in agreement with earlier experiments .
we then proposed a new feedback scheme for gated quantum dots based on edsr which should achieve significantly larger pump rates and thus feedback gains .
our model predicts dephasing times of up to @xmath186s for this scheme when considering only the diffusive fluctuations of the z - component of the overhauser field .
interestingly , this value is comparable to the measured value of the hahn - echo coherence time , which was found to be on the order of @xmath187s at sufficiently high magnetic fields.@xcite achieving such long dephasing times would significantly enhance the qubit control fidelity . in practice , one may however find that effects not considered here impose further limitations .
in particular , the faster dynamics of transverse nuclear field components have pronounced effects on hahn echo amplitudes at applied fields of a few @xmath188mt and below , and are expected to affect fid measurements even more.@xcite considerably shorter intrinsic fid times have been predicted when taking transverse terms into account.@xcite to partially overcome this limitation , it may turn out important to synchronize both the feedback scheme and measurements to the relative larmor precession of different spin species . on the other hand , it is conceivable that even better results can be obtained at high fields because of an expected cut - off of the spin diffusion spectrum at frequencies exceeding the dipolar nuclear coupling strength . in any case , experimentally probing the limits of the proposed edsr feedback scheme is likely to yield new insights into electron spin dephasing by the nuclear spin bath as it may give access to the so called `` intrinsic fid '' , which is so far a purely theoretical construct.@xcite we note @xmath0 times for gated dots approaching @xmath189s have already been demonstrated by a measurement based narrowing procedure.@xcite compared to this approach , our method would have the advantage of achieving a predetermined center of the overhauser field distribution rather than narrowing around a random value .
this work was supported by the alfried krupp von bohlen und halbach foundation and dfg grant bl 1197/2 - 1 .
we thank a. yacoby , t. botzem and p. cerfontaine for discussions .
32ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevlett.103.046601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.246602 [ * * , ( ) ] link:\doibase 10.1038/nature08120 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.205302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.216803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.107401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.197403 [ * * , ( ) ] link:\doibase 10.1038/ncomms6156 [ * * , ( ) ] link:\doibase 10.1038/nphys1363 [ * * , ( ) ] link:\doibase 10.1126/science.1146850 [ * * , ( ) ] link:\doibase 10.1038/nphys1366 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.235304 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.140403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.165301 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.79 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.235319 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.236803 [ * * , ( ) ] link:\doibase 10.1038/ncomms4268 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.246601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.035315 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.195301 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.104.236802 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) _ _ , @noop ph.d
. thesis , ( ) link:\doibase 10.1103/physrevb.77.165319 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.78.195302 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.075416 [ * * , ( ) ] link:\doibase 10.1038/nphys1856 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.165315 [ * * , ( ) ] | in many electron spin qubit systems coherent control is impaired by the fluctuating nuclear spin bath of the host material .
previous experiments have shown dynamic nuclear polarization with feedback to significantly prolong the inhomogeneous dephasing time @xmath0 by narrowing the distribution of nuclear overhauser field fluctuations .
we present a model relating the achievable narrowing of the overhauser field to both the pump rate and the noise magnitude and find reasonable agreement with experimental data .
it shows that former experiments on gated gaas quantum dots were limited by the pump rate of the pumping mechanism used .
here we propose an alternative feedback scheme using electron dipole spin resonance .
sequentially applying two ac electric fields with frequencies slightly detuned from the desired larmor frequency results in a pump curve with a stable fixed point .
our model predicts that @xmath0 values on the order of microseconds can be achieved .
one limiting factor of spin qubits based on semiconductor quantum dots is the fluctuation of the effective magnetic field arising from nuclear spins .
the interaction of these nuclear spins in the host material with the electron spin in the quantum dot induces decoherence .
several techniques have been developed to reduce decoherence.@xcite narrowing the distribution of the nuclear fluctuations was shown to prolong the inhomogeneously broadened dephasing time for gate defined gaas quantum dots@xcite and for self - assembled inas quantum dots.@xcite to reduce fluctuations the nuclear field is regulated by dynamic nuclear polarization ( dnp ) .
dnp controls the overhauser field by transferring the electron spin angular momentum to the nuclear spins via the hyperfine interaction .
if the effectiveness of the polarization step depends on the current state of the nuclear spin bath , the qubit can be used as a closed feedback loop . in gate
defined gaas double quantum dots , a roughly tenfold enhancement of @xmath0 has been achieved by landau - zener - sweep driven electron - nuclear spin flip - flops that were preceded by a free evolution of the qubit to create a feedback effect.@xcite in self - assembled inas quantum dots a dnp feedback effect was also observed as the quantum dot resonance exhibited locking to frequency @xcite or repetition rate @xcite of the incident laser . driving gaas gate defined quantum dots with electron dipole spin resonance ( edsr ) showed locking to the spin - resonance condition as well.@xcite the narrowing effect has been examined theoretically in various studies but these were focussed on microscopic models and optical manipulation.@xcite a phenomenological rate - equation model to describe the effectiveness of dnp in reducing the overhauser field variance was derived by vink _
et al .
_ @xcite and by latta _
et al_.@xcite the model relates the maximum nuclear field achievable by polarization to the nuclear field variance by treating the polarization as a random variable influenced by dnp and relaxation proportional to the current value .
this model was further refined by yang _
et al_.@xcite however , the dominant relaxation process in gated dots is spin diffusion of the local polarization to its surroundings , which is not a simple relaxation process.@xcite as a result , the maximum field is determined by the slow long range nuclear dynamics , while the variance of the fluctuations also depends on short range fluctuations .
overall , diffusive dynamics have received little attention , but are of particular interest for gated dots . in this paper
we present a model relating the achievable narrowing of the nuclear field distribution to the pump rate and the noise magnitude by explicitly considering the different relaxation rates for polarization modes with different wave length . to this end
we solve the spin diffusion equation including langevin dynamics .
the results are in reasonable agreement with experimental data and show that the variance of nuclear fluctuations depends directly on the pump rate via the feedback gain .
previous experiments were limited by the pump rate , for example in gate defined gaas dots the pump rate was set by slow landau - zener sweeps at the avoided crossing of the singlet and @xmath1 state ( st@xmath2 scheme).@xcite thus , to overcome the limitation of the st@xmath2 scheme , we propose a different polarization scheme based on the work of laird _
et al_,@xcite the edsr scheme. here an ac electric field is used to flip the spin of the electron in the dot via the hyperfine interaction .
simultaneously a nuclear spin is flipped .
high spin flip rates can be achieved as no time is lost in landau - zener sweeps and bidirectional feedback is possible , thereby eliminating the feedback - free reverse pumping necessary for the st@xmath2 scheme .
these high spin flip rates promise further narrowing of the nuclear field distribution .
our calculations indicate that dephasing times of up to @xmath3s are possible with the edsr scheme .
the remainder of the paper is organized as follows . in section
[ sec : model ] we first briefly review the nuclear dynamics and then present the spin diffusion model relating pump rate and noise . section [ sec : comp ] compares the results of the model to experimental data .
edsr and the new edsr scheme for a single quantum dot are described and the achievable narrowing is evaluated in section [ sec : polscheme ] . section [ sec : dd ] shows how the edsr scheme can be applied to a double dot system with and without exchange coupling .
a summary and discussion of results is given in section [ sec : sum ] . |
investigating the relation between chiral symmetry breaking and physical observables has been the subject of great interest up to this date as these effects can be probed in heavy ion collisions and/or nuclear target experiments .
the restoration of chiral symmetry has been linked to vector meson masses , @xmath1 meson masses , the quenching of @xmath2 and even to the @xmath3 mass ( see , e.g. refs .
@xcite ) . of particular interest is the heavy - light quark meson as it embodies both chiral symmetry breaking and heavy quark symmetry .
recently , one of us in collaboration with other authors employed qcd sum rules to show that the @xmath0 meson mass increases as the chiral symmetry is partially restored @xcite . a qualitatively similar conclusion was obtained in ref .
@xcite ( see however refs .
@xcite for other views ) .
these results from qcd sum rules are consistent with the results obtained in analyses using effective chiral models in refs .
@xcite which are based on the chiral partner structure proposed in ref . @xcite .
at a first glance , an increasing @xmath0 meson mass with a decreasing chiral parameter however seems counterintuitive . within the naive heavy quark limit
, the @xmath0 meson can be thought of as a heavy quark playing the role of a color source , with a constituent quark around it , so that when chiral symmetry is partially restored , the light quark would become lighter making the @xmath0 meson also light .
however , as we will see , if confinement persists , the decreasing light quark will allow the light quark to probe larger distances and hence a higher confining energy inducing a competition between the kinetic term and the confinement term .
the combined effect produces a minimum energy value for a certain constituent quark mass so that when the quark mass decreases below the minimum point , the mass of the @xmath0 meson will increase . in this note
, we would like to highlight this effect and discuss why partial chiral symmetry restoration will not necessarily decrease the mass of the heavy - light system .
we will first try to understand this effect with the help of a simple argument and then confirm the validity of our claim in a realistic and more quantitative constituent quark model calculation . in the constituent quark model , the energy , or the mass , of a meson made of an anti - charm quark and a light quark
can be roughly expressed as @xmath4 where @xmath5 and @xmath6 are masses of the charm quark and the light quark , respectively .
@xmath7 denotes a typical relative momentum between the two quarks , @xmath8 stands for the typical size of the hadron , and @xmath1 is the string tension of the potential between the two quarks .
note that we omit the coulomb part of the potential since it will not change the following naive analysis , as can be seen in the explicit quark model computation of the next section .
@xmath9 is a constant which is used to fit the total energy to the physical mass spectrum .
furthermore , we use @xmath6 as the reduced mass in the kinetic energy term , since @xmath5 is much larger than @xmath6 . for a rough estimation of the mass of the heavy - light meson ,
we first replace @xmath8 in eq .
( [ energy ] ) with @xmath10 , and minimize the energy with respect to the typical moment @xmath7 .
the minimization condition provides @xmath11 which leads to the minimum energy as @xmath12 using a conventional value of @xmath13 , we plot the @xmath6-dependence of @xmath14 as a solid curve in fig .
[ fig1 ] . as can be observed in the figure
, @xmath14 takes a minimum value around @xmath15 .
-dependence of @xmath14 according to eq.([eq : emin ] ) with @xmath16 gev .
the red dots represent the @xmath6-dependence of the @xmath0 meson mass in the full constituent quark model . ]
suppose that @xmath15 in the vacuum and that it reduces when the chiral symmetry is partially restored .
it is natural to assume that @xmath1 and @xmath9 does not change its value so much as it is related to the confinement effect and changes at most by less than 5% at normal nuclear matter density @xcite or at finite temperatures around @xmath17 @xcite . as can be seen in fig .
[ fig1 ] , the mass of the heavy - light meson hence increases with a decreasing chiral condensate .
the increase of the mass for smaller quark masses originates from the increase in the relative separation between the quark , as seen in eq .
( [ rel - p ] ) , that leads to a larger contribution to the confining potential in eq .
( [ energy ] ) .
this result is consistent with that from qcd sum rules @xcite , and from effective chiral models @xcite .
let us next perform a more detailed analysis using a non - relativistic constituent quark model .
we start from the following nonrelativistic hamiltonian .
@xmath18 where @xmath19 and @xmath20 are the quark masses and the color operators of the @xmath21th quark , and @xmath22 and @xmath23 are the confinement and the hyperfine potential between quark @xmath21 and @xmath24 , respectively .
we adopt the following form for the confinement and hyperfine potential @xcite , @xmath25 @xmath26 here , @xmath27 is the distance between quark @xmath21 and @xmath24 , and @xmath28 is chosen to depend on the respective masses as follows : @xmath29 the parameters are determined to give an overall fit to the meson systems in the light and heavy quark sector @xcite . it should be noted that including the pion and sigma exchange potentials is important for a consistent description of three - quark and six - quark states as discussed in refs .
we use the variational method with a single gaussian trial wave function and determine the parameters to reproduce the light - light and light - heavy meson masses . using multiple gaussian functions for the trial wave function leads to negligible change to the fitted masses @xcite .
the resulting parameter values are shown in table [ normalmeson_mass-01 ] .
.parameters fitted to the experimental meson masses using the variational method with a single gaussian .
the respective units are given in the third row .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] [ normalmeson_mass-01 ] the @xmath0 meson mass results of the full quark model calculation with a changing constituent quark mass are shown in fig.[fig1 ] as red dots .
it can be observed in this figure that the behavior of the naive formula of eq.([eq : emin ] ) is qualitatively reproduced by our more accurate calculation . in fig.[d - meson ] , the @xmath0 meson ( blue circles ) and @xmath30 mesons ( red triangles ) masses are given as a function of the constituent quark mass , while fig.[b - meson ] shows the same for @xmath31 and @xmath32 mesons .
-dependence of @xmath0 meson ( blue circles ) and @xmath30 meson ( red triangles ) masses in the constituent quark model . ]
-dependence of @xmath31 meson ( blue circles ) and @xmath32 meson ( red triangles ) masses in the constituent quark model . ]
as can be seen in figs.[d - meson ] and [ b - meson ] , the full constituent quark model calculation shows that starting from the fitted constituent quark mass of 324 mev for the light quark , one finds that for both the @xmath0 and @xmath31 mesons , the meson masses increase when the constituent quark mass decreases .
the main reason for this effect is that with a decreasing constituent quark mass , the kinetic term forces the relative momentum to become smaller and hence the light quark to probe larger distances .
this will thus lead to a larger potential energy coming from the string tension .
this effect is well reflected in the wave function obtained for different constituent quark masses .
the blue solid curve in fig.[d - meson2 ] shows the wave function obtained with the vacuum constituent mass of 324 mev , while the red dashed curve represents that with a mass of 160 mev ; the black dot - dashed curve shows the coulomb and confining part of the potential .
as can be seen in the figure , the wave function for the smaller mass spreads out more and thus probes higher values of the potential energy .
therefore , while a smaller quark mass will decrease the heavy - light meson mass , the contribution from probing larger confinement properties will increase its total mass .
-dependence of the @xmath0 meson radial wave function with @xmath33 mev ( blue solid line ) and with @xmath34 mev ( red dashed line ) .
the black dot - dashed line represents the sum of the coulomb and confining potential . ]
a few comments are in order here .
first , in this simple model we did not introduce any modification of the string tension in relation to the ( partial ) restoration of chiral symmetry .
this is so because in nuclear matter , while the chiral order parameter is expected to be reduced by more than 20% at normal nuclear matter density , the gluon condensate is expected to change by less than 5% . while there is no direct relation between confinement and gluon condensate
, some connection can be made between the wilson loop and gluon condensate via the operator product expansion ( ope ) @xcite .
moreover the abrupt ( slow ) change of the electric ( magnetic ) condensate across the phase transition can be associated with the critical ( soft ) change of the space - time ( space - space ) wilson loop across the same phase transition temperature @xcite .
second , we did not consider any potential @xmath35 splitting in medium .
such an effect could be introduced by including a vector meson exchange type interaction of the @xmath0 mesons with the surrounding medium .
hence , the present mass change should be understood as the average mass shift of these mesons .
third , we expect that the present framework will break down for too small quark masses as the non - relativistic approximation will become invalid .
fourth , the present argument will not be reliable for light - light meson systems as the hyperfine potential @xmath36 in eq .
( [ v - ss ] ) will become unrealistically large as the light quark masses decrease .
@xmath36 in fact will increase at a @xmath37 rate with a decreasing constituent quark mass , eventually leading to a negative pion mass , which does not seem to be realistic . as a last point ,
let us discuss the results of the @xmath30 and @xmath32 mesons , shown as red triangles in figs.[d - meson ] and [ b - meson ] .
it is seen in these figures that these mesons within our model behave almost exactly in the same way as their pseudoscalar counterparts .
this means that our model predicts the @xmath30 and @xmath32 meson masses to increase as chiral symmetry is restored and the constituent quark mass decreases .
the existence of the heavy quark symmetry implies the degeneracy between pseudscalar and vector heavy - light mesons in the vacuum , which is expected to be intact in the dense matter .
then , the analysis here is consistent with the heavy quark symmetry similarly to the result obtained in ref .
it is not yet clear whether this behavior will be consistent with the predictions of qcd sum rules , in which the operator product expansion ( ope ) relates various order parameters of chiral symmetry to certain properties of the spectral function . in the @xmath30 ( or @xmath32 )
channel sum rules , the ope term involving the quark condensate has the same sign as in the @xmath0 ( or @xmath31 ) case @xcite , which could indicate an increasing mass as chiral symmetry is restored , consistent with our model .
there are , however , more terms present in the ope , which could modify this naive expectation .
therefore , only a full qcd sum rule analysis done in the same way as in refs.@xcite will make a consistency check with our model possible .
we leave this topic for future work .
in this work , we have studied masses of heavy - light meson in a dense environment such as nuclear matter , in which chiral symmetry is partially restored . for this purpose ,
we have made use of a simple quark model with one heavy and one constituent light quark . this constituent quark becomes lighter as chiral symmetry is restored and therefore likely changes the mass of the whole heavy - light meson .
naively , one would expect the heavy - light meson to decrease its mass as one of its constituents becomes lighter .
we have however shown in this note that this is not necessarily the case .
assuming that the confining potential remains approximately constant at normal nuclear matter density , we have demonstrated that the wave function spreads out as the constituent quark mass decreases and therefore receives a larger potential energy due to the linearly rising confining potential . for constituent quark masses below about @xmath38 gev , this effect leads to an increase of the heavy - light meson , as shown for instance in fig.[fig1 ] . in our quark model calculations ,
we have examined states containing both a charm and bottom quarks and found the same behavior for both cases ( see figs.[d - meson ] and [ b - meson ] ) . our results for @xmath0 and @xmath31 mesons are consistent with recent sum rule analyses which obtain increasing masses with increasing density @xcite and with calculations based on the skyrme crystal model @xcite . for the future
, it will be interesting to check whether the predictions of our model for the @xmath30 and @xmath32 mesons can be confirmed and reproduced by these more qcd - related approaches .
_ acknowledgements _ shl ap and wp were supported by the korea national research foundation under the grant number krf-2011 - 0020333 and krf-2011 - 0030621 .
the work of mh was supported in part by the jsps grant - in- aid for scientific research ( c ) no .
the work of cn is supported in part by the jsps grant - in- aid for scientific research ( s ) no .
mh and cn are grateful for the hospitality of the members of the institute of physics and applied physics during their stay in yonsei university where the main part of this work was done .
m. harada and k. yamawaki , phys .
lett . * 86 * , 757 ( 2001 ) ; m. harada and c. sasaki , phys .
b * 537 * , 280 ( 2002 ) ; m. harada , y. kim and m. rho , phys .
d * 66 * , 016003 ( 2002 ) ; m. harada and k. yamawaki , phys .
rept . * 381 * , 1 ( 2003 ) ; m. harada and c. sasaki , phys .
d * 73 * , 036001 ( 2006 ) .
r. s. hayano and t. hatsuda , rev .
* 82 * , 2949 ( 2010 ) . s. leupold , v. metag and u. mosel , int .
j. mod .
e * 19 * , 147 ( 2010 ) .
s. h. lee and s. cho , int .
j. mod .
e * 22 * , 1330008 ( 2013 ) .
p. gubler and k. ohtani , phys .
d * 90 * , 094002 ( 2014 ) .
w. park , a. park and s. h. lee , phys .
d * 92 * , 014037 ( 2015 ) .
h. huang , j. ping and f. wang , phys .
c * 89 * , 034001 ( 2014 ) .
a. j. buchmann , g. wagner and a. faessler , phys .
c * 57 * , 3340 ( 1998 ) .
w. park and s. h. lee , nucl .
a * 925 * , 161 ( 2014 ) . | we probe effects of the partial chiral symmetry restoration to the mass of heavy - light mesons in a constituent quark model by changing the constituent quark mass of the light quark . due to the competing effect between the quark mass and the linearly rising potential ,
whose contribution to the energy increases as the quark mass decreases , the heavy - light meson mass has a minimum value near the constituent quark mass typically used in the vacuum .
hence , the meson mass increases as one decreases the constituent quark mass consistent with recent qcd sum rule analyses , which show an increasing @xmath0 meson mass as the chiral order parameter decreases . |
with the new experimental data from the large hadron collider ( lhc ) one will be able to scrutinise the standard model ( sm ) boundaries with unprecedented precision .
flavour - changing neutral currents ( fcnc ) top decays such as @xmath0 , where @xmath1 and @xmath2 are highly suppressed in the sm due to the well known glashow - iliopoulos - maiani ( gim ) mechanism @xcite .
top - quark fcnc branching ratios are of the order of @xmath3 and @xmath4 for the electroweak and strong case respectively .
however , some versions of two - higgs doublet models ( 2hdms ) predict quite larger values .
in fact , for some regions of the 2hdm parameter space the br(@xmath5 ) could reach values of the order of @xmath6 @xcite , which is eight orders of magnitude above the sm value .
the high statistics expected in lhc data forces an increase in the precision of all relevant theoretical calculations .
hence , at least next - to - leading ( nlo ) calculations should be made available whenever possible .
the experimental searches performed both at lhc and at the tevatron on the direct top fcnc process have already used nlo cross sections from @xcite .
however , the nlo calculation was used only as a normalization factor and the events were generated using lo generators like toprex @xcite or protos @xcite . in this work we present a new monte carlo tool that generates fcnc direct top events with nlo precision .
the fcnc interactions were implemented via an effective lagrangian @xcite where a complete set of dimension six operators was used .
the events are generated in the les - houches event ( lhe ) format @xcite and can therefore be easily interfaced with shower algorithms such as the one in pythia @xcite . since the generator calculates the total inclusive cross section , the fcnc single top process was also included and can be generated independently .
not only strong fcnc interactions , but also electroweak and four - fermions dimensions six operators were included . each operator can be turned on independently allowing for independent studies of the different lorentz structures . in section [ sec : effective ] the effective formalism is presented while in section [ sec : processes ] we show the physical processes already included in metop . in section [ sec : nlo ] we give a description of the algorithm used by the monte carlo generator in order to obtain a final nlo result . in section [ results ]
we discuss a few selected results and in section [ sec : conclusion ] we sum up our conclusions .
in the effective lagrangian formalism we assume that the sm is no more than a low energy limit of some more general theory to be revealed at tev scale . in this context
, the top - quark fcnc interaction is seen at low energies as a contact interaction that can be parametrized with higher dimension effective operators .
the complete lagrangian is written as @xmath7 where we have truncated the expansion series at dimension @xmath8 .
we follow the formalism in @xcite , where the sm symmetries are imposed and all operators are built using the already known sm fields .
this results in a reduced list of the possible dimension six operators .
additionally , baryon and lepton number conservation is imposed , which restricts the operators to dimension six only . even imposing the sm symmetries ,
the number of all possible fcnc operators is quite vast .
however , the list can be reduced by using appropriate equations of motion and fierz transformations applied to the physical operators we are dealing with .
the final minimal complete set of fcnc operators for top - quark physics was obtained in @xcite . with this procedure , a single operator survives in the strong sector @xcite . following the notation in @xcite this operator
can be written as @xmath9 where @xmath10 is the gluonic field tensor , @xmath11 right - handed @xmath12 or @xmath13-quark field and @xmath14 is the left - handed quark doublet .
the same operator also appears in the literature as a dimension 5 operator @xmath15 where @xmath16 is a real parameter , @xmath17 , @xmath18 is the strong coupling constant and @xmath19 and @xmath20 are complex numbers which respect @xmath21 .
this operator is responsible for the effective fcnc vertices @xmath22 and @xmath23 .
the minimum set of operators in the electroweak sector is @xmath24 + @xmath25 + @xmath26 + @xmath27 where @xmath28 and @xmath29 are the @xmath30 and @xmath31 field tensors , respectively .
the field @xmath32 is the usual higgs boson doublet .
in addition to this electroweak set of operators , we have included subsets of four - fermion operators which are described in @xcite .
the physical processes included in metop are fcnc direct top production and fcnc single top production .
direct top is available both at lo and at nlo while fcnc single top is availble only at lo .
all processes can be generated for both @xmath33 and @xmath34 hadron colliders .
furthermore , the full top decay chain is included so that spin correlations are preserved .
also , in order to perform spin correlations studies , a on / off switch for the top decay chain was made available . in figures [ fig : dtop ] and [ fig : lqtop ] we show the generic feynman diagrams for the physical processes implemented in metop .
the production process depicted in figure [ fig : dtop ] has only strong fcnc operator contributions . born level fcnc direct top is shown in figure [ fig : dtop ] ( left ) together with the @xmath35 process ( right ) . in figure
[ fig : lqtop ] we present the diagrams for the lo @xmath36 production process from gluon fusion ( right ) and with quarks in the initial state ( left ) .
the former is only affected by strong fcnc operators while the later has contributions from the strong , electroweak and four - fermion sectors .
as mentioned in the previous section , metop is able to generate fcnc direct top events with nlo precision .
the implementation was performed by adopting an nlo effective approximation @xcite . a general nlo calculation of an n - particles final state process
can be represented by @xmath37 where @xmath38 stands for born , @xmath39 for virtual and @xmath40 for real .
it is well known that the virtual term has infrared divergences that will only be cured by the inclusion of the real radiation term @xmath41 .
we show in figure [ fig : dtopnlo ] the class of diagrams contributing to nlo direct top production .
the first process @xmath42 correspond to the born term .
the second , @xmath42 , corresponds to the virtual term while the third , @xmath43 , is the real radiation term .
this last term gives rise to an infrared divergence which correspond to a low @xmath44 and/or to a collinear external gluon .
kinematically , the divergence can be parametrized by the gluon transverse momentum , @xmath44 , and it obviously occurs in the limit @xmath45 . in order to avoid the divergence one could just impose a @xmath44 cut .
the resulting region has to be excluded from the event space generation with some care - since the problem arises only in the collinear / soft limit , we can fill this phase space region with a branching mechanism . by doing so
, we assume a collinear factorization , where the qcd radiation is emitted by one of the external legs of the born configuration .
this approximation can be translated by @xmath46 where @xmath47 and @xmath48 are the real radiation for ( @xmath35 ) and direct top amplitudes , respectively ; @xmath49 is the phase space for the @xmath50th body processes and @xmath51 is the altarelli - parisi splitting function .
the variable @xmath52 is a resolution parameter which in our case is the top quark transverse momentum @xmath53 .
the hard @xmath44 region will then be filled by using the appropriate transition amplitude and the low @xmath44 region by the use of a shower mechanism . in this nlo effective approximation
@xcite , the @xmath44 phase - space is then divided in two regions parametrized by @xmath53 , which we call @xmath54 .
this parameter plays the role of a matching parameter where the criterion for a good transition is the smoothness of the final @xmath44 distribution .
the final cross section free of divergences can be written as @xmath55 where @xmath56 is the transverse momentum of the parton shower emission .
the contribution from the virtual term @xmath57 in equation [ eqn : sigmanlo ] is included via a k - factor applied to the born term and because born and virtual terms share the same @xmath58 topology we assume they have similar kinematics .
it is from equation [ eqn : kfactor ] that events in metop are generated . since metop is a parton level generator , the events must then be submitted to a parton shower simulator in order to include initial ( isr ) and final ( fsr ) state radiation .
therefore , since these parton showers are based on a branching mechanism , we leave the collinear and/or soft @xmath44 phase space region to be filled by the corresponding parton shower . in order to avoid double counting
we just have to ensure that the parton shower will not populate the @xmath59 region in the first emission . to prevent it ,
the shower mechanism must be performed with the pt - ordered scheme with the first emission starting at @xmath60 . for the nlo direct top process
, metop assumes that the pythia shower will be used with its pt - ordered scheme @xcite flag on .
finally , we note that because this is a @xmath58 process , the shower mechanism implemented in pythia will not add fsr but only isr .
this does not pose any problems because fsr is negligible when compared to isr due the large top - quark mass .
in figure 4 we show the diagram flow of the metop generation process .
it was drawn specifically for the nlo direct top case .
the remaining processes follow the same flow except for the k - factor calculation step .
the amplitudes used to evaluate the cross sections were generated with calchep @xcite , whit feynman rules derived with lanhep @xcite and all integrations were performed using the cuba library @xcite .
the k - factors are calculated `` on the fly '' and therefore each sub process has its own normalization factor .
= [ rectangle , draw , fill = blue!20 , text width=20em , text centered , rounded corners , minimum height=1.3em ] = [ draw , -latex ] ( init ) initialize : read paramt.txt card ; ( identify ) calculate possible subprocesses ; ( evaluate ) evaluate each cross section : born , virtual and real radiation ; ( kfactor ) evaluate each subprocess k - factor ; ( generate ) generate events according to equation [ eqn : kfactor ] ; ( save ) write evetns in lhe format ; ( init )
( identify ) ; ( identify ) ( evaluate ) ; ( evaluate )
( kfactor ) ; ( kfactor ) ( generate ) ; ( generate ) ( save ) ; [ flow : nlodtop ] in figure [ fig : nomatch ] we present the top - quark @xmath44 distribution for direct top after the first emission and with the starting scale shower at @xmath61 ( black solid line ) .
the blue dashed line represents the top @xmath44 distribution of the real radiation process .
as explained before , both regions overlap and therefore a matching must be introduced in order to avoid double counting .
the matching process must be chosen so that the transition between the rescaled direct top distribution and the real radiation process is smooth . in figure
[ fig : ptmatch510 ] we show the top @xmath44 distribution for @xmath62 gev where direct top is the gray dashed line and @xmath35 is the blue dashed line .
the first one corresponds to the top distribution after the first isr emission , where the pt - shower starting scale was set to 10 and 15 gev , respectively ; the second one is the top distribution from the t+g process where a @xmath63 of 10 and 15 gev was imposed .
we have studied several @xmath54 scenarios in the 5 gev to 20 gev range finding no significant differences .
we therefore recommend that this matching value should be included in any analysis as a systematic error .
distribution of top quark after the first isr branching with a @xmath64 of 10 gev ( left ) and 15 gev ( right ) .
, title="fig:",width=288 ] distribution of top quark after the first isr branching with a @xmath64 of 10 gev ( left ) and 15 gev ( right ) .
, title="fig:",width=288 ] in figure [ fig : lovsnlotop ] we show the final nlo direct top quark @xmath44 and @xmath65 distribution ( solid line ) , as well as the lo result ( dashed line ) after full isr , fsr and multiple interactions ( mp ) .
it is clear that in both distributions the final nlo result does not amount to simply multiplying the lo result by a k - factor .
in fact , the distributions show that the nlo contribution concentrates more events in the low @xmath44 region and produces top quarks at higher angles . in figure
[ fig : lovsnlolepton ] we show the equivalent distribution for the lepton coming from the decaying w. the @xmath66-quark and neutrino distributions show similar differences between the nlo and lo results . for the top decay products , the nlo contribution does not result in a major change in distributions making a k - factor adjustment suitable .
however , one must pay special attention to analysis where the top quark momentum is reconstructed .
for example , analysis which include the study of the top spin polarisation usually requires full reconstruction of its momentum . in this case , figure [ fig : lovsnlotop ] show that nlo events should be considered .
( left ) and @xmath65 ( right ) distributions of the top quark at the partonic level after the full shower ( isr+fsr ) and multiple interaction . ,
title="fig:",width=288 ] ( left ) and @xmath65 ( right ) distributions of the top quark at the partonic level after the full shower ( isr+fsr ) and multiple interaction .
, title="fig:",width=288 ] ( left ) and @xmath65 ( right ) distributions of the lepton from @xmath67 at the partonic level after the full shower ( isr+fsr ) and multiple interaction . ,
title="fig:",width=288 ] ( left ) and @xmath65 ( right ) distributions of the lepton from @xmath67 at the partonic level after the full shower ( isr+fsr ) and multiple interaction . , title="fig:",width=288 ] finally , when studying the inclusive nlo direct top production one must add the @xmath68 ( single top fcnc ) process as well .
the distribution of the sum of the two processes is shown in figure [ fig : ptdtopnloq ] for the top @xmath44 ( right ) and the top @xmath65 ( left ) .
the fcnc single top process was generated with a @xmath69 gev and only subprocesses where a fcnc interaction takes place were included .
processes like @xmath70 which are pure sm single top production were discarded . in tables
[ tbl : dtopnlo ] and [ tbl : lqtop ] we show the total contributions from each lo and nlo processes .
they were calculated for @xmath71 and for lhc @ 7 tev .
as we can see from figure [ fig : ptdtopnloq ] , the fcnc single top contribution does not produce considerable shape modifications , although its contribution to the total cross section is approximately 13% .
( left ) and @xmath65 ( right ) distributions of the top quark for nlo direct top ( solid line ) and nlo direct top plus @xmath72 with @xmath73 gev and jet @xmath74 gev . , title="fig:",width=288 ] ( left ) and @xmath65 ( right ) distributions of the top quark for nlo direct top ( solid line ) and nlo direct top plus @xmath72 with @xmath73 gev and jet @xmath74 gev .
, title="fig:",width=288 ] lll subprocess & lo & nlo + @xmath75 & 6.12 & 8.74 + @xmath76 & 0.91 & 1.67 + lll subprocess & @xmath77 & @xmath78 + @xmath79 & 1.12 & 0.40 +
we have presented a new monte carlo generator dedicated to top fcnc physics .
we have included the main fcnc top production channels : direct top and single top .
the direct top channel is available at nlo and the single top at lo .
a complete set of dimension six operators are already included .
we have shown that the nlo top distributions obtained for the direct top can not be built via a k - factor from the lo ones .
therefore nlo events should be used especially in the cases where top - quark momentum reconstruction is required .
99 glashow s l , iliopoulos j and maiani l 1970 weak interactions with lepton - hadron symmetry _ phys .
d _ * 2 * 1285 aguilar - saavedra j a 2004 top flavor - changing neutral interactions : theoretical expectations and experimental detection _
acta phys .
b _ * 35 * 2695 [ arxiv : hep - ph/0409342 ] .
liu j j , li c s , yang l l and jin l g 2005 next - to - leading order qcd corrections to the direct top quark production via model - independent fcnc couplings at hadron colliders _ phys .
d _ * 72 * 074018 slabospitsky s r and sonnenschein l 2002 toprex generator ( version 3.25 ) : short manual _ comput
commun . _ * 148 * 87 aguilar - saavedra j a 2010 zt , gamma t and t production at hadron colliders via strong flavour - changing neutral couplings _ nucl . phys .
b _ * 837 * 122 buchmller w and wyler d 1986 effective lagrangian analysis of new interactions and flavor conservation _ nucl .
b _ * 268 * 621 alwall j _ et al .
_ 2007 a standard format for les houches event files _ comput .
commun . _ * 176 * 300 sjostrand t , mrenna s and skands p 2006 pythia 6.4 physics and manual _ j. high energy phys . _ jhep0605(2006)026 ferreira p m , oliveira o and santos r 2006 flavour changing strong interaction effects on top quark physics at the lhc _ phys .
d _ * 73 * 034011 ferreira p m and santos r 2006 strong flavour changing effective operator contributions to single top quark production _ phys . rev .
d _ * 73 * 054025 aguilar - saavedra j a 2009 a minimal set of top anomalous couplings _ nucl .
b _ * 812 * , 181 grzadkowski b , iskrzynski m , misiak m and rosiek j 2010 dimension - six terms in the standard model lagrangian _
j. high energy phys . _
jhep1010(2010)085 coimbra r , onofre a , santos r and won m 2012 metop - a generator for single top production via fcnc interactions _ eur .
j. c _ * 72 * 2222 boos e e , bunichev v e , dudko l v , savrin v i and sherstnev a v 2006 method for simulating electroweak top - quark production events in the nlo approximation : singletop event generator , _ phys .
nucl . _ * 69 * 1317 sjostrand t and skands p 2005 transverse - momentum - ordered showers and interleaved multiple interactions _ eur .
j. c _ * 39 * 129 [ hep - ph/0408302 ] .
pukhov a 2004 calchep 2.3 : mssm , structure functions , event generation , batchs , and generation of matrix elements for other packages _ preprint _ arxiv : hep - ph/0412191 semenov a 1998 lanhep : a package for automatic generation of feynman rules from the lagrangian _ comput .
commun . _ * 115 * 124 semenov a 2009 lanhep - a package for the automatic generation of feynman rules in field theory .
version 3.0 _ comput .
commun . _ * 180 * 431 hahn t 2005 cuba : a library for multidimensional numerical integration _
commun . _ * 168 * 78 | in this work we present a new monte carlo generator for direct top and single top production via flavour - changing neutral currents ( fcnc ) .
this new tool calculates the cross section and generates events with next - to - leading order precision for the direct top process and leading - order precision for all other fcnc single top processes .
a set of independent dimension six fcnc operators has been implemented - including four - fermion operators - where at least one top - quark is present in the interaction . |
age - dating the stellar populations of galaxies provides us with a cosmic timescale which is independent of cosmological models , and with a mean of reconstructing their star - formation history .
moreover , the same population synthesis tools used for age dating provide also an estimate of the stellar mass of galaxies .
thus , deriving star formation histories and masses for large numbers of galaxies at various redshifts allow us to attempt an empirical reconstruction of the formation epochs and mass - assembly history of galaxies .
theoretical models for galaxy formation and evolution can then be adjusted in order to comply with the galaxy evolution scenario emerging from the observations . clearly , the evolutionary population synthesis tools play a critical role in this process , and we must pay as much attention as possible to ensure that they provide ages and masses as accurate as possible . in this paper we focus on the problem of age and mass determinations for galaxies which are dominated by stars in the age range @xmath5 gyr , because this is a range where different population synthesis tools give quite discrepant results .
moreover , at redshifts beyond 2 ( or 3 ) the universe itself is younger than @xmath6 ( or @xmath7 ) gyr , and galaxies at such high redshift are bound to contain stars younger than @xmath7 gyr .
hence , the use of an accurate stellar population tool is essential for the correct interpretation of the properties of high redshift galaxies , right at a cosmic time when different renditions of the hierarchical galaxy formation paradigm diverge dramatically ( see e.g. , fig . 9 in
* ) . in stellar populations with ages @xmath5 gyr , stars in the advanced evolutionary stage known as
the thermally - pulsing asymptotic giant branch ( tp - agb ) contribute up to @xmath8 of the bolometric light , and up to @xmath9 of the near - ir ( * ? ? ?
* hereafter m98 ) .
the first appearance of tp - agb stars in the life of a stellar population has been called the agb-_phase transition _
@xcite , and its direct effect is a strong increase of the near - ir flux , as tp - agb stars emit most of their radiation at @xmath10 m . with the advent of the spitzer space telescope ( sst ) ,
the rest - frame near - ir is now accessible to very high - redshifts , which allows one an improved analysis of the spectral energy distribution ( sed ) of high redshift galaxies , and makes it even more important than before to dispose of reliable stellar population tools .
the following concrete example illustrates the case . for a sample of high-@xmath0 galaxies with photometric redshifts @xmath11 and with data extending from the rest - frame uv to near - ir ( from sst ) yan et al .
( 2004 ) estimated an age of 1.5 to 3.5 gyr for the dominant stellar population ( @xmath12 by mass ) , with the residual trace of young stars as due to a recent star - formation episode .
the yan et al .
result was obtained using the synthetic population tool by ( * ? ? ?
* hereafter bc03 ) .
some of the same galaxies were re - analysed by ( * ? ? ?
* hereafter m05 ) using her own synthetic stellar populations , obtaining ages @xmath13 gyr , right in the middle of the epoch when tp - agb stars dominate . moreover , the derived stellar masses are smaller when using maraston models ( @xmath14 versus the typical value of @xmath15 quoted by yan et al . 2004 ) . whereas both results may be astrophysically plausible , worth emphasising is that their implications concerning the formation redshift for the bulk of the stars and the galaxy mass assembly are vastly different .
the difference is likely the result of the different weight given to the tp - agb in the two sets of models , with the tp - agb contribution being systematically lower in the bc03 models compared to the models of maraston ( 2005 ) .
when using bc03 models a massive contribution of old stars is required in order to match the observed near - ir fluxes , and in the meantime a sprinkle of young stars is necessary to account for the rest - frame near - uv . on the contrary , a two - component stellar population
is not required by the m05 models , as a single age ( younger ) population accounts at once for the sed from the near - uv to the near - ir .
it is interesting to note in this context the early result of @xcite , who found for a sample of @xmath16 cluster galaxies that synthetic near - ir colours not including the agb phase were not red enough to match the data at given optical colour .
the problem was cured by adding agb effects on top of these early models using the fuel consumption theorem . besides the treatment of the tp - agb , the bc03 and m05 models also differ for the set of stellar isochrones used to construct the synthetic populations , which may account for part of the difference in the results . in this paper
we investigate further on the effect of stellar population recipes on the modeling of galaxy seds , and we do so by applying the two sets of models to infer the mass and age ( or best - fit star - formation histories ) to a specific sample of high - redshift galaxies .
note that the results obtained with the bc03 models are representative of what would be obtained with other widely used models , like pgase @xcite and starburst99 @xcite due to similar recipes for the tp - agb phase and identical input stellar evolutionary tracks ( m05 ) .
in section 2 we recall the main differences between the bc03 and the m05 models , whereas section 3 describes the main observational characteristics of the galaxies selected for this study . in section 4
we describe the fitting procedure used to estimate the properties of these galaxies and the results are presented and discussed in section 5 .
finally , section 6 is dedicated to a general discussion of the main results of this investigation .
we adopt the current _ concordance cosmology _ , with @xmath17 , @xmath18 and @xmath19)$ ] equal to 0.7 , 0.3 and 0.7 , respectively .
the age of the best - fit model is required to be lower than the age of the universe at the given spectroscopic redshift .
the bc03 and the m05 stellar population models differ by the following aspects : i ) the stellar evolutionary models used to construct the isochrones ; ii ) the treatment of the tp - agb phase ; and iii ) the procedure used for computing the integrated spectra .
this latter point will not be discussed separately here ( see m05 ) . for full details on what below
, we refer the reader to m05 .
* input stellar models*. the bc03 models are based on the padova stellar tracks ( e.g. , * ? ? ?
* ) , whereas the m05 models are mostly based on the frascati stellar tracks ( e.g. , * ? ? ?
the two set differ mostly for : a ) the padova tracks include a certain amount of convective overshooting on the main sequence ( ms ) , whereas the frascati tracks were constructed assuming no overshooting ; b ) the temperature distribution of the red giant branch ( rgb ) phase , that is shifted to cooler temperatures in the padova tracks for solar metallicity and above , a result of the different calibration of the mixing - length theory used for the modeling of envelope convection .
convective - core overshooting has two effects on stellar models .
first , it prolongs the ms lifetime for given stellar mass , while leaving the effective temperature almost unchanged .
this results from the extended convective core ensuring more hydrogen fuel to the central regions where nuclear burning takes place .
thus , for given age of the ms turnoff , isochrones constructed from models with overshooting are hotter , therefore bluer , than those constructed from models without overshooting .
second , overshooting delays the appearance of stars with a degenerate helium core , and the accompanying development of the rgb .
for example , the rgb develops at an age of @xmath20 gyr in the frascati models , and at an age of @xmath21 gyr in the padova models ( * ? ? ?
the development of the rgb makes the sed redder , so in the age range @xmath22 gyr models with overshooting are bluer , as they lack a well - developed rgb .
the net effect is that theoretical seds based on overshooting models are bluer therefore give systematically older ages compared to those based on models without it .
furthermore , by having a reduced contribution by the rgb , population models using stellar tracks with overshooting have lower luminosity for given age and mass in stars , i.e. , have higher @xmath23 ratios , hence higher galaxy masses are derived for given luminosity . in summary ,
the net effect of overshooting is that for a given sed the bc03 models indicate older ages and higher masses compared to m05 models .
of course , this applies to population ages such that ms stars do have a convective core , i.e. , for ages younger than a few gyr . while it is likely that overshooting makes convective cores somewhat more extended in mass , the size of this extension remains somewhat conjectural . on the one hand there exists no firm prediction from first principles , like in most situations involving turbulence . for stellar populations between @xmath24 and @xmath25 gyr , the corresponding stellar masses at the ms turnoff range from @xmath7 to @xmath26 , with the mass of the convective core decreasing with decreasing mass , and eventually vanishing for @xmath27 , the precise value depending on composition .
thus , what one would need to know is the amount of convective overshooting ( e.g. , expressed as a fraction of the mass of the convective core ) as a function of the stellar mass . this unknown function can only be empirically constrained by observations . in a recent attempt , based on fitting the shape of the color - magnitude of some open clusters ,
@xcite concluded that overshooting would amount to @xmath28 of a pressure scale height , of course vanishing as the convective core itself vanishes .
an overshooting about this size is also indicated by the asteroseismology of a @xmath29 cep star @xcite , and by and large is consistent with the amount assumed in the padova models . on the other hand , observations of lmc star clusters
rule out the late rgb development implied by the padova tracks @xcite . to decide what is the size of overshooting as a function of stellar mass goes well beyond the aims of this paper , and we shall limit ourself to exploring the differences in the results when using different sets of models . *
the temperature of the rgb*. whereas for ages less than @xmath24 gyr the m05 code gives redder seds compared to the bc03 code , after the completion of the rgb phase transition this tendency is reversed and the mo5 seds are bluer ( figure 27 in m05 ) .
this is due to the warmer rgbs in the frascati stellar tracks compared to the padova ones ( figure 9 in m05 ) , which in turn results from different calibrations of the mixing - length parameter used in modeling envelope convection .
the ages derived with the m05 models are _ older _ than those obtained from the bc03 models . *
the tp - agb phase*. the treatment of the tp - agb phase is perhaps the main difference between the m05 and the bc03 models .
the contribution of the tp - agb stars to the integrated light of a synthetic stellar population critically depends on what is adopted for the stellar mass loss taking place during this phase .
the higher the assumed mass loss , the sooner the star loses its envelope , and the sooner the tp - agb phase is terminated .
once more , there is no theory relating the mass loss rate to the basic stellar parameters , and once more one has to rely on empirical calibrations . bc03
adopt the mass loss calibration by @xcite , tuned to reproduce the maximum tp - agb luminosity of a sample of magellanic cloud clusters from @xcite . in m05 instead , what is calibrated is the fractional contribution of the tp - agb to the total bolometric light , and this calibration is made essentially on the same mc clusters using the frogel et al .
sharing basically the same observational calibrating data , the two codes should give consistent results .
as shown in m05 they do not , and the effective contribution of the tp - agb is much larger in m05 compared to bc03
. full illustration of the differences is given in m05 , where it is also shown that other population synthesis models , namely pgase and starburst99 behave very much like the bc03 ones .
again , rather than tracing the origin of the discrepancy , we focus here on the effects on the study of high redshift galaxies .
summing up the effects of both overshooting and tp - agb , the m05 models are _ brighter _ and _ redder _ than the bc03 models for ages between @xmath30 and @xmath7 gyr . here
the use of the m05 templates implies the derivation of lower ages and stellar masses . at older ages
the m05-templates are instead _ bluer_.
for the present test we use the sample of seven galaxies in the hubble ultra deep field ( hudf ) which were singled out by the bzk criterion for @xmath31 passively -evolving galaxies @xcite .
these galaxies show early - type morphologies on hudf images and have spectroscopic redshifts between 1.4 and 2.5 ( * ? ? ?
* hereafter d05 ) .
redshifts were determined from the mg+fe absorptions at rest - frame @xmath32 producing a characteristic shape dubbed @xmath33 feature by d05 , on the hst+acs grism spectra acquired in the framework of the grism acs program for extra - galactic science project ( grapes , * ? ? ? * ) .
the presence of the @xmath33 feature imply that the uv / optical spectra of these galaxies are consistent with being dominated by a- or f - type stars .
therefore these objects are not dominated by ob type stars like typical star - forming galaxies and are predominantly passive .
we choose these particular sample for this study because we might expect that the typical ages of stars in these galaxies are close to the critical 0.2 - 2 gyr range where the different evolutionary synthesis models differ by the largest amount .
this is therefore an ideal sample of high redshift galaxies to search for direct evidence of the effects from tp - agb stars .
three of the seven galaxies were assigned a lower quality ( `` b '' class ) redshift by d05 , partly for the non perfect agreement with the photometric redshifts .
this is now improved in most cases with the addition of the irac photometry to the seds and the m05 models ( see table 2 and figure 9 ) .
stellar population properties for these galaxies were derived in d05 by fitting their seds ( ten bands from @xmath34 to @xmath35 ) to bc03 templates spanning a wide range of parameters .
reddening was restricted to be @xmath36 and the calzetti s law @xcite was adopted . in summary ,
the galaxies were found to be @xmath37 1 gyr old , to be in passive evolution since at least @xmath38 gyr , to have de vauculeurs light profiles and large stellar masses ( @xmath39 ) .
they were therefore interpreted as progenitors of massive ellipticals .
the physical sizes of some of the objects ( @xmath40 kpc ) were found to be significantly smaller than objects with similar masses in the local universe , which remained as an open issue in d05 . in the following
we take advantage of the spitzer irac ( and mips ) imaging obtained over the ultra deed field ( udf ) as a part of the great observatories origins deep survey ( goods ) survey ( dickinson et al .
, in preparation and chary et al .
, in preparation , respectively ) .
all the seven galaxies were detected with high signal - to - noise in the four irac bands ( see figure [ images ] ) .
irac photometry was obtained using @xmath41 aperture magnitudes , corrected to total using aperture corrections for unresolved sources .
the optical to near - ir seds of d05 were based on sextractor mag_auto apertures .
we measured large @xmath42 diameter magnitudes in the @xmath35-band to account for the flux lost by mag_auto , to perform the match with the irac magnitudes . in order to account for the uncertainties in this matching process , and in the irac aperture corrections
, we added a term of 0.1 mags in quadrature to the errors of the irac magnitudes .
figure [ colcol ] shows the observed @xmath43 vs. @xmath44 colour - colour diagram for our sample .
overplotted are some stellar population models from the m05 , bc03 and pegase libraries , with ages between 0.1 and 3.5 gyr , in the observed frame at redshift 1.9 .
simple stellar populations ( ssps ) with solar metallicity are indicated by solid lines , while dashed lines display exponentially declining models with e - folding time of 0.3 gyr ( for the pgase models , only the ssp is shown ) . as in analogous plots by @xcite and @xcite ,
the very red @xmath44 colours displayed by the galaxies are not reached by the bc03 ( or pgase ) templates , unless one adds strong dust reddening to the models .
the m05 models match instead the observed galaxy colours at the ages corresponding to the agb phase - transition .
the present database ( see the appendix for the photometric data ) allows the sampling of the galaxy spectral energy distributions up to the rest - frame @xmath35-band .
the whole wavelength range from the rest - frame @xmath45 to @xmath35 will be analysed in the next section .
we use an adapted version of the code _ hyper - z _
@xcite , kindly made available to us by m. bolzonella , which performs sed fitting at fixed spectroscopic redshift .
the fitting procedure is based on maximum - likelihood algorithms and the goodness of the fit is quantified via @xmath2statistics ( see the _ hyper - z _ manual for details ) .
the code computes the @xmath2for a certain number of templates , which differ for star formation ( sf ) histories , metallicities and ages , and finds the best - fitting template among them .
it is important to note that the code does not interpolate on the template grids .
this implies that the template set must be densely populated .
we have constructed templates that cover a wide range of stellar population parameters , namely : i ) metallicities from @xmath46 to 2 @xmath47 ; ii ) ages from 10@xmath48 yr to 5 gyr ( the maximum possible age for the lowest redshift galaxy ) ; iii ) sf histories as : ( a ) instantaneous and chemically homogeneous bursts ( i.e. simple stellar populations , ssps ) ; ( b ) exponentially - declining modes ( sfr @xmath49 , or @xmath50-models , with @xmath51 gyr ) ; ( c ) _ truncated _ models , where sf is constant for a finite time interval @xmath52 and zero thereafter ( @xmath53 gyr ) ; constant star formation.maraston ] these composite templates ( with prolonged sf histories ) were constructed using the bc03 software package , for consistency with d05 and in order to be able to run the _ hyper - z _ code . as in d05
, we determined 95% confidence ranges for ages and masses following @xcite , marginalizing over e(b - v ) , metallicity and sf history .
this implies that the determination of the latter parameters is formally less constrained and the resulting values should be regarded as indicative . to allow a direct comparison with other works , d05 in particular ,
the templates refer to a straight salpeter imf down to 0.1 @xmath54 .
however , several pieces of evidence , from the redshift evolution of the fundamental plane to the @xmath23 of local es to the metal content of the intergalactic medium strongly favour a salpeter - like imf down to 1 @xmath54 but flatter below as required to agree with the dynamically estimated @xmath23 ratios of early - type galaxies @xcite .
the actual shape of the imf does not affect appreciably the sed since at any age most of the light is produced by stars in a narrow mass interval around the turnoff mass .
however , it impacts on the _ amount _ of light emitted per unit mass turned into stars , thereby affecting the stellar @xmath55 ratio .
for example , when using the imf of @xcite the @xmath23 ratio is a factor @xmath56 smaller than in the straight - salpeter case ( m05 ) . for the objects classified of class a redshift by d05 the fitting procedure included also the strength of the @xmath4 absorption feature , as defined in d05 and measured on the grapes spectra .
the difference between observed and model @xmath4 strength was included in the calculation of the @xmath57 when searching for the best - fit model , with the same weight as the photometric data . finally , the entire sed fitting was repeated with bc03-based templates spanning the same stellar population parameters as those used for the m05 models .
the output of the procedure includes : the age @xmath58 , i.e. , the time since the start of sf ; the metallicity @xmath59 $ ] ; the star - formation history sfr ; the reddening @xmath60 ; and the stellar mass @xmath61 .
the stellar mass @xmath61 is evaluated by comparing the observed and the synthetic sed , which provides the mass that went into stars by the age of the galaxy .
this approach was retain in order to allow a comparison with d05 .
however , such mass overestimates the true stellar mass , as stars die leaving remnants whose mass is smaller than the initial one . adopting the prescriptions by @xcite for the initial mass - remnant mass relation , m98 evaluated that the stellar mass of a ssp is reduced by @xmath62 15 gyr after formation
, most of which happens within the first few gyrs .
for given mass turned into stars , such reduction is smaller in stellar populations with extended star formation histories , due to a relatively higher fraction of living stars .
the actual @xmath63 for some illustrative cases of star - formation histories is given in figure [ remnants ] .
tables 1 and 2 give the stellar mass prior of such reduction , thus enabling a direct comparison with other results .
table 2 also reports the mass decrement ( in percent ) that corresponds to the given sfh and age for several star formation histories can be found at www-astro.physics.ox.ac.uk/@xmath64maraston . ] .
this mass decrement should be applied to the masses listed in column 12 .
we first discuss the solutions that are obtained in the assumption of zero reddening .
this allows us to single out the effect of stellar population modeling , thus highlighting some interesting differences between the two sets of models .
the corresponding best - fit solutions ( lowest @xmath57 ) for using the bc03 ( blue lines ) and the m05 ( red line ) models are shown in figure [ fitnored ] .
the observed seds are shown as solid symbols with errorbars , and information on the individual galaxies and best fit sfhs is given in table 1 .
overall , the m05 models allow for better fits ( lower @xmath57 ) for all galaxies , in the rest - frame near - ir as well as in the blue ( e.g. , the objects at @xmath0=1.39 , 1.55 and 2.47 ) .
given the ages required to fit the observed seds , it is likely that the different recipes for the tp - agb phase are primarily responsible for such effect . for the objects at @xmath66 and
@xmath67 none of the bc03-based templates was able to match the near - ir fluxes . in the case of the @xmath67 galaxy
not even an age as old as 2.3 gyr is sufficient to obtain enough near - ir flux .
note also that the ssp solution gives a fairly good fit with m05 models ( which are tp - agb dominated at this age ) , whereas an extended sf history ( constant until @xmath68 ) was required when using the bc03 models . in this case
the m05 single - burst models supply both the high optical and the high near - ir fluxes , which come from the warm turnoff and cool tp - agb stars , respectively .
the stellar masses @xmath61 are typically smaller for the m05 models , being on average @xmath69 of those derived with the bc03 models . in part
this is a consequence of the lower ages , with the average age of m05 models being also @xmath69 of that derived for the bc03 models .
the differences are not as high as they would be in case of single bursts , as the more extended sfhs required by the bc03 models have lower @xmath55 ratios .
the most relevant difference in the results from the two sets of models is about the implied formation redshifts ( see table 1 ) . using bc03 models the stellar populations form at higher redshifts ( sometime at much higher redshifts , see e.g. , the object at @xmath67 ) and , since the galaxies are also more massive compared to the case when using m05 models , an appreciably faster mass growth of galaxies
is implied .
the best - fit solutions when also reddening is treated as a free parameter are displayed in figure [ fitred ] , and the parameters of the best - fits are given in table 2 .
the reddening was allowed to vary among the options offered in the _ hyper - z _ package , namely , milky way , smc , lmc and the so - called calzetti s law for local starbursts ( references in table 2 ) .
all these prescriptions refer to a dust - screen configuration . for this reddening case , in order to exploit all the available information and to control the degeneracies introduced by the larger number of free parameters , we have additionally included the strength of the @xmath4 feature in the computation of the @xmath2 .
this was done only for galaxies with class a redshift , namely for objects 4950 , 1025 , 3650 , 3574 , because for the class b objects the @xmath4 is a less robust measurement ( d05 ) . by allowing the best - fit procedure to treat @xmath60 as an additional free parameter , and allowing it to choose among four different reddening laws , it is no surprise that solutions with appreciably better
@xmath2are found .
however , this does not ensure that such solutions are more likely than those assuming @xmath65 .
it is also no surprise that some best - fits require the calzetti s law , whereas others prefer the smc law . in just one case
the lmc law was preferred , while in no cases the mw law was chosen .
it is worth noting that the best - fit for the @xmath67-object exhibiting the strongest tp - agb contribution is reddening - independent for the m05 models . in general
, the inclusion of reddening reduces the ages of the best - fits obtained with the bc03 models , making them closer to those obtained with the m05 models .
however , reddening does not help to improve substantially the fits to the irac bands , like in the case of objects 1446 and 3650 .
tp - agb effects can not be traded with dust or metallicity effects , which helps reducing the degeneracies in the age determination .
detailed comments on the fits for individual galaxies are given in the appendix .
+ figure [ grapes ] compares the grapes hst+acs spectrum of object 3650 at @xmath70 , with the best - fit solutions corresponding to the bc03 and m05 models ( see table 2 ) .
we use this plot for illustrating the effect of the age / metallicity degeneracy that affects the optical , but is alleviated in the near - ir due to the distinctive spectral features of tp - agb stars .
the best - fit solutions recover similarly well the strength of the @xmath4 line and the spectral shape around it , in spite of having different population parameters , namely young and metal - rich for the m05 models and older and metal - poor for the bc03 models .
however , the bc03-based solution gives an appreciably poorer fit to the near - ir sed ( cf .
figure [ fitred ] ) , showing how the inclusion of the near - ir fluxes can help breaking the age / metallicity degeneracy .
however in most cases spectral absorptions and photometric data on a wide spectral range are not available simultaneously .
how realistic are then the best - fits obtained from the sole photometry ?
to get insight on this issue , we have repeated the fitting procedure without including the @xmath4 line when searching the best - fit model .
we have then checked how well these photometric best - fit models recover the observed @xmath4 .
this is shown in figure [ mguvnochi ] , in which the @xmath4 strengths predicted by the best - fit ` photometric ' models are compared to the observed ones .
although based on a very small number of objects , the predicted values from m05 models are somewhat more accurate than those from the bc03 models ( variances are 0.04 and 0.08 , respectively ) . as a final exercise , we have checked the effect on photometric redshifts derived using one or the other set of population models , and of the inclusion of the irac fluxes .
the results are given in table 2 and shown in figure [ photoz ] .
the photometric redshifts obtained with the m05 templates are in better agreement with spectroscopic redshifts than those obtained with the bc03 models .
the photometric redshifts obtained with the bc03 models and the irac data appear to be more discrepant than those in which only magnitudes up to @xmath35 were used ( cf .
also this effect likely originates from the different tp - agb recipes .
the different star - formation histories derived from the different models have an impact on the derived stellar masses , which is summarised in figure [ mass ] .
the upper panels display the effects of the models ( i.e. , m05 vs bc03 ) on the derived stellar mass of galaxies , when using only the optical and near - ir bands ( left - hand panel ) and when also the irac 3.58 @xmath71 m data are used ( right - hand panel ) .
the masses predicted by the m05 templates are lower , which is a consequence of the lower ages .
the effect is appreciably more pronounced when the irac bands are included , as the tp - agb phase ( the most discrepant ingredient between the two models ) has its largest contribution in the rest - frame near - ir .
however , lower masses can also be derived when the observed - frame @xmath35 is the reddest band used in the best fit .
in fact , for the objects at @xmath72 the @xmath35 band samples the rest - frame @xmath73 , which is already affected by the tp - agb ( m05 , figure 18 ) .
in addition , the later onset of the rgb phase in the overshooting tracks adopted by bc03 helps to make the bc03 templates fainter at ages around 0.5 gyr .
the corresponding @xmath23 ratio is higher and a higher @xmath74is derived ( see section 2 ) .
it is embarrassing to realize to which extent the recipes for the convective overshooting impact on the stellar masses that are derived for high-@xmath0 galaxies , and therefore on the implied galaxy formation scenarios .
the lower panels in figure [ mass ] show the effects ( for given stellar population models ) of the wavelength range included in the fitting procedure : only up to the @xmath35 band ( on the x axis ) and with the irac bands ( on the y axis ) .
the results based on bc03 are not appreciably affected by the inclusion of the irac bands ( as recently pointed out by * ? ? ?
* ) . instead , for the m05-based solutions ,
three galaxies are found to have appreciably lower masses when the irac fluxes are considered , again the likely effect of the inclusion of the tp - agb contribution .
finally , figure [ bzk ] compares the masses obtained in this paper ( including the irac data ) with the calibration for @xmath75 @xmath76-selected galaxies proposed by @xcite .
this calibration was obtained using the observed @xmath35-band magnitude and @xmath77 colour only , along with the bc03 templates .
the figure shows that the calibration is well consistent with the bc03-based masses derived in this paper , which is no surprise .
the m05 models suggest instead a calibration lower by 0.2 dex , on average .
only two galaxies are significantly detected with spitzer+mips at 24@xmath78 ( objects 1025 and 4950 , fig .
[ images ] ) .
the mips 24@xmath79 fluxes of about 2530@xmath71jy convert to star - formation rates of about 20 - 30 @xmath80yr@xmath81 at the redshifts of the sources , using the models from @xcite .
the two mips detections are in fact the galaxies with the bluer uv spectra ( see d05 ) and more extended star formation histories ( see table 2 ) . the mips derived sfrs
are comparable , given the uncertainties , with the residual ongoing sfrs ( of @xmath82yr ) that are obtained through the sed modeling of these 2 galaxies ( see also d05 ) .
for the rest of the galaxies the 3 @xmath83 limits on the 24@xmath78 flux imply sfr below 10 @xmath80yr@xmath81 .
we recall that typical @xmath84 galaxies with similar k - band fluxes to the sample studied here have much stronger 24@xmath78 emission , of order of 120@xmath71jy @xcite .
the low mips fluxes from these sources underline once again , in agreement with their spectral features , that these objects are predominantly passive . in obtaining the best - fitting models we have assumed that the observed seds are the results of just stars and dust .
the possible presence of an agn would obviously affect stellar population modeling in general .
however , differences between the physical properties of galaxies inferred with the m05 or bc03 models remain , as these models have a different ratio of optical to near - ir flux .
therefore , even if at play , agns would not alter the sense of this investigation .
more specifically , agns probably play a minor role on our analysis .
first , no clear agn features are observed in the spectra of these objects .
also , the shape of the sed over irac is clearly stellar with a secure identification of the peak emission at 1.6 @xmath78 .
two of the galaxies ( 1025 and 1446 ) have @xmath85-ray detections , but as discussed in d05 these are extremely hard and absorbed .
coupled to the evidence of the faintness of the galaxies at 24 @xmath78 , we conclude that the agn should not influence appreciably the photometry from the uv to the near - ir rest - frame . for example , if we adopt the sed shapes of ngc1068 or mrk231 and normalise them to the 24@xmath78 photometry of the sources , the contribution of the agn to the optical to near - ir light of the galaxies would be entirely negligible .
similarly , from the narrow @xmath35-band to @xmath85-ray flux correlation of agn shown by @xcite , we conclude that possible agn in our sources are expected to provide a negligible contribution to the @xmath35-band light of our galaxies .
finally , the overall good agreement between predicted and observed @xmath4 line - strengths ( figure 7 ) and the good match between spectroscopic and photometric redshifts ( figure 8) suggest that a significant agn contribution is highly unlikely .
the aim of this paper is to explore how much the star formation histories of high-@xmath0 galaxies , that are derived by fitting observed to synthetic spectral energy distributions depend on the adopted stellar population models . in particular , we focused on the tp - agb evolutionary phase , because tp - agb stars are the dominant bolometric and near - ir contributors in stellar populations with ages @xmath37 1 gyr ( m98 ; m05 ) . therefore , such phase must be of primary importance at high-@xmath0 , when the universe itself was just a few gyrs old @xcite , and especially in the interpretation of _ spitzer _ data for high-@xmath0 objects .
the recipes for the tp - agb phase are found to differ significantly among the various models , which has motivated us to check the robustness of the results obtained for high-@xmath0 galaxies . to this aim
we have analysed a sample of mostly passively evolving high-@xmath0 ( @xmath1 ) galaxies with optical and _ spitzer_/irac photometry , modeling the whole spectral energy distribution from the rest - frame @xmath34 to @xmath35 .
the analysis was performed with two libraries of stellar models : the m05 models in which the tp - agb phase is calibrated on magellanic cloud globular clusters , and the bc03 models in which , like in most other models in the literature ( e.g. pegase , starburst99 ) the energy contribution from the tp - agb phase is significantly lower than in the maraston models .
furthermore , the two models adopt stellar evolutionary tracks , in which the onset of the red giant branch is predicted at different ages , which also impact on the time evolution of the synthetic spectral energy distribution .
the differences between stellar population models are widely discussed in m05 and recalled in section 2 .
besides the 14 photometric bands spanning from rest - frame @xmath34 to @xmath35 , the procedure introduced in this paper also includes the strength of a near - uv absorption feature ( the @xmath4 at 2800 , see d05 ) for searching the best - fit galaxy models . in this way the effect of the age / metallicity degeneracy is substantially reduced .
we have also checked that even without using the spectroscopic information , but only with the photometric sed , the best - fits obtained with the m05 models recover well the strength of the @xmath4 feature , hence provide consistent models for distant galaxies .
we then show to which extent the derived star - formation histories of the sample galaxies are model dependent .
the m05 models typically match the rest - frame near - ir and optical very well , without invoking strong reddening and without violating the constraint set by the age of the universe .
the bc03-based solutions exhibit in some cases near - ir fluxes that are significantly lower than the data , which requires the addition of old stellar populations or dust reddening or a combination of the two .
this is reminiscent of recent results on high-@xmath0 galaxies in which spitzer / irac data are interpreted with the bc models ( e.g. , * ? ? ? * ) or the pegase models ( e.g. , * ? ? ?
the young solutions required to match the blue side of the sed could not match the rest - frame near - ir , which led the authors to assume an underlying old population to be responsible for the near - ir fluxes .
while a two - component model is not implausible in itself , the proportions by mass of the young and old component ( 1 vs. 99 % ) as well as the age of the old component , comparable to the age of the universe , make the star formation history contrived .
the lower masses derived from the m05 models may help understanding the origin of an apparent discrepancy noted by d05 and concerning the relationship between the physical sizes and the masses of the galaxies in the present sample .
the galaxies were found to be too small for the given masses in comparison to local objects .
the lower masses derived here already help in alleviating the discrepancy .
in particular , the @xmath67 object , with a stellar mass around @xmath86 , is more likely the precursor of a bulge rather than a massive elliptical .
its rather small size ( @xmath87 kpc , d05 ) supports this conclusion , as it can be inferred from comparisons with local samples @xcite .
also its formation redshift and metallicity compares well with similar estimates of bulges in the local universe @xcite . according to the bc03 models , three galaxies have rather low metallicities ( @xmath46 ) and at the same time high masses ( @xmath88 ) .
such objects do not match local relationships , even taking metallicity gradients into account .
only one such object is found with the m05 models . finally , we find that the stellar population templates impact on photometric redshift determination .
the use of irac bands and m05 templates give photometric redshifts that are in excellent agreement with the spectroscopic ones .
again , we ascribe this effect to the prescriptions for the tp - agb phase . in this work we have focused on high-@xmath0 galaxies , where the stellar populations are young
therefore tp - agb stars must exist .
our spectral fitting represents the first direct evidence for such stars in the early universe . going to even higher redshifts might reveal further interesting aspects of galaxy evolution , a scope we are pursuing in the future
however , tp - agb stars may matter at any redshift depending on the specific star formation history .
for example , @xcite measured the evolution of the rest - frame @xmath35-band fundamental plane from @xmath89 to the present , by using irac imaging of a sample of early - type galaxies .
they find that the @xmath35-band luminosities of the low - mass objects are too large ( at given @xmath34 luminosity ) , or the @xmath90 colours too red , with respect to the prediction of the bc03 or @xcite models .
the m05 models were found to provide a better match the data .
this effect , which is correctly attributed to the treatment of the tp - agb phase , implies that those low - mass early - type galaxies experienced a burst of star formation at redshift close to one .
in conclusion , the recipe for the stellar population modeling plays a central part in the derivation of the star formation histories of high-@xmath0 galaxies , which should be kept in mind when the derived parameters are used to constrain galaxy formation .
we thank micol bolzonella for assistance with the @xmath91 code . cm ackwnoledges discussions with reinhard genzel .
cm is a marie curie fellow and holds grant meif - ct-2005 - 011566 of the training and mobility of researchers programme financed by the european community .
ed acknowledges nasa support through the spitzer fellowship program , award 1268429 .
some of the data used here are part of the goods spitzer space telescope legacy science program , that is supported by nasa through contract number 1224666 issued by the jpl , caltech , under nasa contract 1407 .
ccccccccc 8238 & 1.39 & m05 & 1.0 & 1.8 & 2 @xmath92 & @xmath93 & 1.5 & @xmath94 + & & bc03 & 2.3 & 2.9 & 2.5 @xmath92 & @xmath95 & 2.6 & @xmath96 + 4950 & 1.55 & m05 & 1.7 & 2.7 & @xmath97 & @xmath95 & 1.5 & @xmath98 + & & bc03 & 1.7 & 2.7 & 2.5 @xmath92 & @xmath95 & 4.9 & @xmath99 + 1025 & 1.73 & m05 & 1.7 & 3.2 & @xmath92 & @xmath95 & 2.2 & @xmath100 + & & bc03 & 1.7 & 3.2 & 2.5 @xmath92 & @xmath95 & 3.8 & @xmath101 + 3523 & 1.76 & m05 & 1.4 & 2.9 & @xmath92 & @xmath102 gyr & 2.2 & @xmath103 + & & bc03 & 2.3 & 4.6 & 2.5 @xmath92 & @xmath104 gyr & 3.1 & @xmath100 + 3650 & 1.91 & m05 & 0.5 & 2.2 & 2 @xmath92 & @xmath105 gyr & 2.5 & @xmath106 + & & bc03 & 2.3 & 5.5 & 2.5 @xmath92 & @xmath104 gyr & 3.5 & @xmath107 + 3574 & 1.98 & m05 & 2.3 & 6 & @xmath108 & @xmath102 gyr & 2.6 & @xmath109 + & & bc03 & 2.6 & 8.2 & 0.4 @xmath92 & @xmath104 gyr & 4.0 & @xmath110 + 1446 & 2.47 & m05 & 0.4 & 2.8 & 2 @xmath92 & ssp & 6.2 & @xmath111 + & & bc03 & 2.3 & 15 & @xmath92 & @xmath104 gyr & 10.4 & @xmath112 + lccccccccccccl 8238 & 1.39 & 1.39 & m05 & 0.7 & 1.7 & 2 @xmath92 & @xmath113 gyr & 0.15 & 1.1 & [email protected] & @xmath115 & [email protected] & 21 + & & 1.30 & bc03 & 2.3 & 2.9 & 2.5 @xmath92 & @xmath104 gyr & 0.15 & 2 . &
& @xmath116 & + 4950 & 1.55 & 1.55 & m05 & 3.5 & 7.5 & @xmath97 & @xmath117 & 0.15 & 0.6 & [email protected] & @xmath118 & [email protected] & 26 + & & 1.67 & bc03 & 3.5 & 7.5 & @xmath108 & @xmath117 & 0.25 & 0.94 & & @xmath119 & + 1025 & 1.73 & 1.72 & m05 & 1.4 & 3.1 & @xmath92 & @xmath120 & 0.07 & 1.1 & [email protected] & @xmath121 & [email protected] & 20 + & & 1.77 & bc03 & 1.4 & 3.1 & @xmath122 & @xmath120 & 0.15 & 0.64 & & @xmath123 & + 3523 & 1.76 & 1.68 & m05 & 0.3 & 1.9 & @xmath92/5 & @xmath105 gyr & 0.4 & 1.2 & [email protected] & @xmath124 & [email protected] & 18 + & & 1.63 & bc03 & 0.5 & 2.1 & @xmath92 & @xmath125 & 0.4 & 1.2 & & @xmath126 & + 3650 & 1.91 & 1.88 & m05 & 0.3 & 2.1 & 2 @xmath92 & ssp & 0.2 & 1.3 & 0.33 - 0.85 & @xmath127 & 1.3@xmath1142 & 19 + & & 1.81 & bc03 & 1.0 & 2.7 & @xmath92/5 & ssp & 0.2 & 1.9 & & @xmath128 & + 3574 & 1.98 & 1.94 & m05 & 0.2 & 2.1 & 2 @xmath92 & ssp & 0.3 & 2.5 & [email protected] & @xmath129 & [email protected] & 18 + & & 1.94 & bc03 & 1.0 & 2.9 & @xmath108 & @xmath113 gyr & @xmath130 & 2 . & & @xmath131 & + 1446 & 2.47 & 2.74 & m05 & 0.4 & 2.8 & 2 @xmath92 & ssp & 0.00 & 3.1 & [email protected] & @xmath132 & [email protected] & 20 + & & 2.86 & bc03 & 1.0 & 3.8 & @xmath92 & @xmath125 & 0.35 & 4.7 & & @xmath133 + * 8238 , z=1.39*. the @xmath4 line was not included in the @xmath2 .
the object is @xmath134 gyr old , metal - rich and in passive evolution since 0.4 gyr .
the reddening follows an smc - like law .
the stellar mass is @xmath135 with small dispersion . for the bc03 models ,
the best solution has a worse @xmath2 , a significantly older age and a significantly longer formation timescale .
the stellar mass is a factor 2.2 larger and consistent with what obtained in d05 from fitting the sole optical bands .
it is interesting to note that the same , smaller stellar mass would have been derived by fitting the m05 models up the sole observed - frame @xmath35 ( see figure [ mass ] ) .
this is due to two effects .
first , at redshift @xmath136 the observed @xmath35 samples into the rest - frame @xmath73 , which is already affected by the tp - agb prescriptions ( m05 ) .
second , as pointed out in section 2 , another important difference between the two sets of models exist , namely the energetics of the different input stellar evolutionary tracks .
the m05 models are based on tracks in which the rgb phase starts developing at younger ages with respect to the padova tracks used in bc03 .
the delay in the latter is due to the overshooting ( see discussion in m05 ) .
the result is that the m05 ssp models are brighter around 1 gyr ( cf .
m05 , fig .
7 ) , which implies a lower stellar mass , an effect anticipated in m05 .
it should be noted that the earlier onset of the rgb phase compares very well with the observed rgb contribution in magellanic cloud globular clusters @xcite , while the predictions based on the padova tracks display a severe discrepancy with the data .
+ * 4950 , z=1.55*. the @xmath4 line was included in the @xmath2 .
the best - fit solution is 3.5 gyr old , is forming stars according to a _ tau _ model with e - folding time of 1 gyr , the metallicity is half - solar and @xmath137 ( smc law ) .
the extended star formation history is consistent with the irregularities and blobs in the @xmath34-band image and with its spiral - like morphology ( d05 ) .
this galaxy has a huge de vauculeurs bulge and turned out to be the most massive object of the sample .
the observed @xmath4 is remarkably well reproduced by this model .
@xmath138 is large and very well constrained .
it is consistent with both the value of d05 and that derived here with the bc03 models .
the best solution obtained with the bc03 models has similar overall parameters , but the reddening is larger and the @xmath2 is worse .
+ * 1025 , z=1.73 . *
the @xmath4 line was included in the @xmath2 .
the best - fit is @xmath37 1.4 gyr old , still forming stars at a very modest rate ( @xmath50 model , with @xmath139 gyr ) .
the stellar mass is @xmath140 .
the solution obtained with the bc03 model has a lower metallicity and a larger mass . + * 3523 , z=1.76*. the @xmath4 line was not included in the @xmath2 .
the @xmath4 for this galaxy is very small ( @xmath141 ) , which implies either very young ages or very low metallicities .
the solution allowing for the best comparison to the observed line turned out to be young and rather metal - poor .
its @xmath4is 1.07 .
the galaxy is passively evolving since 0.2 gyr .
the solution obtained with the bc03 models allows a worse match to the @xmath4 line ( 1.14 ) because of the higher metallicity .
a higher metallicity is found because it provides more flux to the near - ir . + * 3650 , z=1.91*. the @xmath4 line was included in the @xmath2 .
the galaxy is young and metal - rich , and in passive evolution for the m05 models .
it has formed stars at a remarkable rate as @xmath142 is @xmath143 .
the spectrum displays a nice near - ir excess that is very well fitted .
the match to @xmath4 is remarkable good ( 1.25 versus the observed 1.25 ) .
the solution obtained with the bc03 models is older and significantly more metal - poor , though the stellar mass of the galaxy if higher ( @xmath144 ) .
this mass is too large for a metallicity @xmath46 in comparison with what found in the local universe .
this solution matches worse to the observed sed .
+ * 3574 , z=1.98*. the @xmath4 line was included in the @xmath2 .
the best - fit is young , metal - rich , and rather highly reddened
. the model is passively evolving .
the best - solution obtained with the bc03 models is older and metal - poor . + * 1446 , z=2.47*. the @xmath4 line was not included in the @xmath2 .
this is perhaps the strongest case of agb - dominated object .
the fit requires a super - solar metallicity .
however , the amount of tp - agb in the m05 models at this metallicity was even slightly too low around the j band with respect to the observations . since the tp - agb fuel and spectral - types at super - solar metallicities
are not well constrained given the absence of local gcs as calibrators ( see discussion in m05 ) , we could accept to modify the model .
an excellent fit is obtained when the fuel is augmented by 30% and the o - rich stars have a slightly warmer temperature . both changes are certainly inside the uncertainties of the tp - agb phase at high - metallicity .
noteworthy , this fit does not require any reddening . the stellar mass is less than 1/3 the lower bound of the d05 solution .
the bc03-based solution is older and highly reddened , yet the match to the observed sed is not satisfactory , in the near - ir as well as in the optical .
we have verified that older ages ( 2.6 gyr ) would provide a better fit , but these ages are above the limit set by the age of the universe .
note that the visual quality of the fit did not correspond to the @xmath2 values when the f110 band was included in the @xmath2 .
this is due to this photometric data having a very small error , which drives the @xmath2 .
while the best - fit model was found using all bands for consistency with the other galaxies , the @xmath2values given in table 2 were computed by excluding the f110 band .
the data in f110 and j bands are off the models and are responsible for the significantly higher values of @xmath2that are obtained in case of this galaxy with respect to the other objects .
we have noted that significantly better values of @xmath2(@xmath145 ) are obtained when the redshift is higher than the spectroscopic value and close to the redshift estimated photometrically with the m05 models ( 2.74 , cf .
figure [ photoz ] ) . in this case ,
the best - fit model , that has similar overall parameters as that presented in figure [ fitred ] , did not require any modification of the fuel consumption in the tp - agb phase .
rccccccc 8238 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 + 4650 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + 1025 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath159 + 3523 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath172 + 3650 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 + 3574 & @xmath180 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 + 1446 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath193 + cccccccc @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 + @xmath202 & @xmath159 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 + @xmath209 & @xmath159 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 + @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 + @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 + @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 & @xmath239 + @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 + | we explore the effects of stellar population models on estimating star formation histories , ages and masses of high redshift galaxies .
the focus is on the thermally - pulsing asymptotic giant branch ( tp - agb ) phase of stellar evolution , whose treatment is a source of major discrepancy among different evolutionary population synthesis . in particular , besides the models usually adopted in the literature , we use models ( by maraston 2005 ) , in which the contribution of the tp - agb phase is calibrated with local stellar populations and is the dominant source of bolometric and near - ir energy for stellar populations in the age range 0.2 to 2 gyr .
these models also have an underlying different treatment of convective overshooting and red giant branch stars . for our experiment
we use a sample of high-@xmath0 ( @xmath1 ) galaxies in the hubble ultra deep field held to be mostly in passive evolution , with low - resolution uv - spectroscopy and spectroscopic redshifts from grapes , and _ spitzer _
irac and mips photometry from the great observatories origins deep survey .
we choose these galaxies because their mid - uv spectra exhibit features typical of a- or f - type stars , therefore tp - agb stars ought to be expected in post - main sequence .
we find that indeed the tp - agb phase plays a key role in the interpretation of _ spitzer _ data for high-@xmath0 galaxies , when the rest - frame near - ir is sampled .
when fitting without dust reddening , the models with the empirically - calibrated tp - agb phase always reproduce better the observed spectral energy distributions ( seds ) , in terms of a considerably smaller @xmath2 .
allowing for dust reddening improves the fits with literature models in some cases . in both cases ,
the results from maraston models imply younger ages by factors up to 6 and lower stellar masses ( by @xmath3 on average ) .
the observed strengths of the @xmath4 spectral feature compare better to the predicted ones in the case of the maraston models , implying a better overall consistency of sed fitting . finally , we find that photometric redshifts improve significantly using these models on the seds extending over the irac bands .
these results are primarily the consequence of the treatment of the tp - agb phase in the maraston models , which produces models with redder rest - frame optical to near - ir colors .
this work provides the first direct evidence of tp - agb stars in the primeval universe . |
surely one of the most striking features of matter refers to the fact that one encounters two distinct types of particles : fermions and bosons .
as safely as this fact is established from both the conceptual and experimental point of view ( for a review see ref .
@xcite ) , it seems unclear for most people what is the precise theoretical origin of this matter dichotomy and why there do not exist more than two kinds of material particles . in view of the lack of a right understanding of the reason why there are only two particle species ,
some people hold that one must first `` _ _ understand why nature makes such a particular choice from among a host of other possibilities _ _ '' @xcite . consequently for that purpose , they believe
, one should `` _ _ pursue every theoretical possibility we may conceive at present _ _ '' @xcite and these endeavours should refer to the rather general framework of paraquantization .
other people do not trouble themselves about the true origin of the fermion - boson dichotomy but take this simply as an experimentally based fact , and consequently they are content with rephrasing this fact in mathematical terms ( i.e. the postulates of symmetrization for bosons and anti - symmetrization for fermions ) . in this sense , e.g. , dirac simply states that `` _ _ other more complicated kinds of symmetry are possible mathematically , but do not apply to any known particles _ _ '' ( ref .
@xcite , p. 211 ) .
one could go even one step further by resorting to the well - known _ anthropic principle _
@xcite which may be evoked whenever a well - established observation can not be based upon a sound theoretical foundation .
such a situation typically occurs in cosmology @xcite but the present fermion - boson dichotomy could also evoke that principle , e.g. by asking : `` _ did god - for lack of a better word - build a series of failed worlds which spottered and died , or exploded and disintegrated , before discovering the stabilizing effect of anti - commutation relations for half - integral spin fields ? _ ''
@xcite , p.
however it rather seems reasonable to assume that the fermion - boson dichotomy `` _ could conceivably be an essential ingredient of a more fundamental view of the world this could be the case in fundamental string theories or their successors _ '' @xcite .
in any case , such an unclarified situation with the matter dichotomy should provide sufficient motivation for studying this problem also within the context of non - standard quantum theories . in this sense
the present paper is devoted to the study of the matter dichotomy within the framework of relativistic schrdinger theory ( rst ) , an alternative form of quantum theory being based upon fluid - dynamic concepts rather than probabilistic ones @xcite -@xcite .
indeed it has turned out that a matter dichotomy emerges in this theory in the form of positive and negative mixtures @xcite resembling very much the symmetrized ( bosonic ) and anti - symmetrized ( fermionic ) states of the conventional quantum theory : the positive mixtures imply a certain kind of _ fusion _ of two particles into one charge density , and the negative mixtures describe the recession of the two charge densities from one another such that the density of one particle is zero in that region of space - time where the density of the other particle is non - zero ( charge separation @xcite ) . consequently both particles in a negative - mixture configuration are forbidden to occupy the same space - time points which evidently yields a `` stabilizing effect '' for composite matter coming about in the conventional quantum theory through the anti - symmetrization postulate ( or anti - commutation postulate for operators , resp . ) .
however before plunging deeply into the physical implications of this rst matter dichotomy , one first would like to understand better both the differences and the common features of both types of rst mixtures from a purely _ mathematical _ point of view . in this sense ,
the present paper presents a detailed study of the geometric and topological properties of both types of mixture configurations , deferring the discussion of the physical implications to a separate treatment . for the purpose of a survey of the corresponding results ,
it is very instructive to first recall some of the basic ( probabilistic ) concepts of the conventional quantum theory and oppose these to the analogous kinematical elements of the ( fluid - dynamic ) rst .
since , on the conventional side , there exists no successful _ relativistic _ quantum mechanics for many - particle systems ( in contrast to the very successful quantum field theory ) , we have to restrict ourselves here to a comparison of the _ non - relativistic _ conventional theory with rst ( as a fully relativistic theory ) .
furthermore we are satisfied in the present paper with a treatment of 2-particle systems ( neglecting also the particle spin ) which may be sufficient to display already the essential many - particle effects .
( for including also the spin into rst see , e.g. ref .
@xcite ) . since in the conventional theory
the 2-particle hilbert space is the tensor product of the two 1-particle spaces , a _ disentangled _
2-particle state @xmath2 is simply the product of two 1-particles states @xmath3 and @xmath4 , i.e. @xmath5 in contrast to this product construction , a disentangled 2-particle state @xmath6 in rst is the direct sum of the two 1-particle states @xmath7 or reformulated in mathematical terms : the 2-particle vector bundle ( with the rst 2-particle wave function @xmath6 as bundle section ) is the whitney sum of both 1-particle bundles .
if some interaction between the particles is switched - on ( e.g. the electromagnetic interactions , especially the static coulomb interaction ) the conventional state ( [ eq : prod ] ) can not retain its simple product form ( [ eq : prod ] ) but develops into a general 2-particle state @xmath8 obeying the symmetrization postulate @xmath9 in contradiction to certain claims in the older literature , the modern textbooks clearly state that _ this postulate _ ( expressing the matter dichotomy in the conventional theory )
_ can not be deduced from other principles of quantum mechanics _ ( ref .
@xcite , p. 475 ) .
if the interparticle interactions are switched - off again , the generally entangled 2-particle state @xmath10 ( [ eq : syp ] ) does not return to the original disentangled form ( [ eq : prod ] ) but develops to a simply entangled state @xmath11 which is build up again by two 1-particle states but still obeys the symmetrization postulate ( [ eq : syp ] ) @xmath12 this form of the wave function persists even when the particles are localized at large spatial separation and thus gives rise to the well - known einstein - podolsky - rosen paradoxes @xcite . according to the whitney sum construction ( [ eq : sum ] ) ( in place of the conventional product arrangement ( [ eq : prod ] ) ) , the concept of _ entanglement _ is defined in rst in a somewhat different way , namely by describing matter by means of a ( hermitian ) intensity matrix @xmath13 in place of a wave function @xmath14 .
the disentangled rst states @xmath14 ( [ eq : sum ] ) can also be written in form of an intensity matrix @xmath13 but here @xmath13 adopts a special form , namely the tensor product of @xmath14 and its hermitian conjugate @xmath15 , see equation ( [ eq : ipsi ] ) below .
a disentangled rst state @xmath14 ( [ eq : sum ] ) obeys the relativistic schrdinger equation ( [ eq : rse ] ) and an entangled rst configuration @xmath13 ( a `` mixture '' ) satisfies the relativistic von neumann equation ( [ eq : cali ] ) .
the hamiltonian @xmath16 , governing both these field equations ( [ eq : rse ] ) and ( [ eq : cali ] ) , is itself a dynamical object of the theory and has to be determined from its field equations ( [ eq : intcond ] ) and ( [ eq : coneq ] ) .
all these rst prerequisites , being indispensable for the subsequent mathematical elaborations , have been collected into * section ii*. however , the crucial point with the matter dichotomy is now that it arises in rst as a direct consequence of the rst dynamics itself and not as an additional kinematical postulate as in the conventional theory ( [ eq : syp ] ) .
this is proven in * section iii * for an arbitrary number @xmath17 of ( scalar ) particles .
the rst dichotomy consists in the fact that the determinant of the @xmath18 intensity matrix @xmath13 is always of definite sign for any mixture type , see the corresponding discussion of the equation ( [ eq : dndetin ] ) below , together with fig . 1 .
thus , in rst there occur exactly two forms of matter , namely the positive and negative mixtures ; this is in contrast to the conventional theory which has its problems with the _ theoretical _ exclusion of the higher - order multiplets of the symmetric group of identical particle permutations @xcite , leaving over only the symmetric and anti - symmetric states as its lowest - order representations .
after the matter dichotomy has thus been safely established for a general n - particle system , we restrict ourselves for the remainder of the paper to the _ 2-particle systems ( n=2)_. therefore in * section iv * all relevant rst mathematics is specialized down to the 2-particle case which requires as the typical fibre of the vector bundle the two - dimensional complex space @xmath19 ( and its tensor - product associates ) . the interesting point with the rst
2-particle kinematics is here that one can reparametrize the whole field system by single - particle variables and by exchange fields .
this is again in contrast to the conventional theory which ( for an entangled state ) does attribute one single 2-particle wave function @xmath8 ( [ eq : syp])-([eq : syp2 ] ) to both particles but does not attribute a 1-particle wave function to either particle . as a consequence of this kinematic peculiarity
, one can clearly identify in rst the exchange forces ( different from the gauge forces ) which are responsible for the fusion effects of the positive mixtures ( @xmath20 bosonic states ) and separation effects of the negative mixtures ( @xmath20 fermionic states ) .
indeed the mixture variable @xmath21 enters the single - particle rst dynamics ( [ e4:44 ] ) together with the exchange fields @xmath22 in form of a potential term so that one arrives at the disentangled 2-particle situation ( i.e. pure rst state @xmath6 ( [ eq : sum ] ) analogous to the conventional @xmath2 ( [ eq : prod ] ) ) when one puts the exchange fields to zero .
such a disentangled rst situation can then be described by two ordinary klein - gordon equations ( [ e4:46 ] ) , namely one equation for either particle .
this rst picture of how the exchange forces produce the fusion and separation effects has no counterpart in the conventional theory , where it is not possible to trace back the pauli exclusion principle to the action of some kind of force preventing , say , a fermion from occupying the same quantum state as an other identical fermion ( see the discussion of this problem in ref .
@xcite ) . but with a convenient parametrization of the 2-particle systems being now at hand , one can study the geometric and topological differences of both kinds of mixture configurations ( * section v * ) . here , the point of departure is the observation that the dynamical equations ( of section iv ) ensure the existence of a closed 2-form @xmath23 which is composed of the mixture variable @xmath24 and the exchange fields @xmath25 and @xmath26 , see equation ( [ e5:1 ] ) below .
it is just the process of revealing the origin of the closedness of @xmath27 which yields a detailed demonstration of the mathematical peculiarities of the mixtures : first it can be shown that @xmath27 is the euler class of an appropriately constructed fibre bundle over space - time ( density bundle ) so that the closedness relation for @xmath27 ( [ e5:3 ] ) appears as nothing else than the bianchi identity for the bundle curvature .
next it is proven that the winding numbers of the euler class @xmath27 are always zero and thus @xmath27 is an _ exact _
2-form over space - time ( * * e**=*de * , see equation ( [ e5:5 ] ) ) .
the origin of the exactness of @xmath27 lies in the fact that @xmath27 is the pullback of the surface 2-cell of the one - parted hyperboloid @xmath28 for the negative mixtures and of the two - parted hyperboloid @xmath29 for the positive mixtures , see fig . 1 .
since these two surfaces @xmath30 carry different topologies ( @xmath28 is infinitely connected , the connected components of @xmath29 are simply connected ) , a positive - mixture configuration can never be continously deformed into a negative mixture ( and vice versa ) .
this may be considered as the rst counterpart of the fermion - boson superselection rule of the conventional theory which is of purely phenomenological character .
the 2-particle configurations admit a concrete geometric realization if the underlying space - time manifold ( as the base space of the applied fibre bundles ) is parallelisable .
as is well - known in differential geometry , any parallelisable manifold admits a _ global _ frame of tangent vectors and thus admits the introduction of a _ flat _ connection for its tangent bundle
( though this does not imply in general the flatness of the tangent metric @xcite ) . on the other hand ,
the rst 2-particle mixtures require the existence of a trivial principal lorentz bundle over space - time , whose sections generate a flat lorentz connection via the ( pullback of the ) maurer - cartan form over the lorentz group @xmath31 .
the reduction of this 4-frame to a 2-frame then generates the bundle geometry of the 2-particle mixtures ( * section vi * ) .
in order to present the subsequent discussions in a sufficiently self - contained form , we briefly sketch the rst fundamentals with emphasis of those aspects which lead to the emergence of the matter dichotomy , the main subject of the present paper . according to the rst claim of representing a general theory of matter
, we evoke the mathematical framework of fibre bundles over the pseudo - riemannian space - time as the base space and we identify the corresponding bundle connection with the physical gauge potentials according to the principle of minimal coupling . thus the coordinate - covariant derivative @xmath32 refers to the levi - civita connection @xmath33 of the pseudo - riemannian fibre metric @xmath34 in the tangent bundle of space - time , the gauge potential @xmath35 acts as the bundle connection in the complex vector bundles over space - time ( with typical fibre @xmath36 ) .
the gauge - plus - coordinate covariant derivative of a bundle section is thus @xmath37 .
the n - component wave function @xmath14 represents a section of the vector bundle and the intensity matrix @xmath13 ( generalizing the concept of wave function ) is a hermitian section @xmath38 of the product bundle whose typical fibre is the tensor product @xmath39 of the vector fibres @xmath36 .
further examples of operator - valued sections are the anti - hermitian field strength @xmath40 as the curvature of the connection @xmath41 @xmath42 \ ; .\ ] ] the hamiltonian @xmath43 ( being neither hermitian nor anti - hermitian ) may be split up into its ( anti- ) hermitian parts as follows @xmath44 here , both the kinetic field @xmath45 and the localization field @xmath46 are hermitian @xmath47 and build up the ( anti-)hermitian constituents of the hamiltonian @xmath16 .
the basic dynamical equations of the theory may be considered here in a purely formal view as mathematically consistent links between those rst objects mentioned above .
the first equation to be mentioned is the relativistic schrdinger equation ( rse ) for the wave function @xmath14 @xmath48 or its generalization , resp . , the relativistic von neumann equation ( rne ) for the intensity matrix @xmath13 : @xmath49\ .\ ] ] contrary to schrdinger s non - relativistic quantum mechanics , the relativistic hamiltonian @xmath16 is here not a rigid object , to be fixed a priori , but is a dynamical variable obeying its own field equations .
the first of these is the `` _ integrability condition _ ''
@xmath50=i\hbar c { { \mathcal{f}}}_{\mu\nu}\ ] ] and the second is the `` _ conservation equation _ ''
@xmath51 with @xmath52 being the mass operator , i.e. for identical particles of mass m @xmath53 the meaning of the integrability condition ( [ eq : intcond ] ) is to ensure the existence of ( local ) solutions for the rst dynamics , but the conservation equation ( [ eq : coneq ] ) implies the existence of certain conservation laws , such as the charge conservation for @xmath17 particles @xmath54 @xmath55 or the energy - momentum law @xmath56 here the lorentz force density @xmath57 is being composed of the field strength @xmath58 and the currents @xmath59 as usual @xmath60 where the ( hermitian ) force operator @xmath61 is given by @xcite @xmath62 clearly if the energy - momentum content of the gauge field is included in the energy - momentum density @xmath63 , one has the usual energy - momentum conservation law @xcite for the total system @xmath64 the last equation , closing the whole dynamical system , is taken as the ( generally non - abelian ) maxwell equation @xmath65 ( @xmath66 ) .
thus in the last step , it remains to specify the current operator @xmath67 in terms of the matter variable @xmath13 and then one has a closed dynamical system . the desired relationship between the intensity matrix @xmath13 and current @xmath67 is obtained by decomposing the current operator @xmath67 with respect to the generators @xmath68 of the gauge group @xmath69 and then defining the current components @xmath70 as @xmath71 where the velocity operators read in terms of the hamiltonian @xmath16 for scalar particles @xmath72 ( the velocity operator for a spinning particle coincides with the dirac matrices : @xmath73 @xcite ) .
one can easily show that this arrangement automatically guarantees the validity of the charge conservation laws ( [ eq : chcon ] ) as a consequence of the rst dynamics with no need of additional requirements .
such a pleasant effect occurs also for the energy - momentum laws ( [ eq : enlaw ] ) and ( [ eq : encon ] ) .
indeed it is possible to define the energy - momentum content of matter @xmath63 in such a way that the laws ( [ eq : enlaw ] ) and ( [ eq : encon ] ) are _ automatically _ obeyed again as an implication of the rst dynamics itself .
more concretely , one defines the energy momentum density @xmath63 as @xmath74 with the energy - momentum operator @xmath75 for scalar particles being given through @xmath76\ , \ ] ] ( see ref .
@xcite for the energy - momentum operator of dirac particles ) .
if the gravitational interactions are to be included , one adds the einstein equations to the rst dynamics which however enforces the energy - momentum conservation law ( [ eq : encon ] ) as a consistency condition , in place of the weaker source equation ( [ eq : enlaw ] ) .
the stronger condition ( [ eq : encon ] ) is attained by adding the energy - momentum density of the gauge field to that of the matter field ( [ eq : ttrit ] ) which then yields a closed system with the strong law ( [ eq : encon ] ) playing the part of a closedness condition @xcite .
the preceding examples of _ automatic _ conservation laws have been presented in such a great detail because , for the present purpose of demonstrating the emergence of a rst matter dichotomy , it is important to have a further such automatic consequence of the rst dynamics : namely the emergence of a scalar amplitude field @xmath77 from the localization operator @xmath78 ( [ eq : hsplit ] ) . in order to see more clearly the reason why the general rst dynamics gives rise to such an amplitude scalar @xmath77 , one first transcribes the integrability condition ( [ eq : intcond ] ) to the kinetic and localization fields @xmath79 ( [ eq : hsplit ] ) and thus finds for the hermitian part @xmath45 of @xmath16 @xmath80-i \left[{{\mathcal{l}}}_\mu , { { \mathcal{l}}}_\nu\right ] = i{{\mathcal{f}}}_{\mu\nu } \ ; , \ ] ] and similarly for the localization field @xmath46 @xmath81 + i \left[{{\mathcal{k}}}_\mu , { { \mathcal{l}}}_\nu\right ] = 0 \ ; .\ ] ] by inspecting the trace of the localization field @xmath82 one finds for the trace @xmath83 of @xmath46 the following curl relation from ( [ eq : lherm ] ) : @xmath84 consequently , this vector field @xmath83 must be a gradient field generated by some scalar ( @xmath85 , say ) @xmath86 here , @xmath17 is chosen as the dimension of the typical vector fibre ( i.e. @xmath0 for two - particle systems , see below ) . the significance of this _ amplitude field _
@xmath77 for the matter dichotomy will readily become evident .
the very general features of rst mentioned up to now are already sufficient in order to conclude that the rst matter emerges in two different forms which are dynamically kept apart : the positive and negative mixtures , being separated by the pure states as a certain kind of borderline configuration . indeed , as will become evident through the following arguments , we even need only the rne ( [ eq : cali ] ) together with the integrability condition ( [ eq : intcond])in the form ( [ eq : curll])-([eq : lscal ] ) in order to establish the desired matter dichotomy .
this will be readily proven explicitly for the lowest - order cases @xmath87 and from here it becomes evident by means of some inductive arguments that the dichotomy must exist for any fibre dimension @xmath17 ( which is the minimal number of fibre dimensions necessary for the description of @xmath17 scalar particles ) .
the point of departure is the recursive construction of certain polynomials @xmath88 of degree @xmath89 for the intensity matrix @xmath13 . introducing here the _ deviation densities _ @xmath90 as the trace of the _ deviators _
@xmath88 , i.e. @xmath91 the deviators @xmath88 are defined through @xmath92 the starting point of the recursion is the lowest - order deviator @xmath93 which we put identical to the intensity matrix @xmath13 itself [ eq : d1 ] @xmath94 then comes the first ( non - trivial ) deviator @xmath95 @xmath96 correspondingly the second deviator reads @xmath97 @xmath98 and finally let us mention the third deviator @xmath99 @xmath100 it should be evident that this recursive construction of deviators becomes trivial when the deviation order @xmath101 coincides with the fibre dimension @xmath17 , i.e. we have @xmath102 thus the last non - trivial deviator for fibre - dimension @xmath17 is @xmath103 and the next one vanishes , @xmath104 , i.e. by means of equation ( [ eq : devi ] ) @xmath105 however , for a regular intensity matrix @xmath106 this implies for fibre dimension @xmath17 @xmath107 which is nothing else than the well - known _ hamilton - cayley identity _ of matrix calculus @xcite .
thus for fibre dimension @xmath0 we conclude for the @xmath108-matrix @xmath13 @xmath109 with the deviation density @xmath110 being identical to the determinant of the intensity matrix , i.e. @xmath111 similarly for the fibre dimension @xmath112 one deduces from equation ( [ eq : d3 ] ) for the @xmath113-matrix @xmath13 @xmath114 where the second deviation density @xmath115 is now the determinant of the intensity matrix @xmath13 @xmath116 or finally , for fibre dimension @xmath117 , the closing condition ( [ eq : dn ] ) yields on account of equation ( [ eq : d4 ] ) the corresponding hamilton - cayley identity @xmath118 with @xmath119 these few examples ( in place of a rigorous induction scheme ) may be sufficient in order to see the general formalism working well for any fibre dimension @xmath17 .
it will be used now in order to establish the desired dichotomy effect .
the interesting point with the rst dynamics is now that it subdivides the configuration space of matter ( i.e. the space of regular @xmath120-matrices ) into two dynamically separated subsets : the positive and negative mixtures , with the pure states appearing as an asymptotic limit of both kinds of mixtures .
here the criterium of membership refers to the determinant @xmath121 of the intensity matrix @xmath13 such that for @xmath122 one has a `` positive '' ( `` negative '' ) mixture . on the other hand
the pure states are uniquely characterized by the well - known fierz identity @xcite for the intensity matrix , i.e. @xmath123 which restricts @xmath13 to the tensor product of some wave function @xmath14 : @xmath124 indeed , this is easily verified by simply substituting the latter form of @xmath13 into the definition of the first deviator @xmath125 ( [ d2 ] ) .
for the special case of two fibre dimensions @xmath0 the hamilton - cayley identity ( [ eq : df ] ) says that the fierz identity upon @xmath13 ( [ eq : df0 ] ) is even equivalent to the one real condition of vanishing fierz deviation @xmath126 ( see below for a detailed treatment of the two - particle systems ) .
however , by what circumstance does this dichotomic mixture effect come about ?
the answer comes from the combination of the rne ( [ eq : cali ] ) with the existence of the amplitude field @xmath127 ( [ eq : lscal ] ) . here , the first step consists in splitting up the @xmath17-order deviator @xmath103 into the highest - power term of @xmath13 and the remainder @xmath128 : @xmath129 for instance from the definition of the first deviator @xmath125 ( [ d2 ] ) we obtain @xmath130 or similarly for the second deviator @xmath131 ( [ eq : d3 ] ) @xmath132 and finally for the third deviator @xmath133 ( [ eq : d4 ] ) @xmath134 the general procedure should be self - suggesting from these few examples .
in the next step , one considers the derivative of the determinant @xmath135 ( [ eq : deltan ] ) which is found to consist of two terms according to the splitting ( [ eq : dny ] ) @xmath136 the first term on the right is due to the derivative of the first term @xmath137 of the splitting ( [ eq : dny ] ) when the rne ( [ eq : cali ] ) has been used and after this that splitting ( [ eq : dny ] ) is applied again in the opposite direction leading back to @xmath103 .
however since @xmath103 is of that simple form ( [ eq : dn ] ) , namely proportional to the identity @xmath138 , the derivative of the determinant @xmath135 ( [ eq : split ] ) becomes @xmath139 provided the second term on the right of ( [ eq : split ] ) vanishes , i.e. concretely if we have the trace identity @xmath140 but this identity actually does hold , from which one is easily convinced by the above mentioned examples @xmath141 and the corresponding induction arguments .
indeed , one merely observes for the first term on the left of ( [ eq : trid ] ) @xmath142 with the trace of @xmath128 being easily computed by means of the definition of the deviation densities @xmath135 ( [ eq : deltan ] ) for @xmath143 .
thus for @xmath117 , e.g. , one immediately finds from equation ( [ eq : y4 ] ) @xmath144 on the other hand , the computation of the second term on the left of ( [ eq : trid ] ) just meets with this result because for the lower - order deviation densities @xmath145 one also has @xmath146 cf .
( [ eq : split ] ) , and this then validates definitively the trace identity ( [ eq : trid ] ) .
observe here that the latter relation ( [ eq : trdnl ] ) , though being seen now to hold for all deviation orders @xmath101 up to the maximal value @xmath147 , had to be exploited only for the maximal order @xmath17 ( in order to yield the desired result ( [ eq : ddeltan ] ) because the closing relation ( [ eq : dn ] ) does apply exclusively to this maximal order @xmath17 ! once the decisive derivative relation ( [ eq : ddeltan ] ) is now firmly established , the matter dichotomy emerges by simply observing the former gradient property for the location field @xmath46 ( [ eq : trl])-([eq : lscal ] ) . indeed
the existence of the amplitude field @xmath77 ( [ eq : lscal ] ) admits to formally integrate the differential equation ( [ eq : ddeltan ] ) for @xmath135 to yield @xmath148 thus , at any event @xmath149 of space - time the determinant of the intensity matrix @xmath13 has the same sign as at the initial event @xmath150 provided both events @xmath149 and @xmath150 can be connected by some path along which the amplitude field @xmath77 is a regular scalar .
the conclusion is that whenever the rst dynamics admits a regular solution over some region of space - time , then this field configuration must be either a positive mixture @xmath151 or a negative mixture @xmath152 or a `` quasi - pure state '' @xmath153 .
the latter kind of mixtures are not ordinary pure states ( which have @xmath154 , see ( [ eq : df0 ] ) ) but they are not forbidden by the rst dynamics to decay to such an ordinary pure state ( with vanishing deviator @xmath125 ) ; however the positive and negative mixtures must strictly preserve their mixture character over the whole domain of regularity in space - time ( _ `` matter dichotomy '' _ ) .
when matter appears in two different forms , one clearly wants to see the common features and also the differences . for a study of such questions in the subsequent sections ,
we restrict ourselves to the two - particle systems ( @xmath0 ) and we specialize first the general dynamical equations for this situation . here
it is important to parametrize the rst field system in such a way that a clear separation between single - particle variables and exchange fields comes about .
such a conceptual separation of the two field subsystems facilitates afterwards the discussion of the exchange effects occuring between the single - particle subsystems . as attractive as the covariant equations may look like in their abstract form , for concrete computations one mostly resorts to a special reference system and decomposes the covariant objects with respect to such a reference frame .
the corresponding components of the ( gauge and/or coordinate ) covariant objects are then treated as the proper dynamical objects entering the component form of the abstract dynamical equations . for the present rst operator dynamics ,
one thus chooses first some orthonormal basis for the operators acting over the typical vector fibre @xmath155 of the two - particle systems .
this may be done by choosing two _ single - particle projectors _
@xmath156 ( @xmath157 ) [ e4:1 ] @xmath158 and by complementing these by two _ permutators _
@xmath159 , @xmath160 which obey the following algebraic relations : [ e4:2 ] @xmath161 & = 2i({{\mathcal{p}}}_1-{{\mathcal{p}}}_2 ) \\ \big\{\hat{\pi},{{\mathcal{p}}}_a \big\ } & = -i\big[\tilde{\pi},{{\mathcal{p}}}_1\big]=i\big[\tilde{\pi},{{\mathcal{p}}}_2\big]=\hat{\pi } \\
\big\{\tilde{\pi},{{\mathcal{p}}}_a \big\ } & = - i\big[\hat{\pi},{{\mathcal{p}}}_1\big]=i\big[\hat{\pi},{{\mathcal{p}}}_2\big]=\tilde{\pi } \ ; .
\end{aligned}\ ] ] the four ( hermitian ) operators @xmath162 , @xmath159 , @xmath163 form a complete basis over the 2-particle fibre @xmath155 and consequently the generators @xmath164 ( [ eq : jtau ] ) of the gauge group must be certain linear combinations in this basis set ; for the ( abelian ) electromagnetic interactions one has @xcite @xmath165 and thus the ( anti - hermitian ) bundle curvature @xmath166 and its connection @xmath167 are decomposed as follows [ e4:4 ] @xmath168 once the reference section for the operators has been fixed , any operator - valued section can be decomposed with respect to this _ rotating basis _ @xmath162 , @xmath159 , @xmath163 , e.g. for the intensity matrix @xmath13 @xmath169 properly speaking , any hermitian ( @xmath170)-matrix ( such as @xmath13 ) in general has @xmath171 components but the rotating basis is co - moving with the intensity matrix in such a way that one of the four components of @xmath13 is kept zero permanently . due to the orthonormality of the rotating operator basis
the components @xmath172 , @xmath173 of @xmath13 can be extracted from @xmath13 in the usual way applicable to any hilbert - space structure , namely the _ single - particle densities _ @xmath172 as @xmath174 and the _ overlap density _ @xmath173 as @xmath175 similarly , the decomposition of the kinetic field @xmath176 ( [ eq : hsplit ] ) reads @xmath177 or for the localization field @xmath178 @xmath179 the coefficients @xmath180 , @xmath181 in front of the projectors @xmath156 are the _ single - particle _ components and correspondingly the coefficient fields in front of the permutators @xmath159 , @xmath160 are the _ exchange fields_. from ( [ e4:9 ] ) one immediately recognizes the former amplitude vector @xmath182 ( [ eq : trl ] ) as the sum of the corresponding single - particle vectors @xmath183 one could now write down the whole rst dynamics ( [ eq : rse])-([eq : coneq ] ) in component form referring to those component fields just introduced through ( [ e4:5])-([e4:9 ] ) but for the interpretation of the physical effects it is very favorable to pass over to a new combination of these component fields .
the first hint upon the advantage of dealing with a new combination of the exchange fields @xmath184 ( [ e4:8])-([e4:9 ] ) comes from the density dynamics , i.e. the component form of the rne ( [ eq : cali ] ) [ e4:11 ] @xmath185 here we have made use of the total density @xmath186 ( [ eq : d1]b ) as the trace of the intensity matrix @xmath13 ( [ e4:5 ] ) @xmath187 and also of the relative density @xmath188 as well as the difference ( @xmath189 ) of localization fields @xmath181 @xmath190 the interesting point with the density system ( [ e4:11 ] ) is now that it admits a conserved quantity ( @xmath191 , say ) emerging from the following relation which is a consequence of that system : @xmath192 this may be formally integrated to yield @xmath193 clearly this result is nothing else than the 2-particle specialization ( @xmath0 ) of the general result ( [ eq : dndetin ] ) , since for two fibre dimensions the first deviation density @xmath194 ( [ d2 ] ) is found as @xmath195 in this way it is once more seen very clearly that the emergence of the rst matter dichotomy is an immediate consequence of the dynamical equations , not a kinematical postulate as in the conventional quantum theory .
now the latter result ( [ e4:16 ] ) suggests to reparametrize the densities @xmath186 , @xmath196 , @xmath173 by means of the amplitude field @xmath77 and certain renormalization factors @xmath197 , @xmath198 , @xmath199 as follows [ e4:18 ] @xmath200 as a consequence of this arrangement the following constraint upon the renormalization factors is obtained @xmath201 with the _ mixing index _
@xmath191 taking the values : @xmath202 ( positive mixtures ) , @xmath203 ( pure states ) , and @xmath204 ( negative mixtures ) ( fig.1 ) . thus
, for such a small fibre dimension ( @xmath0 ) the quasi - pure states coincide with the ordinary pure states , see the remarks below equation ( [ eq : dndetin ] ) .
clearly with that arrangement ( [ e4:18 ] ) for the renormalization factors , their dynamical equations are deduced from the original density dynamics ( [ e4:11 ] ) and are thus found to be of the following form : [ e4:20 ] @xmath205 the second step of reparametrization refers now to the exchange fields @xmath206 , @xmath207 . from the third equation ( [ e4:20]c ) of the renormalization dynamics one concludes that the special combination of exchange fields @xmath208 , @xmath209 emerging there on the right - hand side must be a gradient field ( @xmath210 , say ) : @xmath211 this suggests to introduce a new combination ( @xmath210 , @xmath212 ) in place of the original ( @xmath207 , @xmath206 ) in the following way : [ e4:22 ] @xmath213 with respect to these new exchange fields , the renormalization dynamics ( [ e4:20 ] ) adopts now the corresponding new form [ e4:23 ] @xmath214 this reparametrization embraces different kinds of mixtures via the new scalar field @xmath215 which is defined through @xmath216 and which obeys the field equation @xmath217 from its definition ( [ e4:24 ] ) one immediately concludes that @xmath218 for the positive mixtures ( @xmath202 ) and that @xmath219 for the negative mixtures ( @xmath204 ) while @xmath220 for the pure states ( @xmath203 ) . since the different kinds of mixtures are kept separated for dynamical reasons ( matter dichotomy ) , the field equation for the _ mixture variable _
@xmath24 must have solutions which strictly respect this dynamical separation . however
this requirement is easily seen to be satisfied because the new exchange field @xmath210 ( [ e4:22]a ) is found to be a gradient field , cf .
( [ e4:21 ] ) @xmath221 and thus the field equation for @xmath24 ( [ e4:25 ] ) is easily integrated to yield @xmath222 \displaystyle \frac{s^4}{s^4-s_c{}^4 } \quad\mbox { { \rm , $ \sigma_{\ast}=-1 $ } } \end{array } \right.\ ] ] with @xmath223 being an arbitrary integration constant .
though this result merely reproduces the original definition of @xmath24 ( [ e4:24 ] ) through the identification @xmath224 the differential version of which is just the former equation ( [ e4:21 ] ) , we thus can be sure that the matter dichotomy is correctly incorporated also into the dynamics of the new set of variables : the field equation for @xmath24 ( [ e4:25 ] ) admits either solutions ( [ e4:26_2 ] ) with @xmath225 ( @xmath226 positive mixtures : @xmath202 , @xmath227 ) or solutions with @xmath228 ( @xmath226 negative mixtures : @xmath204 , @xmath229 ) .
this is the two - particle realization of the general result explained at the end of section iii , below equation ( [ eq : dndetin ] ) .
a special situation is encountered for vanishing exchange fields , especially @xmath230 . for this case , the new scalar field @xmath231 ( [ e4:26 ] ) must be a constant over space - time ( @xmath232 , say ) which then implies the constancy of the new variable @xmath24 ( @xmath233 ) through the present result ( [ e4:26_2 ] ) @xmath234 thus , the vanishing of the exchange fields leaves us alone with the mixture configurations which then are found to be degenerate in the sense that they generate the same physics as the pure states ( _ mixture degeneracy _
@xcite ) . in this case
there exists a group of mixing transformations ( _ mixing group _ ) which admits to construct the mixture configurations from the pure states and the group parameter is just the present constant @xmath235 ( [ e4:28 ] ) .
thus we clearly recognize the effect of switching - on the exchange fields , namely to make the mixing group parameter space - time dependent : @xmath236 .
the resulting cooperation of mixture and exchange effects is just the essence of the rst quantum effects ( see a separate paper ) .
the reparametrization of exchange fields ( [ e4:22 ] ) is not complete because it refers to only two of the four exchange fields .
however a corresponding transformation for the remaining exchange fields @xmath237 and @xmath238 suggests itself when we consider now the physical densities . turning first to the charge densities
@xmath239 ( [ eq : jtriv ] ) , one substitutes the decompositions of the intensity matrix @xmath13 ( [ e4:5 ] ) and of the hamiltonian @xmath43 ( [ e4:8])-([e4:9 ] ) into that definition ( [ eq : jtriv ] ) for the currents @xmath239 and then finds the following result @xcite [ e4:29 ] @xmath240 the specific shape of this result suggests now to introduce two new fields \{@xmath241,@xmath242 } in place of \{@xmath237 , @xmath238 } , namely in the following way [ e4:30 ] @xmath243 the reason for this arrangement is because both currents @xmath239 ( [ e4:29 ] ) adopt a very simple form , namely [ e4:31 ] @xmath244 and this form even simplifies further to [ e4:32 ] @xmath245 when one simultaneously introduces the _ effective kinetic fields _
@xmath246 ( @xmath247 ) through [ e4:33 ] @xmath248 thus the final form of the currents @xmath239 ( [ e4:32 ] ) is the same which is obtained also by putting all the exchange fields to zero in the original version ( [ e4:29 ] ) and replacing the original kinetic fields @xmath180 by their effective counterparts @xmath246 ( [ e4:33 ] ) .
such an absorption of the exchange fields into the single - particle fields in order to build up the effective fields is not only a purely formal advantage but yields also a deeper insight into the physical essence of the theory .
for the special case of pure states ( @xmath220 ) , the present reparametrization of the exchange fields in terms of @xmath210 , @xmath212 , @xmath241 , @xmath242 has been described in a preceding paper @xcite where the new parametrization was helpful for a discussion of the exchange degeneracy .
once the kinetic fields have been transcribed to a new form , it is strongly suggestive to try a similar transcription for the localization fields @xmath181 ( [ e4:9 ] ) .
the precise form of the desired transformation to the _ effective localization fields _
@xmath249 ( @xmath247 ) comes from a closer inspection of the energy - momentum density @xmath250 ( [ eq : ttrit ] ) which however proceeds along the same line of arguments as applied also to the currents @xmath239 .
thus one is led to introduce the effective localization fields @xmath249 through [ e4:34 ] @xmath251 and these new fields build then up the energy - momentum density @xmath250 , albeit in a somewhat more complicated way as was the case with the kinetic fields in connection with the currents @xmath239 ( [ e4:32 ] ) . in order to display the general structure of @xmath250 ,
it is convenient to split it up into three parts : @xmath252 where the first two terms @xmath253 ( @xmath247 ) are the single - particle contributions and the third term @xmath254 contains the exchange and mixture effect .
similarly as for the currents @xmath239 , the single - particle contributions @xmath253 do no longer contain here explicitly the exchange fields since these have been absorbed into the effective single - particle fields @xmath246 , @xmath249 : [ e4:36 ] @xmath255 \bigg\ } \\[3 mm ] { } ^{(2)}t_{\mu \nu } = & \frac{\hbar^2}{m}\ , \rho_2 \bigg\ { { \mathbb{k}}_{2 \mu}{\mathbb{k}}_{2 \nu}+{\mathbb{l}}_{2 \mu}{\mathbb{l}}_{2 \nu } \nonumber \\ & -\frac{1}{2}g_{\mu \nu } \bigg[{\mathbb{k}}_{2 \lambda}{\mathbb{k}}_2{}^{\lambda } + { \mathbb{l}}_{2 \lambda}{\mathbb{l}}_2{}^{\lambda } - \left(\frac{mc}{\hbar } \right)^2 \bigg ] \bigg\ } \ ; .\end{aligned}\ ] ] on the other hand , the exchange term @xmath254 is built up itself from two contributions @xmath256 which do not contain the single - particle fields and thus look as follows [ e4:38 ] @xmath257 \big\ } \\ { } ^{(2)}t_{\mu \nu } = & \frac{\hbar^2}{m}\ , \lambda(1-\lambda)\rho_2 \big\ { ( f_{\mu}+g_{\mu})(f_{\nu}+g_{\nu } ) + ( s_{\mu}+t_{\mu})(s_{\nu}+t_{\nu } ) \nonumber \\[3 mm ] & -\frac{1}{2}g_{\mu \nu } \big[(f_{\lambda}+g_{\lambda})(f^{\lambda}+g^{\lambda } ) + ( s_{\lambda}+t_{\lambda})(s^{\lambda}+t^{\lambda } ) \big ] \big\ } \ ; .\end{aligned}\ ] ] observe here that , in contrast to the currents @xmath239 ( [ e4:32 ] ) , the energy - momentum density @xmath250 ( [ e4:35 ] ) contains an extra term ( namely @xmath254 ) which even after the absorption of the exchange fields into the new single - particle fields @xmath246 , @xmath249 is built up explicitly by all the exchange fields ( @xmath210 , @xmath212 , @xmath241 , @xmath242 ) together with the mixture variable @xmath215 . such a term in @xmath250 is necessary in order to validate the energy - momentum conservation law ( [ eq : encon ] ) ( or ( [ eq : enlaw ] ) , resp . ) whereas the charge conservation law ( [ eq : chcon ] ) works without such an _ explicit _ exchange term . observe also that the latter term @xmath254 is built up by _ both _ the mixture variable @xmath215 _ and _ the exchange fields and thus represents the cooperation of both kinds of field degrees of freedom .
if one wishes to convince oneself of the consistency of those conservation laws for the physical densities , one has to use the source and curl relations for the hamiltonian component fields which of course are to be deduced from the abstract integrability condition ( [ eq : intcond ] ) and conservation equation ( [ eq : coneq ] ) .
however , with the present reparametrization one will readily transcribe the field equations to the new variables , i.e. the single - particle fields \{@xmath246 , @xmath249 } and exchange fields \{@xmath210 , @xmath212 , @xmath241 , @xmath242}. in this way there emerges a kind of _ exchange dynamics _ for the latter fields and a _ single - particle dynamics _ for the first fields .
_ single - particle dynamics _ in fact , the integrability condition yields first for the original localization fields @xmath181 ( [ e4:9 ] ) @xmath258=2 \lambda\ , [ s_{\mu}t_{\nu}-s_{\nu}t_{\mu}-f_{\mu}g_{\nu}+f_{\nu}g_{\mu}]\;.\ ] ] this once more verifies the previous claim ( [ eq : trl])-([eq : lscal ] ) that the total localization field @xmath182 @xmath259 actually is a gradient field , in contrast to the individual fields @xmath181 ( @xmath247 ) .
though this result forbids us to conceive the latter fields @xmath260 as gradient fields , we nevertheless can introduce individual amplitude fields @xmath261 ( @xmath247 ) for the effective localization fields @xmath249 ( [ e4:34 ] ) through putting @xmath262 for the verification of this claim one merely recalls the density dynamics ( [ e4:11 ] ) and uses therein the new exchange fields \{@xmath210 , @xmath212 } in place of the old ones \{@xmath209 , @xmath208 } , cf .
( [ e4:22 ] ) . in this way
one finds for the single - particle densities @xmath172 ( @xmath247 ) [ e4:42 ] @xmath263 which actually identifies the effective localization fields @xmath249 as gradient fields @xmath264 in agreement with the claim ( [ e4:41 ] ) . once the transition from the localization fields @xmath181 to the amplitude fields @xmath265 ( [ e4:41 ] ) has been made , one will not be satisfied with transcribing the abstract conservation equation ( [ eq : coneq ] ) to the localization field @xmath181 ( or @xmath249 , resp . ) but one will readily proceed to the corresponding wave equations for the scalar amplitudes @xmath265 ( _ amplitude equations _ ) [ e4:44 ] @xmath266 \box \ , { \mathbb{l}}_2 & + { \mathbb{l}}_2 \big\ { \big ( \frac{mc}{\hbar } \big)^2-{\mathbb{k}}_{2 \mu}{\mathbb{k}}_2{}^{\mu } \big\ } \nonumber\\ & = \lambda(1-\lambda ) { \mathbb{l}}_2 \big\ { ( s^{\mu}+t^{\mu})(s_{\mu}+t_{\mu})+(f^{\mu}+g^{\mu})(f_{\mu}+g_{\mu } ) \big\ } \ ; .\end{aligned}\ ] ] observe here that these single - particle amplitude equations degenerate to their trivial forms ( @xmath247 ) @xmath267 when _ either _ one deals with pure states ( @xmath268 ) _ or _ when one neglects the exchange fields ( @xmath269 ) .
for such a situation , the degenerate equations ( [ e4:45 ] ) can be shown to be equivalent to two decoupled klein - gordon equations @xcite for two single - particle wave functions @xmath270 ( @xmath247 ) : @xmath271 ( @xmath272 ) .
evidently , the two particles are now disentangled and thus can interact exclusively via the ordinary gauge forces mediated by the corresponding gauge potentials @xmath273 .
either of the two wave functions @xmath274 is then of the usual form [ e4:47 ] @xmath275 where the phases @xmath276 are given by @xmath277 but clearly the typical quantum effects are expected to occur in connection with the simultaneous presence of non - trivial exchange fields and mixtures ( @xmath278 ) . here the two independent klein - gordon equations ( [ e4:45])-([e4:46 ] ) become modified to their `` entangled '' forms ( [ e4:44 ] ) yielding a coupling of the single - particle amplitudes @xmath279 to the exchange fields with the occurrence of certain _ exchange effects _ ( separate paper ) .
a similar coupling of single - particle and exchange fields is naturally to be expected also for the kinetic fields .
indeed , the following curl relations for the effective kinetic fields @xmath246 ( @xmath247 ) are deduced from the abstract integrability condition ( [ eq : intcond ] ) in a straightforward manner : [ e4:49 ] @xmath280\\ \nabla_{\mu}{\mathbb{k}}_{2 \nu}-\nabla_{\nu}{\mathbb{k}}_{2 \mu}&= f_{2 \mu \nu}+2\lambda ( 1-\lambda)\,\big [ ( s_{\mu}+t_{\mu})(f_{\nu}+g_{\nu})-(s_{\nu}+t_{\nu})(f_{\mu}+g_{\mu } ) \big ] \;.\end{aligned}\ ] ] from here it becomes obvious that the effective kinetic fields @xmath246 do not only `` feel '' the corresponding components @xmath281 of the bundle curvature @xmath282 ( [ e4:4]a ) but they feel an additional _ exchange field strength _ ,
namely the second terms on the right of equation ( [ e4:49 ] ) which is built up essentially by the exchange fields @xmath210 , @xmath212 , @xmath241 , @xmath242 .
however for the pure states ( @xmath283 ) this exchange effect vanishes again and thus we arrive at the same result as mentioned previously in connection with the energy - momentum density @xmath250 ( [ e4:35 ] ) and the amplitude equations ( [ e4:44 ] ) , namely that the exchange effects can become active only for the mixtures but not for the pure states .
observe however that the present exchange effect , occuring for the effective kinetic fields @xmath246 ( [ e4:49 ] ) , is somewhat different in comparison to the preceding situations of energy - momentum density @xmath284 ( [ e4:38 ] ) and amplitude equation ( [ e4:44 ] ) ; the reason for this is that the kinetic exchange term ( [ e4:49 ] ) vanishes if one of the two exchange pairs \{@xmath210 , @xmath212 } or \{@xmath241 , @xmath242 } vanishes whereas the other two exchange effects can survive also with only one of the two exchange pairs being present . in order to close the single - particle dynamics for the effective fields
, one has to write down the source relations for the kinetic fields .
the desired relations are obtained again in a straightforward way from the abstract conservation equation ( [ eq : coneq ] ) : [ e4:50 ] @xmath285 observing here the gradient property of the effective localization fields @xmath249 ( [ e4:41 ] ) , one easily recognizes the present source relations ( [ e4:50 ] ) as nothing else than the former charge conservation laws ( [ eq : chcon ] ) when the result ( [ e4:32 ] ) for the currents @xmath239 is observed together with the density derivatives ( [ e4:42 ] ) .
_ exchange dynamics _
as the preceding single - particle dynamics demonstrates , there is a coupling of the single - particle fields to the exchange fields , see the amplitude equations ( [ e4:44 ] ) and curl relations ( [ e4:49 ] ) for the effective kinetic fields .
this coupling shows that the single - particle dynamics decouples from the exchange fields only for the pure states ( @xmath283 ) but not for the mixtures ( @xmath278 ) .
thus , when dealing with the pure states , one can first solve the autonomous single - particle system and substitute this as a kind of rigid background into the exchange dynamics @xcite .
however for the mixtures , one must solve the single - particle system in one step together with the exchange system because also the latter system couples back to the first system .
this coupling is explicitly seen by writing down the source relations for the exchange fields ( deducible again from the conservation equation ( [ eq : coneq ] ) ) : [ e4:51 ] @xmath286 \nabla^{\mu}t_{\mu } = & ( s^{\mu}-t^{\mu}){\mathbb{l}}_{1 \mu}-(s^{\mu}+t^{\mu}){\mathbb{l}}_{2 \mu}-(f^{\mu}-g^{\mu}){\mathbb{k}}_{1 \mu}+(f^{\mu}+g^{\mu}){\mathbb{k}}_{2 \mu } \nonumber\\ & + 2(2\lambda-1)(s^{\mu}t_{\mu}-f^{\mu}g_{\mu})\\[3 mm ] \nabla^{\mu}f_{\mu } = & -(f^{\mu}-g^{\mu}){\mathbb{l}}_{1 \mu}-(f^{\mu}+g^{\mu}){\mathbb{l}}_{2 \mu}-(s^{\mu}-t^{\mu}){\mathbb{k}}_{1 \mu}-(s^{\mu}+t^{\mu}){\mathbb{k}}_{2 \mu } \nonumber\\ & + 2(\lambda-1)g^{\mu}t_{\mu}+2(3\lambda-1)f^{\mu}s_{\mu}\\[3 mm ] \nabla^{\mu}g_{\mu } = & ( f^{\mu}-g^{\mu}){\mathbb{l}}_{1 \mu}-(f^{\mu}+g^{\mu}){\mathbb{l}}_{2 \mu}+(s^{\mu}-t^{\mu}){\mathbb{k}}_{1 \mu}-(s^{\mu}+t^{\mu}){\mathbb{k}}_{2 \mu } \nonumber\\ & + 2(3\lambda-2)s^{\mu}g_{\mu}+2\lambda f^{\mu}t_{\mu } \;.
\end{aligned}\ ] ] observe here that even for the pure states ( @xmath283 ) the exchange fields do not decouple from the single - particle fields .
such a decoupling effect occurs exclusively for the curl relations ( to be deduced from the integrability condition([eq : intcond ] ) ) : [ e4:52 ] @xmath287 + 2[g_{\mu}f_{\nu}-g_{\nu}f_{\mu}]\\ \nabla_{\mu}f_{\nu}-\nabla_{\nu}f_{\mu}&=2(\lambda-1)[s_{\mu}f_{\nu}-s_{\nu}f_{\mu}+g_{\mu}t_{\nu}-g_{\nu}t_{\mu}]\\ \nabla_{\mu}g_{\nu}-\nabla_{\nu}g_{\mu}&=2\lambda [ f_{\mu}t_{\nu}-f_{\nu}t_{\mu}+s_{\mu}g_{\nu}-s_{\nu}g_{\mu } ] \ ; .\end{aligned}\ ] ] here , the first equation ( [ e4:52]a ) says that @xmath210 must always be a gradient field which however is clear from its very definition ( [ e4:21 ] ) .
furthermore the vector field @xmath241 becomes also a gradient field for the pure states ( @xmath283 ) , cf .
( [ e4:52]c ) .
thus the new parametrization yields not only a formal simplification of the whole rst dynamics but makes also the theory more transparent by providing us with a clear subdivision of the field variables into single - particle and exchange objects . as a consequence , one becomes able to discuss the rst predictions in terms of single - particle concepts and exchange interactions .
whenever a superselection rule is seen to be in action , such as the present dynamical separation of positive and negative mixtures , one tends to guess that such a strict rule might have to do something with the global topological features of the corresponding field configurations .
indeed it is easily seen from fig.1 that the one - parted hyperboloid @xmath28 , being infinitely connected , can not be continuously deformed into the two - parted hyperboloid @xmath29 whose individual parts are simply connected .
therefore it is self - suggesting to guess that this topological arrangement is somehow encoded into the field variables of positive and negative mixtures .
thus there arises an attractive mathematical problem , namely to identify the place where the two different topologies are buried in the space - time dependence of the field variables .
the answer to this question is the following : take the renormalization factor @xmath288 ( [ e4:18]c ) and the mixture variable @xmath215 ( [ e4:24 ] ) together with the 1-forms @xmath289 and @xmath290 , cf . ( [ e4:30 ] ) , and construct the 2-form @xmath291 @xmath292 i.e. in components @xmath293 \ ; .\ ] ] then use the field equations for the mixture variable @xmath24 ( [ e4:25 ] ) and renormalization factor @xmath199 ( [ e4:23]c ) together with the curl relations for @xmath294 and @xmath295 ( [ e4:52 ] ) in order to show that the 2-form @xmath296 is closed : @xmath297 , i.e. in components @xmath298 actually @xmath296 is an element of the derham cohomology group @xmath299 over space - time ( @xmath300 ) , and its value upon some 2-cycle @xmath301 is a fixed integer @xmath302 $ ] @xmath303=\oint_{c^2}{\mathbf e}\ ] ] being independent of continuous deformations of the closed 2-surface @xmath301 . furthermore , for the present 2-particle case ( @xmath0 ) the `` winding numbers '' @xmath302 $ ] are found to be always zero for both types of mixtures , no matter which element @xmath301 of the corresponding homology group of the underlying space - time is taken as the integral surface .
thus @xmath296 turns out to be an exact 2-form : @xmath304 , in components @xmath305 with @xmath306 some appropriate ( non - unique ) 1-form .
though both mixtures have the same ( trivial ) winding numbers @xmath302 $ ] , the 2-form @xmath296 carries nevertheless different topological information .
it can namely be shown that @xmath296 is the _ euler class _ of appropriate fibre bundles ( _ density bundles _ , to be constructed below ) .
this means concretely that there exists a regular `` density map '' ( @xmath307 ) from space - time to the hyperboloids @xmath30 in 3-dimensional , pseudo - euclidian density configuration space ( fig.1 ) [ e5:6 ] @xmath308 such that the 2-form @xmath296 ( [ e5:1 ] ) yields just the 2-volume form on those hyperboloids ( up to an exact 2-form @xmath309 ) : @xmath310 therefore the winding numbers ( [ e5:4 ] ) measure the surface content ( on the hyperboloids @xmath30 ) of the image of the 2-cycle @xmath301 with respect to the density map @xmath311 ( [ e5:6 ] ) .
but since both hyperboloids @xmath30 as non - compact 2-surfaces have trivial second homotopy group ( @xmath312 ) , the image of the compact @xmath301 ( e.g. 2-sphere @xmath313 ) under @xmath311 yields a multiple covering ( with alternative orientations ) of some subspace of @xmath30 and this must necessarily imply the vanishing of the winding numbers @xmath302 $ ] ( [ e5:4 ] ) .
would the density map @xmath311 lead into some image space with non - trivial second homotopy group @xmath314 ( , as it may be expected for particle number @xmath315 ) , the winding numbers of positive and negative mixtures could well be different in a non - trivial way . in what follows a proof is presented for all these claims , leading directly to the central result ( [ e5:7 ] ) . for a deeper understanding of the mixture topology it is highly instructive to evoke the mathematical tool of fibre bundles . within this framework ,
one recovers not only the decisive curl relation ( [ e5:3 ] ) ( which one could find also by merely applying the dynamical equations in a straight - forward manner ) but one becomes capable to identify that relation as the _ bianchi identity _ for the bundle curvature @xmath296 . in this way
one arrives at the true origin of the winding numbers @xmath302 $ ] ( [ e5:4 ] ) . for the construction of the fibre bundle in question ,
recall first that the intensity matrix @xmath13 , as an hermitian @xmath316-matrix , properly speaking has four independent components .
these however could be reduced to only three real components by resorting to a co - moving reference system ( the _ rtb basis _
@xcite ) , so that the intensity matrix @xmath13 reappears as a 3-component object , cf . ( [ e4:5 ] ) . as the three remaining components one may take the overlap density @xmath173 ( [ e4:7 ] ) , the total density @xmath186 ( [ e4:12 ] ) and the relative density @xmath196 ( [ e4:13 ] ) which obey the density dynamcis ( [ e4:11 ] ) and give rise to the introduction of the renormalization factors @xmath317 , @xmath318 , @xmath199 ( [ e4:18 ] ) .
the interesting point here is now that the latter three objects may be collected into a 3-vector @xmath319 ( @xmath320 ) as an element of a 3-dimensional real vector fibre over any event @xmath149 of space - time : [ e5:8 ] @xmath321 this 3-dimensional vector fibre is equipped with a minkowskian fibre metric @xmath322 so that the normalization constraint ( [ e5:6]b ) simply says that @xmath311 ( [ e5:8 ] ) is a unit vector @xmath323 namely either `` time - like '' ( @xmath202 ) or `` space - like '' ( @xmath324 ) , resp .
`` light - like '' ( @xmath203 ) .
the next interesting point comes into play when one wishes to equip this density bundle ( @xmath311 ) with a connection ( @xmath325 , say ) . in order to preserve the normalization ( [ e5:9 ] )
, the connection 1-form @xmath325 must take its values in the 3-dimensional lorentz algebra @xmath326 .
the latter algebra may be thought to be spanned by three anti - hermitian generators @xmath327 where the concept of hermiticity refers here to the minkowskian metric @xmath328 .
this means that when @xmath329 are the elements of the matrix @xmath330 and @xmath331 are the elements of @xmath332 then we have @xmath333 the standard solution of this requirement for the generators is the well - known permutation symbol @xmath334 as the completely antisymmetric unit tensor object of the third rank @xmath335 the corresponding commutation relations are easily found as @xmath336=\epsilon^l{}_{jk}{{\mathcal{b}}}_l \ ; .\ ] ] thus the desired bundle connection @xmath325 of ( @xmath311 ) , being defined through @xmath337 ( in components : @xmath338 ) , may be decomposed with respect to the anti - hermitian basis @xmath339 as @xmath340 ( in components : @xmath341 ) . in order to explicitly determine the connection @xmath325
, one could think now that the connection coefficients @xmath342 could simply be read off from the renormalization dynamics ( [ e4:23 ] ) which would yield the following result [ e5:15 ] @xmath343 however the problem is here that this connection is not unique since one can add to any @xmath325 some @xmath326 element ( @xmath344 , say ) which annihilates the 3-vector @xmath311 , i.e. @xmath345 : @xmath346 here , the @xmath347-valued 1-form @xmath348 is arbitrary and @xmath344 is easily seen to decompose with respect to the standard basis ( [ e5:11 ] ) as follows @xmath349 this non - uniqueness of the connection @xmath325 in the density bundle ( @xmath311 ) must not necessarily be considered as a deficiency of the theory but may rather be interpreted as the presence of a gauge degree of freedom inherent in the bundle formalism .
this means that we admit @xmath350-valued gauge transformations @xmath351 which act homogeneously over the vector fibre in the following way [ e5:18 ] @xmath352 ( i.e. in components @xmath353 ) .
these gauge transformations induce an inhomogeneous transformation law for the bundle connection @xmath325 : @xmath354 now we try to restrict the gauge group to a subgroup of @xmath350 in such a way that the original non - uniqueness relation ( [ e5:16 ] ) becomes identical to the present gauge transformation ( [ e5:19 ] ) .
this requirement fixes the elements @xmath355 of the restricted gauge group as @xmath356 \ ; , \ ] ] i.e. the abelian rotations in the orthogonal 2-space of the density vector @xmath311 which itself is invariant under the restricted gauge group : @xmath357 furthermore , it can be shown by some simple bundle analysis that the 1-form @xmath348 , emerging in the non - uniqueness relation ( [ e5:16 ] ) , is changed by the derivative of the group parameter @xmath358 ( [ e5:20 ] ) into @xmath359 : @xmath360 this analysis becomes quite easy by resorting to the fact that the generator of rotations @xmath344 ( [ e5:17 ] ) is covariantly constant @xmath361 ( @xmath362 $ ] ) , or in components : @xmath363 ( @xmath364 ) . it is important to remark here that the restricted gauge transformations ( [ e5:20 ] ) change the 1-form @xmath348 in the non - uniqueness relation ( [ e5:16 ] ) by a gradient field @xmath365 , cf . ( [ e5:22 ] ) , and therefore @xmath348 can be gauged off exclusively when the original @xmath348 is also a gradient field . on the other hand , one can take advantage of the non - uniqueness of @xmath325 by letting vanish its component @xmath366 relative to the gauge algebra \{@xmath344 } @xmath367 in this way one is lead to an other connection ( the _ projective connection _
@xmath368 ) which is obtained from the original @xmath325 by subtracting off its non - unique @xmath344 part , i.e. one puts @xmath369 this does not change the topological properties of the bundle connection because the euler class @xmath370 ( [ e5:7 ] ) is modified merely by some exact 2-form @xmath371 and this of course can not change the winding numbers @xmath302 $ ] ( [ e5:4 ] ) .
the reason is that the integral of an exact 2-form over some cycle @xmath301 always must vanish @xmath372 since this projective connection @xmath368 ( [ e5:26 ] ) plays an important part for the subsequent method of bundle trivialization , it is convenient to display some of its geometric properties . first the projector @xmath373 is defined by the usual relations [ e5:28 ] @xmath374 and furthermore it is related to the square of the rotation generator @xmath344 ( [ e5:17 ] ) through @xmath375 it can also be expressed in terms of the 3-vector @xmath311 as @xmath376 and finally the corresponding projective connection @xmath368 in the density bundle ( @xmath311 ) can then be written down in terms of the projector @xmath377 as @xmath378=[{{\mathcal{p}}},[{{\mathcal{p}}},{{\mathcal{b}}}_{\mu}]]\ ] ] from which it is immediately seen that @xmath368 has vanishing component relative to the gauge algebra spanned by the generator @xmath344 : @xmath379 the projector @xmath377 is also helpful for considering the nature of the restricted gauge group generated by the elements @xmath355 ( [ e5:20 ] ) which read in terms of @xmath377 and @xmath344 @xmath380 this result says that the restricted gauge group for the positive mixtures ( @xmath202 ) is the ordinary rotation group @xmath381 and for the negative mixtures ( @xmath204 ) it is the lorentz group @xmath382 of ( @xmath383 ) dimensional minkowskian space . but apart from this difference with respect to the gauge group , the projective bundle geometry is the same for both kinds of mixtures . though the preceding projection mechanism for constructing the density bundle ( @xmath311 ) may tempt one to think that both kinds of mixtures share the same geometric structures , the subsequent discussion of the bundle curvature will clearly reveal some essential differences
however in the first instance , the bundle curvature ( @xmath384 , say ) looks still identical for both situations .
the general definition of the curvature operator is @xmath385\ ] ] and this operator measures the non - commutativity of the gauge covariant derivatives , e.g. in form of the bundle identity for the generator @xmath344 ( [ e5:17 ] ) @xmath386{{\mathcal{z}}}\equiv [ { { \mathcal{c}}}_{\mu \nu},{{\mathcal{z } } } ] \ ; .\ ] ] now according to the well - kown ambrose - singer theorem @xcite , the curvature operator @xmath384 spans the algebra of the holonomy group which in general is a subgroup of the gauge group . since , for the present situation , the latter group does not have non - trivial subgroups
, the curvature operator must take its values in the gauge algebra itself , i.e. in the rotation algebra @xmath387 for the positive mixtures ( @xmath202 ) and in the lorentz algebra @xmath388 for the negative mixtures ( @xmath204 ) , see equation ( [ e5:33 ] ) .
this however implies that the curvature is proportional to that generator ( i.e. @xmath389 ) ; and indeed if one substitutes the present bundle connection @xmath325 ( [ e5:15 ] ) into the general curvature definition ( [ e5:34 ] ) , a straightforward calculation yields with the help of the curl relations ( [ e4:39 ] ) and ( [ e4:52 ] ) the following result involving the euler class @xmath390 ( [ e5:2 ] ) @xmath391 clearly this result makes the bundle identity ( [ e5:35 ] ) consistent with the covariant constancy of the generator @xmath344 ( [ e5:23 ] ) .
furthermore , the general structure of curvature @xmath384 ( [ e5:34 ] ) implies the well - known bianchi identity @xmath392 which is immediately transcribed to the curvature coefficient @xmath390 as shown by the former equation ( [ e5:3 ] ) . in this way one has revealed the true origin why that 2-form @xmath390 must be closed
this closedness property must appear as a mere incident when one can resort exclusively to the dynamical equations .
the curvature result ( [ e5:36 ] ) says that the curvature 2-form @xmath384 takes its values in the 1-dimensional algebra \{@xmath344 } of ( pseudo-)rotations and thus sweeps out a more restricted range than its connection @xmath325 .
the latter takes its values in the lorentz - algebra @xmath326 and therefore one can on principle apply @xmath350-valued gauge transformations .
the subgroup of gauge transformations @xmath355 ( [ e5:20 ] ) , applied so far , thus turns out as the _ holonomy group _ of the density bundle ( @xmath311 ) .
so far , there evidently appears no difference in the geometric structure of both types of mixtures , but actually the curvature coefficient @xmath390 carries different topological characteristics for either case .
in order to see this more clearly , one identifies that curvature coefficient @xmath390 as the pullback ( with respect to the density map @xmath311 ( [ e5:6 ] ) ) of the volume cell on the hypersurface @xmath30 in density configuration space ( fig.1 ) .
the desired identification is obtained by embedding the non - trivial density bundle ( @xmath311 ) into some higher - dimensional _ trivial _ bundle such that ( @xmath311 ) reappears as a reduced bundle @xcite . according to this line of arguments
, one adds to the non - trivial connection @xmath325 in the density bundle ( @xmath311 ) some element @xmath393 of the embedding lie algebra such that the resulting connection ( @xmath394 , say ) is trivial .
more concretely , we put @xmath395 where @xmath393 must transform _ homogeneously _ @xmath396 in order that the new connection @xmath394 can transform _ inhomogeneously _ as is generally required for bundle connections ( see ( [ e5:19 ] ) ) @xmath397 the point with this change of connection is here that the new curvature ( @xmath398 , say ) @xmath399\ ] ] is required to vanish ( @xmath400 ) , i.e. @xmath401 \ ; .\ ] ] this allows us to recast the original bundle curvature @xmath384 into the following form @xmath402)\ ] ] ( @xmath403 $ ] ) .
this structure is then inherited also by the curvature coefficient @xmath390 ( [ e5:36 ] ) .
indeed , the general structure of @xmath390 follows directly from the latter form ( [ e5:43 ] ) of the curvature @xmath384 ; this is immediately seen by inverting the former curvature relation ( [ e5:36 ] ) as @xmath404 and then substituting herein the new form of @xmath384 ( [ e5:43 ] ) .
this procedure finally yields for the curvature coefficient @xmath390 @xmath405+\frac{1}{2 } \sigma_{\ast } tr({{\mathcal{z}}}\cdot [ { { \mathcal{e}}}_{\mu},{{\mathcal{e}}}_{\nu}])\ ] ] with the 1-form @xmath406 being given by ( [ e5:25 ] ) and expressed here in terms of @xmath393 @xmath407 this latter relation hints at the possibility of projective generation of the non - trivial density bundle ( @xmath311 ) from a trivial embedding bundle , namely through the fact that the component of the trivial connection @xmath394 ( [ e5:38 ] ) with respect to the holonomy algebra \{@xmath344 } is zero : @xmath408 just as was the case with the projective connection @xmath368 , see equation ( [ e5:32 ] ) . evidently the present result ( [ e5:45 ] ) is nothing else than the former claim for @xmath390 ( [ e5:7 ] ) provided we can prove the coincidence of both second terms on the right - hand sides of equations ( [ e5:7 ] ) and ( [ e5:45 ] ) , i.e. it remains to prove the following identity @xmath409 ) \equiv
\theta_j \epsilon^j{}_{kl}(\partial_{\mu}\theta^k)(\partial_{\nu } \theta^l)\ ; .\ ] ] now it should be clear that this final step may well bring into play some differences between the positive and negative mixtures .
indeed , we will readily see that the expected difference consists in the specific way of embedding the density bundle ( @xmath311 ) into some _ trivial _ bundle of higher dimension .
thus the goal is now to find the bundle of smallest fibre dimension which can carry a _
trivial _ but non - vanishing connection @xmath394 , whose curvature @xmath398 ( [ e5:42 ] ) vanishes together with the projection of its connection @xmath394 to the holonomy subalgebra , cf .
( [ e5:47 ] ) . in order to get some ansatz for the desired @xmath394 ,
one recalls that any trivial connection can be written as a `` pure gauge '' , i.e. the pullback of the maurer - cartan form over the gauge group : @xmath410 here @xmath411 represents an element of that lie group which is generated by the lie - algebra valued range of @xmath394 .
of course the most simple guess for @xmath411 would be just the element @xmath355 ( [ e5:20 ] ) of the holonomy group itself ( @xmath412 ) ; in this case one would obtain for @xmath394 ( [ e5:49 ] ) the following result for _ positive _ mixtures ( @xmath202 ) : @xmath413 thus , in order to obtain the desired bundle reduction requirement ( [ e5:38 ] ) , one fixes the group parameter @xmath358 as a constant over space - time ( @xmath414 ) such that @xmath415 and in this way one arrives at the desired result ( for @xmath202 ) @xmath416 the @xmath326-valued 1-form @xmath417 , emerging herein , is the desired complement @xmath393 ( [ e5:38 ] ) of the density connection @xmath325 , i.e. more concretely @xmath418= \epsilon^l{}_{jk}b^j{}_{\mu}\theta^k { { \mathcal{b}}}_l \;,\ ] ] and thus it obeys the orthogonality condition in a trivial way @xmath419 indeed , this is necessary in order that our present result @xmath394 ( [ e5:52 ] ) has vanishing component in the holonomy algebra , see the former constraints ( [ e5:47 ] ) and ( [ e5:32 ] ) . in the last step ,
it remains to prove the surface relation ( [ e5:48 ] ) for the two - parted hyperboloid @xmath29 ; this however is easily achieved by simply comparing the general trivialization relation ( [ e5:38 ] ) for the change of connections with the present result ( [ e5:52 ] ) , being due to our guess ( [ e5:49 ] ) - ( [ e5:50 ] ) for the positive mixtures .
this comparison immediately yields namely for the algebraic complement ( @xmath393 ) of @xmath368 in @xmath326 @xmath420 and finally it is a nice exercise to substitute this back into the left - hand side of the surface relation ( [ e5:48 ] ) in order to be convinced of the validity of this formula .
simultaneously , the present method of bundle trivialization demonstrates that the euler class @xmath390 is an element of the _ integral _ cohomology group @xmath421 , since the winding numbers @xmath302 $ ] ( [ e5:4 ] ) must necessarily appear as integers .
the reason for this is that the winding numbers may be normalized by means of the surface content of the ( compact ) range of the density map ( @xmath311 ) .
( if the image space of @xmath311 is non - compact , as for the present situation of fig.1 , nomalization of the winding numbers is not possible and they become all zero ) .
thus the validity of the topological claim ( [ e5:7 ] ) has been proven for the positive mixtures ( @xmath202 ) but not for the negative ones ( @xmath204 ) .
but why does the present mechanism of bundle trivialization not work for the negative mixtures in an analogous way ?
the reason is that the negative - mixture bundle ( @xmath311 ) can not be embedded into an @xmath350 bundle with trivial connection @xmath394 ( [ e5:38 ] ) .
this is most easily seen by trying for the principal - bundle section @xmath422 ( [ e5:49 ] ) the same @xmath350-valued ansatz @xmath351 ( [ e5:20 ] ) but with the density vector @xmath423 being now `` space - like '' , i.e. one takes @xmath204 for the normalization condition ( [ e5:9 ] ) .
this would namely yield for the corresponding pure gauge @xmath394 ( [ e5:49 ] ) @xmath424 from here it is immediately seen that one has to take again the group parameter @xmath358 as a constant over space - time in order that the projection of @xmath394 onto the holonomy algebra \{@xmath344 } be zero , cf .
( [ e5:47 ] ) .
however one evidently can have no non - trivial value for this constant @xmath358 so that the decomposition ( [ e5:38 ] ) does apply with both the connection forms @xmath394 and @xmath325 obeying the _ inhomogeneous _ law ( [ e5:40 ] ) and @xmath393 obeying the _ homogeneous _ law ( [ e5:39 ] ) !
this is the true reason why the negative - mixture bundles can not be embedded into a trivial @xmath350 bundle , as is the case with the positive - mixture configurations , and therefore we have to look for an even higher - dimensional trivial bundle for the purpose of embedding .
the solution to this problem is obtained by considering a trivial lorentz bundle over space - time which has the proper lorentz group @xmath31 as its structure group . the corresponding lorentz algebra
is easily obtained by simply complementing its @xmath326 subalgebra , being spanned by the generators @xmath330 ( [ e5:10 ] ) - ( [ e5:12 ] ) , by three additional generators @xmath425 which obey the following algebra [ e5:57 ] @xmath426&=-\epsilon^l{}_{jk}{{\mathcal{b}}}_l \\ [ { { \mathcal{t}}}_j,{{\mathcal{b}}}_k]&=\epsilon^l{}_{jk}{{\mathcal{t}}}_l \ ; .\end{aligned}\ ] ] the lorentz algebra @xmath427 is usually not referred to the present @xmath326 basis \{@xmath330 } and its complement \{@xmath425 } but rather to the ordinary @xmath428 rotation generators @xmath429 ( @xmath320 ) @xmath430=l_3 \ ; \mbox { , ( cycl.)}\ ] ] and the `` boost operators '' @xmath431 : [ e5:59 ] @xmath432&=-l_3 \ ; \mbox { , ( cycl.)}\\ [ l_1,l_2]&=-[l_1,l_2]=l_3 \ ; \mbox { , ( cycl.)}\end{aligned}\ ] ] ( the remaining commutators of any subset are obtained by cyclic permutation ( cycl . ) of the generators ) .
however it is easy to see the correspondence between both basis sets , namely : @xmath433 this embedding of the 3-dimensional @xmath326 algebra into the 6-dimensional lorentz algebra @xmath427 leaves unchanged the bundle reduction formulae ( [ e5:38])-([e5:48 ] ) with the curvature being still given by ( [ e5:36 ] ) and the projector relation ( [ e5:29 ] ) ; and thus it is merely the definition of hermiticity ( [ e5:10 ] ) which must be extended to the _ four - dimensional _ standard representation of the generators @xmath434 and @xmath435 ( @xmath436 ) , i.e. @xmath437 ( @xmath438 ) ) . according to this bundle embedding we try now again to find a trivial connection form @xmath394 which takes its values in the embedding lorentz algebra @xmath427 and whose projection to the @xmath326 subalgebra just yields the @xmath326-valued connection @xmath325 ( [ e5:15 ] ) for the _ negative _ mixture configurations ( @xmath204 ) . in other words , the desired connection @xmath394 is taken again as the maurer - cartan form ( [ e5:49 ] ) but with the _ global _ section @xmath422 now taking its values in the poper lorentz group @xmath31 : @xmath439.\ ] ] the associated maurer - cartan form @xmath394 ( [ e5:49 ] ) is then easily computed as @xmath440 and thus is obviously the @xmath31 analogue of the former @xmath350 result ( [ e5:50 ] ) for the positive mixtures . the projective connection @xmath368 is defined here again by the same prescription as for the lower - dimensional case ( [ e5:26 ] ) whereas the complement @xmath441 , as the analogue of @xmath417 ( [ e5:53 ] ) for the positive mixtures , is composed now of the complementary elements @xmath425 alone : @xmath442 clearly , in order to get the connection @xmath325 for the negative - mixure configurations as the @xmath326 projection of the trivial @xmath394 ( [ e5:63 ] ) , one resorts again to the former arrangement ( [ e5:51 ] ) for the group parameter @xmath358 , i.e. we have then the desired result for the negative mixtures @xmath443 and thus we obtain for the algebraic complement @xmath393 ( [ e5:38 ] ) @xmath444 if this is substituted back into the general euler class @xmath445 ( [ e5:45 ] ) one arrives again at the original claim ( [ e5:7 ] ) whose validity for both kinds of mixtures is thus ensured , together with the topological conclusions being implied by that claim .
it is thus true that both kinds of mixtures share the triviality of their euler class @xmath445 , yet there are nevertheless some geometric differences being worth to be considered in some more detail now .
it is just the emergence of trivial principal bundles in connection with the 2-particle configurations which admits a concrete geometric picture of these configurations .
this appears in form of a local fibre distribution in the tangent bundle .
let us consider the negative mixtures first , for which a _
section @xmath422 ( [ e5:62 ] ) exists for the embedding lorentz bundle .
however since the latter bundle has the same fibre space @xmath31 as the principal bundle associated with the tangent bundle of the underlying space - time , one may look upon @xmath422 also as a global section of the ( tangent ) principal bundle , provided the existence of such global sections is admitted by the topology of space - time . in order
that this condition be satisfied , space - time must namely be a _
parallelisable _ manifold ( such as , e.g. the closed robertson - walker universe @xcite which is topologically the direct product of the 3-sphere @xmath446 and real line @xmath447 ) . by virtue of their very definition ,
parallelisable manifolds admit the existence of a _ global _
frame @xcite , i.e. for the present case of four - dimensional space - time a tetrad of linearly independent vector fields @xmath448 .
these can be chosen to be orthonormal with respect to the presumed space - time metric @xmath449 @xmath450 the tetrad @xmath451 may further be taken to be a solution of the transport equations @xmath452 with the connection coefficients @xmath453 being due to the trivial connection @xmath454 ( [ e5:63 ] ) , which in general will be a non - metric connection .
the vanishing of the curvature @xmath455 ( [ e5:41 ] ) of @xmath394 then guarantees that any global tetrad solution @xmath451 of ( [ e6:2 ] ) is unique over the whole space - time and thus it establishes a global isomorphism of all the tangent fibres .
the restriction of the trivial lorentz connection @xmath394 to the non - trivial @xmath326 connection @xmath325 ( [ e5:65 ] ) geometrically means to omit one of the ( space - like ) tetrad vectors and to consider the remaining triad @xmath456 @xmath457 this triad defines a local ( 1 + 3 ) splitting of the tangent fibres .
since however the curvature @xmath458 ( [ e5:36 ] ) of @xmath325 is proportional to the @xmath388 generator @xmath344 , it annihilates the same vectors as does the projector @xmath377 ( [ e5:29 ] ) , which is nothing else than the square of the generator @xmath344 . but this two - dimensional projector @xmath377 itself annihilates a 2-dimensional subspace of any tangent fibre and thus induces a local ( 2 + 2)-splitting . in the corresponding 2-distribution , as a sub - distribution of the former ( @xmath459 ) splitting ( [ e6:3 ] ) , the projector @xmath377 acts as the identity and if this 2-distribution is integrable , there arises a foliation of space - time into a system of 2-surfaces . the euler class @xmath460 ( [ e5:1 ] ) appears then as the intrinsic curvature of these 2-surfaces . in this sense ,
the negative two - particle mixtures are geometrically represented by a system of 2-surfaces which are equipped with a lorentzian tangent metric .
the existence of such a metric is guaranteed by the vanishing of the euler number @xmath461 ( [ e5:4 ] ) @xcite ; and its invariance group consists of the @xmath382 tranformations being specified by equation ( [ e5:33 ] ) .
on the other hand , the positive mixtures are geometrically characterized by a trivial @xmath326 connection @xmath394 ( [ e5:50 ] ) and thus start with a global triad of tangent vectors @xmath462 from the very beginning @xmath463 however such a triad can always be complemented to a tetrad ( by selection of some linearly independent fourth vector ) and then the same arguments do apply as for the negative mixtures .
the geometric difference between both mixtures is that the representative 2-surfaces for the positive mixtures are space - like ( such that the gauge element @xmath355 ( [ e5:33 ] ) acts as an ordinary rotation of the space - like tangent plane of the 2-surfaces ) whereas the negative mixtures have a ( 1 + 1)-dimensional minkowski plane as their tangent space . | within the framework of relativistic schrdinger theory ( an alternative form of quantum mechanics for relativistic many - particle systems ) it is shown that a general n - particle system must occur in one of two forms : either as a `` positive '' or as a `` negative '' mixture , in analogy to the fermion - boson dichotomy of matter in the conventional theory .
the pure states represent a limiting case between the two types of mixtures which themselves are considered as the rst counterparts of the entangled ( fermionic or bosonic ) states of the conventional quantum theory . both kinds of mixtures are kept separated from dynamical as well as from topological reasons .
the 2-particle configurations ( @xmath0 ) are studied in great detail with respect to their geometric and topological properties which are described in terms of the euler class of an appropriate bundle connection .
if the underlying space - time manifold ( as the base space of the fibre bundles applied ) is parallelisable , the 2-particle configurations can be thought to be generated geometrically by an appropriate ( @xmath1 ) splitting of the local tangent space . |
it is one of the most important subjects to understand mechanism of color confinement in qcd .
the abelian projection of qcd is to extract an abelian theory with charges and monopoles performing a partial gauge - fixing .
thooft conjectured an interesting idea @xcite that condensation of the abelian monopoles produces color confinement due to the dual meissner effect .
there are infinite ways of extracting an abelian theory out of qcd .
maximally abelian(ma ) gauge is exciting . in this gauge , many interesting facts have been found .
there are phenomena called abelian dominance @xcite .
the string tension from abelian wilson loops is produced only by monopoles @xcite .
the effective monopole action is derived from vacuum configulations @xcite . in the infinite volume limit
, entropy dominance over energy induces the condensation of the monopoles , which leads us to the confinement of the abelian charges , i.e. , the color confinement .
therefore the ( extended ) monopoles control the confinement mechanism in ma gauge .
the following subjects are to be clarified : \1 ) relation of the monopole dynamics to the deconfinement transition and \2 ) gauge dependence of the monopole dynamics . from this point of view , we investigate the polyakov loop which is an important order parameter of the deconfinement transition . in this note , we show the abelian polyakov loop is written by a product of monopole dirac string part and photon part and vanishing of the abelian polyakov loop comes only from the monopole dirac - string contributions not only in ma gauge but also in unitary gauges .
we adopt the usual @xmath1 and the @xmath2 wilson actions . to study gauge dependence , we consider three types of abelian projection .
the matrices to be diagonalized are @xmath3 $ ] in ma gauge , polyakov loop in polyakov gauge and plaquette in @xmath4 gauge .
after any gauge fixing is over , we can extract an abelian link variable @xmath5 and an angle variable @xmath6 from it@xcite . now let us show that an abelian polyakov loop operator of an external current @xmath7 @xmath8 , \label{apol}\end{aligned}\ ] ] is given by a product of monopole and photon contributions . using the definition of a plaquette variable @xmath9
, we get @xmath10 , \label{t4}\ ] ] where @xmath11 is the lattice coulomb propagator . since @xmath12 , the second term in the right - hand side of ( [ t4 ] ) does not contribute to the abelian polyakov loop ( [ apol ] ) .
epsf = 0.5 the plaquette variable can be decomposed into three terms : @xmath13 which @xmath14 , @xmath15 and @xmath16 are an integer - valued dirac strings @xcite , a dual potential and a photon terms , respectively .
the dual potential term has no contribution because of the antisymmetric tensor .
therefore we have @xmath17 $ ] where @xmath18 we observe separately the real parts @xmath19 and @xmath20 of the photon @xmath21 and the dirac - string @xmath22 , respectively .
we first check that the separation works well in villain form of the compact @xmath23@xciteon a @xmath24 lattice .
since there are natural monopoles and degrand - toussaint(dgt ) monopoles in the villain case of qed , we observe @xmath25 in terms of the two types of the monopoles .
the results are shown in fig .
[ villain ] .
1)the monopole dirac - string data vanish in the confinement phase , whereas the photon data remain finite and change gradually for all @xmath26 .
the characteristic features of the polyakov loops are then due to the behaviors of the dirac - string contributions alone .
2)monopole polyakov loops show more enhancement than the total ones for @xmath27 .
3)both types of monopoles give almost the same results .
epsf = 0.5 the monte - carlo simulations were done in @xmath1 on @xmath28 lattice ( @xmath29 ) in the ma and the unitary gauges . in @xmath2
qcd , we adopted @xmath30 lattice ( @xmath31 ) .
all measurements were done every 50 sweeps ( 40 sweeps in the @xmath2 case ) after a thermalization of 2000 sweeps .
we took 50 configurations totally for measurements .
1)we plot the @xmath1 data in the ma gauge in fig .
the abelian polyakov loops remain zero in the confinement phase , whereas they begin to rise for @xmath32@xcite .
the monopole dirac string contribution is zero for @xmath33 , whereas it begins to rise rapidly and it reaches @xmath34 for large @xmath26 . on the other hand ,
the photon part has a finite contribution for both phases and it changes only slightly .
characteristic behaviors of the abelian polyakov loops are then explained by the dirac - string part of monopoles alone .
2)the same results are obtained also in pure @xmath2 qcd in the ma gauge as shown in fig .
there is a clear hysteresis behavior showing the first order transition .
epsf = 0.5 epsf = 0.5 the data in the polyakov gauge are plotted in fig .
it is very interesting to see that the abelian and the dirac - string polyakov loops are zero in the confinement phase , which suggests occurrence of flux squeezing in the unitary gauge , too .
they show finite contribution above the critical temperature @xmath35 .
photon contributions are finite and change gradually in both phases . in the case of @xmath4 gauge ,
the results are very similar to these in polyakov gauge .
our analyses done here show that abelian monopoles are responsible for confinement in @xmath1 and @xmath2 qcd , and condensation of the monopoles is the confinement mechanism .
_ these are the first phenomena suggesting gauge independent realization of the thooft conjecture . _ 9 g.thooft , nucl.phys.*b190*,(1981)455 .
t.suzuki , nucl.phys.b(proc . suppl . ) * 30 * , ( 1993 ) 176 and see also references therein .
h.shiba and t.suzuki ,
phys.lett.*b333 * , + ( 1994)461 . s.ejiri,s.kitahara,y.matsubara and t.suzuki , kanazawa university report kanazawa 94 - 14 , 1994 .
j.d.stack , s.d.neiman and r.j.wensley , phys .
rev.*d50*,(1994)3399 . h.shiba and
t.suzuki , nucl.phys.b(proc .
j.villain , j.phys.(paris)*36*,(1975)581 .
t.a.degrand and d.toussaint , phys .
j.engels , j.jersak , k.kanaya , e.laermann , c.b.lang , t.neuhaus and h.satz , nucl.phys.*b280*,(1987 ) 577 . | monte - carlo simulations of abelian projection in @xmath0 pure lattice qcd show that 1 ) polyakov loops written in terms of abelian link fields alone play a role of an order parameter of deconfinement transition , 2 ) the abelian polyakov loops are decomposed into contributions from dirac strings of monopoles and from photons , 3 ) vanishing of the abelian polyakov loops in the confinement phase is due to the dirac strings alone and the photons give a finite contribution in both phases .
moreover , these results appear to hold good in unitary gauges .
this suggests that monopole condensation as the color confinement mechanism is gauge independent . |
the pairing properties of nuclear systems - finite nuclei and the bulk nuclear matter - play an important role in the physical manifestations of these systems . in most cases of interest
the pairing occurs at finite isospin asymmetry .
for example , charge neutral bulk nuclear matter with an admixture of leptons in @xmath0 equilibrium implies proton abundancies at about @xmath1 in supernova matter and @xmath2 in neutron star matter .
the dominant partial - wave channel for the pairing depends on the density , temperature , and the isospin asymmetry in general . for large enough asymmetries
the matter is paired in the isospin triplet channels - the @xmath3 channel at low densities @xcite and @xmath4-@xmath5 or @xmath6 channels at high densities@xcite . for weakly asymmetric systems
the isospin singlet attractive @xmath7-@xmath8 and @xmath9 channels@xcite dominate the pairing interaction at low and high densities , respectively .
in bulk nuclear matter of compact stars the asymmetries are most likely too large to support isospin singlet pairing .
however the dilute nuclear matter in supernovas and the low - density tails of the exotic nuclei can support @xmath7-@xmath8-channel pairing , since in general the isospin asymmetry is less effective in destroying the pair correlations in dilute matter@xcite .
the mechanism of the suppression of the bcs pairing of the kind found in asymmetric nuclear matter was initially studied in metallic superconductors in a spin - polarizing magnetic field . since a homogenous magnetic field
is screened beyond the london penetration depth in the bulk of a superconductor , in real materials the electron spin polarization is induced by low - density paramagnetic impurities , which interact with an electron by spin nonconserving forces .
the effect of collision induced polarization is commonly modeled by an average polarizing field which implies an asymmetry in the populations of spin - up and -down quasiparticles .
the resulting electron pauli paramagnetism suppresses the @xmath10-wave pairing when the fermi surfaces of the electrons with spin - up and -down are apart on a scale of the order of the energy gap in the quasiparticle spectrum
. a close analogy to the ordinary superconductors can be seen in the effect of the large @xmath11 fields in the neutron stars on the @xmath10-wave pairing of neutrons .
our general discussion below also applies to the pairing in spin - polarized neutron matter in the @xmath3 channel , although the numerical computations are carried out for the isoscalar pairing in asymmetric matter .
the understanding of the isospin singlet pairing as a function of the isospin asymmetry requires an extension of the standard bcs theory to describe phases with spontaneously broken space symmetries .
larkin and ovchinnikov@xcite and fulde - ferrell@xcite ( loff ) argued , in the context of metallic superconductors , that a transition from the bcs paired to the normal state occurs via a phase where the cooper pairs carry a finite total momentum .
the isospin polarization in nuclear matter with pairing in the @xmath7-@xmath8 channel leads to a nuclear loff phase for a range of isospin asymmetries@xcite . unlike the metallic superconductors where the ratio of the pairing gap to the average chemical potential is of the order 10@xmath12 , the nuclear loff phase emerges in the strong coupling regime , where this ratio is of the order 0.1 .
the spatially periodic structure of the nuclear loff phase corresponding to oscillations of the condensate wave function are on the scales of the order of several fm .
note that , independent of the detailed structure of the lattice , the loff phase breaks _ both _ the translational and rotational invariance of the original bcs state .
the formation of the loff phase is a robust feature of the fermionic systems with broken time - reversal symmetry and has been studied in detail in the flavor asymmetric high - density qcd ( crystalline color superconductivity ) @xcite .
the evolution of the condensate from the bcs to paired states with broken space symmetries , as the isospin asymmetry is increased , can take a different path if we give up the assumption that the intrinsically homogeneous matter features spherical fermi surfaces@xcite . as shown in ref .
@xcite , for quadrupole deformed fermi surfaces ( dfs ) the bcs condensate evolves to a state with spontaneously broken rotational symmetry [ in group theory terms the @xmath13 symmetry breaks down to @xmath14 .
such a state is still axially symmetric and , unlike the loff phase , preserves the translational symmetry .
we shall refer to the superconducting state with deformed fermi surfaces as the dfs phase with the understanding that the expansion describing the deformation of fermi surfaces is truncated at quadrupole order ( the higher order terms in the multiple expansion have not been studied so far ) .
why do the phases with broken space symmetries dominate the bcs state ?
the bcs quasiparticle spectrum ( for homogeneous systems ) is isotropic ; when the polarizing field drives apart the fermi surfaces of fermions with spin / isospin up and down the phase space overlap is lost , hence the pairing is suppressed .
the finite momentum of the cooper pairs of the loff phase implies an anisotropic quasiparticle spectrum ; the magnitude of the anisotropy is controlled by the value of the net momentum of the pairs which is treated as a parameter .
this new degree of freedom , which can be viewed as an additional variational parameter for the minimization of the ground state of the system , ( partially ) overcomes the phase - space loss caused by the departure of the fermi surfaces from the perfectly overlapping configuration . in effect
the gain in the pairing energy through increase of the pairing field dominates the increase of the kinetic energy . to motivate the dfs phase , note that formally the loff spectrum can be viewed as a dipole ( @xmath15 ) perturbation of the spherically symmetrical bcs spectrum , where @xmath16 are the legendre polynomials , and @xmath17 is the cosine of the angle between the particle momentum and the total momentum of the cooper pair .
the dfs phase assumes an expansion of the quasiparticle spectrum up to the next - to - leading order , quadrupole ( @xmath18 ) deformation ; it also assumed that the total momentum of the cooper pairs is zero , i.e. , there are quadrupole deformations only . although the perturbations from the spherical symmetry cost kinetic energy , they are compensated by the gain in the potential energy . for a range of asymmetries the dfs phase becomes the ground state of the paired system@xcite .
the purpose of the present paper is a combined treatment of the loff and the dfs phases in bulk isospin asymmetric nuclear matter within the same model as that in refs .
the model assumes a pairing force given by a phase - shift equivalent ( i.e. , realistic ) interaction and nonrenormalized quasiparticle spectrum .
the effect of the mean field in the brueckner - hartree - fock approximation is to reduce the range of isospin asymmetries where the pairing exists@xcite .
we shall follow the strategy of disentangling the effects of the mean field from the pairing since already at the level of approximations of refs .
@xcite the loff and dfs phases show a complicated behavior .
the renormalization of the quasiparticle spectrum for the pairing interaction is a problem of its own which has not been solved to date .
issues such as the consistency between the vertex and propagator renormalization and fulfillment of the spectral sum rules , etc . need further attention . in section 2
we derive the bcs equations , which include the effects of the finite momentum of the cooper pairs and fermi surface deformations , within the finite - temperature real - time green s functions formalism .
the numerical solutions of these equations are presented in section 3 , where we discuss the phase diagram of the combined loff and dfs phases and their properties at finite temperatures .
section 4 contains a summary of the results and an outlook .
the starting point of the model is the hamiltonian of fermions interacting via two - body forces : [ h ] h&=&_d^3r ^(r , t , ) ( r , t , ) + & + & _
d^3r d^3r ^(r , t , ) ^(r,t , ) v(r , t,;r,t , ) ( r , t , ) ( r , t , ) , where @xmath19 are second quantized operators in the heisenberg representation , @xmath20 is the space and time local bare interaction , and @xmath21 , @xmath0 stand for discrete quantum numbers ( spin , isospin , etc . ) .
we choose to formulate the finite - temperature pairing theory in terms of nonequilibrium real - time green s functions .
the single - particle propagator in this formalism is combined in a @xmath22 matrix , [ 2 ] ( x_1,x_2 ) =( cc g^c(x_1,x_2 ) & g^<(x_1,x_2 ) + g^>(x_1,x_2 ) & ^a(x_1,x_2 ) ) =( cc -it(x_1 ) ^(x_2 ) & i ^(x_2 ) ( x_1 ) + -i ( x_1)^(x_2 ) & -it ( x_1 ) ^(x_2 ) ) , where @xmath23 , and the averaging is over an arbitrary nonequilibrium state of the system ; @xmath24 and @xmath25 are chronological and antichronological time ordering operators for the nambu spinors [ nambu ] ^ = = .
the nambu spinors span the particle - hole space and satisfy the fermionic equal - time anticommutation relations \{_(x , t),_^(x,t ) }
= _ ^3(x - x),\{_(x , t),_(x,t)}=0 , \{_(x , t)^,_^(x,t ) } = 0 .
starting from the equations of motion for the heisenberg operators and the hamiltonian ( [ h ] ) the martin - schwinger hierarchy can be constructed for a matrix green s function involving an arbitrary number of field operators . for our purposes in
is sufficient to truncate the hierarchy at the @xmath26 level .
the equation of motion of a one - particle greens function is then given by the time - dependent dyson equation [ dyson01 ] g_(0)^-1(x_1)g(x_1,x_2 ) = _ z(x_1-x_2 ) - v(x_1,x_6;x_5,x_4)(x_5,x_4,x_2,x_6^+ ) , where the subscript @xmath27 refers to the free propagator , @xmath28 stands for summation / integration over the repeated discrete / continuous variables , @xmath29 are the components of the vector of the pauli matrices , and the superscript @xmath30 means an infinitesimal increment in the time argument .
the hierarchy at a given level can be decoupled ( formally ) by introducing the self - energy matrix : [ dyson2 ] g_(0)^-1(x_1)g(x_1,x_2 ) = _ z(x_1-x_2 ) + ( x_1,x_3 ) ( x_3,x_2 ) , which is defined as [ sigma1 ] ( x_1,x_2 ) = - v(x_1,x_6;x_5,x_4)(x_5,x_4,x_7,x_6^+ ) g^-1(x_7,x_2 ) .
the first of these equations defines the quasiparticle spectrum of the superconducting state , the second , the self - energies , e.g. , the gap function . to close the set of exact equations we need to specify the approximation to the two - particle green s function .
the bcs theory follows in the hartree approximation , where the two - particle green s function is approximated by a product of single - particle green s functions .
note that the anomalous contributions of the type @xmath31 and @xmath32 are automatically built in in the hartree approximation .
this can be verified from the explicit form of the propagator in eq .
( [ 2 ] ) , when the spinors ( [ nambu ] ) are substituted .
each element of the @xmath22 matrix ( [ 2 ] ) is in turn a @xmath22 matrix in the particle - hole space : [ 1 ] g_12^j =( cc g^j(x_1,x_2 ) & f^j(x_1,x_2 ) + f^ j(x_1,x_2 ) & ^j ( x_1,x_2 ) ) , where @xmath33 .
the matrix form of the self - energies in the particle - hole space is analogous to that of the propagators : ( x_1,x_2 ) = ( cc ^c(x_1,x_2 ) & ^<(x_1,x_2 ) + ^>(x_1,x_2 ) & ^a(x_1,x_2 ) ) , ^j ( x_1,x_2 ) ( cc ^j ( x_1,x_2 ) & ^j(x_1,x_2 ) + ^j ( x_1,x_2 ) & ^j(x_1,x_2 ) ) , where the off - diagonal elements of the @xmath34 matrix correspond to the pairing amplitude . note that the inverse free - particle propagator is diagonal in the particle - hole , @xmath35 and spin / isospin spaces : ^-1_(0 ) = ( i_t + ^2/2m_- _ ) _ .
the dyson eq .
( [ dyson2 ] ) is a @xmath36 matrix equation in general if we include the spin and isospin .
the size of the matrix to be diagonalized can be reduced in several successive steps . here
we note that in the time - independent stationary limit the number of the propagators can be reduced by applying the unitary transformation @xmath37 to eqs .
( [ dyson2 ] ) and ( [ sigma1 ] ) , in the @xmath35 space , and introducing the retarded / advanced propagators by the relations [ retarded ] g^r , a(x_1,x_2 ) = g^c(x_1,x_2 ) - g^<,>(x_1,x_2 ) = g^>,<(x_1,x_2 ) - g^a(x_1,x_2 ) .
further reductions will be carried out in the next section .
the dyson equation , e.g. , for the retarded propagator is invariant under the rotation in the @xmath35 space affected by the transformation above , i.e. , [ dyson3 ] g^r(x_1,x_2 ) = g_(0)^r(x_1,x_2)+ g_(0)^r(x_1,x_2)^r(x_1,x_3 ) ^r(x_3,x_2 ) ; the retarded self - energies are related to the matrix components in the @xmath35 space by relations analogous to eq .
( [ retarded ] ) . the integral eq .
( [ dyson3 ] ) for the retarded propagators can be solved without specifying the approximation to ( and hence the form of ) the self - energies , in the quasiclassical approximation . when the characteristic length scales of the spatial variations of the macroscopic condensate are much larger than the inverse of the momenta involved in the problem @xmath38 , where @xmath39 is the fermi momentum , the two - point correlation functions can be approximated by their quasiclassical counterparts .
going over to the center of mass @xmath40 and relative @xmath41 coordinates in the two - point functions and carrying a fourier transform with respect to the relative coordinates we arrive at the mixed representation for these functions , e.g. , [ mix ] g^<(p , x)=e^ip g^<(x+,x-)d^4 , where @xmath42 are the relative frequency and momentum , and @xmath43 .
since the variations of the propagators and self - energies are slow on the scales of the order of @xmath44 , keeping the leading order terms in the gradient expansion is accurate to order @xmath45 ) .
the quasiclassical counterpart of the dyson eq .
( [ dyson3 ] ) is [ qc ] _
( cc - _ ^+ & -_^r ( , ) + - _ ^r ( , ) & + _ ^- ) ( cc g_^r ( , ) & f_^r ( , ) + f_^ , r ( , ) & g_^r ( , ) ) = _ , where [ sp ] _
^= [ ( /2)^2/2m_- _ ] _ & & ^r,(+)_(,/2 ) -^r,(- ) _ ( , /2 ) , and @xmath46 is a unit matrix in the particle - hole space . here
we have defined the symmetric and antisymmetric combinations of the particle / hole retarded self - energies _
^r,(+ ) ( , ) & & , + ^r,(- ) _
( , ) & & .
the dependence on the center - of - mass time is dropped in the above expressions in the stationary limit .
the retarded self - energies now can be expanded in legendre polynomials : [ exp1 ] ^r,(+/- ) _ ( , ) & = & _ l = 0^^r,(+/- ) _ , 2l(,p ) p_l(cos ) , where @xmath47 is the angle between the vectors @xmath48 and @xmath49 . in practice
the expansion will be truncated at order @xmath50 ; the @xmath51 term renormalizes the chemical potential , as discussed below . for
translationally invariant interactions the odd terms do not contribute in the expansion ( [ exp1 ] ) ; the interactions are assumed to be timelocal . for translationally noninvariant interactions the effect of the @xmath52 term would be a renormalization of the dipole deformation in the spectrum due to the kinetic energy terms .
it can be included in the renormalization of the particle mass entering the anisotropic contribution to the kinetic energy , although it would not contribute to the net isotropic increase of the kinetic energy due to the finite momentum .
since the total momentum is treated as a variational parameter in the problem , the effect of the self - energy renormalization at @xmath52 can be accounted for by a simple rescaling of this parameter ; therefore we anticipate that the @xmath52 terms would not change the results below . for spin and isospin conserving forces
the normal green s functions and self - energies are diagonal in the spin and isospin spaces .
it is sufficient to consider the anomalous propagators , e.g. , in the isospin space , since the resulting spin structure , for @xmath10-wave interactions , is uniquely determined for each isospin combination . if we restrict ourselves to the neutron - proton pairing in the @xmath7-@xmath8 channel , which is justified when @xmath53 ( subscripts @xmath54 and @xmath55 refer to protons and neutrons ) then @xmath56 .
the spectrum , in this case , takes the form [ spectrum2 ] _
= e_a , where the symmetric and asymmetric parts of the spectrum ( which are even and odd with respect to the time - reversal symmetry ) are defined as e_s = ( _ + + _ - ) , e_a = ( _ + -_- ) .
the eigenvalues ( [ spectrum2 ] ) are valid for an arbitrary approximation to the self - energies .
further , we shall approximate the symmetric and antisymmetric parts of the spectra using ( i ) the quasiparticle and ( ii ) the momentum - independent self - energy shift approximations .
the first approximation implies [ qpa ] _
n / p^r,(-/+ ) ( , ) _
n / p^r,(-/+)(,_p ) _ n / p^r,(-/+)(,_p)_p_f + _ n / p^r,(-/+)(,_p ) _ p_f ( p - p_f ) ; the second approximation keeps the leading order ( constant ) term in the expansion ( [ qpa ] ) . within these approximations
we obtain [ es ] e_s & = & + --(+ ) ^2 , + [ ea ] e_a & = & - + ( + ) ^2 , where we defined the average chemical potential and the conformal deformation [ mu1 ] & = & - , + [ mu2 ] & = & , as the @xmath51 and @xmath50 contributions from the isospin symmetric part of the quasiparticle spectra ; similarly the relative chemical potential and the relative deformation are defined as the @xmath51 and @xmath50 contributions from the isospin antisymmetric part of the quasiparticle spectra : [ mu3 ] & = & - , + [ mu4 ] & = & , where we used the abbreviation ^-/+_n / p(p_f)=re .
the self - energies on the right - hand side of eqs .
( [ mu1])-([mu4 ] ) are functionals of the pairing field , i.e. , they are nonzero even in the case where all the interactions apart from the pairing interaction are switched off .
they depend on the quasiparticle momentum via the momentum dependence of the gap function ; if the latter is approximated by its value at the fermi momentum the self - energies become constants . the values of the average and relative chemical potentials
@xmath57 and @xmath58 can be adjusted to reproduce the particle density at a given temperature and isospin asymmetry .
the conformal and relative deformations @xmath59 and @xmath60 along with the finite momentum of the pairs @xmath61 are treated as variational parameters to be determined from the ground state of the superconducting system . the solution of the dyson eq .
( [ qc ] ) for the propagators can be written in terms of the eigenvalues as g_n / p^r ( , ) & = & + , + f^r ( , ) & = & u_p v_p ( - ) , where the bogolyubov amplitudes are u_p^2 = + , v_p^2 = - .
the remainder green s functions can be reconstructed using the spectral representation of the retarded green s functions [ spforg ] g_n / p^r ( , ) & = & _ -^ d , + [ spforf ] f^r ( , ) & = & _ -^ d , where @xmath62 and @xmath63 are the spectral functions .
the kadanoff - baym ansatz@xcite provides the link between the the retarded components and the remainder green s functions in the @xmath35 space .
we have [ kbforg ] g_n / p^ > , < ( , ) & = & a_n / p ( , ) f^ > , < ( ) , + [ kbforf ] f^ > , < ( , ) & = & b ( , ) f^ > , < ( ) , where , in equilibrium , the wigner distribution functions of the kadanoff - baym ansatz@xcite reduce to the fermi distribution function @xmath64^{-1}$ ] and @xmath65 $ ] ; here @xmath0 is the inverse temperature .
note that the relations above are exact in the equilibrium limit .
the causal and acausal green s functions are then obtained from the relations ( [ retarded ] ) .
the bcs gap equation follows from the hartree approximation for the two - particle green s function in eq .
( [ dyson01 ] ) : ( x_1,x_2;x_3,x_4 ) = ( x_1,x_2 ) ( x_3,x_4 ) , which implies , according to eq .
( [ sigma1 ] ) , ( x_1,x_2 ) = - ( x_1,x_3;x_2,x_4 ) ( x_4,x_3^+ ) . upon applying the quasiclassical approximation to the above equation in the case of time local interaction
, we obtain the retarded ( off diagonal in the particle - hole space ) self - energy [ gap2 ] ^r ( , ) = v(, ) f^ < ( , , ) , or using the relation @xmath66 , which follows from eqs .
( [ spforf ] ) and ( [ kbforf ] ) , [ gap3 ] ^r ( , ) = - v(, ) imf^r(,,)f^<( ) .
further progress requires partial - wave decomposition of the interaction , which can be performed after angle averaging the remainder functions on the right - hand side of eq .
( [ gap3 ] ) .
the result of this procedure is [ gap4 ] _
l^r(p , p)=-_l v_ll(p , p ) , where @xmath67 denotes the average over the angle between the relative and total momenta and @xmath68 is the angle averaged gap . here the pairing interaction is approximated by the bare neutron - proton interaction @xmath69 in the @xmath7-@xmath8 channel .
as discussed above the average and relative chemical potentials can be fixed by adjusting the pairing field to reproduce the matter density @xmath70 and the isospin asymmetry @xmath71 .
the total momentum , the relative and conformal deformations , are treated as variational parameters to be determined from the ground state energy of the system . the corresponding expression for partial densities of protons and neutrons is provided by the relation _
n / p= g_n / p^ < ( , ) , or , using eqs .
( [ spforg ] ) and ( [ kbforg ] ) , [ density ] _
n / p= -2 _ g_n / p ( , ) f ( ) = _ \{u_p^2
f^<(_)+v_p^2 f^ < ( _ ) } , where @xmath72 stands for quasiparticle spin and the second relation follows in the quasiparticle approximation
. the phase diagram of the paired state at fixed finite temperature and density can be obtained from the free energy in the mean field approximation .
the free energy is given by the thermodynamic relation [ gr ] f _ , = u - ^-1s .
where @xmath73 is the internal energy and @xmath10 is the entropy .
a thermodynamically stable paired state minimizes the difference of the free energies of the superconducting and normal states , @xmath74 .
the mean field entropy of the superfluid state is [ entropy ] s_s = - 2 k_b \{f^<(_+)ln f^<(_+)+ f^>(_+)ln f^>(_+ ) + f^<(_-)ln f^<(_- ) +
f^>(_-)ln f^>(_- ) } , where @xmath75 is the boltzmann constant and the sum is over the momentum states in the quasiparticle approximation .
the corresponding expression for the entropy of the normal state @xmath76 is obtained by taking the limit @xmath77 in eq .
( [ entropy ] ) .
the internal energy of the superconducting state in the mean field approximation is [ u ] u_s= 2\ { + _ ll v_ll(p , p ) _
l(p)_l(p ) } , where the normal and superconducting occupation probabilities are defined as n_n
/ p(p)u_p^2 f^<(_)+v_p^2 f^ < ( _ ) , ( p ) u_pv_p . the first term in eq .
( [ u ] ) includes the kinetic energy of quasiparticles which is a functional of the pairing gap . in the normal state it reduces to the kinetic energy of noninteracting quasiparticles .
the second term describes the bcs mean - field interaction among the particles in the condensate and vanishes in the normal state .
the nuclear loff and dfs phases were studied numerically using the paris nucleon - nucleon interaction .
our results did not depend on this choice of the interaction , since there are no significant deviations in the scattering phase shifts in the @xmath78 channel among various realistic potentials and these reproduce the properties of the deuteron at the same level of accuracy .
the net density of the matter was fixed at the empirical saturation density of the nuclear matter @xmath79 fm@xmath12 . at this density
the typical energies of the nucleons are below the threshold laboratory energy @xmath81 mev at which the interaction in the @xmath78 channel becomes repulsive .
lower densities could provide a more realistic ( from the physical point of view ) setting , however , they have the disadvantage that the effects of the bose - einstein condensation of deuterons may start to play a role@xcite .
the computations were carried out for several fixed values of the density asymmetry @xmath21 that are above the critical values at which the nuclear loff and dfs phases set in ( see refs .
the qualitative behavior of quantities of interest is generic for different values of @xmath21 , therefore , we shall show the results for a fixed value @xmath82 at which the loff and dfs phases ( separately ) dominate the ordinary bcs state .
the results of the combined treatment of the loff and dfs phases will be presented in two steps : in the next section we fix temperature and vary simultaneously the net momentum of the pairs @xmath61 ( loff phase ) and the relative deformation @xmath60 ( dfs phase ) . in the subsequent section we study the temperature dependence of these phases at either constant @xmath61 or @xmath60 .
figure 1 shows the pairing gap as a function of the total momentum of cooper pairs @xmath61 and the relative deformation @xmath60 for vanishing conformal deformation ( @xmath83 ) .
the temperature is fixed at the value @xmath84 mev .
the ratios of the temperature over the fermi energy and the pairing gap in the symmetric state are 0.1 and 0.4 respectively , i.e. , they correspond to the low - temperature regime .
the gap is normalized to its value @xmath85 in the rotationally / translationally invariant state at @xmath82 .
although @xmath86 changes in the interval @xmath87 $ ] in general , the symmetry of the equations with respect to the indices labeling the species reduces the range of @xmath21 to @xmath88 $ ] .
the relative deformation obviously is not bounded and can assume both positive and negative values .
positive values of @xmath60 imply an oblate deformation for the fermi surface of neutrons and a prolate deformation for the fermi surface of protons .
the behavior of the pairing gap can be understood by examining the effect of the symmetric and antisymmetric parts of the quasiparticle spectrum on the gap eq .
( [ gap4 ] ) .
the antisymmetric part @xmath89 appears only in the fermi distribution functions .
if @xmath90 one recovers the bcs limit , where the fermi distribution functions become identical , i.e. , the fermi surfaces perfectly coincide . in this limit
the phase - space overlap between the paired states is maximal .
consider the effect of a finite @xmath89 when @xmath91 .
it increases / decreases the energies of the isospin up / down particles ; the shift in the fermi - levels of the isopin up / down particles samples different momentum regions of the phase - space in the kernel of the gap equation .
this blocking effect or phase - space decoherence in the pair states in turn reduces the magnitude of the gap . in particular
this allows for pairs with small @xmath89 . switching on a finite @xmath61 and/or @xmath92 acts to restore partially the phase space coherence of the isospin up / down states when @xmath93 [ see eq .
( [ ea ] ) ] .
therefore , the magnitude of the gap increases in this case , i.e. the phase - space overlap of the fermi surfaces is ( partially ) restored .
increasing @xmath61 and/or @xmath92 further acts in a manner similar to pure energy shift @xmath58 - the decoherence increases and the pairing is eventually destroyed .
this picture is seen in fig . 1 in the region @xmath94 .
small perturbations in @xmath61 and @xmath60 from the asymmetric bcs state ( @xmath95 but @xmath96 ) increase the pairing gap . for large perturbation in either @xmath61 _ or _
@xmath97 the gap vanishes as the decoherence increases .
the combined effect of the finite momentum and the quadrupole deformations is seen in the minimum plateau in the region of small @xmath61 and @xmath60 , followed by a maximum in the gap with increasing @xmath61 and/or @xmath60 , and a rapid falloff beyond the maximum .
the behavior of the gap in the limits of pure loff and dfs phases is the same as the one found in the previous work @xcite . in the case of @xmath98
the @xmath99 term in eq .
( [ ea ] ) combines with the chemical potential shift @xmath58 . to satisfy a given particle number asymmetry
there are now two parameters ( while there was only one , @xmath58 , for positive @xmath60 ) . as a result with increasing deformation and hence @xmath99 ( the changes in @xmath57 are insignificant ) the density asymmetry
can be maintained at a cost of smaller @xmath58 which tends to zero and eventually changes the sign ( @xmath100 at @xmath101 this marks the `` turnover '' in the gap equation which eventually vanishes for @xmath102 we shall see that the regions of the negative values of the deformation parameter are physically not relevant , since these states have larger kinetic energies . as in the case of positive @xmath60
the effect of finite momentum is to suppress the pairing for a large deformation ; the combined phase increases the pairing only for small perturbations from the asymmetric bcs state in the vicinity of the `` minimum plateau '' ( @xmath103 @xmath104 ) .
the symmetric part of the quasiparticle spectrum ( [ es ] ) controls the location of the maximum ( corresponding to @xmath105 ) and the width of the kernel in eq .
( [ gap4 ] ) in the momentum space through its contribution to the denominator of the gap equation .
a positive contribution to the average chemical potential effectively decreases the density of states ( which can be easily seen analytically in the zero temperature and weak coupling limits ) , and therefore , the gap function .
inversely , a negative contribution to the average chemical increases the density of states .
the changes in the @xmath106 due to the deformation do not affect qualitatively the results above .
in summary of fig .
1 , the combination of the loff and dfs phases promotes the pairing only in the region of small @xmath60 and @xmath61 ( weak perturbation from the bcs state ) . for large quadrupole deformations
the finite momentum is not favorable .
viewed from the loff phase , for large values of @xmath61 the quadrupole deformations are disfavored .
figure 2 displays the difference between the free energies of the superconducting and normal states @xmath107 normalized to its value in the asymmetric bcs state @xmath108 . the thermal contribution due to
the finite - temperature entropy is numerically irrelevant for the net budget of free energies of both states .
whether the superconducting state is thermodynamically favored depends on the relative magnitude of the potential energy of the pair interactions [ second term in eq .
( [ u ] ) ] and the difference in the kinetic energies of the normal and superconducting states .
since the energy of the pair interactions scales as the square of the pairing gap , the shape of the @xmath107 surface closely resembles that of the pairing gap in fig .
1 . there are , however , significant quantitative differences due to the contribution from the kinetic energy of the quasiparticles .
the asymmetric bcs state is the stable ground state of the system ( @xmath109 ) , however its perturbations for finite @xmath60 and @xmath61 are unstable towards evolution to lower energy states . for the pure loff phase ( @xmath110 ) the ground state corresponds to finite momentum @xmath111 ( in units of @xmath39 ) . for the pure dfs phase ( @xmath112 )
there are two minima corresponding to @xmath113 and @xmath114 , i.e. , prolate and oblate deformations of neutron fermi spheres , respectively . in general the position of the minimum of @xmath107 in the @xmath60-@xmath61 plane ( passing through the minima of the limiting cases )
prefers either large deformations or large finite momenta .
the absolute minimum energy state corresponds to @xmath114 _ and _ @xmath112 ; that is , while the loff phase is a local minimum state , it is generally unstable towards evolution to a pure dfs phase with oblate / prolate deformations of neutron / proton fermi spheres .
the effect of the kinetic energy contribution is to suppress the large pairing contribution for negative values of the deformation parameter .
the position of the true minimum of the free energy coincides with the maximum in the pairing gap in the positive @xmath60 regime . when the system is in an isospin asymmetric state the effect of the temperature on the pairing gap and related characteristics of the superconducting state is twofold : first , at high temperatures the pairing is suppressed due to the thermal excitation of the quasiparticle states , in analogy to the classical bcs superconductors ; second , at low temperatures the pairing is enhanced by the temperature , since the thermal excitation of the quasiparticle states at different fermi surfaces reduces the blocking effect and increases the phase - space overlap .
figure 3 shows the pairing gap in the loff state as a function of the temperature and finite momentum . for @xmath115
the asymmetric bcs state exists between two critical temperatures corresponding to the thermal stimulation and suppression of the pairing . for constant asymmetry
the temperature reduces the antisymmetric part of the quasiparticle spectrum [ @xmath89 in eq .
( [ gap4 ] ) ] , and hence the decoherence between the quasiparticle states on different fermi surfaces .
this is responsible for the gap _ reentrance _ phenomenon at lower critical temperature with increasing temperature .
the reduction of the contribution from the symmetric part @xmath106 of the spectrum with increasing temperature is responsible for the quenching of the superconducting state at the upper critical temperature . at @xmath116 , but for finite @xmath61 , analogous reentrace of the gap function is seen in fig .
the existence of the lower and upper critical momenta can again be understood from the contributions of the finite momentum to the quasiparticle spectrum . in the general case @xmath117 and @xmath118 the reentrace phenomenon extends in the @xmath119 plane :
the pairing region of the phase diagram is bounded by the curves for the lower and upper critical values @xmath120
, @xmath121 , or equivalently @xmath122 , @xmath123 .
the duality between the effect of finite @xmath24 on the pairing gap at fixed @xmath61 to that of finite @xmath61 at fixed @xmath24 is evident . in fig .
4 we show the difference between the free energies of the superconducting and normal states @xmath107 for the loff phase .
the true ground state here corresponds to the zero - temperature limit for pair momentum @xmath124 .
the effect of the temperature on the loff phase depends on the momentum of the pair : for low momenta there is a minimum as a function of the temperature ( as is also the case for the asymmetric bcs state ) . for large fixed momenta
the temperature reduces the pairing gap .
the same applies for the perturbations from the @xmath115 state at fixed temperature : for low temperatures the system prefers finite momenta , while for large temperatures the finite momentum only reduces the energy .
the overall dependence of the free energy on the temperature and the finite momentum reflects the dependence of the pairing gap and the potential energy of the pairing interaction on these quantities .
the features seen in the phase diagram of the dfs phase at finite temperatures are analogous to those already discussed for the loff phase . here
we reiterate this discussion with the emphasis on the qualitative differences of these phases .
the pairing gap as a function of the temperature and the relative deformation ( @xmath60 ) is shown in fig .
the reentrace effect for @xmath125 is seen for both signs of @xmath60 ; the lower critical deformation corresponds to the onset of the pairing due to the restoration of the phase - space coherence via the deformation of the spherical fermi surface into quasiellipsoidal form .
the upper critical @xmath60 corresponds to the loss of coherence due to the large shift in the fermi surfaces of the paired particles .
the @xmath126 duality is seen in the case where deformation vanishes ( asymmetric bcs state ) .
the lower critical temperature corresponds to the thermal stimulation of the pair correlations due to the smearing of the fermi surfaces , while the upper critical temperature corresponds to the ordinary thermal suppression of the pairing in the bcs - normal state transition .
the effect of @xmath60 at finite - temperatures can be understood in terms of its contribution to the antisymmetric part of the quasiparticle spectrum , in complete analogy to the loff phase .
the reentrace effect extends in the plane @xmath126 , and the pairing region is bounded by the critical values @xmath127 and @xmath128 , or equivalently @xmath129 and @xmath130 . unlike the loff phase these curves
do not terminate at @xmath110 , rather they show a continuous crossover from the oblate to the prolate deformation of the fermi surface of the paired particles . for either sign of @xmath60
the temperature thermally stimulates the superconducting phase for small perturbations from the asymmetric bcs state corresponding to the @xmath110 cut ; for large deformations the pairing is largest in the zero - temperature limit .
the effect of finite @xmath60 at fixed temperature is a suppression of the pairing in the high - temperature limit , while at lower temperatures there is a pronounced maximum in the pairing gap as a function of @xmath60 .
finally , the difference between the free energies of the superconducting and normal states @xmath107 for the dfs phase is shown in fig .
6 . there are two minima corresponding to the oblate or prolate deformations of the fermi surfaces of neutrons / protons .
the shape of @xmath107 is controlled by the interplay between the pairing potential energy and the kinetic energy ; the contribution from the entropy difference is numerically negligible .
the minima correspond to large deformations of the order 0.5 which optimize the gain in the potential energy over the loss in the kinetic energy due to the deformation ; the true minimum occurs for positive @xmath60 at zero - temperature .
while temperature stimulated pairing correlations are seen for small deviations from the bcs state , this effect is not essential , since the minimum of the @xmath107 surface occurs at large @xmath60 .
the dfs phase is most stable in the zero - temperature limit .
the fermionic condensation in nuclear matter under isospin asymmetry leads to superconducting states which spontaneously break the space symmetries .
we studied two realizations of such condensates - the loff phase , where the cooper pairs carry finite momentum , and the dfs phase where the fermi surfaces of the paired nucleons are deformed , to the lowest order into ellipsoidal form .
a unifying feature of these phases is the deformation of the spectrum of the quasiparticle excitations at the leading ( loff ) and next - to - leading ( dfs ) order in the expansion in legendre polynomials with respect to the angle between the particle momenta and the axis of symmetry breaking .
when these deformations are treated simultaneously , we find that the pure dfs phase has lower energy than the loff phase .
the combination of these phases still lower the energy of the normal system and can exists in a metastable state .
in particular , the pure loff state could be realized in the case where the unpaired state features some intrinsic spatial asymmetries ( e.g. , due to the lattice in the metallic superconductors ) . of course
, the axis of the symmetry breaking for the loff and dfs phases need not be the same and , more generally , the patterns of symmetry breaking could be more complicated than those studied here ( e.g. , higher order multipole deformations of the fermi surfaces ) .
the phase diagrams of the loff and dfs phases in the temperature / deformation plane show a number of unusual features : ( i ) two critical temperatures for the onset of the superconducting state ( the reentrace phenomenon ) .
the upper critical temperature is the analog of the bcs temperature for the superconducting - normal phase transition .
the lower critical temperature is specific to asymmetric systems and results from the thermally stimulated phase - space overlap between the pairing states .
( ii ) there exists duality between the temperature effects and the finite - momentum or deformation effects .
for example , at low temperatures the loff phase is characterized by two critical momenta of the cooper pairs at which the superconducting phase sets in .
( iii ) the free energy of the combined loff and dfs phases shows local minima corresponding to negative and positive relative deformations at zero momentum of the pairs .
the lowest energy corresponds to a phase with oblate neutron and prolate proton fermi surfaces ( @xmath60 is positive ) .
the realization of the @xmath78 pairing in bulk nuclear matter depends on a number of factors : ( i ) the ratio of the pairing gap to the average chemical potential , which controls the density - temperature - dependent critical asymmetries at which the pairing vanishes ; this ratio depends sensitively on the quasiparticle renormalization effects which need further study in this context .
( ii ) a second factor involves the asymmetries implied by the @xmath0 equilibrium of the matter with leptons in the stellar matter ; while the asymmetries derived for the normal state may serve as order of magnitude estimates , the loff and dfs phases under @xmath0 equilibrium might require asymmetries largely different from those deduced for the unpaired system .
( iii ) the matter could prefer a state where the superconducting and normal phases coexist in a mixed phase , with superconducting `` islands '' immersed in the unpaired matter ; the analogy to the type - i superconductors in the magnetic fields is a useful counterpart of this type of ordering .
( iv ) the extreme low - density asymmetric nuclear matter can support pairing in the @xmath78 channel at much larger asymmetries than the matter , say , at the nuclear saturation density ; however , in this strong coupling regime one deals with a bose - einstein condensate of deuterons coexisting with a low - density neutron gas@xcite and one needs an understanding of the crossover from the loff and dfs phases to their bose - einstein condensed counterparts .
we now briefly comment on possible realization of the phases with broken space symmetries in other nuclear systems .
the neutron - proton pairing in finite nuclei@xcite is an obvious candidate .
the odd - number projected bcs equations contain the decoherence effects in the gap equation introduced by the extra odd particle .
the resulting blocking effect is quite analogous to the phase decoherence due to the imposed neutron - proton asymmetry in the bulk nuclear matter .
the problem of realization of the loff phase is facilitated by the fact that the pairing instability occurs for total momenta @xmath131 , where @xmath132 is the radius of the nucleus .
therefore , for already medium size nuclei the influence of the surface on the condensation energy should not be large and the solutions found for the interior of the nucleus should go over to those corresponding to infinite systems .
the translational symmetry is not broken by the dfs phase , therefore the finite size effects ( while being generally of the same order of magnitude as for the loff phase ) would affect the structure of the condensate to the same extent as those for the case of the pairing with undeformed fermi surfaces . quite generally , the phases with broken space symmetries arise in the physical situations where the time - reversal symmetry between the paired fermion states is broken .
this is the case , for example , in finite nuclei under rotation .
the @xmath3 pairing among neutrons and/or protons in such nuclei can support the loff or dfs like states .
similarly , the magnetic field breaks the time - reversal invariance due to the pauli paramagnetism of neutrons and protons . the loff and dfs phases can arise for the case of the @xmath3 pairing in strong magnetic fields , for example , in neutron star crusts .
these phases will emerge when the interaction energy of a neutron spin with the magnetic field @xmath133 , where @xmath134 is the anomalous magnetic moment of a neutron , is of the order of the pairing gap in the zero field limit .
this condition is satisfied for the field strength of the order @xmath135-@xmath136 g ( depending on the value of the pairing gap ) which could occur in the highly magnetized subpopulation of neutron stars ( magnetars ) .
as is well know , the breaking of the continuous symmetries in the infinite systems leads to collective excitations with vanishing minimal frequency ( goldstone s theorem ) .
the loff phase breaks the translational and rotational symmetries ; the existence of the preferred direction in the dfs phase breaks the rotational symmetry of the system .
the breaking of continuous symmetries in the ground of these phases , therefore , implies new collective bosonic modes in asymmetric nuclear matter . in finite systems
the goldstone theorem implies collective modes which tend to zero as a certain power of the radius of the system .
the observation of the goldstone modes associated with the loff and/or dfs phases could provide information on the true ground state structure of the condensate under laboratory conditions , for example , in finite nuclei .
we would like to acknowledge the financial support by the sonderforschungsbereich 382 of dfg .
99 j. w. clark , c .- g .
kllman , c .- h .
yang , and d. a. chakkalakal , phys . lett . *
b61 * , 331 ( 1976 ) .
j. a. niskanen and j. a. sauls ( unpublished ) .
l. amundsen and e. stgaard , nucl .
a437 * , 487 ( 1985 ) .
j. m. c. chen , j. w. clark , e. krotschek , and r. a. smith , nucl .
a451 * , 509 ( 1986 ) .
t. l. ainsworth , j. wambach , and d. pines , phys .
b222 * , 173 ( 1989 ) .
m. baldo , j. cugnon , a. lejeune , and u. lombardo , nucl .
a515 * , 409 ( 1990 ) .
j. wambach , t. l. ainsworth , and d. pines , nucl .
a555 * , 128 ( 1993 ) . v. a. khodel , v. v. khodel , and j. w. clark , nucl . phys . * a598 * , 390 ( 1996 ) .
schulze , j. cugnon , a. lejeune , m. baldo , and u. lombardo , phys .
b375 * , 1 ( 1996 ) .
. elgary and m. hjorth - jensen , phys . rev . * c57 * , 1174 ( 1998 ) .
j. kuckei , f. montani , h. mther , a. sedrakian , nucl - th/0210010 .
m. hoffberg , a. e. glassgold , r. w. richardson , and m. ruderman , phys .
* 14 * , 775 ( 1970 ) .
t. takatsuka and r. tamagaki , prog .
. phys . * 46 * , 114 ( 1971 ) .
t. takatsuka , prog .
theor . phys . *
48 * , 1517 ( 1972 ) .
r. w. richardson , phys . rev . * d5 * , 1883 ( 1972 ) .
l. amundsen and e. stgaard , nucl . phys . *
a442 * , 163 ( 1985 ) .
elgary , l. engvik , m. hjorth - jensen , and e. osnes , nucl . phys . *
a607 * , 425 ( 1996 ) .
m. baldo , .
elgary , l. engvik , m. hjorth - jensen , and h .- j .
schulze , phys .
* c58 * , 1921 ( 1998 ) . v. a. khodel , v. v. khodel , and j. w. clark , phys .
lett . * 81 * , 3828 ( 1998 ) .
v. a. khodel , j. w. clark , and m. v. zverev phys .
lett . * 87 * , 031103 ( 2001 ) . v. j. emery and a. m. sessler , phys
. rev . * 119 * , 248 ( 1960 ) .
t. alm , b. l. friman , g. rpke , and h. schulz , nucl .
phys . * a 551 * , 45 ( 1993 ) .
a. sedrakian , th .
alm , and u. lombardo , phys . rev .
* c55 * r582 ( 1997 ) a. sedrakian and u. lombardo , phys .
. lett . * 84 * , 602 ( 2000 ) .
u. lombardo , p. nozieres , p. schuck , h .- j .
schulze , and a. sedrakian , phys .
* c64 * 064314 ( 2001 ) .
elgary , l. engvik , m. hjorth - jensen and e. osnes , phys . rev . *
c57 * , r1069 ( 1998 ) .
a. i. akhiezer , a. a. isayev , s. v. peletminsky , and a. a. yatsenko , phys . rev . *
c63 * 021304 ( 2001 ) .
a. a. abrikosov and l. p. gorkov , zh .
. fiz . * 39 * 1781 ( 1960 ) [ sov .
jetp * 12 * , 1243 ( 1961 ) ] .
a. m. clogston , phys .
. lett . * 9 * , 266 ( 1962 ) .
b. s. chandrasekhar , app .
phys . lett . * 1 * , 7 ( 1962 ) .
l. p. gorkov and a. i. rusinov , zh .
exsp . teor
* 46 * 1363 ( 1964 ) [ sov .
jetp * 19 * , 922 ( 1964 ) ] .
a. i. larkin and yu .
n. ovchinnikov , zh .
47 * 1136 ( 1964 ) [ sov .
jetp * 20 * , 762 ( 1965 ) ] .
p. fulde and r. a. ferrell , phys .
rev . * 135 * , 550 ( 1964 ) .
a. sedrakian , phys . rev . *
c63 * 025801 ( 2001 ) .
a. a. isayev , phys .
* c65 * 031302 ( 2002 ) . | isoscalar cooper pairing in isospin asymmetric nuclear matter occurs between states populating two distinct fermi surfaces , each for neutrons and protons . the transition from a bcs - like to the normal ( unpaired ) state , as the isospin asymmetry
is increased , is intervened by superconducting phases which spontaneously break translational and rotational symmetries .
one possibility is the formation of a condensate with a periodic crystallinelike structure where cooper pairs carry net momentum ( the nuclear larkin - ovchinnikov - fulde - ferrell - phase ) .
alternatively , perturbations of the fermi surfaces away from spherical symmetry allow for minima in the condensate free energy which correspond to a states with quadrupole deformations of fermi surfaces and zero momentum of the cooper pairs .
in a combined treatment of these phases we show that , although the cooper pairing with finite momentum might arise as a local minimum , the lowest energy state features are deformed fermi surfaces and cooper pairs with vanishing total momentum . |
the vortex matter of cuprate superconductors has a rich phase diagram in the @xmath7-@xmath8 plane ( see e.g. ref . ) .
not long after the discovery of the cuprate superconductors it has been established that there are at least two phases , a fluid ( liquid or gas)@xcite phase at high temperatures and fields and a solid ( lattice or glass ) phase at low temperatures and fields .
it was shown that the transition between these two phases is of first order , at least in low fields up to a critical point .
@xcite recently , it has been shown@xcite that there are two distinct solid phases . in low fields
the vortices form a quasi - ordered lattice in a so - called bragg glass,@xcite which is stable against the formation of dislocations . in higher fields they form a highly disordered , entangled@xcite solid .
the nature of the high field phase is not well understood .
it could be anything from a highly viscous fluid to a vortex glass with unbounded barriers against vortex movement.@xcite we will simply term it `` glass '' henceforth .
the high - field glass phase is caused by weak , random and uncorrelated disorder due to point - like defects , such as oxygen vacancies.@xcite this type of disorder is present even in the most pure crystals of cuprate superconductors .
strong , correlated disorder due to extended defects , such as twin boundaries or columnar irradiation defects , has got a different influence on the vortex matter and can lead to additional phases , e.g. a bose glass " phase@xcite located at low fields.@xcite correlated disorder may not be present in all crystals . in the following , when we simply speak of `` disorder '' ,
we are referring to weak random point - like disorder .
the pinning induced by the disorder has two effects on the vortex lattice.@xcite firstly , it localizes vortices by trapping them in low - lying metastable states , hardening the solid .
secondly , it promotes transverse wandering of vortices from their ideal lattice positions .
the mechanism of the destruction of the lattice , triggering the transition to the glass phase,@xcite is very similar to the mechanism of melting at higher temperature .
similar to the way the lattice looses translational order and melts into a fluid , when thermal fluctuations become of the order of the lattice spacing , it looses translational order and `` melts '' into a glass , when the disorder - induced line wandering becomes of the order of the lattice spacing . in both cases
the stability against the formation of topological defects , such as dislocations , is lost .
the massive proliferation of dislocations accompanying the destruction of translational order leads to an entangled configuration of vortices , i.e. topological order is lost as well . in highly anisotropic superconductors ,
the lattice - glass transition may fall together with a decoupling between the layers,@xcite of which there is some evidence in bi@xmath9sr@xmath9cacu@xmath9o@xmath10 ( bscco).@xcite this is not expected in less anisotropic systems , like yba@xmath9cu@xmath11o@xmath12 ( y123).@xcite there is considerable evidence that the phase transition between lattice and glass , like the melting transition between lattice and fluid , is of first order@xcite and is even a continuation of the melting line.@xcite since the solid - solid transition is an order to disorder transition with the disordered phase located at high fields , the transition line @xmath4 can be expected to shift to lower fields with the introduction of additional disorder in the crystal structure .
calculations@xcite of @xmath6 , based on a lindemann criterion , confirm this expectation . indeed
, there seems to be a correlation between the position of @xmath6 and the purity of a sample , as can be seen by comparing different measurements on y123 crystals .
also , studies of the effect of electron irradiation on the transition between lattice and glass have found a systematic decrease of @xmath6 with increasing irradiation dose.@xcite an alternative way to introduce structural disorder into a system is to partially substitute an element of the compound .
however , the generally varying level of oxygen within samples , which is difficult to measure and control , constitutes a serious problem .
yba@xmath9cu@xmath2o@xmath3 ( y124 ) is a compound very well suited for studying the dependence of @xmath6 on chemically introduced disorder , because in this compound the oxygen content is fixed to 8 per unit cell .
another advantage of y124 is that twinning , which leads to strong correlated disorder , does not exist in this compound.@xcite finally , y124 has got an intermediate anisotropy @xmath13,@xcite which , as compared to y123 , shifts the glass phase down to fields more easily attainable experimentally . to exclude or at least to minimize concurrent sources of disorder in the system
it is advisable to use single crystals for any measurements .
if vortices are not ordered in a lattice , they can better adapt to the local pinning potential .
therefore , critical current densities can be expected to be higher in the glass phase .
high critical currents in high magnetic fields are desirable for many applications .
investigations of the disordered solid state and it s boundaries are thus very important in view of practical applications . many irradiation studies found that critical current densities can increase by orders of magnitude upon irradiating a sample.@xcite studies on compounds with chemical disorder have also found increases of @xmath14 and even the development of a second peak.@xcite in this work , we study the influence of a substitution of isovalent sr for ba in yba@xmath0sr@xmath1cu@xmath2o@xmath3 on the lattice to glass transition line @xmath4 , the irreversibility line @xmath15 and the critical current densities @xmath14 . in the paper , we are going to give evidence of an enhancement of the maximum critical current density in the glass phase @xmath5 by more than an order of magnitude due to structural disorder introduced by sr substitution .
the change of the vortex matter phase diagram due to the influence of relatively strong quenched disorder will be presented .
[ samples ] single crystals of y124 , with and without sr substitution , were grown using a high - pressure technique , the details of crystal growth are reported elsewhere.@xcite single crystals of yba@xmath0sr@xmath1cu@xmath2o@xmath3 with sr content up to @xmath16 or 32% were obtained .
the crystals were checked with edx and their structure analyzed with single crystal x - ray analysis.@xcite the transition temperatures @xmath17 of the crystals were determined with a squid magnetometer by temperature sweeps ( both zero field and field cooled ) in an applied field of @xmath18 .
all samples chosen for further magnetic measurements had a well - resolved transition with a transition width ( 10% to 90% ) smaller than @xmath19 ( see fig . [ fig1 ] ) .
curves of the crystals used in this work showing sharp superconducting transitions .
the curves shown were measured in a magnetic field of @xmath20 , under zero field cooled conditions . ] [ magn_exp ] measurements of the dc magnetization were performed with a quantum design mpms5 squid magnetometer , additional measurements were performed with a non - commercial squid magnetometer with a sensitivity better than @xmath21 .
all measurements were done with the magnetic field applied parallel to the @xmath22-axis of the crystal and generally the magnetization was measured at fixed temperatures as a function of the external field being swept up and down . after each change of the field
, we waited for a time of @xmath23 before measuring the magnetic moment four times using a scan length of @xmath24 .
the relatively short scan length was chosen to ensure maximum homogeneity of the applied field and was necessary to ensure a sufficient resolution of all features of interest discussed below .
the wait time of @xmath23 was chosen to avoid effects of the initial relaxation of the field of the superconducting magnet without extending the time of measurements too much . throughout the paper @xmath7
denotes the applied field , i.e. no correction of the demagnetization was made .
this is not a problem , however , since the magnetization is always much smaller than the external field except in the low field range , which is not the particular interest in this paper .
the critical current density @xmath14@xcite was calculated from the width of the hysteresis loop using an extended bean model.@xcite the irreversibility field @xmath25 was determined as the field where increasing and decreasing field branches of the @xmath26 hysteresis loop meet , or equivalently as the field where the critical current density vanishes
. a uniform criterion of @xmath27 was used . for the crystals
studied it corresponds to a criterion of in between @xmath28 and @xmath29 , which is not far from the accuracy limit of our experimental setup .
in the following we will discuss the influence of sr substitution on disorder and structure , critical current densities , the irreversibility line , and the transition from lattice to glass .
sr substitution increases the disorder in the structure , evidenced for example by a dramatically increased nqr line width.@xcite no evidence for an introduction of correlated disorder was found and since twin boundaries are absent in yba@xmath0sr@xmath1cu@xmath2o@xmath3 the assumption of the disorder in our crystals to be purely random and point - like seems justified .
the sr substitution also induces a variation of structural parameters , causing a charge redistribution .
bond valence sum calculations and nuclear quadrupole resonance measurements indicate a transfer of holes from oxygen to copper atoms in the cuo@xmath9 planes and from copper to oxygen in the chains.@xcite the charge redistribution can explain the increase of @xmath17 upon substituting more than 10% sr for ba .
the initial decrease of @xmath17 may be due to the increased site disorder.@xcite an important structural modification due to sr substitution is a decrease of the thickness of the blocking layer @xmath30 separating the superconducting cuo@xmath9 planes , linear with sr content .
a substitution of 32% sr for ba causes the blocking layer to shrink by about 1%,@xcite which leads to an improved coupling between the superconducting planes .
torque measurements indeed indicate a corresponding decrease of the anisotropy @xmath31.@xcite figure [ fig2]a ) shows one of the half hysteresis loops measured .
the second peak is clearly discernible and dominates @xmath26 . in all crystals ,
the second peak was observed over the whole temperature range measured , even just @xmath32 below @xmath17 .
the form of the hysteresis loop and thus of @xmath33 can be linked to the different phases of vortex matter . in low fields hysteresis
is small and the hysteresis loop is rather anisotropic , which indicates that the major source of hysteresis in low fields is not of bulk origin ( see also sec . [ il_disc ] ) .
the origin of the low bulk hysteresis is probably that the vortices are in the lattice phase in small @xmath7 .
weak random point disorder is not effective in pinning an ordered dislocation - free lattice of vortices .
the magnitude of the critical current density in low fields gives an indication of the degree of disorder in the system .
for our samples , measurements of the critical current density at reduced temperatures of @xmath34 and @xmath35 in fields of @xmath36 and @xmath37 suggest that the disorder is increased monotonically up to the highest substitution level measured , with the biggest rate between @xmath38 and @xmath16 . between them
are , denoted as @xmath39 , @xmath40 and @xmath6 , respectively .
note that there are small differences between the respective fields between the two directions of field change ( see text and fig .
[ fig10 ] ) .
the irreversibility field , where field increasing and decreasing branches meet , is denoted as @xmath25 .
b ) additional parts of hysteresis loops showing the kink region at three different temperatures . ]
when , upon rising the field , the lattice is destroyed in the glass state and the vortex system has an additional `` dislocation degree of freedom'',@xcite individual vortices can much better adapt to local minima in the pinning potential .
this results in a higher effective pinning force and thus a higher critical current density @xmath14 and relaxation barrier.@xcite additionally , the entanglement itself was also suggested to increase critical current densities , due to an increased intervortex viscosity,@xcite provided the barriers to flux cutting are high enough.@xcite in our case , a steep rise of the magnetization occurs just below the kink field , denoted @xmath6 in fig .
[ fig2]a ) ( see also sec.[hss_disc ] ) .
local magnetic measurements on a bscco crystal revealed a sharp increase in @xmath41 , followed by a monotonic decrease afterwards.@xcite the so - called `` second peak '' ( or `` fishtail '' ) effect in bscco can thus be accounted for by the order - disorder transition . on untwinned y123 crystals ,
a similar sharp increase in @xmath41 was observed by local magnetic measurements,@xcite but contrary to the situation for bscco , @xmath41 continues to increase upon increasing field , although with a slower rate , which leads to a sharp kink in @xmath26 . only in even higher fields
the magnetization starts to decrease with increasing field .
independently of vortex matter phase transitions , the second peak or fishtail effect in y123 was attributed to a crossover from elastic to dislocation - mediated plastic creep.@xcite this mechanism is closely connected to the lattice - glass transition though , since in the lattice phase dislocations are strongly suppressed while in the glass phase dislocations proliferate , as mentioned above.@xcite from collective creep theory a rise of the experimentally observable critical current density with the field is expected,@xcite as long as elastic creep can be assumed and the bulk pinning is relevant . from the shape of the hysteresis loops shown in fig .
[ fig2 ] follows that in yba@xmath0sr@xmath1cu@xmath2o@xmath3 the situation is similar to the one in y123 . after rising the field even more , above @xmath25 ,
the vortex matter finally enters the fluid phase , where critical current densities are zero .
the dependence of the hysteresis loop of an unsubstituted y124 single crystal on defects created by fast neutron irradiation was studied in ref . .
fast neutron irradiation leads to improved pinning properties due to collision cascades ( strongly pinning extended defects of spherical shape ) and agglomerates of smaller defects .
a clearly discernible second peak was observed in the untreated crystal . in the weakly irradiated crystal
the second peak was less pronounced and after irradiation to a fluence of @xmath42 it disappeared . at the same time
, @xmath14 drastically increased in low to medium fields and the hysteresis loops became more symmetrical .
the effect of fast neutron irradiation to remove the second peak may be understood from the different defect structure .
the disorder resulting from the collision cascades is to some extent correlated and fast neutron irradiation can therefore create a bose glass phase in low fields .
bulk pinning is also strong in the bose glass phase , and if the irradiation induced correlated disorder is strong enough , the bose glass phase may well destroy all of the bragg glass phase . in high fields
fast neutron irradiation is not as effective in enhancing @xmath14 and the irreversibility fields @xmath25 are not increased . vs. reduced temperature , for different sr contents .
it can be seen that unsubstituted y124 has much lower maximum critical current densities than substituted y124 , at all temperatures .
inset : @xmath43 vs. sr content @xmath44 at @xmath45 . ] on sr content , at a temperature of @xmath46 . the maximum pinning force density increases by more than an order of magnitude upon substituting 32% of sr for ba . upper left inset : second peak field vs. sr content at @xmath45 .
lower right inset : maximum pinning force density vs. sr content , at @xmath45 .
] in our case of sr substitution the significant fields shift , but the qualitative form of the hysteresis loop remains the same . as with fast neutron irradiation , sr substitution increases @xmath14 , but most pronounced in medium fields , in the second peak region .
the maximum critical current densities there , @xmath47 , increase dramatically upon sr substitution , as can be seen in fig .
[ fig3 ] . at the temperature of @xmath48 ,
the critical current density reaches almost the values of critical current densities for y123 : for the crystal with 32% sr substitution @xmath49 , while for a y123 single crystal @xmath50.@xcite this is not the case , however , at the lower temperature of @xmath51 , where y123 crystals have a maximum critical current density roughly three times higher@xcite than the one measured on sr substituted yba@xmath0sr@xmath1cu@xmath2o@xmath3 .
the main panel of fig.[fig4 ] shows the dependence on sr content of the maximum pinning force density @xmath52 corresponding to the critical current densities measured at @xmath53 .
it can be seen that @xmath54 increases by more than an order of magnitude upon sr substitution .
the lower right inset of fig.[fig4 ] shows the sr dependence of @xmath54 at a fixed _ reduced _ temperature .
the initially approximately linear rise of the maximum pinning force density saturates for @xmath55 indicating that increasing the substitution level to @xmath56 does not increase the pinning drastically any more .
the critical current density itself also shows saturation behaviour , but for lower sr content ( see inset of fig .
[ fig3 ] ) .
the difference is due to the substitution dependence of the second peak field , shown in the upper left inset of fig .
the dependence @xmath57 is similar to the substitution dependence of @xmath6 , discussed in sec .
[ hss_disc ] the irreversibility fields @xmath25 as a function of temperature are , for different sr substitution levels , presented in fig .
the theoretical meaning of the irreversibility field is not very clear , often parts of it are controlled by surface or geometrical barriers ( see below ) . in high fields , it was often found to be located near the transition from glass to fluid ( see e.g. refs . ) . however , whether a true phase transition between a glass and a fluid phase even exists , is not a settled question .
not much is known about the properties of the highly disordered entangled solid located above @xmath6 , apart from properties that distinguish it from the lattice phase .
the feature differentiating the phase from the liquid phase is a rapid freezing of the dynamics , but it is not sure whether this freezing of the dynamics is a proper phase transition or merely a crossover . the existence of a separate so - called vortex glass ( vg ) phase , distinguished by an unbounded distribution of the heights of barriers between metastable states , was proposed early after the discovery of high @xmath17 superconductivity.@xcite vortex glass scaling of current voltage relations was used to determine the vortex glass transition line @xmath58 experimentally , and also provided strong support for the vg theory ( see e.g. refs . ) . however , later experiments and numerical calculations found discrepancies ( see ref . and references therein ) , and suggest a glass phase more akin to window glass . despite these uncertainties , @xmath15 is certainly important from a more practical point of view , as it limits the field range for applications .
( open symbols ) , for different sr contents in yba@xmath0sr@xmath1cu@xmath2o@xmath3 : circles correspond to @xmath59 , diamonds to @xmath60 , down triangles to @xmath61 , up triangles to @xmath38 and squares to @xmath16 .
error bars have been omitted for clarity .
full lines are fits of the @xmath4 data to eq.@xmath62([deltatc ] ) . except for @xmath59 ,
fitting was restricted to the range @xmath63 .
dashed lines are guides for the eye indicating an exponential temperature dependence of @xmath6 at lower temperatures . variations of @xmath15 generally are relatively small and the details depend on the temperature .
@xmath4 of the unsubstituted crystal is located in much higher fields than @xmath4 of all substituted crystals and for high temperatures is practically located in the region of the irreversibility fields . ]
it is hard to see a clear trend in the sr substitution dependence of the measured irreversibility fields .
differences between @xmath25 of the different samples measured are relatively small , and the details depend on the temperature as well . now it is important to remember that bulk pinning is not the only source of magnetic hysteresis .
there are additional contributions due to bean - livingston surface barriers@xcite and due to geometrical barriers.@xcite the significance of surface and geometrical barriers depends on several factors .
one of them is the strength of the bulk pinning - the smaller the ( bulk ) critical current density the bigger the relative importance of barriers . in yba@xmath0sr@xmath1cu@xmath2o@xmath3
the critical current densities depend , as discussed above , strongly on @xmath44 , and are very low for the unsubstituted crystal .
indeed , while the hysteresis loops of substituted crystals are generally rather symmetric , this is not the case for the unsubstituted crystal , which is an indication that barrier hysteresis is more important than bulk hysteresis.@xcite another factor is the sample shape - bulk hysteresis depends linearly on the sample diameter , while the barrier hysteresis scales with the ratio of thickness to diameter .
thus , barrier effects are more important for thicker and smaller ( in diameter ) crystals .
the thickness to diameter ratio varies from @xmath64 to @xmath65 for the crystals used in this study .
finally , an important factor is the temperature , since the bulk hysteresis is depressed by thermal depinning , while the one due to geometrical barriers is less affected by temperature.@xcite in the second peak region , the main part of the hysteresis is clearly of bulk origin , as follows from the shape of the hysteresis loops ( see fig . [ fig2 ] ) .
however , since the total hysteresis observed is simply the sum of bulk and barrier hysteresis , the irreversibility field may still be determined by surface or geometrical barriers , since the bulk hysteresis goes to @xmath66 faster at higher fields.@xcite in high fields , bulk critical current densities were found to go to zero exponentially.@xcite a linear relationship between @xmath14 and @xmath67 is indeed followed in the crystals we measured .
figure [ fig6 ] shows this relationship for yba@xmath68sr@xmath69cu@xmath2o@xmath3 .
the figure shows also , however , that there are systematic deviations below a certain temperature dependent threshold of critical current densities . for higher temperatures the deviation appears at larger critical current densities
this may indicate that the irreversibility lines are indeed influenced considerably by barrier hysteresis at higher temperatures . .
] on the other hand , measurements of the influence of fast neutron irradiation on the high temperature part of the irreversibility field on a y124 single crystal found a slight decrease of @xmath25 upon irradiating the crystal.@xcite irradiation may reduce bean - livingston type surface barriers , but those are , like bulk pinning , strongly temperature dependent and were generally not found to be important near @xmath17 .
geometrical barriers , on the other hand , should not change upon irradiation . as the irradiation did change the irreversibility field at high temperatures
, it would indicate that @xmath25 at high temperatures can not be due to bean - livingston or geometrical barrier hysteresis .
however , as mentioned above , the importance of barriers strongly depends on the sample shape and we can not be completely sure to what extent barriers influence @xmath25 , especially in the case of the unsubstituted crystal , where bulk pinning is very weak . in order to explain the effect of the sr substitution on the bulk properties , it is therefore advisable to compare @xmath15 at not too high reduced temperatures .
a comparison of @xmath15 at @xmath70 is shown in fig .
the position of the irreversibility field is not changed much , but the general trend that @xmath25 rises with @xmath44 is consistent with the also rising maximum critical current density .
the only irreversibility field that does not fit into this trend is the one of the unsubstituted crystal , which may be attributed to barrier effects that still are not negligible at this temperature for this weakly pinned compound . at temperatures of @xmath71 ( open diamonds ) and @xmath72 ( open circles ) , and of @xmath73 ( full diamonds ) . for both temperatures @xmath6 drops significantly by a small substitution of sr for ba , but remains essentially constant upon further sr substitution .
@xmath25 also drops by a small sr substitution , but less significantly .
it rises again upon further substitution and is higher for @xmath16 than for @xmath59 .
the @xmath44 error bars have been omitted for clarity . ] an increasing glass to liquid transition field can be justified theoretically in a similar way as the decrease of the lattice to glass transition field . since the glass to liquid transition line depends on the relation between the pinning energy and the thermal energy , and
because the pinning energy is increased upon introducing more disorder into the system , it would seem natural that the transition occurs at higher temperatures .
the dependence of the glass to fluid transition on disorder was studied , for electron irradiated y123 , by nishizaki _
et al._.@xcite they measured both the glass transition temperature @xmath74 resistively , and the irreversibility field @xmath25 , and found the two lines @xmath75 and @xmath15 to coincide .
_ find the opposite disorder dependence of the glass to fluid transition , in their case @xmath58 and @xmath25 decrease with increasing fluence of electron irradiation .
there are different possible explanations for the opposite behaviour of the glass to liquid transition line when the disorder is increased between the present measurements and those of nishizaki _
et al._. firstly , the disorder introduced by substituting 10% or more sr for ba in y124 is much stronger than the disorder introduced by an irradiation of @xmath76 electrons with a fluence of up to @xmath77 in untwinned y123 .
this can be seen by comparing the phase diagrams .
the irradiated y123 crystals show a first order melting transition in fields up to @xmath78 , while all of our sr substituted y124 crystals show no first order melting transition at all up to temperatures very near @xmath17 .
et al . _
note that the decrease of @xmath58 upon increasing disorder they measured is consistent with a theoretical prediction made for systems with weak pinning.@xcite the disorder in our substituted crystals may be too big for this to be applicable .
secondly , the rise of @xmath25 with sr substitution could also be explained by a decreased anisotropy of the substituted crystals , since @xmath25 was found to scale with @xmath79 , where @xmath80 is the distance between two adjacent superconducting planes.@xcite if the rise of @xmath25 in yba@xmath0sr@xmath81cu@xmath2o@xmath3 with @xmath44 for @xmath82 is due to a decreased anisotropy caused by the decreased blocking layer thickness , the initial drop for @xmath83 may be due to the disorder , in which case the dependence of @xmath25 on disorder would be qualitatively the same as the one found by nishizaki _
et al._. as noted in sec .
[ jc_disc ] , a kink in @xmath26 ( denoted @xmath6 ) is visible between onset and maximum of the second peak of the hysteresis loop shown in fig .
[ fig2 ] . according to previous studies performed on y123,@xcite
it is likely that the kink corresponds to the transition line from the lattice to the glass state of vortices . in ref .
it was argued that the equilibrium phase transition corresponds to the kink in @xmath26 in decreasing field , while the kink in increasing field corresponds to the upper limit of metastability of the lattice phase .
however , since the pinning is much stronger in the glass phase , the equilibrium transition can be expected to be located quite close to the upper limit of metastability .
this was indeed found from the measurements of ref .. also in our case , while we have more data on the field increasing branch , the kink in @xmath7 decreasing was checked at representative temperatures and the difference in the two kinks was found to be rather small indeed ( see fig.[fig2 ] ) .
we should mention that the actual equilibrium transition is a little bit lower than the field of the kink in @xmath84 , which we call @xmath6 ( see fig.[fig10 ] ) .
it is interesting to note that generally the kink positions ( for both directions of field changes ) do not depend on the waiting time ( we tested between @xmath85 and @xmath86 ) after changing the the field .
additionally the kink field location does not depend on the field change step size and not on the small inhomogeneities of the magnetic fields ( @xmath87 ) tested by using different scan lengths , but the _ shape _ of the @xmath26 curve around the kink _ does _ depend on step size and scan length . on the other hand , the onset fields @xmath39 , defined as the fields where the absolute value of the magnetization has a minimum , have a slight tendency to shift to lower fields upon increasing waiting time , while the absolute value of the magnetization at the onset is decreased significantly upon increasing waiting time .
additionally the onset measured in decreasing fields can be much lower ( up to 50% at low temperatures ) than the onset measured in increasing field .
figure [ fig8 ] shows a selection of magnetization curves measured , in the region where the kink is located .
the sharpness of the kink varies with temperature and depends on the sample as well .
the inset of fig .
[ fig8 ] shows the onset region of one of the curves enlarged and the derivative of the magnetization ( average of the four scans ) , which can help to locate @xmath6 . within a range of a few kelvin below @xmath17
the unambiguous determination of @xmath6 becomes difficult nonetheless .
curves for increasing field near the onset of the second peak .
dashed lines are guides for the eye pinpointing the kink in the magnetization @xmath6 .
the inset shows a magnification of the onset region of a @xmath26 curve and it s derivative at @xmath88 .
@xmath6 is located at the position of the steepest increase of @xmath89 . ]
the increase of the absolute value of the magnetization around @xmath6 observed in our measurements is hardly sharp enough to be called a `` jump '' .
one reason for this is the spatial averaging of the inhomogeneous induction inside the sample .
additionally , the difference between the onset of the second peak in increasing fields @xmath90 , defined as the minimum of the absolute value of the magnetization before the second peak , and the kink field @xmath6 can be explained by the existence of a region of metastability around the phase transition.@xcite recent experiments@xcite indicate that an abrupt change of the field injects a transient disordered vortex phase at the sample edges .
if the thermodynamically stable phase is the lattice phase , the transient disordered vortex phase then decays .
this decay happens with a rate decreasing to 0 as the field reaches @xmath6 . in our measurements ,
the fields @xmath90 , @xmath91 and @xmath92 was found to follow the same dependence on temperature and sr content as @xmath6 ( @xmath93 ) . below the onset , both the magnitude of the critical current density and the average value of the magnetization for the two branches of the hysteresis loop suggest that in this region surface and geometrical barriers are more important than bulk pinning in our samples , i.e. bulk pinning seems to be relevant only if the vortex matter is in the glass phase . the @xmath4 lines , as determined by the kink in the @xmath84 curves for the measured samples , are also shown in fig .
for all samples , @xmath6 decreases monotonically with increasing temperature .
a monotonically decreasing @xmath6 was also found on nd@xmath94ce@xmath95cuo@xmath96,@xcite but measurements on y123@xcite showed an @xmath6 increasing with temperature .
the position of the phase transition between quasi - ordered lattice and highly disordered glass is determined by the interplay between elastic and pinning energy . for the case of weak , random disorder due to point - like defects the theory of collective pinning
was developed by larkin _
et al._.@xcite the defects can interact with the vortices in two ways.@xcite they can cause a spatial variation of the transition temperature ( @xmath97-pinning ) , described by a modulation of the linear term of the ginzburg - landau free energy functional . alternatively , they can cause a spatial modulation of the mean free path ( @xmath98-pinning ) , described by a modulation of the gradient term of the free energy functional . in both cases ,
the influence of disorder is described by a disorder parameter @xmath99 , proportional to the defect density . however , the temperature dependence of @xmath99 is different for the two cases . for @xmath100-pinning , @xmath101 , while for @xmath102-pinning , @xmath103 , where @xmath104 and @xmath105 are the penetration depth and the coherence length.@xcite the order - disorder transition position was calculated analytically by using a lindemann criterion.@xcite for the case of not too big anisotropy and disorder , following the calculation of ref . , we get @xmath106 with @xmath107 , @xmath108 and the collective pinning energy @xmath109 . here , @xmath110 is the flux quantum and @xmath111 the lindemann number .
@xmath6 is inversely proportional to both the anisotropy @xmath31 and the disorder parameter @xmath99 .
concerning the temperature dependence , as long as we are below the depinning temperature @xmath112 , we get @xmath113 in the case of @xmath100-pinning and @xmath114 in the case of @xmath102-pinning.@xcite the calculation of ref
. leads to qualitatively the same results .
since @xmath6 decreases with temperature in our case , the pinning in yba@xmath0sr@xmath1cu@xmath2o@xmath3 can not be @xmath102 pinning .
the temperature dependence of @xmath6 of the unsubstituted crystal agrees satisfactorily with the formula proposed for @xmath115 pinning . in the case of the substituted crystals ,
however , the agreement is limited to values of @xmath116 between about @xmath117 and @xmath118 ( see fig .
[ fig5 ] , the full bold lines are fits to eq.([deltatc ] ) ) .
the lowered @xmath6 in the vicinity of @xmath17 may be attributed to the finite transition width .
however , at low temperatures , the temperature dependence is rather exponential in the case of the substituted crystals , as indicated by dashed lines in fig .
[ fig5 ] , with some indications of a flattening at the lowest temperatures measured .
this behaviour is clearly at odds with eq .
( [ deltatc ] ) and to the best knowledge of the authors is not predicted by any present theory of the order - disorder transition .
an exponential upturn in the _ onset _ field @xmath90 at low temperatures was also found on a nd@xmath0ce@xmath1cuo@xmath119 single crystal by andrade _
et al._,@xcite who attributed it to bean - livingston surface barriers .
however , since the observed magnetic hysteresis is the sum of bulk and barrier hysteresis , it is difficult to imagine why the kink in @xmath26 at @xmath6 should be influenced by surface or geometrical barriers , and in our case a clear kink is observable at low temperatures where @xmath6 has already clearly departed from the @xmath120 dependence observed at higher temperatures . and @xmath25 in the vicinity of @xmath17 , for substitutions of 10% and 32% sr .
] a possible cause for a changed temperature dependence is a dimensional crossover . due to the layered structure of the cuprate superconductors , vortex lines
should be thought of being composed of stacks of `` pancake '' vortices.@xcite only below the 2d/3d crossover field @xmath121 , where @xmath80 is the distance between two adjacent layers , is the interlayer interaction between pancake vortices larger than their intralayer interaction and the pancakes form well defined vortex lines . for all our crystals ,
@xmath122 is estimated well above the upper limit of the fields attainable in our magnetometers .
also , @xmath123 for 2d and @xmath124 pinning,@xcite i.e. the temperature dependence should be even flatter in the 2d regime .
a dimensional crossover can therefore not be responsible for the observed upturn of @xmath6 at low temperatures . in principle a second pinning mechanism , which is very effective at low temperatures , could be overlaid .
however , point - like disorder should rather ( additionally ) suppress @xmath6 , while the influence of correlated disorder should be more visible at high temperatures or low fields.@xcite figure [ fig9 ] shows @xmath6 and @xmath25 of two of the substituted crystals in the vicinity of the transition temperature .
no sign of a tricritical point , where the two lines would meet , can be seen .
a tricritical point would have to be located very near @xmath17 , where reliable measurements become increasingly difficult .
it can be seen in fig .
[ fig5 ] that any substitution of sr lowers @xmath6 very significantly . for two values of @xmath116 ,
@xmath6 vs. strontium content @xmath44 is shown in fig .
[ fig7 ] . after a substitution of just 10% sr for ba ,
@xmath6 drops roughly by a factor of 5 , at both temperatures .
a further increase of sr substitution , however , does not reduce the magnitude of @xmath6 any more .
rather , @xmath125 remains essentially constant .
we do not consider the peak at @xmath55 as big enough to be significant .
the initial decrease of @xmath6 upon sr substitution does not come unexpected .
it follows from eq .
( [ hss_pos ] ) that @xmath126 . sr substitution increases the disorder parameter and decreases the anisotropy ( see sec .
[ sr_disc ] ) .
the large decrease of @xmath6 upon substituting 10% of sr indicates that the influence of the sr substitution on disorder is much bigger than it s influence on the anisotropy . ,
a ) ) and the highest ( @xmath16 , b ) ) sr substitution . the quasi - ordered lattice phase at low temperatures and fields is separated from the highly disordered glass phase by the transition line @xmath92 ( full down triangles , thick line ) . around the transition
is a region of metastability,@xcite where both phases can coexist ( shaded ) .
the lower limit of metastability is marked by @xmath91 ( open down triangles ) and the upper limit , very close to the equilibrium transition , by @xmath92 ( full up triangles ) . also shown is @xmath90 ( open up triangles ) . the irreversibility field @xmath25 ( full circles ) is assumed to correspond roughly to the transition line between glass and fluid phases .
both @xmath6 and @xmath25 are decreasing monotonically with increasing temperature .
for the crystal with @xmath59 the region of a possible tricritical point was not measured and is blurred in the figure . for the crystal with @xmath16 a tricritical point could not be detected and would have to be located very near @xmath17 . ]
what is more surprising is the apparent saturation of the influence of the additional disorder for a substitution level @xmath127 , while the maximum pinning force is still linearly increasing up to @xmath55 . in the case of weak point - like pinning @xmath128 , where @xmath99 is the disorder parameter and @xmath129 the defect density , is expected.@xcite it would seem natural to assume the defect density to be proportional to the sr substitution level @xmath44 .
the blocking layer thickness on the other hand , which is the structural factor with the biggest influence on the anisotropy , decreases linearly with @xmath44.@xcite a large initial decrease of @xmath6 followed by an almost independence on @xmath44 is therefore difficult to explain .
a downward shift of @xmath6 upon introducing more disorder was also found in electron irradiated crystals of bscco@xcite and y123.@xcite the notable difference between both of these irradiation experiments and sr substitution in y124 is that in both cases of electron irradiation there is no sign of any saturation of the disorder induced lowering of @xmath6 . as discussed above , the disorder induced by sr substitution
is probably , even for the lowest substitution level of @xmath130 , already much higher than the disorder induced by the electron irradiation with the largest fluence measured in refs . and .
it may be that in our substituted crystals the disorder is already too big for the applicability of formulas derived for weak disorder .
the resulting phase diagrams for yba@xmath131cu@xmath2o@xmath3 and yba@xmath132sr@xmath133cu@xmath2o@xmath3 are contrasted in fig.[fig10 ] .
the glass phase of yba@xmath132sr@xmath133cu@xmath2o@xmath3 has , as compared to yba@xmath9cu@xmath2o@xmath3 , expanded dramatically from a rather small part of the easily experimentally accessible phase diagram to the phase covering the biggest area in fields @xmath134 .
the most dramatic shift is the lowering of @xmath6 , which is to be expected as the ( quasi-)ordered phase should become smaller upon introducing more disorder into the system .
the region of metastability is much smaller for the substituted crystal , indicating that strong disorder tends to reduce over - heating and especially under - cooling effects .
we speculate that this may be connected to our observation that for unsubstituted y124 with very weak disorder @xmath6 depends strongly on the exact amount of disorder , while for substituted y124 with rather strong disorder it almost does not depend on the exact level of disorder . in real crystals ,
the disorder density depends on the position on a mesoscopic scale .
et al._@xcite observed that for melting supercooling exists only at local maxima of the transition field . in a first approximation the width of metastability can be linked to the difference between maximal and minimal transition fields within the sample . in our case
the same spatial variation of disorder should produce a much bigger variation of the local transition field for the unsubstituted crystal than for the substituted ones .
we investigated the influence of structural disorder , introduced by a substitution of sr for ba in yba@xmath9cu@xmath2o@xmath3 , on the critical current density at elevated fields and on the borders of the glass phase of vortices .
we stress that the disorder introduced chemically by means of sr substitution is random and point - like , rather than disorder due to extended defects .
this is shown by the hysteresis loops keeping their general form , similar to the case of electron irradiation and unlike the case of neutron irradiation , where at least partially the effect is to introduce extended defects .
our main conclusion is that the introduction of even a small amount of disorder into very clean systems changes the phase diagram drastically by lowering the order - disorder transition line @xmath4 and also changing it s temperature dependence .
when the disorder reaches a certain threshold , however , introducing additional disorder does not continue this tendency . with other words , in highly disordered systems ( including probably most systems with partial chemical substitutions ) , the vortex matter phase diagram is relatively robust with respect to variations in the exact degree of disorder .
this conclusion is supported by the rather small variation of the measured kink location as an indication of the order - disorder transition at fixed reduced temperature for any sr substitution level @xmath135 , especially as compared to the corresponding transition in a clean , unsubstituted crystal . that the disorder is still increasing with increasing substitution level for @xmath82 is indicated both by an increasing nqr line width@xcite and increasing critical current densities at low fields and the maximum pinning force density . in strongly disordered systems ,
metastability is less important than in very clean systems and especially the region of undercooling of the glass phase is much smaller .
this might be linked to the observation that the position of the transition line does not depend much on the exact amount of disorder in strongly disordered systems , since variations of the disorder density on a mesoscopic scale do not produce large positional differences of the local transition field in this case .
similar to the robustness of the phase diagram above a disorder threshold , both the maximum pinning force density and the maximum critical current density in the glass phase increase initially with increasing disorder , the critical current density increases by an order of magnitude upon substituting 10% sr .
but they then show saturation behaviour , i.e. the pinning force density and critical current densities can be raised only up to a certain point by introducing random point - like disorder .
this may be relevant for possible applications where high critical current densities are required .
it should be noted that current detailed theories of the order - disorder transition generally assume the disorder to be very weak .
our measurements indicate that in systems with a high amount of disorder , the phase diagram and in particular the order - disorder transition differ qualitatively from the weak disorder case .
theoretical investigations beyond the weak disorder limit would be very helpful .
both the pinning of unsubstituted y124 and sr substituted y124 is likely to be of @xmath115 , rather than @xmath102 type , indicated by the temperature dependence of the order - disorder transition at intermediate temperatures .
for the substituted crystals this can be expected , since there are variations of @xmath17 upon sr substitution . however , the exponential - like temperature dependence of the order - disorder transition of substituted crystals at low temperatures is also emphasized .
the dependence can not be linked to the influence of surface barriers , at least not in any straightforward way , the same dependence of all significant fields ( onsets and clearly discernible kinks for both field directions ) suggests it is a true bulk property of the transition .
we stress that the observed temperature dependence is at odds with currently published theoretical formulas .
[ ack ] this work was supported by the swiss national science foundation , by the european community ( program ica1-ct-2000 - 70018 , centre of excellence celdis ) and by the polish state committee for scientific research ( kbn , contract no .
5 p03b 12421 ) .
h. safar , p. l. gammel , d. a. huse , d. j. bishop , j. p. rice , and d. m. ginsberg , phys .
lett . * 69 * , 824 ( 1992 ) ; h. safar , p. l. gammel , d. a. huse , d. j. bishop , w. c. lee , j. giapintzakis , and d. m. ginsberg , _ ibid_. * 70 * , 3800 ( 1993 ) .
y. paltiel , e. zeldov , y. n. myasoedov , h. shtrikman , s. bhattacharya , m. j. higgins , z. l. xiao , e. y. andrei , p. l. gammel , and d. j. bishop , nature * 403 * , 398 ( 2000 ) ; y. paltiel , e. zeldov , y. myasoedov , m. l. rappaport , g. jung , s. bhattacharya , m. j. higgins , z. l. xiao , e. y. andrei , p. l. gammel , and d. j. bishop , phys .
lett . * 85 * , 3712 ( 2000 ) .
j. karpinski , g. i. meijer , h. schwer , r. molinski , e. kopnin , k. conder , m. angst , j. jun , s. kazakov , a. wisniewski , r. puzniak , j. hofer , v. alyoshin , and a. sin , supercond .
* 12 * , r1 ( 1999 ) .
see e.g. r. b. van@xmath136dover , e. m. gyorgy , l. f. schneemeyer , j. w. mitchell , k. v. rao , r. puzniak , and j. v. waszczak , nature * 342 * , 55 ( 1989 ) ; l. civale , a. d. marwick , m. w. mcelfresh , t. k. worthington , a. p. malozemoff , f. h. holtzberg , j. r. thompson , and m. a. kirk , phys .
* 65 * , 1164 ( 1990 ) .
actually rather the shielding current density @xmath137 than the true critical current density @xmath138 , but since the time scale used is a typical one for squid measurements we still denote it @xmath14 .
e. zeldov , a. i. larkin , v. b. geshkenbein , m. konczykowski , d. majer , b. khaykovich , v. m. vinokur , and h. shtrikman , phys .
lett . * 73 * , 1428 ( 1994 ) ; e. zeldov , a. i. larkin , m. konczykowski , b. khaykovich , d. majer , v. b. geshkenbein , and v. m. vinokur , physica c * 235 - 240 * , 2761 ( 1994 ) .
e. zeldov , d. majer , m. konczykowski , a. i. larkin , v. m. vinokur , v. b. geshkenbein , n. chikumoto , and h. shtrikman , europhys . lett . * 30 * , 367 ( 1995 ) ; c. d. dewhurst , d. a. cardwell , a. m. campbell , r. a. doyle , g. balakrishnan , and d. m. paul , phys .
b * 53 * , 14594 ( 1996 ) . | a series of yba@xmath0sr@xmath1cu@xmath2o@xmath3 single crystals was measured to study the influence of site disorder on the transition line @xmath4 between quasi - ordered vortex lattice and highly disordered vortex glass , as well as on the maximum critical current density within the glass phase , @xmath5 .
when 32% of ba is replaced by sr , @xmath5 is an order of magnitude higher than in the unsubstituted compound .
in contrast , the transition field @xmath6 first drops by a factor of about five with a substitution of just 10% of sr for ba , and then remains approximately constant for higher sr contents .
our results indicate that in very clean systems the order - disorder transition is affected very strongly by any crystallographic disorder , while above a certain threshold it is relatively robust with respect to additional disorder . in all substituted crystals @xmath6 monotonically decreases with an increase of temperature . |
by now , there is a rich ` phenomenology ' of strong - weak coupling duality constructions in supersymmetric yang - mills theory .
pure n=2 yang - mills theory has been analyzed , following the seminal work of seiberg and witten @xcite , for the groups @xmath6 @xcite , @xmath7 @xcite , and @xmath8 @xcite .
for each group , one finds a riemann surface @xmath0 , together with a preferred differential @xmath9 , such that the spectrum of bps - saturated states of electric and magnetic charges @xmath10 satisfies @xmath11 the moduli of the riemann surface are casimirs of the gauge group constructed out of the expectation value of the higgs field @xmath12 .
its degenerations , and the monodromy of @xmath13 , @xmath14 yield a great deal of information , enabling one to reconstruct the effective gauge pre - potential , @xmath15 .
the construction of the surface and preferred differential has been a somewhat _ ad hoc _ procedure , and one would like to find a uniform presentation for all gauge groups .
we will outline such a construction , in which the differential , @xmath9 , and the proper period integrals can be obtained in a canonical and universal manner .
the complete details will be postponed for another occasion .
it was noticed independently by one of the present authors , and by gorsky @xcite , that the @xmath6 curve appears as the spectral curve of an integrable system the periodic toda lattice .
the latter group went further , pointing out that the seiberg - witten differential appears naturally in the so - called _ whitham dynamics _ of adiabatic perturbations of integrable systems , and analyzing the case in detail .
we will show that , with a few twists , this framework of whitham dynamics for the toda lattice generates the effective pre - potential for all gauge groups .
the theory of integrable systems has become a rich arena for the interplay of dynamical systems , ( loop ) group theory , and algebraic geometry .
the existence of a complete set of integrals of the motion guarantees the existence of a canonical transformation to action - angle variables in which time development is straight - line motion on a torus . in the lax formulation
, one realizes the flows generated by these integrals as = [ _ i , ] , [ laxeqs ] with the dynamical variables appearing as parameters in the lax operator @xmath16 .
integrability is a consequence of regarding as the condition for a flat gauge potential , so that @xmath16 , @xmath17 are elements of the lie algebra @xmath18 of a lie group @xmath2 .
group theoretically , the phase space of the system is a coadjoint orbit of @xmath2 , typically reduced under the action of some subgroup ( frequently borel ) .
particularly interesting cases arise when @xmath2 is a loop group ; then @xmath16 , @xmath17 are laurent polynomials in the loop variable @xmath19 .
given a representation @xmath4 , the riemann surface @xmath5 defined by the characteristic polynomial = 0 [ curve ] is invariant under the flows .
the torus promised by complete integrability is a subspace of the jacobian @xmath20 @xcite .
the specific integrable system whose invariant curve(s ) arises in n=2 yang - mills theory is the periodic toda lattice .
each affine dynkin diagram @xcite defines a periodic toda lattice @xcite via lax operators @xmath21 + z\ee_{\alpha_0}+z^{-1}a_0\ee_{-\alpha_0}\nonumber\\ \b_k&=&{\rm tr}_{\sst l } ( \a^k \ip)\ . \label{todalax}\end{aligned}\ ] ] the notation is as follows : @xmath22 is the loop variable ; @xmath23 denotes the trace on the left factor of the @xmath24-matrix @xcite = _ \ { affine rootsystem } sign ( ) _ _ - .
the phase space variables are @xmath25 , @xmath26 ; a certain product of the @xmath25 is left invariant under the flow , so that the true dimension of the phase space is @xmath27 ) .
the @xmath28 , @xmath29 , are the simple root generators of @xmath18 ; the @xmath30 span the cartan subalgebra .
the affine root generator @xmath31 in is the highest ( long ) root for the usual untwisted kac - moody algebras @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , given by _ 0=_i=1^n_i_i with the integer coefficients @xmath39 for each simple root shown on the corresponding dynkin diagram in table 1 .
we will introduce the toda lattices for twisted kac - moody algebras in a moment .
one might wonder how much of this data used in the linearization of the hamiltonian flows is invariantly specified . in particular , the construction of the spectral curve @xmath5 requires the choice of a representation @xmath4 that appears neither in the original hamiltonian equations of motion , nor in the coadjoint orbit description of the phase space .
indeed , it is known @xcite that different choices of @xmath4 lead to different but equivalent descriptions of the dynamics , via an _ algebraic correspondence _ ( roughly , a one - to - many map in both directions ) between any given pair of choices .
this fact will be important to us below , for instance one can use it to relate the curves for @xmath40 and @xmath41 yang - mills that have appeared in the literature @xcite .
one of the special things about the toda lattice is that its coadjoint orbit is of minimal dimension @xmath42 ; thus the constants of motion are in one - to - one correspondence with the casimirs of @xmath18 .
the root system of @xmath18 respects a grading determined by the eigenvalue of the coxeter element of the weyl group of @xmath18 , or equivalently by the grading determined by the principal subalgebra whose cartan subalgebra generator is @xmath43 , where @xmath44 is the weyl vector ( @xmath45 )
. the equation will respect this grading ; giving weight one to @xmath26 and two to @xmath25 , the coefficient of @xmath46 in must be an invariant of the lie algebra of the appropriate grade , hence is a polynomial in the fundamental casimirs of @xmath18 .
the powers of @xmath47 will occur at grade equal to twice the coxeter number @xmath48 , multiplied by the trivial invariant @xmath49 alluded to above ; in the yang - mills theory , this parameter sets the quantum scale .
it is also useful to note that by conjugating @xmath50 , one can find a matrix @xmath51 that satisfies @xmath52 . from this
it follows that is invariant under @xmath53 .
the list of these curves for the classical groups with lax operators in the fundamental representation is : @xmath54 for @xmath6 and @xmath8 , the substitutions @xmath55 and @xmath56 , respectively , convert these curves to the forms presented in @xcite for @xmath6 and @xcite for @xmath8 ; however , we will find that the natural toda variable @xmath22 is more convenient .
we immediately see a problem if we are to relate the invariant curves of toda dynamics to those of n=2 yang - mills theory : the former have the affine root term @xmath57 appearing at a grade @xmath58 , whereas the latter have the instanton - generated term @xmath57 appearing at a grade @xmath59 .
.2 cm [ cols="<,^,^,^,^,^,^,^,^,^",options="header " , ] .2 cm for the simply - laced groups @xmath6 , @xmath8 , and @xmath60 , this distinction is irrelevant ; @xmath61 . however ,
for the toda curves of the nonsimply - laced groups @xmath7 , @xmath62 , @xmath63 and @xmath64 , the two differ : @xmath65 .
fortunately , for each nonsimply - laced affine dynkin diagram , there is a dual diagram with long roots @xmath66 short roots , and @xmath67 ; these are listed in table 2 .
the ade algebras are self - dual under this transformation .
these affine algebras may be obtained as the orbifold of an untwisted kac - moody algebra by the obvious symmetry of its ordinary dynkin diagram ; thus @xmath68 , @xmath69 , @xmath70 , and @xmath71 , where the superscript @xmath72 denotes the order of the diagram automorphism .
the ordinary lie algebra @xmath73 arising from the untwisted sector is the dual of @xmath18 under the operation roots @xmath66 coroots .
the affine root @xmath74 is the highest ( short ) root in the twisted sector _
0=_i=1^m_i _ i , where the coefficients @xmath75 may be read from the table . with these definitions ,
the toda curves for the dual groups are @xmath76x^4+\nonumber\\ & \qquad [ u_4
+ 2u_2(z+\mu / z)]x^2 = 0\nonumber\ .\end{aligned}\ ] ] the fact that the twisted kac - moody algebra respects a grading under the ` dual ' coxeter element @xcite guarantees that the ` instanton ' terms involving @xmath77 will have grade @xmath78=2\hdual$ ] . with the substitution @xmath79 in the @xmath80 curve ,
one readily finds the form obtained in @xcite . to summarize , to the montonen - olive dual of each simple lie algebra @xmath18 , we may associate a family of toda curves @xmath5 whose moduli are the casimirs of @xmath18 , and which respect the discrete @xmath24-symmetry of n=2 yang - mills
. we will now show how one may pick out a preferred rank(@xmath18)-dimensional subspace of the jacobian of @xmath5 on which the flow linearizes , and the correspondence between them for different @xmath4 .
the associated cycles on @xmath5 will be the special ones needed for the seiberg - witten construction .
to find the spectrum of bps - saturated states one must compute @xmath25 , @xmath81 from the periods of a seiberg - witten differential @xmath1 on the riemann surface @xmath5 .
the problem is that for all but the fundamental representation of @xmath6 , the genus of @xmath5 is much larger than rank(@xmath18 ) .
thus , apart from finding @xmath1 , one also needs to determine rank(@xmath18 ) preferred @xmath82- and @xmath83-cycles .
this issue has already been encountered in the study of orthogonal groups , where the preferred cycles can be isolated using the involution @xmath84 @xcite .
this method generalizes neither to higher representations nor to the exceptional groups .
however , the general solution to this problem can be obtained using toda theories .
the key is to realize that the coadjoint orbit , and not so much the surface or its jacobian , that contains the essential physics .
it is easy to see why this must be so in the toda context : there are only @xmath85 dynamical variables , and since the equations linearize on the jacobian , the hamiltonian flows select a complex rank(@xmath18 ) dimensional sub - variety of the jacobian of any given surface .
this sub - variety turns out to be a prym variety , is universal , and in particular is independent of the representation , @xmath4 . we believe that this jacobian sub - variety is the one that solves the seiberg - witten problem . the purpose in this section is to give a far more usable characterization of this special prym variety . for @xmath86
, the fundamental representation happens to give a curve of genus equal to rank(@xmath18 ) , and thus characterizes the special prym : it is the complete jacobian of this curve .
one can then find how this jacobian sits inside the jacobian of other curves by setting up a _
correspondence_. given two algebraic curves , @xmath87 and @xmath88 , defined by : f(x , z ) = 0 f(x^,z^ ) = 0 , [ corresp ] a correspondence between these curves is a multivalued holomorphic map given by ( one or more ) polynomial equations of the form @xmath89 , @xmath90 .
a correspondence can be used to map homology cycles and holomorphic differentials from one curve to another , and in particular maps the special prym of @xmath87 to the special prym of @xmath88 therefore the fundamental representation of @xmath86 generates the special prym in any representation .
this is discussed extensively in @xcite , where the example of @xmath40 in the @xmath91 and in the @xmath92 is discussed in detail . indeed , suppose that @xmath93 defines the curve for the fundamental of @xmath86 , and @xmath94 defines the curve for the two index anti - symmetric tensor of @xmath86 .
then a correspondence between these curves is given by z = z^ f(x - x^,z ) = 0 .
for @xmath95 this correspondence maps the jacobian of the fundamental curve of genus @xmath96 , to a three dimensional sub - variety of the the jacobian of the genus @xmath97 curve derived using the @xmath92 of @xmath40 .
one can check that the corresponding cycles are precisely those identified in @xcite as the solution of the seiberg - witten problem for @xmath41 , and that @xmath1 for one curve maps to @xmath1 for the other .
this method of correspondences is subsumed in a more general group theoretic procedure .
the first step is to view the riemann surface ( [ curve ] ) as a foliation over the riemann sphere for @xmath22 , and extend @xmath98 to an analytic function of @xmath22 .
the number of leaves is then equal to the dimension of the representation @xmath4 .
leaves are connected when @xmath22 takes a value at which two of the eigenvalues of @xmath99 become the same . since the determinant ( [ curve ] ) is invariant under @xmath100 , such values of @xmath22 occur in pairs related by @xmath53 . at a generic higgs vev @xmath101 , and for finite @xmath98 , the eigenvalues come together in pairs , and the result is a square - root branch cut running between the two values of @xmath22 at which this happens .
it is now convenient to consider the element @xmath102 of the cartan sub - algebra of @xmath18 that is conjugate to @xmath99 . by using the action of the weyl group
, we will take @xmath102 to be in the fundamental weyl chamber .
if @xmath103 are the weights of the representation @xmath4 , then the eigenvalues of @xmath99 , and hence the leaves of the foliation , are given by @xmath104 , @xmath105 . since
casimir invariants are equivalent to weyl invariant polynomials in the weights , it follows that in general ( [ curve ] ) will decompose into disconnected component riemann surfaces associated with the weyl orbits of the weights of @xmath4 .
so we will focus on one of these components .
the sheets of the foliation will generically come together when @xmath102 ceases to be a regular element of the lie algebra , that is , when it hits a wall of the fundamental weyl chamber . at generic higgs vevs in the fundamental weyl chamber
, this happens when @xmath106 , for some simple root @xmath107 .
there can be also be `` accidental '' singularities where two eigenvalues come together , and yet @xmath102 is regular . whether this happens or not depends upon the details of the specific example , and even if it happens , there is a method of constructing a cover of the riemann surface with accidental singularities such that the cover only has branch points when @xmath102 ceases to be regular @xcite .
we will therefore ignore accidental singularities , apart from a brief comment later on .
there are thus rank(@xmath18 ) branch cuts over the @xmath22-plane , each associated with a simple root of @xmath18 . above the cut corresponding to @xmath108 ,
leaves labelled by @xmath103 and @xmath109 will be joined together along the cut if and only if the weyl reflection generated by @xmath108 exchanges the the weights @xmath103 and @xmath109 .
this characterizes the foliation at finite values of @xmath98 .
there is another cut that runs from @xmath110 to @xmath111 which determines the structure at @xmath112 . taking the limit as @xmath113 in ( [ curve ] ) , and keeping the dominant terms
, one finds that the result has the coxeter symmetry of the dual affine lie algebra , and that the leaves in the foliation are connected under the action of this symmetry .
the whole point of this construction is that the resulting surface has a symmetry action of the weyl group acting upon it .
this can be used to induce a symmetry on the cycles , upon the holomorphic differentials and hence upon the jacobian variety . however , even though the leaves in the foliation are a weyl orbit , the weyl representation can be ( and in fact is ) reducible ; moreover , it generically contains the fundamental rank(@xmath18 ) dimensional reflection representation .
the preferred prym is obtained by passing to the part of the jacobian variety that corresponds to the reflection representation of the weyl group . in a moment
we will give an example , but we first wish to make the foregoing statement more precise by quoting the results of @xcite .
consider a principal @xmath2-bundle @xmath114 over an algebraic variety @xmath115 , together with a section @xmath12 of the adjoint bundle @xmath116 ( , a higgs vev ) can be _ any _ algebraic variety , leading to the notion of an _ algebraically completely integrable _ hamiltonian system .
] . given this data
, one can construct @xcite a kind of ` universal spectral cover ' @xmath117 with galois group @xmath118 , the weyl group of @xmath2 .
the points of @xmath117 over @xmath119 can typically be identified with chambers of the dual of the unique cartan subalgebra @xmath120 containing @xmath121 .
for instance , when @xmath2 is @xmath86 considered as a matrix in the fundamental representation , a point of the fiber is an ordering of the eigenvalues of @xmath121 .
the universal cover @xmath117 is independent of a choice of representation . given a representation @xmath122 , with highest weight @xmath123 , we can construct a surface @xmath124 as outlined above ; or we can construct a quotient space @xmath125 , where @xmath126 is the parabolic subgroup that fixes @xmath127 . if there are no accidental singularities then @xmath124 is isomorphic to @xmath128 , otherwise @xmath124 is birational to @xmath128 , with @xmath128 covering @xmath124 .
consider the group ring ( the regular representation ) of @xmath118 and its decomposition into irreps @xmath129
~_i _ i__i^ * . using @xmath118-equivariant maps from @xmath129 into @xmath130 , @xmath131 , one obtains decompositions of @xmath130 and @xmath132 @xmath133 here @xmath134 is the subspace of @xmath129 left fixed by @xmath135 .
so to any spectral curve @xmath128 , there is a natural decomposition of its prym variety under irreducible representations of the weyl group . in particular
, one has @xcite : * _ the prym variety @xmath136 corresponding to the reflection representation @xmath137 ( the weight space of @xmath2 ) occurs with positive multiplicity in @xmath138 for any proper subgroup @xmath139 . _ in other words
, we can always find the cartan torus inside the picard variety of the desingularization @xmath128 of @xmath124 .
kanev@xcite and donagi@xcite go on to construct a projection operator onto this preferred prym . finally , the integrable flows linearize on @xmath136 @xcite .
consider now the example of @xmath140 in the ten dimensional representation .
we present this example because it displays most of the features of the generic case ( and provides a useful warmup for @xmath141 in the @xmath142 ) .
the polynomial ( [ curve ] ) is quadratic in @xmath143 , and can be factored as @xmath144 one can proceed directly and think of the surface as a four - fold cover of the @xmath98-plane .
counting the branch points , a simple application of riemann - hurwitz shows that the surface has genus 11 . following the discussion above ,
it is more natural to think of the surface as a ten - sheeted cover over the @xmath22-plane , with four branch cuts of order @xmath145 ( _ i.e. _ square root cuts ) at finite @xmath98 , and one branch cut of order @xmath146 ( _ i.e. _ @xmath147 ) between @xmath110 and @xmath111 .
label the weights of the @xmath148 , and hence the leaves in the foliation , by vectors of the form @xmath149 ( see figure 1 . ) .
the coxeter element , @xmath150 , is the cyclic permutation of order @xmath97 , and collects the weights of the @xmath148 into two groups of @xmath97 .
this specifies the connections between the foliations at the cut between @xmath110 and @xmath111 . for each simple root @xmath108 , @xmath151
, there are exactly three pairs of weights that are interchanged by the weyl reflection @xmath152 , and corresponding leaves are connected in pairs , leading to @xmath153 pieces of interconnecting plumbing .
if one takes into account the connections from the cut @xmath110 and @xmath111 , the ten sheets connect into two disjoint two spheres , and the remaining @xmath153 connections , associated with the simple roots , convert this into a genus @xmath154 surface . the fact that we get the same answer viewing the surface as a foliation over the @xmath22-plane , and as a branched cover of the @xmath98-plane using indicates that there are no accidental singularities .
the resulting picture is shown in figure 1 .
it is relatively easy to find the reflection representation of the weyl group of @xmath140 inside the ten - dimensional representation of @xmath140 : one defines @xmath155 , where @xmath156 is the basis vector of the @xmath148 with weight @xmath123 . under the action of the weyl group on the @xmath156 , the vectors @xmath157
transform as if they were the vectors @xmath158 in the cartan sub - algebra indeed , the inner product matrix of the @xmath157 is a multiple of the cartan matrix .
one can easily check that combining leaves in the foliation in this manner is very closely related to action of the correspondence mapping ( [ corresp ] ) .
there is a natural method for identifying a preferred set of @xmath82-cycles .
define @xmath159 to be lifts of the cycles that run around the @xmath108-cuts in the @xmath22-plane . specifically , above each such cut in the @xmath22-plane there are three connections between leaves , about each of which we can draw a cycle .
( the orientation of the cycles in the lift may be fixed by choosing it to be anti - clockwise in the leaf whose weight has positive inner product with @xmath108 . ) the @xmath159 transform in the reflection representation of the weyl group , and thus define the preferred prym .
the associated @xmath83-cycles also arise from the @xmath22-plane ; the obvious guess works
one lifts a curve that runs around @xmath110 and one of the @xmath108 branch points .
that is , one can construct a curve @xmath160 that only passes through the plumbing at @xmath112 , and through the @xmath108 branch cuts ( for fixed @xmath161 ) , in such a manner that it has an intersection number @xmath162 with the three disjoint pieces of @xmath159 .
one can describe this cycle , @xmath160 , group theoretically .
an unreduced product of elements of the weyl group ( a `` word '' ) can be viewed as defining a path by starting on some leaf and then connecting to other leaves in a manner defined by the basic weyl reflections that make up the element .
consider @xmath163 , where @xmath152 is the simple reflection of the root @xmath108 and @xmath150 is the coxeter element ; use it to define a path which only makes connections through the @xmath108 cut , or through infinity . since the element @xmath164 has order @xmath146 , the word @xmath165 defines a closed loop .
the result is four closed cycles , @xmath160 , whose intersection numbers with the @xmath166 are @xmath167 .
( it is important to choose the proper coxeter element here . )
while we have only verified this construction in some examples , we believe that this will provide an explicit construction of the preferred prym in all generality .
having constructed a toda spectral curve with the symmetries of the n=2 yang - mills effective theory , and a special prym variety with the structure of yang - mills electric and magnetic weights , the next logical step is to look for natural objects in the toda theory which correspond to the seiberg - witten differential @xmath9 and the gauge pre - potential .
the answers are quite striking : first , @xmath9 is the action differential ` @xmath168 ' for a natural conjugate pair of toda variables , so that the integrals @xmath169 are action coordinates for toda ; second , the n=2 gauge pre - potential is the free energy @xmath170 of the ( dubrovin - krichever ) topological field theory of adiabatic deformations on the moduli space of toda dynamics !
the transformation of the toda evolution to action - angle variables proceeds in two steps .
first one defines an ` auxiliary spectrum ' the zeroes @xmath157 of a fixed eigenvector @xmath171 of the lax operator @xmath16 .
the analysis of @xcite shows that these zeroes travel along the special cycles of the curve , and hence provide a local system of coordinates on @xmath172 ( via the abel map ) .
the canonical conjugate momenta @xmath173 take the values @xmath174 $ ] on the classical motion .
the solution @xmath175 to @xmath176 , eq .
, defines the spectral curve @xmath5 ; this equation may then be interpreted as a kind of dispersion relation .
the orbit invariants ( action integrals ) are j_i=-_i^th a - cycledx_i .
[ orbit - invt ] in the second step , one derives the angle variables as @xmath177 from the action s(x , j)=^x dx , as well as the oscillation frequencies @xmath178 . the intermediate variables @xmath98 , @xmath22 are not merely useful in solving for the toda flows ; in the yang - mills theory the orbit invariants are precisely the @xmath25 that enter the bps mass formula ! it is easy to verify that the formulae for @xmath9 in the literature on the particular groups @xmath6@xcite , @xmath7@xcite , and @xmath8@xcite all take the form _
sw = xd(z)-(z)dx when expressed in toda variables .
hence we may identify _
swds_toda .
we should mention that much of the work on these conjugate variables is specific to @xmath6 toda ( the other classical groups are considered in @xcite ) ; however we are confident that the results are general . the final entry in our toda / yang - mills lexicon is the effective n=2 pre - potential @xmath179 .
its characteristic properties are @xmath180 where @xmath181 is the period matrix of the riemann surface @xmath5 . given an integrable hierarchy , krichever@xcite and dubrovin@xcite have constructed a topological field theory having precisely these properties ; this topological field theory describes the dynamics of the _ whitham - averaged _ hierarchy . starting with the periodic motion of the original integrable system
, one may ask whether it is possible to adiabatically perturb the integrals of motion ( the spectral curve @xmath5 ) in a consistent manner such that the slow perturbation of the integrals of motion is ` orthogonal ' to the fast orbital motion .
in other words , one wants to find a dynamics on the moduli space of the original system ( the procedure thus has strong analogies to the quantization of collective coordinates , and also to low energy monopole scattering @xcite ) .
this nonlinear version of the wkb method starts with an eigenfunction of the unperturbed problem _
0(t_)=_0(k_t_|u_1, ... ,u _ ) , where the @xmath182 are the moduli of the spectral curve ( the constants of the motion ) ; then one takes an ansatz for the perturbed solution ( t_)=(^-1s(t_)|u_i(t _ ) ) .
it is useful to keep in mind the usual linear wkb method , where @xmath183 $ ] is replaced by @xmath184 $ ] .
taking @xmath185 and @xmath186 , the ansatz satisfies the equations of motion to lowest order in @xmath187 .
expanding to next order gives a set of consistency conditions on the averaged densities @xmath188 of the conservation laws _ _
= _ _ , [ whitham - flows ] and implies @xmath189 ; , the averaged hierarchy is again an integrable hierarchy . dubrovin@xcite and krichever@xcite
show that the tau function for this hierarchy satisfies the axioms of topological field theory , and has the properties .
thus we may identify @xcite _
n=2 ym(_toda - whitham ) .
we have seen that there exists a uniform structure underlying the effective action of n=2 yang - mills theory , involving integrable systems and topological field theory .
we have not yet understood how to derive this structure from the underlying yang - mills dynamics .
the situation is rather reminiscent of the @xmath190 matter - gravity theories@xcite .
there one has an underlying field theoretic description involving liouville field theory coupled to the matter system ( at least near fixed points ) .
it is very difficult to compute in this theory .
however , if one is interested in global questions such as integrated correlations of order parameters , there exists an elegant solution in the form of @xmath191-functions , the kp hierarchy , and topological field theory@xcite .
the connection between this construction and the field - theoretic formulation is quite indirect .
answers to ` local ' ( nevertheless gauge invariant ) questions , such as the ising correlation function at fixed geodesic distance or the partition function at fixed surface moduli , are harder to come by ; they are not encompassed by the integrable - topological formalism and the only tool at present is to resort to the field theoretic description .
on the other hand , this zero - momentum sector contains a great deal of information about the phase structure of the theory .
a great deal is known about the singularities of toda dynamics @xcite which might be adapted to the study of degenerations in n=2 yang - mills @xcite .
one may contemplate a much broader applicability of the above ideas .
curves and seiberg - witten differentials for theory with fundamental and adjoint matter@xcite , and for @xmath86 with fundamental matter@xcite , have been proposed .
they have natural forms when expressed in toda - style variables . due to the discrete @xmath24-symmetry , the effect on the spectral curve of adding matter is to reduce the degree of the @xmath57 term from @xmath192 to @xmath193 $ ] , with the deficit in homogeneity weight made up in these examples by a polynomial @xmath194 of degree @xmath195 .
the obvious modification of the toda lax operator to try is to lower the grade of the highest weight element @xmath196 .
this will simultaneously increase the size of the coadjoint orbit , so that the moduli of the curve may include the @xmath75 .
the appearance of kac - moody algebras in the solution of n=2 gauge theories is a profound mystery .
the fact that the dual @xmath3 of the gauge group @xmath2 appears suggests a connection to monopole dynamics .
the appearance of the spectral parameter ( loop coordinate ) @xmath22 begs a physical explanation .
this sort of structure has been encountered in @xmath197 supersymmetric models in two dimensions , where the dependence of certain two - point functions upon couplings and scales , and even two - dimensional instanton corrections , is determined by some classical integrable equations ( the @xmath198 equations ) @xcite . following similar lines of thought ,
one is naturally led to consider the four - dimensional version of the chiral ring , and its action on the bundle of yang - mills vacua over the base of higgs vevs .
one finds many close parallels , and some significant differences , between the two and four - dimensional structures , and this approach is still under investigation .
the fact that there is a duality between two and four dimensional instanton corrections in string theory @xcite strongly suggests that there should indeed be a four dimensional version of @xmath199 .
the appearance of affine lie algebras also suggests that the level - rank dualities of wzw models @xcite might ` explain ' n=1 yang - mills dualities . in topological field theories closely related to ours @xcite , a useful set of coordinates localizes the dynamics on the critical points ( branch points ) of the curve ; this may be the convenient way to describe soft breaking to n=1 @xcite .
the soft breaking terms could be added to the whitham dynamics as a secular perturbation of the flow equations @xcite .
one might also imagine extending the scope of our construction to string theory .
kachru @xcite have found the spectral curve of rigid n=2 yang - mills theory as a limit in the calabi - yau moduli space of some examples , but not others .
the role of the periods of the seiberg - witten differential is played by periods of a three - form on the calabi - yau .
this is not inconsistent with our results ; the invariant specification of toda as coadjoint orbit dynamics on a loop group makes no reference to riemann surfaces .
notice also that the notions of spectral cover and analytically completely integrable hamiltonian system are quite general , and not at all restricted to @xmath115 being a riemann surface ; it is not out of the question that one could construct an integrable system based on calabi - yau manifolds that ` explains ' the results of kachru @xcite , and that includes the effects of coupling to supergravity .
donagi and markman @xcite have built an integrable system canonically associated to any calabi - yau threefold , and conjecture that it is related to mirror symmetry ( and , one might expect , all of t - duality ) .
this calabi - yau integrable system bears many similarities to its riemann surface counterpart employed above .
indeed , could there not be a grand integrable system that governs all the miraculous stringy dualities that have emerged ?
* acknowledgements : * we are grateful to k. intriligator for discussions of supersymmetric gauge theory , and especially i. krichever for explanations of his work .
e.m . would like to thank u. of paris vi and cern , and both authors thank the aspen center for physics , for their hospitality during the course of this work . e.m and n.w .
are supported in part by funds provided by the doe under grant nos .
de - fg02 - 90er-40560 and de - fg03 - 84er-40168 , respectively .
n. seiberg and e. witten , _ electric - magnetic duality , monopole condensation , and confinement in n=2 supersymmetric yang - mills theory _ , nucl .
b426:19 - 52,1994 , erratum - ibid .
b430:485 - 486,1994 .
e - print archive : hep - th/9407087 .
krichever , _ methods of algebraic geometry in the theory of nonlinear equations _ , russ .
surveys 32:185 - 213(1977);p .
van moerbeke and d. mumford , _ the spectrum of difference operators and algebraic curves _
, acta math .
143:93 - 154,1979 .
m. adler and p.
van moerbeke , _ completely integrable systems , euclidean lie algebras , and curves _ , adv . math .
38:267 - 317,1980 ; _ linearization of hamiltonian systems , jacobi varieties , and representation theory _ , adv .
38:318 - 379,1980 .
krichever , _ whitham theory for integrable systems and topological quantum field theories _ , in _ new symmetry principles in quantum field theory _ , j. frohlich ( eds . ) , nato asi series b , vol .
295 , plenum press , ny ( 1992 ) . for reviews , see p. di francesco , p. ginsparg and j. zinn - justin , _ 2-d gravity and random matrices _ , phys .
254:1 - 133,1995 .
e - print archive : hep - th/9306153 ; + p. ginsparg and g. moore , _ lectures on 2d gravity and 2d string theory _ , lectures given at tasi summer school , boulder co , june 11 - 19 , 1992 .
published in boulder tasi 92:277 - 470 e - print archive : hep - th/9304011 ; + r. dijkgraaf , _ intersection theory , integrable hierarchies and topological field theory _ , in _ new symmetry principles in quantum field theory _ , j. frohlich ( eds . ) , nato asi series b , vol .
295 , plenum press , ny ( 1992 ) .
e - print archive : hep - th/9201003 .
m. adler and p. van moerbeke , _ the toda lattice , dynkin diagrams , singularities and abelian varieties _ ,
inv . math .
103:223 - 278,1991;m .
adler , l. haine , and p. van moerbeke , _ limit matrices for the toda flow and periodic flags for loop groups _ , math . ann .
296:1 - 33,1993 .
s. cecotti and c. vafa , _ topological antitopological fusion _ , nucl .
b367:359 - 461,1991 ; _ exact results for supersymmetric sigma models _ ,
68:903 - 906,1992 .
e - print archive : hep - th/9111016 ; _ on classification of n=2 supersymmetric theories _ , comm . math . phys .
158:569 - 644,1993 .
e - print archive : hep - th/9211097 . s.g .
naculich and h.j .
schnitzer _ duality relations between @xmath200 and @xmath201 wzw models and their braid matrices _ , phys .
b244:235 - 240,1990 ; _ duality between @xmath200 and @xmath201 wzw models _ ,
b347:687 - 742,1990;s.g .
naculich , h.a .
riggs and h.j .
schnitzer , _ group level duality in wzw models and chern - simons theory _ , phys .
b246:417 - 422,1990;e.j .
mlawer , s.g .
naculich , h.a .
riggs and h.j .
schnitzer _ group level duality of wzw fusion coefficients and chern - simons link observables _
b352:863 - 896,1991;d .
altschuler , m. bauer and h. saleur , _ level rank duality in nonunitary coset theories _ , j. phys .
a23:l789-l794,1990;h .
saleur and d. altschuler , _ level rank duality in quantum groups _
b354:579 - 613,1991;t .
nakanishi and a. tsuchiya , _ level - rank duality of wzw models in conformal field theory _ , comm .
144:351 - 372,1992 .
r. donagi and e. markman , _
cubics , integrable systems , and calabi - yau threefolds _ , to appear in the proceedings of the algebraic geometry workshop on the occasion of the 65th birthday of f. hirzebruch , may 1993 ; e - print archive : alg - geom/9408004 . | after the work of seiberg and witten , it has been seen that the dynamics of n=2 yang - mills theory is governed by a riemann surface @xmath0 . in particular
, the integral of a special differential @xmath1 over ( a subset of ) the periods of @xmath0 gives the mass formula for bps - saturated states .
we show that , for each simple group @xmath2 , the riemann surface is a spectral curve of the periodic toda lattice for the dual group , @xmath3 , whose affine dynkin diagram is the dual of that of @xmath2 .
this curve is not unique , rather it depends on the choice of a representation @xmath4 of @xmath3 ; however , different choices of @xmath4 lead to equivalent constructions .
the seiberg - witten differential @xmath1 is naturally expressed in toda variables , and the n=2 yang - mills pre - potential is the free energy of a topological field theory defined by the data @xmath5 and @xmath1 .
epsf * * integrable systems and supersymmetric gauge theory * * * emil j. martinec * .4 cm _ enrico fermi institute and dept . of physics _ _ university of chicago , 5640 s. ellis ave . ,
chicago , il 60637 usa _ .4 cm _ and _ .4 cm * nicholas p. warner * .4 cm _ physics department , university of southern california _
_ university park , los angeles , ca 900890484 usa _ # 1([#1 ] ) *v * # 1#1 # 1#2#1#2 # 1#2 # 1|#1 # 1#1| # 1#1
= 11 /#1 a # 1#2 = 12 .5 cm _ dedicated to the memory of claude itzykson , recalling his gift _ _ for combining elegant physics and beautiful mathematics _ |
current models of cosmic structure formation based on cdm scenarios predict that the first collapsed , luminous ( hereafter pop iii ) objects should form at redshift @xmath2 and have a total mass @xmath3 or baryonic mass @xmath4 ( couchman & rees 1986 , haiman 1997 [ hrl ] , tegmark 1997 ) .
this conclusion is reached by requiring that the cooling time , @xmath5 , of the gas is shorter than the hubble time , @xmath6 , at the formation epoch . in a plasma of primordial composition the only efficient coolant in the temperature range @xmath7
k , the typical virial temperature of pop iii dark matter halos , is represented by h@xmath8 molecules whose abundance increases from its initial post - recombination relic value to higher values during the halo collapse phase .
it is therefore crucial to determine the cosmic evolution of such species in the early universe to clarify if small structures can continue to collapse according to the postulated hierarchical structure growth or if , lacking a cooling source , the mass build - up sequence comes to a temporary halt .
the appearance of pop iii objects is now thought to cause a partial destruction of the available molecular hydrogen either in the intergalactic medium ( igm ) and/or in collapsing structures ; the result is a negative feedback on galaxy formation .
this effect has been pointed out by hrl and it works as follows .
as stars form in the very first generation of objects , the emitted photons in the energy band 11.2 - 13.6 ev are able to penetrate the gas and photodissociate h@xmath8 molecules both in the igm and in the nearest collapsing structures , if they can propagate that far from their source .
this negative feedback and its possible limitations are discussed by ciardi , abel & ferrara ( 1998 , caf ) . here
we propose and investigate instead a possible _ positive _ feedback based on supernova ( sn ) explosions , which , under many aspects , is reminiscent of a scaled version of the explosive galaxy formation scenario introduced by ostriker & cowie ( 1981 ) and put forward by many others .
pop iiis are very fragile due to their low mass and shallow gravitational potential : only a few sne are sufficient to blow - away ( ferrara 1998 ) their baryonic content and drive an expanding blastwave into the igm , which eventually becomes radiative and allows the swept gas to cool in a dense shell .
the cooling transient , as we will see , is characterized by a strong nonequilibrium condition in which recombination lags behind the temperature decrease . as already pointed out by shapiro & kang ( 1987 ) and kang & shapiro ( 1992 )
, this is a favorable condition for h@xmath8 formation .
our conclusions are that the amount of molecular hydrogen thus formed can exceed the one destroyed via photodissociation , yielding a net increase of the h@xmath8 in the universe . as a consequence , not only the galaxy formation process is not halted , but instead it is favored by the effects of pop iii formation .
2 describes the main properties of multi - sn explosions in the early universe ; sec .
3 is devoted to the calculation of h@xmath8 formation in their cooling shell . in sec .
4 we compare the magnitude of positive and negative feedbacks ; some discussion of the results is given in sec .
the mechanical luminosity of an ob association in a pop iii object at redshift @xmath9 can be written as ( ciardi & ferrara 1997 , cf ) @xmath10 where @xmath11 and @xmath12 ergs is the energy of a sn explosion ; we assume a salpeter imf , according to which one supernova is produced for each 56 @xmath13 of stars formed .
the baryon density parameter is @xmath14 , of which a fraction @xmath15 ( abel 1997b ) is able to cool and become available to form stars .
the dark matter halo mass and density are @xmath16 and @xmath17 $ ] g @xmath18 , respectively ( we assume a flat eds universe @xmath19 and @xmath20 km s@xmath21 mpc@xmath21 ) ; the corresponding free - fall time is @xmath22 ; @xmath23 is the star formation efficiency , calibrated on the milky way .
the correlated sn explosions drive a blastwave in the surrounding gas , which eventually propagates into the igm ; due to their low mass , the effects of pop iii ism on the blastwave expansion can be neglected ( maclow & ferrara 1998 ) . the evolution of the shock radius , @xmath24 , can be obtained in the thin shell approximation by solving numerically eqs .
( 5)-(6 ) of cf .
for sake of simplicity we use the following ( sedov ) analytical solution of the same equations which holds for the adiabatic case when the external confining pressure and the gravitational pull of the dm halo are neglected : @xmath25 where @xmath26 , @xmath27 yr is the average lifetime of massive stars , @xmath28 is the total energy of the explosion , @xmath29 the igm density at redshift @xmath9 .
[ rs ] does not take into account the effects of hubble expansion , which are negligible since @xmath30 ( see eq .
[ rsc ] ) .
the corresponding temperature of the postshock gas is @xmath31 k. such hot gas will not be able to cool until its cooling time , @xmath5 , becomes smaller than the hubble time , @xmath6 .
the energy losses , @xmath32 , due to gas - radiative processes in a primordial gas are of order @xmath33 ergs @xmath34 s@xmath21 in the range @xmath35 k ( schmutzler & tscharnuter 1993 ) ; the other relevant cooling agent is inverse compton cooling off cmb photons ( ikeuchi & ostriker 1986 ) . comparing the two rates , it is easily seen that inverse compton dominates the cooling for @xmath36 , for @xmath37 k. the condition @xmath38 requires that @xmath39 ; shocked gas produced at epochs earlier than this will be able to cool and condense in a dense shell behind the shock front .
note that as the temperature drops to lower values with subsequent gas recombination , cooling via hydrogen ly@xmath40 line excitation will start to become important . at @xmath41 gyr ,
the shock radius and postshock gas temperature are respectively @xmath42 @xmath43 these relations hold if the mach number of the shock with respect to the igm sound speed is @xmath44 , a condition verified if the igm has not yet been reheated .
in the following , we calculate the time - dependent , nonequilibrium ionization and molecule abundance evolution behind the multi - sn shock as a function of the object mass and epoch of the explosion .
the postshock cooling gas is assumed to be initially at temperature @xmath45 and density @xmath46 . in hydrogen of density @xmath47
( we neglect he ) , the evolution of the ionization fraction , @xmath48 , molecular fraction , @xmath49 , and temperature , @xmath50 , is determined by the following rate equations : @xmath51 @xmath52 n f\ ] ] @xmath53.\ ] ] the various rate coefficients @xmath54 are labelled according to the nomenclature given in abel 1997a ( appendix a ) ; @xmath55 is the @xmath56 species number density ; @xmath57 is the specific heat ratio and @xmath58 is the gas pressure . in addition to the standard atomic cooling processes ,
@xmath32 includes compton and @xmath0 cooling ; the latter has been adopted from martin ( 1996 ) .
we assume that no heating is provided to the gas after it has been heated by the shock to the initial temperature and neglect adiabatic cooling due to hubble expansion for reasons similar to those given in the previous section .
the first equation describes the usual collisional ionization ( rate @xmath59 ) recombination ( @xmath60 ) balance .
the equation for @xmath61 is slightly more complicated and we describe it shortly . at the low densities considered here , in the absence of dust ,
@xmath0 is formed in the gas phase mainly via the channel @xmath62 , at rate @xmath63 ( formation via the @xmath64 channel , when included , is found to be negligible in our case ) ; it is then destroyed by one of the following mechanisms : ( i ) it is collisionally ionized by proton impacts ( @xmath65 ) ; ( ii ) it is dissociated by collisions with electrons ( @xmath66 ) and/or h - atoms ( @xmath67 ) .
abel 1997a have shown that using the equilibrium value for @xmath55 is a very good approximation since the reactions determining the @xmath56 abundance occur on a much shorter timescale than those responsible for the @xmath0 chemistry ; therefore we have calculated @xmath55 from their eq . (
also , we note that in the above form , the rate equations correspond to the so - called minimal model described in that paper .
the energy equation has been solved in the isobaric limit ; this approximation is known to be appropriate for the evolution of galactic supernova remnants ( cioffi 1988 ) , where most of the radiative phase is spent at constant pressure ; the same result was found to hold by shapiro & kang ( 1987 ) and kang & shapiro ( 1992 ) for a wide range of shock velocities in a cosmological context .
the numerical solution of the above rate equations are shown in fig
. 1 for two different object masses ( @xmath68 ) and initial redshift ( @xmath69 ) . almost independently on the specific values of these parameters a high @xmath0 fraction ( @xmath70 ) is produced by redshift ten , with a steep increase occurring when the gas has cooled down at @xmath71 k. @xmath0 production is faster in high-@xmath9 , high @xmath72 shocks , but the peak abundance is practically the same in all cases ; the final gas temperature is in the range @xmath73 k. we have also experimented a different @xmath0 cooling function ( hollenbach & mckee 1979 ) : this produces a cooler final state ( @xmath74 k ) but only a marginally different asymptotic value @xmath75 . for comparison ,
we have also plotted in fig .
1 the tegmark 1997 `` rule of thumb '' value @xmath76 at which @xmath0 cooling becomes important .
the results obtained in sec . 3 show that typically at least a @xmath0 fraction @xmath77 is formed after explosive events leading to the blow - away of pop iiis . as discussed in the introduction
, hrl have argued that the same objects could suppress the surrounding @xmath0 abundance due to their uv photodissociating radiation . in order to evaluate the impact of pop iiis on the subsequent galaxy formation , largely regulated by the availability of @xmath0 in this mass ( and redshift ) range
, it is useful to compare the @xmath0 mass production vs. destruction .
a lower limit to the amount of @xmath0 produced in an explosion is readily found to be equal to @xmath78 or @xmath79 as shown by caf , the uv / ionizing radiation from pop iii s massive stars previous to blow - away will produce both a hii region and a region of photodissociated intergalactic @xmath0 ( radius @xmath80 ) in which the object is embedded .
the radius @xmath80 can be defined by requiring that the photodissociation timescale ( @xmath81 , notation as in eqs .
[ xrate]-[frate ] ) is shorter than @xmath6 .
this condition yields the definition @xmath82 , where @xmath83 is the uv photon flux in the @xmath0 lyman - werner ( lw ) bands ( 11.213.6 ev ) , assumed here to be proportional to the flux of lyc photons , @xmath84 , just before the massive stars explode .
the value of the constant @xmath85 depends somewhat on the imf and on the evolutionary stage of the stellar cluster , but its value should be reasonably close to unity . paralleling eq .
[ l ] , we estimate @xmath84 to be @xmath86 where @xmath87 is the uv photon production per collapsed proton efficiency ( tegmark 1994 ) and @xmath88 is the escape fraction of such photons ( absorption can be caused both by interstellar neutral h and/or dust ) .
we have checked that this simple estimate is within a factor of 2 of the value obtained from the recently revised version of the bruzual & charlot ( 1993 ) spectrophotometric code .
the adopted reference value @xmath89 is an upper limit derived from observational ( leitherer 1995 , hurwitz 1997 ) and theoretical ( dove & shull 1994 ) studies .
it follows that @xmath90 where @xmath91 .
thus , the ratio between the @xmath0 mass produced and destroyed by pop iiis is @xmath92 ^ 3.\ ] ] the post - recombination relic fraction of intergalactic @xmath0 is estimated to be @xmath93 ( palla 1995 ; anninos & norman 1996 ; lepp , private communication ) . the previous relation is graphically displayed in fig .
2 , along with the values of @xmath94 and @xmath80 . from that plot
we see that objects of total mass @xmath95 produce more @xmath0 than they destroy for @xmath96 ; larger objects ( @xmath97 ) provide a similar positive feedback only for @xmath98 , since they are characterized by a higher @xmath99 ratio .
however , since in a hierarchical model larger masses form later , even for these objects the overall effect should be a net @xmath0 production .
the results obtained in the previous section could be modified by the fact that so far we have neglected the effects of the pop iii stellar cluster radiation impinging on the cooled shell .
in fact , even after all the sne have exploded blowing - away the gas , the coeval low mass stars will continue to produce some residual flux in the energy range @xmath100 ev , relevant to the @xmath0 formation network .
two processes might have some effect in this context : ( i ) @xmath56 photoionization , and ( ii ) @xmath0 two - step photodissociation .
the first process occurs at a rate @xmath101 s@xmath21 per @xmath56 atom , or @xmath102 s@xmath21 using eqs .
[ rsc ] , [ s ] and reference values for the remaining parameters .
the most effective @xmath103 destruction process among the ones included is @xmath0 formation , whose rate in the same case is @xmath104 s@xmath21 .
thus , photoionization is negligible , and we have checked this conclusion numerically .
we can calculate the effects of two - step photodissociation on the shell as follows .
the shell @xmath0 column density is @xmath105 @xmath106 ( we have assumed that all the swept igm mass is in the cool shell ) .
since the critical column density for self - shielding is @xmath107 @xmath106 ( draine & bertoldi 1997 ) , the shell is optically thick to dissociating radiation .
nevertheless , the flux in the lw bands , @xmath108 , where @xmath109 is the planck constant , will induce a dissociation front in the shell ; @xmath110 is very likely dominating on a possible background radiation .
the propagation speed will be @xmath111 , where @xmath112 is the lw photon mean free path , and @xmath113 is a constant obtained by comparison with numerical results .
thus , the time required to dissociate the entire shell is @xmath114 yr , where @xmath115 , and @xmath116 . since theoretical work ( bruzual & charlot 1993 ) suggests that @xmath85 drops to @xmath117 after all the massive stars ( the main contributors to the uv flux ) have died , dissociating the entire shell will require a very long time , certainly larger than the hubble time .
finally , kang & shapiro ( 1992 ) pointed out that any external _ ionizing _ radiation field ( which could be provided by the hot postshock gas and/or by the residual intermediate mass stars in the cluster ) tends to increase the final value of @xmath61 , although introducing a time delay due to a temperature plateau in the evolution characteristic of photoionization heating .
thus , our estimate seems to provide a robust lower limit to the molecular hydrogen abundance . how can this @xmath0 enhanced gas be used for galaxy formation ?
first , regions in which pop iiis are clustered enough for their shells to interact might become sites of active star / galaxy formation .
shell interactions appear to be necessary to make them gravitational unstable , since the temperature required for isolated fragmentation ( couchman & rees 1986 ) is well below the values ( @xmath118 k ) found here .
next , these intergalactic shells can be accreted by neighbor objects , possibly increasing their @xmath0 abundance above the threshold @xmath119 required for efficient cooling ( tegmark 1997 ) .
we conclude that pop iii objects can produce regions of considerably high molecular hydrogen abundance due to multi - sn shocks propagating in the igm .
we have also seen that the @xmath0 thus produced can exceed the amount of relic @xmath0 destroyed inside the photodissociation region surrounding a given pop iii .
this occurrence suggests that these first objects might have a positive feedback on subsequent galactic formation .
such conclusion is certainly valid on a local scale , as defined by the radii of the two influence spheres , @xmath24 and @xmath80 , of a single object .
however , our definition of @xmath80 ( see sec .
4 ) does not exclude that some photodissociating flux is present outside @xmath80 .
if many of these objects are present in the universe at the same epoch , they might contribute to an early soft - uv background which , illuminating an isolated , collapsing cloud , could photodissociate and depress its @xmath0 content ; this possibility has been discussed by hrl . however , caf have demonstrated that , once the _ surviving relic igm fraction between photodissociated spheres _ is taken into account , the universe opacity to lw photons before reionization becomes of order unity on scales larger than @xmath80 , but smaller than the typical interdistance between pop iiis at @xmath120 .
this implies that the soft - uv background is very weak due to intergalactic absorption , and , as a consequence , the proposed negative feedback considerably inhibited .
furthermore , most of the dissociating flux is absorbed by the thick @xmath0 shells discussed above .
given that in general @xmath121 , where @xmath122 is the hii region radius of a pop iii ( caf ) , clearing of intergalactic @xmath0 by photodissociated sphere overlapping will occur before complete reionization .
it is not clear at this stage if that event occurs before the typical mass scale of the collapsing objects is such that cooling can proceed via ly@xmath40 only , thus considerably weakening the arguments in favor of a temporary halt of galaxy formation at high redshift .
2truecm we acknowledge t. abel , b. ciardi , d. cox and g. field for discussions ; e. corbelli , z. haiman , a. loeb , f. palla , o. pantano , p. pietrini , and y. shchekinov for comments ; thanks are due to l. pozzetti for help with the spectrophotometric code . | we study the formation of molecular hydrogen in cooling gas behind shocks produced during the blow - away process thought to occur in the first collapsed , luminous ( pop iii ) objects in the early universe .
we find that for a wide range of physical parameters the @xmath0 fraction is @xmath1 .
the @xmath0 mass produced in such explosions can exceed the amount of relic @xmath0 destroyed inside the photodissociation region surrounding a given pop iii .
we conclude that these first objects , differently from the suggestion of haiman 1997 , might have a net _ positive _ feedback on subsequent galactic formation .
we discuss the effects of radiation and the implications of our results for the soft - uv background . |
understanding active galactic nucleus ( agn)galaxy interaction and feedback is of great consequence for the study of both galaxy and agn evolution ( e.g. , silk & rees 1998 ; kauffmann et al .
2003 ; hopkins et al . 2006 ) .
the intense ionizing radiation , relativistic jets , and winds from the agns interact with the interstellar medium ( ism ) of their host galaxies , and may either stimulate or suppress star formation ( e.g. , sazonov , ostriker & sunyaev 2004 ) .
visible signatures of direct agn - host interaction include kpc - scale regions with [ oiii ] and h@xmath4 line emission , known as the extended narrow line region ( enlr ) , which have been found in many nearby seyfert galaxies ( schmitt et al .
2003 ; veilleux et al .
soft x - ray emission is seen associated with the enlr , providing an opportunity to obtain x - ray diagnostic of the physical properties of the interacting ism ( e.g. , young et al .
2001 ; yang et al . 2001
; ogle et al .
2000 , 2003 ; bianchi et al . 2006 ; evans et al . 2006 ; kraemer , schmitt & crenshaw 2008 ) . another important link between agns and starbursts is that molecular gas accreted to the nuclear region may induce a starburst ( elmegreen 1994 ) , feed a central agn , or both .
agn - related shocks may further stimulate star formation ( e.g. , scoville 1992 ; gonzlez delgado et al .
1998 ; veilleux et al . 1995 ; veilleux 2001 ) .
star formation in the disks and galactic winds in hot halos of spiral galaxies can be traced by the diffuse x - ray emission as well ( e.g. , tyler et al . 2004 , strickland et al .
2004 ; swartz et al . 2006 ; warwick et al .
ngc 1365 , the object studied in this paper , is ideal for these investigations , since it hosts an agn and also has ir emission coincident with active , intensive nuclear star formation .
ngc 1365 is a nearby ( @xmath5 mpc , @xmath6 pc ; madore et al .
1998 ; silbermann et al .
1999 ) archetype barred spiral galaxy ( sbb(s)i ; sandage & tammann 1981 ) , and likely a member of the fornax cluster ( jones & jones 1980 ) .
a thorough review of the ngc 1365 galaxy is given in lindblad ( 1999 ) .
it hosts vigorous star formation in the circumnuclear region and a variably obscured seyfert 1.5 nucleus ( veron et al .
1980 ; hjelm & lindblad 1996 ; risaliti et al . 2007 ) .
the spectacular symmetric main spiral arms extending from the ends of a strong bar are apparent in the _
xmm_-newton optical monitor ( om ) images ( figure [ om ] ) .
dark dust lanes run across the nuclear region and partially obscure the nucleus .
ngc 1365 has long been known to exhibit a biconical outflow ( burbidge & burbidge 1960 ; phillips et al .
1983 ; storchi - bergmann & bonatto 1991 ) . a @xmath7 long ( @xmath0450 pc )
100@xmath8 wide conical [ oiii]@xmath2 emission line region ( elr ) is present to the southeast ( se ) and to the northwest ( nw ) of the nucleus ( hjelm & lindblad 1996 ; kristen et al .
1997 ; veilleux et al .
2003 ) , which suggests an agn ionization cone , although a starburst driven outflow explanation has also been advanced ( komossa & schulz 1998 ) . in the hubble space telescope ( _ hst _ )
images , the inner @xmath9 of the conical outflow seen in [ oiii ] is resolved into a number of small clouds and larger agglomerations ( kristen et al .
a high excitation [ nii]@xmath10/h@xmath4 region is aligned with the [ oiii ] structure , but appears to be rectangular ( veilleux et al .
sandqvist et al .
( 1995 ) show that the inner parts of the [ oiii ] region contain a nuclear radio jet , extending @xmath11 se along the galaxy minor axis .
note that the spatial properties of the ngc 1365 galaxy and the outflow cones are well determined from these multiwavelength studies ( see lindblad 1999 review and references therein ) : the position angle ( pa ) of the line of nodes is @xmath12 , and the inclination angle of the galaxy is @xmath13 ; the [ oiii ] cone extends from the nucleus out of the galactic plane with its symmetry axis closely aligned with the rotation axis of the galaxy , with a full opening angle of about 100@xmath8 ; the counter cone to the nw is partially obscured by the absorbing dust in the galaxy plane .
another prominent characteristic of ngc 1365 is the circumnuclear star - forming ring with a diameter of @xmath14 ( @xmath15 kpc ; kristen et al .
the kpc - size nuclear vicinity contains bright optical `` hot spots '' ( sersic & pastoriza 1965 ; clearly visible in the uv images shown in figure [ om ] ) .
the circumnuclear regions are resolved into many compact super star clusters ( kristen et al .
1997 ) , forming an elongated ring surrounding the agn .
its minor axis is parallel to both the bi - cone axis and the galaxy minor axis , suggesting the possibility of a collimating torus ( sandqvist 1999 ) .
high resolution mid - infrared imaging of the ring ( galliano et al . 2005 ) also unveils a circumnuclear population of point - like sources coincident with bright co spots ( sakamoto et al .
2007 ) and non - thermal radio continuum sources ( sandqvist et al . 1995
; forbes & norris 1998 ) , which are interpreted as embedded young massive star clusters . the biconical elr / outflow and
the circumnuclear ring should be readily detectable in the _ chandra _ images as hinted from the spatially extended x - ray emission in the rosat hri image ( stevens et al .
1999 ) .
in this paper we present a high spatial resolution x - ray imaging study of the complex circumnuclear region of the active galaxy ngc 1365 with the _ chandra _ x - ray observatory .
ngc 1365 has received a great deal of attention as a result of x - ray observing campaigns to monitor the obscuration of the agn ( e.g. , risaliti et al .
2005a , b ; risaliti et al .
comparing to observations of ngc 1365 from previous generation x - ray satellites and the _ xmm_-newton observatory , the _ chandra _ images offer an unprecedented view of the innermost regions of ngc 1365 , which is critical to disentangle the different emission components .
ngc 1365 was first imaged in 2002 with the back - illuminated chip of the advanced ccd imaging spectrometer spectroscopy array ( acis - s ; garmire et al .
2003 ) positioned at the focal point of _ chandra x - ray observatory _ high resolution mirror assembly ( van speybroeck et al . 1997 ; weisskopf et al .
it was further monitored with six _ chandra_/acis - s observations in `` 1/4 window '' mode during april 2006 , resulting in a total exposure time of @xmath0100 ks ( see table [ obslog ] for details ; the overlap region common to these observations is outlined in figure [ om ] ) .
the data products were analyzed with the _ chandra _ x - ray center ( cxc ) ciao v4.0 software and heasoft v6.4 package .
the new release of calibration files caldb v3.4.3 were used .
the data were processed by the cxc ( ascds version 7.6.7 - 7.6.9 ) ; verification and validation of the data products showed no anomalies . following the standard _
acis data preparation thread , we reprocessed all level 1 data to create new level 2 events for consistency . a preliminary run of ciao tool _ wavdetect _ ( freeman et al . 2002 )
is done for point source detection in the central @xmath16 region of each observation . a @xmath17 radius circle centered at the s3 aimpoint
was used to query the naval observatory merged astrometric dataset ( nomad ) catalogue ( zacharias et al . 2004 ) and the extracted optical sources are searched for counterparts to _ chandra _ detections .
using optical point sources in the field detected in x - rays , the absolute astrometry for observation i d ( obsid ) 3554 is measured to be accurate to within @xmath18 . by comparing the positions of the bright point like x - ray sources in common between observations
, we then corrected the absolute astrometry for any relative offsets between the 2006 data sets and obsid 3554 .
the relative shifts were @xmath19 pixel . to remove periods of high background , for each obsid we extracted a light curve for chip s3 , excluding any bright sources in the field .
the light curves for 2006 observations showed no strong flares .
however , a period of enhanced background was present in the last quarter of obsid 3554 , and 1.6 ks exposure of affected data was screened out . finally all seven observations were merged into a single event file after correcting the relative astrometric offsets . detailed study of the point source population is deferred to a future paper ( j. wang et al . in preparation ) , which will also use the _ xmm_-newton observations of ngc 1365 .
in addition to _ chandra _ data , ancillary uv and mid - ir images were also used in this paper .
ngc 1365 was observed with the _
spitzer space telescope _
( werner et al . 2004 ) as part of a survey of luminous infrared galaxies ( lirgs ; @xmath20 ) .
observations with the infrared array camera ( irac ; fazio et al .
2004 ) were dithered to correct for cosmic ray events and bad pixels , and raster mapping was used to construct mosaics of the galaxy at 3.6 , 4.5 , 5.8 and 8.0 @xmath21 m ( spitzer aor # 12346880 ) .
details of the irac observations and data processing are described by mazzarella et al .
( 2008 , in preparation ) .
the xmm - newton om data were aquired from _
xmm_-newton obsid # 0205590301 ( pi : g. fabbiano ) , using the processing pipeline system ( pps ) products processed with science analysis system ( sas ; version 6.5.0 ) .
the uv images shown in figure 1 used the following om filters ( mason et al . 2001 ) : @xmath22 ( @xmath23 ) , @xmath24 ( @xmath25 ) , and @xmath26 ( @xmath27 ) .
smoothed images of the ngc 1365 region were first created without removing the point sources .
this enabled us to examine the spatial distribution of the point sources with respect to the galaxy .
we extracted images from the merged data in the commonly used soft band ( 0.30.65 kev ) , the medium band ( 0.651.5 kev ) , and the hard - band ( 1.57 kev ) .
note that the medium band encompasses the emission from the fe - l blend .
the 0.3 kev low - energy boundary is the lowest energy where acis remains well calibrated , and the high energy 7 kev cut - off is chosen to limit the contribution from the background . to derive exposure - corrected images , exposure maps were created for individual observations and bands , then reprojected to create the combined exposure maps matching the merged data .
the band - limited images were then divided by the appropriate exposure maps and adaptively smoothed .
this process enhances faint extended features while preserving the high resolution of brighter features . in order to prevent large variations in the adaptive smoothing scales between the bands with different counts
, we applied to the soft and hard bands the same smoothing scales calculated for the medium band image , which has the best signal to noise ratio ( s / n ) .
the significance was chosen between 2.4 and 5@xmath28 above the local background ( corresponding to the smoothing gaussian kernel of varying scales between 1 and 40 pixels ) .
figure [ 3c_diff ] presents the `` false color '' composite image of the central @xmath29 region of ngc 1365 , where the soft , medium , and hard band smoothed images are shown in red , green , and blue , respectively . besides the bright nucleus ( marked as `` n '' in figure [ 3c_diff ] )
, some 40 point sources are present .
the bright off - nuclear x - ray source towards the lower right corner is a well - studied , variable ultraluminous x - ray source ( ulx ) ngc 1365 x-1 ( komossa & schulz 1998 ) , most likely a 50150 accreting black hole ( soria et al .
2007 ) .
figure [ 3c_diff ] shows the circumnuclear ring ( mainly pink ; @xmath30 in diameter ) , intersected by the obscuring dust lane ( blue ) .
this ring is emphasized with an alternative set of energy bands shifted to higher energies images in figure [ 3c_ring ] , in which the soft band ( red ) is 0.31.5 kev , the medium band ( green ) is 1.53.0 kev , and the hard band ( blue ) is 3.07.0 kev ; the same smoothing was applied as for figure [ 3c_diff ] .
the hard emission ( presumably from young supernovae remnants and x - ray binaries ) is less affected by the obscuration from the dust lane , tracing the extent of the star forming ring .
note that with this energy band selection the ( broader ) soft band now has the best @xmath31 of the three bands and so the smoothing scales from this band were used to smooth the images in the other two bands .
this alternate band selection is only used here in this paper .
comparison with _ spitzer _
data clearly relates the hard x - ray ring to the circumnuclear star - forming ring prominent in the infrared .
figure [ spitzer ] shows the _ chandra _ x - ray image of the diffuse emission ( from figure [ truecolor ] ) in ngc 1365 ( green / blue ) , together with the mid - ir ( 8 @xmath21 m ) view of the galaxy ( red ) .
figure [ spitzer2 ] zooms in to the central 1 arcmin , which shows a composite mid - ir image of the circumnuclear ring and a composite mid - ir / uv image of the same region with an overlay of the medium - band x - ray emission .
the circumnuclear ring is compact and maintains regularity close to the center .
such circumnuclear star formation rings are commonly observed in strongly barred spirals and assumed to be dynamically associated with the inner lindblad resonance ( ilr)inflowing gas accumulates to form a massive star forming ring between the two inner resonances ( see lindblad 1999 ) .
the inner ilr and outer ilr in ngc 1365 is @xmath32 and @xmath33 from the nucleus , respectively ( lindblad et al . 1996 ) , hence the ring is located @xmath34 outside of the inner ilr .
we note a hot - spot located @xmath35 ( @xmath36 kpc ) ne of the nucleus , which shows co - spatial x - ray , uv , and mid - ir emission .
ir filaments bridge the ring and this object , suggesting that it is related to the galaxy , instead of being a background object , and may be associated with the bar .
the spatial correlation between the x - ray and uv emission is apparent .
because of the obscuring dust lane , which is traced by its bright ir emission , a bright ridge of x - ray / uv emission with hot - spots appears coincide with the edge of the ir ring @xmath37 ( 1 kpc )
n and nw of the nucleus , probably related to the discrete regions of active star formation associated with the ring . to study the large scale diffuse emission , we followed the procedure of baldi et al .
( 2006a ) to create mapped - color images with point sources removed . given the short individual exposure times , we performed point source detection in the combined exposures in the full band ( 0.37.0 kev ) .
the point - like acis detections including the nucleus were then removed from the data , and the resulting holes were filled by interpolating the surrounding background emission .
the same procedure to define source regions and background regions employed by baldi et al.(2006a ) was followed .
the results were visually compared with the band - limited images and the smoothed images to ensure that no apparent individual point sources were missed . in the circumnuclear ring region , a few `` point '' sources detected by _ wavdetect _ appear extended .
these could be either association of unresolved point sources or bright patches and knots of diffuse emission . to check whether they affect the spectral analysis of the circumnuclear ring , we created separate event files without removing these sources with uncertain classification .
these event files were used in spectral extraction and fitting described later in this paper .
the images with point sources removed and holes filled in the three bands ( 0.30.65 kev ; 0.651.5 kev ; and 1.57 kev ) were adaptively smoothed , then combined to create a `` mapped - color '' image of the diffuse emission shown in figure [ truecolor ] .
the mapped color image shows complex morphology and striking substructures that have not been seen in lower resolution x - ray images previously .
the overall morphology of the x - ray emission is rather consistent with the optical picture described in
[ intro ] : 1 .
extended soft x - ray emission appears to resemble a bicone ( nw and se ) structure with the apex at the nucleus .
the se structure is more extended ( @xmath38 measured from the nucleus ) and visually closer to a cone - shape .
its inner region is brighter and appears to contain substructure .
the nw cone is only bright at the base close to the nuclear region ( @xmath39 ) , with fainter emission extends towards far nw .
both cones are incomplete , possibly due to obscuring material in the galaxy plane and bulge .
the measured projected half opening angle of the cone in the x - ray image is approximately @xmath40 .
the x - ray cones appear to be edge - brightened , which is consistent with the hollow cones seen in the optical ( lindblad 1999 ) .
harder x - ray emission is associated with the circumnuclear star - forming ring , which appears pink and clumpy in the image .
3 . the hardest x - ray emission , which appears as a @xmath41-wide blue band intersecting the top portion of the circumnuclear ring , is coincident with the dust extinction lane of the galaxy .
a quick examination of the spectra extracted from the diffuse emission show the presence of the oxygen+iron+neon ( o+fe+ne ) emission blend ( 0.61.16 kev ) , the magnesium ( mg)-xi ( 1.271.38 kev ) and the silicon ( si)-xiii ( 1.751.95 kev ) lines ( figure [ spectra ] ) .
to identify regions of strong apparent line emission and to select possible anomalous abundance regions for spectral analysis , we generated an emission line map , following fabbiano et al .
( 2004 ) and baldi et al .
( 2006a ) , to highlight regions of prominent emission lines in the hot ism .
three band - limited images were created to encompass the lines noted above .
a `` continuum template '' was extracted using the 1.41.65 kev and 2.053.05 kev bands which are not visibly contaminated by emission lines , and was used to determine the underlying continuum in the emission line bands .
all point - like sources were removed and the holes were filled with background as described in [ truecol ] .
the o+fe+ne 0.61.16 kev band image contains the largest number of counts and so was used to set the smoothing scales for the other line and continuum images . to determine the continuum contribution in each line image
, we extracted the spectrum of the entire diffuse emission of ngc 1365 and fitted the continuum with a simple two - component thermal bremsstrahlung model , excluding the three energy ranges with strong line emission .
the best fit model provides the appropriate scaling factors between the continuum level in our continuum template and the continuum in a given line band .
the scaling factors are 2.5 , 0.2 , and 0.19 for the 0.61.16 kev , 1.271.38 kev , and 1.751.95 kev band , respectively .
each line image was continuum subtracted after weighting the continuum image by these scaling factors , and the resulting images were combined to create a mapped - color line - strength image ( figure [ line ] ) .
figure [ line ] shows that the emission - line structure of the hot ism in ngc 1365 is also quite complex .
silicon emission ( blue ) appears most pervasive in the dust lane ; this could be due to obscuration of the softer energy photons in this region , and this possibility is explored further below .
magnesium lines ( green ) and o+fe+ne emission ( red ) are prevalent in most regions .
part of the nw diffuse region shows enhanced o+fe+ne emission .
noticeably there is an area ( orange in color ) south of the star - forming ring that is lacking in silicon line emission .
the mapped - color x - ray image ( figure [ truecolor ] ) and the line - strength map ( figure [ line ] ) demonstrate that the diffuse emission of the hot ism in ngc 1365 is rich in spatial and spectral features .
these figures suggest structure in the hot ism , possibly the effect of varying absorption columns , plasma temperatures , and metal enrichment . as outlined in figures [ c_diff_reg ] and [ line ] , we identified five separate regions based on morphology and three regions based on apparent abundance pattern , to perform spectral analysis of the x - ray emission from the hot ism .
the region `` ring '' refers to the circumnuclear ring , excluding the nucleus ( see [ truecol ] and figure [ 3c_ring ] ) .
the region `` diffuse '' refers to all the diffuse emission , excluding the area inside the outer limit of the ring .
it is further divided into `` inner diffuse '' ( the bright portion of the cones ) and `` outer diffuse '' ( the faint diffuse emission ) .
in addition , for the ring area we also extracted photons from the `` ring '' region including the _ wavdetect _ selected point sources that appear extended ( see [ truecol ] ) .
we dubbed the region and the corresponding extraction `` ring+ '' .
we did not further divide the diffuse emission into sub - regions ( e.g. , the ring is separated by the dust lane into north - ring and south - ring ) to have good statistics for the spectral fitting .
regions related to the emission line features noted in figures [ line ] were also created , with descriptive names ( `` high - si '' , `` low - si '' , and `` high - fe '' ) .
we emphasize here that , even though the extended emission is referred to as `` diffuse '' , the x - ray emission is not @xmath42 diffuse plasma
. there is evidence for extensive clumpiness and x - ray hot - spots .
the `` diffuse '' x - rays include a mixture of hot ism and unresolved point sources ( xrb / snr ) associated with regions of very active star formation .
the hard power law component seen in the spectral fit ( see [ fitting ] ) is likely related to the xrb / snr components .
we extracted spectra separately from the individual obsids , for each diffuse emission regions described above .
we excluded all the point source areas above @xmath43 described in [ truecol ] .
we also extracted background counts representative of the field from three circular source - free regions ( each with a @xmath44 radius , resulting in a total area of 1.5 arcmin@xmath45 ) well outside the galaxy , but within the acis - s3 chip in all seven observations ( figure [ om ] ) .
a new ciao script @xmath46 was used to extract the region spectra for multiple observations .
this script extracts the spectrum and creates area - weighted response matrix files ( rmf ) and ancillary response files ( arf ) for each region . for each region of interest ,
all source spectra and background spectra were co - added and the response files were combined with their appropriate weights using the ftools software .
the resulting spectra were then grouped so as to have signal to noise ratio @xmath47 in each energy bin .
the extracted spectra were analyzed with xspec v12.0 ( arnaud 1996 ) , using combinations of absorbed optically - thin thermal emission and power law components to fit the data .
we used the astrophysical plasma emission code ( apec ) thermal - emission model ( smith et al .
2001 ) that utilizes improved atomic data from the astrophysical plasma emission database .
the absorption column was modeled with @xmath48 ( wilms et al .
2000 ) , which includes updated photoionization cross section and abundances of the ism , as well as a treatment of interstellar grains and the h@xmath49 molecule .
we first considered absorbed single - temperature apec models as well as two - temperature apec models .
a single line - of - sight absorption column was used in all cases , equivalent to galactic line - of - sight column ( @xmath50 @xmath51 ; kalberla et al .
2005 ) combined with the intrinsic absorption within ngc 1365 , which was free to vary .
when the fit required unmeasurably small absorption ( @xmath52 @xmath51 ) , we froze @xmath53 at the galactic column .
as the spectrum of the integrated diffuse emission clearly shows prominent mgxi ( @xmath01.3 kev ) , ovii - neix / x - fe - l complex ( 0.6 - 1.2 kev ) , and sixiii ( @xmath01.8 kev ) line features , the abundances of these elements were left to vary freely . when fitting with the two component models ,
the oxygen abundance was fixed at the solar value as it is poorly constrained in our data and consistent with solar abundance .
a power law component is introduced to the single - temperature model to allow for the possible integrated emission of unresolved x - ray binaries that are too faint to be detected individually .
a simple absorbed power law model give poor fits ( @xmath54 ) to all the extracted spectra .
to fit the spectra of the three extreme line ratio regions ( figures [ line ] ) , we only allow the abundances of mg , si , and fe to vary , given the limited number of extracted counts . for each of the regions used in the spectral analysis , the net counts ( background subtracted ) in the full band , the reduced @xmath55 ( @xmath55 over the numbers of degrees of freedom @xmath56 , @xmath57 ) , the resulting best - fit temperature(s ) @xmath58 ( kev ) , absorbing column density @xmath53 ( @xmath51 ) , and power law photon index @xmath59 ( if a power law component is invoked ) are summarized in table [ mytable1 ] , together with the abundances of oxygen , neon , magnesium , silicon , and iron relative to the solar values .
the last column of table [ mytable1 ] gives the adopted `` best - fit model '' , which is chosen based on having the smallest @xmath60 .
if more than one model can fit the spectrum well and the @xmath60 values are close ( difference less than 0.1 ) , a simpler model with less free parameters is preferred .
the errors are quoted at @xmath61 ( 90% confidence interval ) for one interesting parameter .
the spectra of the regions , together with the best - fit models and the data@xmath62(best - fit - model ) residuals , are shown in figure [ spectra ] . in summary , these results show variations in different morphological / physical regions : 1 .
the large scale diffuse emission ( `` diffuse '' ) is well fitted with a single temperature thermal model , with a well determined @xmath63 kev .
the absorption may be a little in excess of the galactic column towards ngc 1365 ( @xmath64 @xmath51 ) .
to adequately fit the emission lines in the acis spectrum , the data require sub - solar abundances ( @xmath65 ) .
the sub - regions of the diffuse emission show similar characteristics to the combined emission .
the absorption column of the outer diffuse emission ( `` outer diff '' ) is low , consistent with just galactic absorption , which is reasonable , as the region is further away from the spiral disk .
the spectra of the circumnuclear ring ( and `` ring+ '' ) require higher absorption columns , @xmath66 @xmath51 , and a hard spectral component in addition to the @xmath67 kev thermal component similar to the diffuse emission ( `` diffuse '' ) .
the large @xmath53 is not surprising given the dust lane and higher obscuration towards the nuclear region . in the power - law fit this hard component has a photon index @xmath68 ; using a two - temperature thermal model , we obtain a hard @xmath69 kev , but the fit with power - law component has a better @xmath55 for the same numbers of free parameters . in either case a hard component is required , and maybe due to the unresolved x - ray source population in the ring .
the best - fit model suggests super - solar abundances in the hot ism ( @xmath70 times @xmath71 ) .
3 . for the extreme line ratio regions ,
all the fits have @xmath57 less than 1 indicating the models are over complex for the data .
nevertheless , the `` high - si '' region ( dust lane ) shows high absorption column ( @xmath72 ) as expected .
the fit suggests under - abundant fe , but this result may be affected by the obscuration of the dust lane .
taking the @xmath73 @xmath51 ( table [ mytable2 ] ) and a visual extinction @xmath74 mag ( roy & walsh 1997 ; kristen et al .
1997 ; galactic contribution is removed from @xmath53 and @xmath75 ) , we measured a gas - to - dust ratio @xmath76 @xmath51 mag@xmath77 , which is roughly in agreement with the standard galactic gas - to - dust ratio @xmath78 @xmath51 mag@xmath77 ( maiolino et al .
2001 and references therein ) .
the extracted spectra of `` low - si '' region also suggests extremely low abundance of si ( @xmath79 preferred from the fit ; we consider it unconstrained ) .
we note that a similar low si abundances have been reported in some regions of the antennae ( ngc 4038/9 ; baldi et al 2006b ) .
a possible explanation for this deficiency is that a large fraction of the silicon has cooled and locked in dust grains .
sn ia enrichment may also yield a relatively low si abundance ( nagataki & sato 1998 ) .
lastly , for all the spectral fitting , we emphasize that the temperature and absorption column for the same region determined from various models show little variation , indicating that these parameters are rather robust . in table
[ lumin ] we list the derived emission parameters from the spectral analysis of each region . from these emission parameters and those listed in table [ mytable1 ] , following fabbiano et al .
( 2003 ) and baldi et al .
( 2006b ) , we can further estimate some physical properties of the emitting plasma , such as electron density @xmath80 ( @xmath81 ) , thermal energy content @xmath82 , cooling time @xmath83 , pressure @xmath84 , hot ism mass @xmath85 , and supernovae rate @xmath86 .
these quantities are given in table [ mytable2 ] .
the emitting volume ( @xmath87 ) of the hot gas in the region of interest ( assuming a cylindrical geometry ) is derived using the projected area on the plane of the sky multiplied by a height of @xmath88 pc ( for the typical depth of a spiral disk ) .
considering the approximate nature of the geometry and depth of the hot gas , the inclination of the galaxy ( @xmath89 ) has little effect on the estimated emitting volume .
the filling factor @xmath90 is assumed to be 100% in table [ mytable2 ] , and the derived parameters have weak dependence on @xmath90 ( @xmath91 and @xmath92 ) .
these estimates assumed an @xmath90 of unity and the geometry of the x - ray emitting gas , and should not be treated as precise measurements ( see discussion in baldi et al .
2006b ) .
although a thermal model accounts well for the bulk of the diffuse x - ray emission , some positive residuals are seen around 0.53 kev ( nvii - ovii ) and around 0.70.8 kev ( possibly ovii and oviii radiative recombination continua ) in the thermal fit of diffuse emission ( figure [ spectra ] ) , which hint at the presence of additional possible photoionized , emission lines . emission from a photoionized medium might be the dominant component in the inner region around the nucleus ( guainazzi et al .
, in preparation ) , though not at radii greater than @xmath93 . to explore the line emission from photoionized gas , we used xstar ( kallman & bautista 2001 ) v2.1ln7 and the xspec12 local model @xmath94 to model the photoionized contribution to the extended x - ray emission ( `` diffuse '' and `` inner diffuse '' ) . using a photoionized model alone yielded poor fits ( @xmath95 ) of the spectra
( see further discussed from an energetic viability perspective in [ discuss ] ) . unlike in detailed photoionization modeling studies of high resolution grating spectra ( e.g. ,
ogle et al .
2000 , 2003 ) , we froze the parameters in our best - fit thermal model , and added an emission component from a warm photoionized emitter .
we assume the gas is photoionized by an incident spectrum with a power law index @xmath96 ( consistent with the average spectral index of the nucleus , risaliti et al .
2007 ) , and calculated the photoionized emission with xstar .
we find that the best - fit ionization parameter is @xmath97 where @xmath98 , @xmath99 is the luminosity of the ionizing source , @xmath100 is the ion density , and @xmath101 is the distance to the source .
we estimate an @xmath102 @xmath103 for the photoionized emitter , about 1% of the @xmath104 of the thermal component .
since the x - ray absorbing gas towards ngc 1365 nucleus shows significant variability , we also attempted modeling with an incident spectrum of @xmath105 ( the spectral slope derived from high s / n @xmath106-newton spectra with the lowest absorption column , which represents the best measurements of the continuum parameters ; risaliti et al .
2008 , in preparation ) .
the resulting ionization parameter is poorly constrained , @xmath107 , and the photoionized emitter may contribute @xmath017% of the total @xmath104 .
a biconical enlr is a common feature of nearby seyfert galaxies ( schmitt et al .
2003 ) , and the associated soft x - ray emission has been suggested to originate from the same photoionzed gas seen in [ oiii ] ( e.g. , ogle et al .
2000 ; bianchi et al . 2006 ) . for the ionization cones in ngc 1365 ( in the `` diffuse '' region ) , the soft emission is consistent with a thermal origin , as the thermal model provides adequate fit to the spectra while the pure photoionization model fails .
further morphological evidence from x - ray and [ oiii ] emission will be discussed in the next section .
the presence of the fe - l complex , which is often seen dominant in spectra from hot gas in collisional ionization equilibrium ( phillips 1982 ) supports a thermal origin , although ogle et al .
( 2003 ) cautioned that these lines could arise from recombination and photoexcitation as well .
_ chandra _ observations of ngc 1068 ( ogle et al .
2003 ) demonstrated that agn outflows can show a wide range in ionization ( @xmath10813 ) .
the derived @xmath109 ( or @xmath110 if @xmath105 ) in our simple photoionization fitting ( [ photo ] ) is low but marginally consistent with the lower boundary of @xmath111 . following weaver et al . ( 1995 ) and evans et al . ( 2006 ) , we further examine the energetic viability of the photoionization model with the measured x - ray luminosity , assuming that the extended soft x - ray emitting gas is predominantly photoionized by the nuclear emission . in this case , an ionizing parameter @xmath112 ergs cm s@xmath77 is required to produce the strong fe - l emission feature arising from ionized gas ( kallman & mccray et al . 1982 ; kallman 1991 ) .
adopting an emissivity @xmath113 ergs @xmath114 s@xmath77 for the photoionized gas and the observed soft x - ray emission @xmath115 ergs s@xmath77 , we derived a density @xmath116 @xmath103 .
it is worth noting that this density is in excellent agreement with the @xmath117 @xmath103 that we derived in the hot gas from the thermal model ( table [ mytable2 ] ) , indicating that our estimate is reasonable despite the simple assumption made here . substituting @xmath111 , @xmath100 , and a distance @xmath118 ( 900 pc ) in @xmath98 , the luminosity of the nucleus required to photoionize the gas to the observed level is @xmath119 ergs s@xmath77
this is @xmath01001000 times higher than the unabsorbed @xmath120 ergs s@xmath77 ( 210 kev ) reported in risaliti et al .
( 2005b ) , suggesting that photoionization alone can not account for the extended soft x - ray emission unless the ionizing radiation from the nuclear source is highly anisotropic .
as ngc 1365 is a relatively low luminosity agn , that thermal emission is dominant over an agn - continuum photoionized gas may not be surprising .
we note that ngc 1365 is characterized with a surface density of the star formation rate @xmath121 yr@xmath77 kpc@xmath122 and a mean gas surface density @xmath123 pc@xmath122 ( sakamoto et al . 2007 ;
a @xmath124co - to - h@xmath49 conversion factor is adopted as in kennicutt 1998 ) .
these surface densities follow the kennicutt - schmidt law for a sample of starbursts and normal disk galaxies ( kennicutt 1998 ) .
wang et al.(2007 ) derive a domain in the @xmath125@xmath126 plot where the circumnuclear medium is significantly affected by the agn radiation , and with a sample of 57 seyfert galaxies they show evidence that the circumnuclear star formation of many seyferts are suppressed by the agn feedback .
we find ngc 1365 lies near but outside of the boundary of the agn feedback domain , resembling the starbursts well .
our spectral results are overall consistent with previous x - ray studies of ngc 1365 that found that the extended soft nuclear emission can be well fitted by a thermal plasma model with a very low local absorbing column ( fabbiano et al .
1992 ; iyomoto et al .
1997 ; komossa & schulz 1998 ; stevens et al .
we caution that the low resolution ccd imaging spectroscopy can not distinguish whether the `` continuum '' is actually made of unresolved emission lines from photoionized gas .
higher resolution grating spectra , especially constraints from the line intensities can shed more light on the agn - photoionized emission component ( e.g. , guanazzi et al .
2007 , evans et al .
2007 , longinotti et al .
2007 , kraemer et al . 2008 ) .
additional evidence in favor of a thermal hot ism can be gathered by comparing our results with the [ oiii ] properties of the circumnuclear region .
figure [ oiii ] compares the spatial correspondance between the x - ray emission and continuum subtracted [ oiii ] image from veilleux et al .
( 2003 ) , enabled by the superior arcsec resolution x - ray image and the excellent astrometry .
both x - ray and [ oiii ] are present in the inner @xmath93 region around the nucleus .
most imortantly , the nw cone shows some complementary structure between the two types of emission ( a `` swiss - cheese''-like morphology ) : the x - ray emission is weak where the [ oiii ] emission is strong .
the hot x - ray emitting gas appears delimited by the cooler condensation traced by the [ oiii ] emission .
this spatial anti - correlation can not be attributed to obscuration or projection , and is consistent with a predominantly thermal origin for the x - ray emission ( elvis et al .
1983 , ogle et al .
2000 ) , which was also suggested by our spectral analysis ( [ specanal ] ) .
our data do not have the resolution to perform a detail comparison of the emission properties of the numerous smaller clouds within @xmath127 of the nucleus that were resolved in the _ hst _ [ oiii ] image ( kristen et al .
however , assuming a temperature of @xmath128 k and a density of @xmath129 @xmath103 for the [ oiii ] emitting clouds , which are typical values observed in the optical emission line gas ( e.g. , osterbrock & ferland 2006 , wilson et al .
1985 , ferruit et al . 1999 ) , we estimate the gas pressure of the [ oiii ] clouds to be @xmath130 dyne @xmath51 .
these pressures are comparable to ( or smaller than ) the thermal pressure of the hot ism we calculated in table 5 , implying a possible pressure equilibrium between the x - ray emitting hot gas and the optical line - emitting cool gas .
the x - ray - emitting gas may serve as the hot phase confining intercloud medium to the nlr clouds ( see elvis et al .
1983 ) . as summarized in
[ fitting ] , the results of our spectral fitting of the diffuse ( enlr cone , `` diffuse '' ) and of the circumnuclear ring emission ( `` ring '' ) , suggest a pervasive hot ism with temperatures of about 0.6 kev .
metal abundances are overall larger ( super - solar ) in the ring , and become sub - solar at larger radii in the cone region .
this temperature and abundance pattern is similar to that of the hot ism in many regions of ngc 4038/9 , `` the antennae '' galaxies ( see baldi et al .
2006a and b ) ; not surprisingly , the calculated emission parameters of the hot gas in ngc 1365 , such as electron densities , pressure and cooling times , are also in agreement . in the antennae ,
the metal abundances of the hot ism are consistent with metal enrichment from sn ii ejecta .
as the ring in ngc 1365 is undergoing a starburst , we may similarly expect to see enhanced abundances .
nakataki & sato ( 1998 ) show that if the elemental enrichment is from type ii sn ejecta , the expected [ ne / fe ] and [ mg / fe ] values approach 0.3 and [ si / fe ] approaches 0.5 on average .
in contrast , the expected values from type ia sn is significantly lower .
thus we can explore the metal enrichment in ngc 1365 based on this difference .
we compiled the element abundances relative to iron in different regions in table [ mytable3 ] .
together we also show the abundance values from the averaged hot halo of starburst galaxies ( strickland et al .
2004 ) , the warm galactic halo ( savage & sembach 1996 ) , and type ia , type ii sne from a variety of theoretical models ( nakataki & sato 1998 ) . although there are some uncertainties , the morphologically selected regions in ngc 1365 show [ ne / fe ] , [ mg / fe ] , and [ si / fe ] ratios fully consistent with a type ii sn - dominated enrichment scenario ( see figure [ nemg ] ) . as the circumnuclear ring is where active star formation is occurring and where candidate radio sne are identified ( galliano et al.2005 ) , it is plausible that recent episodes of star formation led to the enrichment of the ism almost exclusively through sne ii injection .
notably the `` ring '' ( and `` ring+ '' ) region has abundance ratios matching very well with the theoretically predicted values .
two sne have been seen in ngc 1365 ( not related to galliano et al .
radio sne ) in four decades ( lindblad 1999 ) , giving a rate about six times higher than our estimated @xmath86 in table [ mytable2 ] but with a large poisson error bar .
the @xmath131 optical study by kristen et al .
( 1997 ) estimate a supernovae rate of @xmath132 yr@xmath77 for the luminous star clusters in ngc 1365 , which is in good agreement with our derived @xmath86 .
however , we note that these measurements in ngc 1365 , similar to x - ray estimates , only place a lower limit on the actual , intrinsic supernova rate as they are severely affected by dust extinction .
it is well established that the extinction in lirgs is high enough to obscure sne in their central regions , even at near - ir wavelengths ( van buren & greenhouse 1994 ; see also the near - ir sn survey by mannucci et al .
fir studies of starburst galaxies appear to yield a higher sne explosion rate assuming that @xmath133 is powered mainly by star formation . using observations of ngc 253 and
m 82 , van buren & greenhouse ( 1994 ) derived the relation @xmath134 yr@xmath135 .
a similar factor ( @xmath136 ) was derived by mannucci et al .
( 2003 ) and independently ( @xmath137 ) by mattila & meikle ( 2001 ) . applying this to ngc 1365 s total
@xmath133 predicts a total sne rate of @xmath00.2 yr@xmath77 , which is @xmath020 times higher than the largest diffuse region in table [ mytable2 ] with @xmath138 yr@xmath77 . within the ring alone ,
using our estimate of @xmath139% of the total @xmath133 ( see [ comp2 ] ) implies @xmath140 yr@xmath135 , which is @xmath0200 times higher than the x - ray estimate in table [ mytable2 ] .
this discrepancy can be attributed to a number of uncertainties .
first , although we used absorption corrected x - ray luminosity to account for the attenuation of intrinsic x - ray luminosity , this absorption correction derived from spectral fitting is still not precise , especially when the absorption column is high enough to obscure the soft x - rays . in comparison ,
fir and radio are known to be reliable quantitative measures of emitted luminosities as most galaxies become transparent at these wavelengths .
second , we estimated @xmath86 assuming that the mechanical energy released in sn explosions heat the ism thermal emission , the efficiency of which depends on the detailed properties of the interstellar environment .
a low efficiency can increase the required @xmath86 by a factor of 100 .
in addition , the assumed x - ray emitting volume and volume filling factor are uncertain and difficult to quantify .
altogether , the x - ray estimated sn rate should be deemed as a lower limit , and does not contradict the fir - derived sn rate .
some caution is warranted as the above analysis relies upon the abundances of the emitting plasma being accurately modeled .
measuring hot gas elemental abundances through x - ray spectral fitting was recognized to be quite uncertain ( especially with the low resolution spectra of ccds ) because of ambiguities in the fits ( e.g. , a degeneracy between the temperature and metallicity ) , and strongly dependent on the model choice ( see dahlem , weaver , & heckman 1998 ; dahlem et al .
2000 ; strickland et al . 2002 , 2004 , baldi et al .
these uncertainties are strongly affected by poor spatial resolution , and spatially resolved spectroscopy is needed to determine reliable constraints .
however , the agreement of all the regions on a set of abundances close to the sne ii values seems unlikely to occur by chance and suggest that _ chandra _ has spatially resolved the major variations . figure [ jet ] shows that the bright nuclear emission in the full band ( 0.37 kev ) shows a clear asymmetry , extending @xmath141 towards the se .
the orientation of this jet - like feature emanating from the nucleus is well aligned with the cone - axis seen at larger scale .
we consider this to be a candidate x - ray counterpart of the radio jet ( ngc 1365:f , radio spectral index -0.94 , sandqvist et al .
however a fit to the extracted spectra ( @xmath0300 net counts ) suggests an absorbed power law emission with @xmath142 , which is unusally soft for a jet ( harris & krawczynski 2006 ) .
if the x - rays and radio emission are produced in synchrotron radiation from the same population of electrons , the steep x - ray spectrum implies that the x - ray emission is in the tail of the spectral distribution of the synchrotron radiation ( @xmath143 , the critical frequency ) .
alternatively , stevens et al .
( 1999 ) pointed out that the radio feature may represent enhanced star formation often seen at the ends of bars in barred galaxies .
this would result in a soft x - ray spectrum similar to the thermal emission from hot gas in the circumnuclear ring , as we see here ( @xmath144 kev ) if we fit the jet spectrum with a single temperature thermal emission model .
we examined the high angular resolution ( 0.07 ) _
f437m(@xmath145 ) image that was presented in kristen et al.(1997 ) .
none of the reported compact continuum source is associated with the jet , with an observed flux completeness limit corresponds to @xmath146 mag .
no apparent mid - ir source is found at the location of the jet in the _ spitzer _ irac images .
the lacking in optical and ir counterparts seems to disfavor the enhanced star formation origin for the x - ray emission . neither the radio observation nor our x - ray image show evidence for a counter jet .
if indeed we are seeing a radio jet , then the observed one - sidedness could be explained by relativistic beaming with a jet velocity 0.7@xmath147 , or with free - free absorption by the circumnuclear disk as suggested by hjelm & lindblad ( 1996 ) .
it is also interesting to note that the x - ray - to-6 cm luminosity ratio @xmath148 is @xmath149 for the jet feature ( 6 cm luminosity of the jet from stevens et al .
1999 ) , similar to that of radio sne and radio quiet quasars ( e.g. , a few hundred , stevens et al . 1999 ; landt et al .
2001 ) .
ngc 1365 is a lirg ( @xmath150 ; ghosh et al .
1993 ; sanders et al .
the central nucleus contributes little ( few percent ; komossa & schulz 1998 ) to the total ir luminosity , most of which could be attributed to the active star forming ring .
sakamoto et al . ( 2007 ) found that starburst in subregions of the ngc 1365 circumnuclear ring alone is comparable to the entire starburst in galaxies like m82 and ngc 253 .
the star formation rate in the inner kpc region is @xmath151yr@xmath77kpc@xmath122 , among the highest in the non - merging galaxies ( kennicutt 1998 ) . for star - forming galaxies , a good correlation between the x - ray and the far infrared ( fir )
luminosities has long been known and extensively investigated ( e.g. , fabbiano , gioia & trinchieri 1988 ; david et al .
1992 ; fabbiano & shapley 2002 ; ranalli et al . 2003 ; colbert et al . 2004
; persic & rephaeli 2007 ) .
the x - ray emission in the circumnuclear starburst mainly comes from the young sne ( also snrs ) and x - ray binaries ( xrbs ; fabbiano 1989 ) , the formation of which requires time for the stellar evolution of o and b stars to complete .
therefore x - ray emission may trace the evolution of a starburst with some delay ( a few myr , the main - sequence life time of a massive star ) , and the x - ray - to - fir luminosity ratio @xmath152 could reflect the age of the starburst event ( see swartz et al .
recently mas - hesse et al .
( 2008 ) computed soft x - ray and fir luminosities for different starburst scenario ( instantaneous or extended ) , using realistic efficiency of converting mechanical energy into soft x - ray emission , and concluded that the soft x - ray emission can be used to trace the star formation rate in relatively young and unevolved starbursts . in table [ starburst ] we compiled @xmath153 ( 0.12.5 kev ; total line - of - sight absorption corrected to represent intrinsic luminosity ) , iras fir luminosity @xmath133 , and measured ages of young clusters associated with starburst events for a number of starburst galaxies ( condon et al .
1998 , sanders et al . 2003 ; see complete reference list in table footnote ) . for ngc 1365 , we calculated the @xmath152 using both total ( non - nuclear ) x - ray emission over total fir emission , and x - ray emission from circumnuclear starburst ring over fir emission from the ring .
given the available data , fir emission from the ngc 1365 ring can not be accurately measured due to the low spatial resolution at 70@xmath21 m and 160@xmath21 m ( saturated ) _ spitzer _ images . to derive an approximate @xmath133 value for the ring
, we estimate the @xmath154 ratio between the nucleus and the nucleus combined with the ring at 24@xmath21 m , which is @xmath050% . together the nucleus and the ring contribute @xmath070% of the fir emission from ngc 1365 , therefore we estimate that @xmath133 of the ring is approximately 35% of the iras @xmath133 for ngc 1365 .
the @xmath155 is in good agreement with that of m82 , whose @xmath152 represents a sample of starburst galaxies ( strickland et al .
2004 ) . in comparison , the ultra - luminous ir galaxy ( ulirg )
arp 220 has a lower @xmath156 ( soifer 1984 ; mcdowell et al .
2003 ) , which hosts very young clusters ( 13 myr old ; wilson et al .
2006 ) . on the other hand
, the merging system antennae shows a higher @xmath152 , where older clusters ( age @xmath013 myr ) have been found ( mengel et al .
another well known starburst galaxy ngc 253 had a starburst event 2030 myr ago ( engelbracht et al .
its @xmath157 is a lower limit because of the high obscuration of soft x - ray emission in the starburst nucleus , which could be similar to that of the antennae .
we also included the galactic center , if viewed from outside of our galaxy , the galactic center will appear as a circumnuclear starburst with a weakly active nucleus ( sgr a@xmath158 ) .
figure [ lxlfir ] shows the resulting @xmath153 vs. @xmath133 plot and the predicted values for 1 myr , 4 myr , and 6 myr old starbursts from mas - hesse et al .
the different assumptions on star formation history and energy conversion efficiency yield similar predictions for such relatively unevolved systems .
their locations relative to the model tracks show that the starburst in ngc 3351 is over 1 myr old .
ngc 1365 , m82 , and ngc 253 lie closely with the 4 myr model tracks , and the older merging system antennae shows values consistent with the 6 myr model . comparing to the ages obtained in the literature
( see table [ starburst ] ) , these values derived from the model tracks appear to be qualitatively consistent but a few myrs younger .
many factors may contribute to this : the assumption of the star formation history in the galaxies is apparently over - simplified ; the predicted age in the model is coupled with the efficiency of re - processing mechanical energy into soft x - ray emission , which could vary in different ism environment ; the measured values from literature also have large range as charting the age of starbursts is challenging ( see gallagher & smith 2005 review ) ; the measured soft band @xmath153 can be significantly underestimated because of the heavily obscured starburst regions ( see related discussion in table [ starburst ] footnotes ) .
this raises caution that the comparison with theoretical evolution may only be used to roughly estimate the age of starbursts .
it is known that the emergent soft x - ray luminosities may be greatly reduced from the star forming galaxies with significant visual extinction ( mag ) ( ryter 1996 ) with the x - ray absorption cross sections from morrison & mccammon ( 1983 ) gives @xmath159 at @xmath160 kev when @xmath161 . ] .
however , accurate attenuation corrections are unrealistic , the intrinsic absorption columns towards the galaxies are inhomogeneous and difficult to measure .
for example , the extinction to the core of arp 220 is estimated at @xmath162 mag from sub - mm continuum measurements ( sakamoto et al . 2008 ) , while the revealed star clusters have @xmath74 mag ( wilson et al.2006 ) . to correct for absorption ,
we adopt @xmath53 from x - ray spectral fit if available , otherwise we convert published @xmath75 that are typical of the star clusters to @xmath53 using the galactic gas - to - dust ratio . for the galaxies with @xmath163 , the uncertainty in the absorption columns results in a @xmath1640.08 dex uncertainty in the absorption corrected @xmath165 from the xspec spectral fit . for arp 220 and ngc 253 , correcting for a @xmath166 column ( the most obscured nuclear region ) can increase @xmath165 by a factor of 100 ( see lutz et al .
2004 for estimated correction in type 2 seyferts ) . in comparison ,
the correlation between the x - ray luminosities ( 0.54.5 kev band ) and the fir luminosities derived in david et al .
( 1992 ) assumes an average @xmath167 @xmath51 for all fir bright galaxies ; the empirical calibration between the soft x - ray luminosities ( 0.52.0 kev ) and the fir luminosities in ranalli et al .
( 2003 ) is corrected for galactic absorption only .
swartz et al . ( 2006 ) reported _ chandra _ observations of the nearby ( @xmath168 mpc ) barred spiral galaxy , ngc 3351 , which has a prominent kpc wide circumnuclear ring in reminiscent of the ngc 1365 ring .
the x - ray emission in the ngc 3351 ring has @xmath169 kev ( @xmath170 ergs s@xmath77 , 0.33.0 kev ) , attributed to hot diffuse gas associated with starburst .
the x - ray luminosity of the central nucleus is estimated to be @xmath171 ergs s@xmath77 ( 0.58.0 kev ) , and x - ray photoionization is considered unimportant .
the extended soft x - ray emission , similar to what we see in ngc 1365 , is interpreted as gas outflow from the star - forming ring out of the galactic plane .
a multiwavelength study of the circumnuclear extended emission in seyfert 2 galaxy ngc 2110 ( evans et al .
2006 ) suggested that shock - heated multiphase gas can account for the observations , although photoionized gas by the nucleus can not be ruled out .
in contrast , in another nearby agn - starburst galaxy ngc 6764 , croston et al .
( 2008 ) detected extended , @xmath172 kev x - ray emission coincident with the radio bubbles , likely inflated by the agn ( hota & saikia 2006 ) .
they rule out a galactic wind as the origin of the x - ray emission and instead favor hot gas heated by jet / ism interaction in the agn outflow , similar to the radio lobes in radio galaxies ( e.g. , centaurus a , kraft et al . 2003 ) .
this is quite a different situation from ngc 1365 , which shows a weak radio nucleus and jet , and no radio bubble ( sandqvist et al .
1995 ) .
comparison of ngc 1365 to these starburst and seyfert 2 galaxies suggests that , in galaxies where the agn is not the dominant energetic factor in the circumnuclear environment ( i.e. , where a starburst plays major role ) , the extended soft x - ray emission shows a thermal origin as sne and massive stars drive outflow into the disk and halo ( galactic winds , see heckman et al .
1990 , strickland et al .
if instead the energetics of nuclear radiation suffice , circumnuclear x - ray emission may be solely attributed to agn photoionization ( e.g. , ogle et al .
2003 , bianchi et al .
when a powerful agn jet / outflow is seen as evidenced by large scale radio bubbles , the interaction with the ism may heat the entrained gas and account for the x - ray emission ( e.g. , kraft et al .
2003 , croston et al .
these galaxy feedback processes are critical in our understanding of galaxy formation and evolution , which can be best studied when high resolution radio , x - ray , and optical nlr data are combined ( e.g. , evans et al .
2006 ; bianchi et al .
we have presented a detailed _ chandra_/acis imaging study of the nearby seyfert galaxy ngc 1365 . based on the high resolution 100 ks images , we confirm results from previous x - ray studies and present new findings in the complex circumnuclear region .
we have created an x - ray broad band color image as well as narrow band line - emission image , showing a biconical soft x - ray emission region extending from the nucleus , coincident with the high excitation outflow cones seen in [ oiii]/h@xmath4 emission line images .
a prominent kpc - scale circumnuclear star - forming ring is seen emitting harder x - rays .
the spectral fitting on the spatially separated components indicates thermal plasma origin for the x - ray emission in the ionization cones , which is also supported by the spatial anti - correlation between x - ray emission and the [ oiii ] emission .
the abundance ratios for regions of ngc 1365 are fully consistent with the theoretical values for enrichment from type ii sne .
x - ray morphology of ngc 1365 is compared with _ spitzer _
mid - ir images and [ oiii ] emission line observations .
we argue that pure photoionization by agn can not account for the extended soft x - ray emission based on the spectral fit , energetic estimate , and morphology evidences .
we attribute the observed soft x - rays to thermal emission from shock heated hot gas , although we can not rule out a photoionized contribution in the line emission .
x - ray emission from a previously reported nuclear radio jet is detected .
further progress will require deeper _ chandra _ images of ngc 1365 , which we plan to obtain in the future to enable studies of the morphology and spectral properties in greater details ( similar to baldi et al .
2006a , b ) , with better statistics and constraints .
isolation of discrete sources in the deep images will allow removal of contribution from the xrb population to the diffuse emission .
this enables investigations to distinguish the nature of x - ray hot - spots from the truly diffuse emission they embedded in , and studies of the x - ray luminosity function of xrbs in ngc 1365 .
spatial complexity of the x - ray emission can be overcome by dividing the circumnuclear ring and diffuse cones to smaller regions .
this will also localize any abundance anomalies . with more counts in the diffuse emission , we will be able to perform spectral modeling on the radial gradient with different models and search for variation in the x - ray structure and spectra .
expecting that a photoionized wind will have constant ionization parameter , and that a thermal wind will cool radially , we plan to use this diagnostic to unambiguously test the photoionized wind model and thermal wind model for the origin of x - ray emission associated with the enlr in ngc 1365 .
radially constant ionization parameters have been found in a sample of type 2 seyferts ( bianchi et al .
2006 ) , suggesting that at least in some cases , soft x - ray and optical nlr come from a single photoionized medium .
deeper x - ray imaging will also clarify the nature of the jet - like feature .
we thank an anonymous referee for helpful comments that improved the clarity of the manuscript .
we are grateful to s. veilleux for providing us the [ oiii ] images in veilleux et al .
( 2003 ) , and tim kallman for the helpful information about xstar . j. w. thanks a. baldi for his help in the data analysis , h. oti floranes for providing the model results in mas - hesse et al .
( 2008 ) , and a. siemiginowska for helpful discussion on radio jets .
g. f. and m. e. acknowledge stimulating discussions at the aspen center for physics workshop on active galactic nuclei .
this work is supported by nasa contract nas8 - 39073 ( cxc ) and _ chandra _ go grant g06 - 7102x ( pi : risaliti ) .
3554&december 24 , 2002 & 53.409638 & -36.147430 & 321.9&13.0 + 6868&april 17 , 2006 & 53.399219 & -36.139588 & 215.0&14.6 + 6869&april 20 , 2006 & 53.399299 & -36.139509&211.8&15.5 + 6870&april 23 , 2006 & 53.399377 & -36.139370&208.1&14.6 + 6871&april 10 , 2006 & 53.399118 & -36.139908&223.9&13.4 + 6872&april 12 , 2006 & 53.399126 & -36.139825&221.3&14.6 + 6873&april 15 , 2006 & 53.399201 & -36.139764&218.4&14.6 + llcccccccccc diff .
& @xmath173&44.7/50&@xmath174&@xmath175 & @xmath176&@xmath177&@xmath178&@xmath179&@xmath180&&1 t + ring&@xmath181&21.8/31&@xmath182&@xmath183 & 1.0 & @xmath184&@xmath185&@xmath186&@xmath187&@xmath188&tp + ring+&@xmath189&34.7/47&@xmath190&@xmath191 & 1.0&@xmath192&@xmath193&@xmath194&@xmath195&@xmath196&tp + inner diff . &
@xmath197&22/32&@xmath198&@xmath199&@xmath200&@xmath201&@xmath202&@xmath203&@xmath204&&1 t + outer diff.&@xmath205&19.8/25&@xmath206&@xmath207 & @xmath208&@xmath209&@xmath210&@xmath211&@xmath212&&1 t + high - si&@xmath213&18/38&@xmath214&@xmath215 & 1.0&1.0&@xmath216&@xmath217&@xmath218&&1 t + low - si&@xmath219&19/39&@xmath206&@xmath220 & 1.0 & 1.0&@xmath221&@xmath222&@xmath223&&1 t + high - fe&@xmath224&10/32&@xmath206&@xmath225 & 1.0&1.0&@xmath226&@xmath227&@xmath228&&1 t + ccccccc diff & & @xmath229 & @xmath230 & & @xmath231 & + ring & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & 2.17 + ring+ & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & 2.30 + inner diff . &
& @xmath242 & @xmath243 & & @xmath244 & + outer diff .
& & @xmath245 & @xmath246 & & @xmath247 & + cccccccc diff .
& 29.2 & 10.8 & 50.9 & 9.26 & 1.97 & 155.3 & 7.3 + ring & 1.1 & 11.9 & 2.3 & 8.64 & 2.34 & 6.5 & 0.3 + ring+ & 1.3 & 15.6 & 3.5 & 6.62 & 3.06 & 10.0 & 0.5 + inner diff .
& 5.8 & 23.5 & 18.9 & 3.95 & 3.69 & 67.1 & 2.7 + outer diff .
& 23.4 & 6.8 & 25.6 & 14.7 & 1.24 & 78.3 & 3.7 + cccc all diff & @xmath248 & @xmath249 & @xmath250 + all ring & @xmath251 & @xmath252 & @xmath253 + all ring+ & @xmath254 & @xmath255 & @xmath256 + inner diff . &
@xmath257 & @xmath258 & @xmath259 + outer diff . &
@xmath260 & @xmath261 & @xmath262 + hot halo emission s04 & 0.44 & 0.49 & @xmath263 + warm galactic halo & & 0.25 & 0.5 + type ii sn a96 & 0.19 & 0.37 & + type ii sn t95 & 0.31 & 0.47 & 0.49 + type ii sn w95a ( @xmath264 ) & 0.15 & 0.18 & 0.61 + type ii sn w95b ( @xmath264 ) & 0.29 & 0.22 & 0.43 + type ii sn w95a ( @xmath71 ) & 0.24 & 0.25 & 0.49 + type ii sn w95b ( @xmath71 ) & 0.31 & 0.31 & 0.46 + type ii sn n98s1 & 0.41 & 0.61 & 0.63 + type i sn tnh93 & -2.14 & -1.44 & -0.23 + type i sn wdd2 & -2.8 & -1.7 & 0.01 + type i sn w7 & -2.18 & -1.42 & -0.25 + llccccc ngc 1365 & 3 - 6 & 20.5 & @xmath265 & @xmath266 & @xmath267 & 1,2,3 + ngc 1365 ring & 3 - 6 & 21.1 & @xmath268 & @xmath269 & @xmath270 & 1,4 + m82 & 4 - 6 & 21.0 & @xmath271 & @xmath272 & @xmath273 & 5,6,7 + ngc 3351 & 4 - 5 & 20.7 & @xmath274 & @xmath275 & @xmath276 & 8,9,3 + antennae & 4 - 13 & 21.2 & @xmath277 & @xmath269 & @xmath278 & 10,11,3 + arp 220 & 1 - 3 & 21.8 & @xmath279 & @xmath280 & @xmath281 & 12,13,14 + ngc 253 & 20 - 30 & 22.0 & @xmath282 & @xmath283 & @xmath284 & 15,16,17 + galactic center & 2 - 7 & 22.8 & @xmath285 & @xmath286 & @xmath287 & 18,19 + | we present the first _
chandra_/acis imaging study of the circumnuclear region of the nearby seyfert galaxy ngc 1365 .
the x - ray emission is resolved into point - like sources and complex , extended emission .
the x - ray morphology of the extended emission shows a biconical soft x - ray emission region extending @xmath05 kpc in projection from the nucleus , coincident with the high excitation outflow cones seen in optical emission lines particularly to the northwest .
harder x - ray emission is detected from a kpc - diameter circumnuclear ring , coincident with the star - forming ring prominent in the _ spitzer _ mid - infrared images ; this x - ray emission is partially obscured by the central dust lane of ngc 1365 .
spectral fitting of spatially separated components indicates a thermal plasma origin for the soft extended x - ray emission ( @xmath1 kev )
. only a small amount of this emission can be due to photoionization by the nuclear source .
detailed comparison with [ oiii]@xmath2 observations shows the hot interstellar medium ( ism ) is spatially anticorrelated with the [ oiii ] emitting clouds and has thermal pressures comparable to those of the [ oiii ] medium , suggesting that the hot ism acts as a confining medium for the cooler photoionized clouds .
the abundance ratios of the hot ism are fully consistent with the theoretical values for enrichment from type ii supernovae , suggesting that the hot ism is a wind from the starburst circumnuclear ring .
x - ray emission from a @xmath3 pc long nuclear radio jet is also detected to the southeast . |
high performance coronagraphs with small inner working angle ( iwa ) are unavoidably very sensitive to small pointing errors and other low order aberrations @xcite .
this property is due to the fact that the wavefront of a source at small angular distance ( typically between 1 and 2 @xmath1 ) from the optical axis is `` similar '' ( in the linear algebra sense of the term ) to an on - axis wavefront with a small ( @xmath2 ) pointing error .
if the coronagraph must `` transmit '' the former , it will also transmit a significant part of the latter , and therefore be extremely sensitive to pointing errors .
this behavior is indeed verified by performance comparison between coronagraph concepts @xcite . for high performance coronagraphs , such as the phase - induced amplitude apodization ( piaa )
coronagraph @xcite used as example in this paper , the coronagraph should ideally be designed to balance iwa against stellar angular diameter , which sets a fundamental limit on the achievable coronagraph performance .
such coronagraphs are therefore pushed to become sensitive to pointing errors corresponding to the angular size of nearby stars , roughly 1 milliarcsecond ( mas ) , and are also highly sensitive to other low order wavefront errors such as focus and astigmatism .
milliarcsecond - level pointing error can increase stellar leakage in the coronagraph to the point where a planet would be lost in photon noise .
even smaller errors can create , if not independantly measured , a signal which is very similar to a planet s image .
a robust and accurate measurement of low order aberrations ( especially tip - tilt errors , which are easily generated by telescope pointing errors and vibrations ) is therefore essential for high contrast coronagraphic observations at small angular separation .
the science focal plane after the coronagraph is unfortunately `` blind '' to small levels of low order aberrations , which can only be seen when already too large to maintain high contrast in the coronagraphic science image .
a better option is to monitor pointing errors by using starlight which would otherwise be rejected by the coronagraph .
this scheme was successfully implemented on the lyot project coronagraph @xcite .
alternatively , the measurement could be performed independently from the coronagraph optical train ( for example , the wavefront sensor in the adaptive optics system upstream of the coronagraph ) .
we propose in this paper an improved solution to obtain accurate measurement of several low - order aberrations including pointing : the coronagraphic low - order wavefront sensor ( clowfs ) .
the wavefront control requirements for a piaa coronagraph are first clearly defined in [ sec : requ ] .
the clowfs principle is presented in
[ sec : principle ] .
wavefront reconstruction algorithms and clowfs sensitivity are discussed in [ sec : wfreconstr ] .
the aberration sensitivity of a piaa coronagraph equipped with a clowfs is discussed in [ sec : aberrsens ] , and the results of a laboratory demonstration on a piaa coronagraph system are shown in [ sec : labexp ] .
simulated science focal plane image in a piaac system with a 0.005 @xmath3/d pointing offset .
the piaa design adopted for this simulation has a 1.9 @xmath3/d iwa . ]
wavefront aberrations can be produced either before the piaa optics ( for example a bending of the telescope primary mirror , or a telescope pointing error ) or between the piaa optics and the focal plane mask . in a space - based telescope free of atmospheric turbulence , the strongest sources of aberration are likely to be telescope pointing errors and low - order aberrations due to structural deformations of the optical telescope assembly .
the piaa coronagraph is especially sensitive to such low order aberrations if they are introduced prior to the piaa optics @xcite , as they will then scatter light in the most scientifically precious area of the science focal plane : the inner part of the field .
as illustrated in figure [ fig : psf21 ] , a very small pointing error ( 0.5% of @xmath3/d ) can be sufficient to create an artefact as bright as an earth - like planet .
extreme sensitivity to pointing is unavoidable in small iwa coronagraphs , and , in the case of the piaa , is due to the fact that a `` remapped '' tip / tilt scatters light outside the focal plane mask .
coronagraphs with larger iwa and better tolerance to pointing errors exist , and even within the piaa `` family '' of coronagraph , sensitivity to pointing errors can be balanced against iwa by changing the size of the focal plane mask and the pupil apodization profile .
the piaa coronagraph is much less affected by aberrations after the beam shaping optics .
for example , a small amount of tip / tilt in the post - piaa optics will simply move the psf on the focal plane mask ( which is most likely slightly oversized ) without introducing scattering at larger separations .
in addition , low order aberrations introduced after the piaa optics will likely be much smaller in amplitude , thanks to the small size of optical elements between the piaa optics and the focal plane mask . for successful detection of a faint source ( assumed here to be an earth - like planet at @xmath4 contrast ) ,
low - order aberrations must simultaneously : * be small enough to avoid loosing the planet signal within the scattered starlight s photon noise .
for the piaa coronagraph design used in this work , a pre - piaa 0.005 @xmath3/d pointing error is sufficient to scatter @xmath5 of the starlight into the science field , and this scattered light produces two wide `` arcs '' on either side of the optical axis , at a @xmath4 contrast level .
although a piaa coronagraph could be designed with reduced sensitivity to pointing errors ( but with larger iwa ) , we assume here that the maximum allowable pointing error for a space piaa coronagraph mission aimed at direct imaging of `` earth - like '' planets is 0.005 @xmath3/d ( corresponding to starlight leak peaking at @xmath4 contrast ) . on a 1.4-m diameter telescope in the visible
, this corresponds to 0.4 mas , a value which is similar to the angular radius of a `` typical '' target ( a main - sequence star at 10 pc ) . on larger telescopes ,
the angular radius of the star ( 0.5 mas for a sun - like star at 10 pc ) will drive the coronagraph design , which will therefore also end up with a @xmath6 0.5 mas rms pointing error requirement . * be stable , or calibrated , to a fraction of the planet s expected flux .
we assume in this paper that coronagraphic leaks due to low order aberrations must be calibrated to 10% of the expected planet s contribution .
this second requirement is therefore much more severe .
light scattered in the science focal plane scales approximately as the square of the aberration amplitude : pointing must be stable ( or calibrated ) to @xmath7 ( 0.13 mas on a 1.4-m telescope in the visible ) for a planet at @xmath4 contrast .
fortunately , high accuracy measurement of low order aberrations by the clowfs scheme proposed in this paper can be used to reliably model the stellar leakage in the coronagraphic science focal plane .
this model can then be numerically subtracted from the science image to reveal much fainter underlying sources down to the photon noise and detector noise limits . pointing errors larger than 0.0016 @xmath1 ( but smaller than 0.005 @xmath1 ) are acceptable , as long as they are measured to 0.0016 @xmath1 accuracy .
a simplified optical layout for a clowfs system for a piaa coronagraph is shown in figure [ fig : lowfsprinciple ] .
clowfs light is extracted by the focal plane mask located after the piaa optics . in the piaa coronagraph design ,
the role of this focal plane is to selectively remove starlight , while transmitting the science field .
the mask is therefore illuminated by a large number of photons , which are freely available , and , if properly used , allow highly sensitive measurement of low order aberrations . the central part of the focal plane mask used in the clowfs design is opaque : only a reflective annulus around this central part sends light to the clowfs optics .
as shown in figure [ fig : lowfsprinciple ] we denote @xmath8 the radius of the inner opaque zone and @xmath9 the radius of the focal plane mask , which is fixed at the iwa of the coronagraph . for the baseline configuration adopted in this paper @xmath10 , with @xmath11 on the sky : the central 40% of the focal plane mask is opaque .
since the pupil before the focal plane mask is strongly apodized by the piaa optics , only a small fraction of the total light is reflected by the reflective annulus covering the @xmath12 range ( see figure [ fig : fluxr1r2 ] ) .
the clowfs detector acquires a defocused image of this annulus , at defocus distance @xmath13 waves ( in this paper , the defocus value is given as peak - to - valley in the post - piaa pupil plane ) .
while this defocus value may seem large , most of the light is in the central part of the apodized pupil , and the `` effective '' defocus is about half to a third of the optical defocus listed in this work .
the clowfs pupil is oversized by a factor 2 ( @xmath14 ) to include most of the light reflected by the clowfs focal plane mask .
fraction of the total starlight collected by the telescope sent to the clowfs detector as a function of @xmath15 . ] the clowfs defocus ( the clowfs detector is not conjugated to the focal plane mask ) is necessary to measure focus , but is not necessary if only tip and tilt are measured . in a clowfs configuration where the detector is conjugated to the focal plane mask , a single clowfs image can be used to recover the amplitude of focus in the incoming beam , but not its sign ( in an optical system free of aberrations , focal plane images acquired inside and outside focus are identical ) .
the clowfs defocus is therefore introduced to remove this sign ambiguity . a mathematical proof of the quadratic ( rather than linear ) response of a `` in - focus '' clowfs to pupil focus aberration is provided in [ ssec : lin ] .
while in an ideal system ( where photon noise would be the only source of noise ) the central opaque zone would reduce the clowfs performance , it is in practice extremely advantageous for two reasons , which are now described .
without the opaque zone , the clowfs detector would be illuminated by a large number of photons and the signal to be extracted would be a tiny relative change in intensity , which would be challenging to measure in practice . to detect this signal , the detector calibration would need to be very accurate and detector saturation would be a serious issue . masking the central part of the psf
greatly reduces the total amount of light in the clowfs but only slightly reduces the amplitude of the signal produced by low order aberrations .
for example , a small pointing error in the post - piaa pupil creates maximal signal where the psf surface brightness slope is the greatest .
the very center of the psf , although it contains a lot of flux , contains very little signal .
thanks to the central dark zone , microscopic changes in the coronagraphic wavefront produce macroscopic changes in the clowfs image , as shown in figure [ fig : lowfsims ] .
images obtained by the clowfs for @xmath16 ( left ) and @xmath10 ( right ) for a defocus distance of @xmath17 . with @xmath10 , the images are significantly fainter but aberrations are more easily seen . in each of the 8 images shown , the inside and outside focus images of the clowfs mask are side - by - side . in the clowfs system
we propose , only one of these two images would need to be acquired . ] the central dark zone of the focal plane mask allows the clowfs to be insensitive to small motions of the clowfs elements ( detector and off - axis parabola or lens to re - image the focal plane mask ) .
if the focal plane mask were fully reflective , a tip - tilt error in the post - piaa pupil ( which should be corrected ) and a tip - tilt error in the clowfs optics ( which should not be corrected ) would look identical .
this is due to the fact that the outer edge of the focal plane mask is dark , and the clowfs would therefore measure tip - tilt as a translation of the image on the clowfs detector .
this sensitivity to small amount of tip - tilt in the clowfs optics would therefore require the clowfs optical elements positions to be accurately known and stable . with the dark zone in the center of the focal plane mask
, however , the same tip - tilt is measured as a macroscopic change in the clowfs image shape ( see figure [ fig : lowfsims ] ) instead of a small translation .
thanks to the dark zone , the clowfs tip - tilt measurement is accurately referenced to the focal plane mask .
the geometry of the focal plane mask is driven by both the coronagraph architecture and the clowfs .
the transmission of the mask must first satisfy the coronagraph s requirements . implementing a clowfs
requires : * a tilted focal plane mask to redirect stellar light to the clowfs imaging optics .
if the tilt angle is large , or if the coronagraph requires a very circular focal plane mask , the focal plane mask could be made elliptical . * a reflective zone on the focal plane mask .
this requirement should have no negative impact on the coronagraph s performance , as a reflective coating may be deposited on the illuminated side of the focal plane mask without affecting its transmission .
the focal plane mask reflectivity defines the clowfs performance .
an `` ideal '' focal plane mask for clowfs , as described in this paper , may be very challenging to manufacture : * it is very difficult to make the central part of the mask truly `` black '' , and some of the light in the central part of the psf will be reflected into the clowfs imaging optics .
* the reflective annulus may not be uniformly reflective and could also introduce wavefront errors ( the reflective surface may not be very flat ) . * coatings , wether optimized to be black ( central part of the mask ) or reflective ( annulus )
, are somewhat chromatic : the reflectivity map of the focal plane mask is wavelength - dependent .
even with these imperfections , the clowfs will still produce a strong response for the aberrations to be measured ( tip , tilt , focus and their remapped equivalents ) .
it may however be difficult to predict what there responses will be if the focal plane mask is poorly calibrated . as detailed in [ sec : wfreconstr ] , we therefore propose to first measure these linear responses by introducing aberrations in the coronagraph optics , and then use these responses to decompose the clows image in a linear sum of aberrations .
this step requires no additional hardware , provided that there are actuators to correct for the aberrations measured by the clowfs .
this is very similar to what is commonly done on adaptive optics systems , where a `` response matrix '' is first acquired by measuring the wfs signal when each actuator of the deformable mirror is moved .
we show in this section that for small wavefront aberrations ( @xmath18 1 rad ) , the clowfs image is a linear function of the wavefront aberration modes to be measured .
this convenient property is due to : * the fact that the clowfs is not operating on a `` dark '' fringe : when the incoming wavefront is perfect , the clowfs images already contain some light . * the non - orthogonality between the aberration - free clowfs complex amplitude distribution in the clowfs detector array and the change introduced on this complex amplitude by the aberration modes to be measured .
we denote @xmath19 and @xmath20 the 2-d coordinate respectively in the post - piaa pupil plane and in the clowfs detector plane .
we denote @xmath21 the 2-d complex amplitude obtained by the clowfs for a perfect wavefront @xmath22 .
the complex amplitude @xmath23 obtained in the clowfs detector plane is a linear function of the pupil plane complex amplitude ( virtually all optical systems are linear in complex amplitude ) .
since by definition , @xmath24 for @xmath25 , we can therefore write @xmath26 where @xmath27 is a linear operator . the corresponding light intensity @xmath28 in the clowfs detector plane is @xmath29 + |m w(u)|^2\ ] ] which is linear function of @xmath30 as long as : @xmath31 , which helps to satisfy condition [ equ : lincond ] . if there is an aberration mode for which @xmath21 and @xmath32 are orthogonal ( @xmath33 = 0 $ ] for all values of @xmath20 ) , condition [ equ : lincond ] will not be satisfied even for small aberration levels , and the clowfs image @xmath34 will be a quadratic function of the wavefront aberration ( @xmath35 ) .
although such a situation is `` unlikely '' because orthogonality would need to occur simultaneously on each pixel of the clowfs detector , it does occur for the focus aberration mode if the clowfs detector is conjugated to the focal plane mask ( no defocus in the clowfs re - imaging ) . in this special case , since the clowfs re - images the focal plane reflective annulus without defocus , we can look at the orthogonality between @xmath21 and @xmath36 directly on the focal plane mask .
the complex amplitude @xmath21 is the fourier transform of the pupil plane complex amplitude and is purely real ( no imaginary part ) over the reflective annulus of the focal plane mask when no wavefront aberration is present .
@xmath32 is the change in the complex amplitude over the focal plane annulus introduced by a focus aberration in the pupil plane , and is therefore the fourier transform of @xmath37 where @xmath38 is the radial coordinate in the pupil plane , @xmath39 is the amplitude profile in the exit pupil of the piaa coronagraph ( @xmath39 decreases with @xmath38 ) , and @xmath40 is the phase in the exit pupil of the piaa coronagraph ( @xmath40 is the focus phase term remapped radially by the piaa optics ) . the approximation in equation [ equ : puppha ] is valid because we are considering small aberrations ( @xmath41 ) . since both @xmath40 and @xmath39 are real functions , the fourier transform of @xmath42 is purely imaginary .
we have therefore demonstrated that in this special configuration , the aberration - free clowfs complex amplitude distribution in the clowfs detector array and the change introduced on this complex amplitude by a focus aberration are perfectly orthogonal .
this problem can be solved by introducing a defocus in the clowfs re - imaging optics , as proposed in this paper .
this defocus term is equivalent to convolving the complex amplitude map in the annulus by a kernel which breaks the orthogonality .
numerical simulations show that the clowfs s response is linear to the modes considered in this paper .
this linearity is only valid for small amplitudes , as shown in figure [ fig : lintip ] for tilt .
linearity of the clowfs response to a tilt aberration . with a linear wavefront reconstruction algorithm ,
the measured tip signal ( solid line ) differs slightly from the true tip aberration ( dashed line ) .
this non - linearity reaches about 20% at 0.2 @xmath3/d . ]
as shown in figure [ fig : lowfs8mif ] , the main steps of the clowfs closed loop are : * acquire a single frame with the clowfs detector array * compute the difference between this frame and the `` reference '' image obtained ( or computed ) with no aberrations . * decompose this difference as a linear sum of modal responses .
these modal responses are either pre - computed or measured prior to starting the clowfs loop .
* the coefficients of the linear decomposition described above are used to drive actuators to remove the aberrations measured by the clowfs the estimation of low - order aberrations therefore requires knowledge of the modal responses ( how aberrations will linearly modify the image acquired by the clowfs ) . as shown in figure [ fig : lowfs8mif ] ( left ) , this information can consist of clowfs responses to a series of wavefront aberration modes , which are here zernike aberrations introduced either before ( m1 to m5 ) or after ( m6 to m8 ) the piaa optics .
they can either be pre - computed by simulation or measured as a response to an aberration introduced in the optical train ( this latter method is more robust as it will accurately account for unknown fabrication errors in clowfs optical components ) .
the wavefront control algorithm must avoid confusion which could arise from the fact that the same low order aberration will produce nearly the same clowfs signal ( although with a different amplitude ) wether it is introduced before or after the piaa optics : in the configuration adopted in this paper , responses to pre - piaa and post - piaa tip / tilt are 99.6% similar .
although post - piaa aberrations are expected to be smaller / slower , they still need to be properly corrected .
one must avoid compensating for a telescope pointing error ( pre - piaa tip - tilt ) by introducing a post - piaa tip - tilt , or compensating for a post - piaa tip - tilt aberration by a pre - piaa tip - tilt correction : both scenarios would create strong diffracted light at and beyond the iwa in the science focal plane even though the `` overall '' tip - tilt signal seen by the clowfs would be zero .
the reconstruction algorithm shown in figure [ fig : lowfs8mif ] addresses this issue by creating `` differential '' tip - tilt and focus modes ( m6 , m7 , m8 ) obtained by subtraction of pre - piaa responses m1,m2 , m3 from post - piaa responses m6 , m7 , m8 .
these differential modes track the difference between pre and post - piaa aberrations , and can only be measured slowly since the corresponding clowfs response is weak ( due to the high degree of similarity between modes m1 , m2 , m3 and m6 , m7 , m8 ) . since most tip - tilt /
focus aberrations originate before the piaa optics , the proposed algorithm offloads all of the tip / tilt signal to telescope pointing corrections ( pre - piaa ) .
differential aberrations are measured more slowly and off - loaded as post - piaa corrections . in the example given here , thanks the large gap between pre - piaa and post - piaa pointing control bandwidths ( @xmath6 khz vs. @xmath6 hz ) the differential pointing signal can be sent to a post - piaa tip / tilt corrector rather than a combination of pre - piaa and post - piaa tip - tilt .
the accuracy of the clowfs loop is ultimately limited by the measurement accuracy ( assuming ideal actuators ) . for small wavefront excursions ,
the linear model is a very good approximation and the measurement accuracy is driven by detector noise and photon noise . for optimal performance ,
the clowfs exposure time and loop gain need to be carefully chosen to match the dynamic properties of the wavefront aberrations . in principle , in a very stable environment , the clowfs loop could correct very small errors by averaging a large number of measurements ( small loop gain ) .
quantitative estimates will be given in
[ sec : aberrsens ] assuming a photon - noise limited detector .
the closed loop controller proposed in this paper is based on the linearity of the clowfs response . for large wavefront excursions ,
the actual clowfs response can differ significantly from this linear model , and the closed loop may `` unlock '' .
while we have not performed a detailed loop stability analysis , we expect the clowfs loop to stay locked as long as pointing excursions are within a few tenths of @xmath1 : * figure [ fig : lintip ] shows that the clowfs is still very close to the linear model at 0.2 @xmath1 , and , more importantly , the curve shown in this figure is smooth and monotonic . with a 0.2 @xmath1 pointing excursion ,
the estimate obtained from our linear model will not be exact , but will be sufficiently good to bring back the pointing significantly closer to on - axis at the next iteration * as illustrated in figure [ fig : lowfsims ] , the clowfs recovers pointing errors by essentially looking for a brightness enhancement on one side of the re - imaged reflective ring and a corresponding decrease in surface brightness on the opposite side of the ring .
as long as pointing errors are small enough for the peak of the stellar image to stay within the inner edge of the reflective ring , this scheme will correctly estimate the direction of the pointing error , and the amplitude of this error will be estimated with sufficient precision to reduce the pointing error after correction . since the inner edge of the reflective
mask is at least 0.5 @xmath1 , the pointing loop is expected to stay stable over this range .
laboratory operation of the clowfs confirm this behavior , and the closed loop locks even if the initial pointing error is 0.5 @xmath1 .
if the initial pointing error is too large for the linear clowfs loop to lock , a non - linear model of the clowfs may be used .
even with a non - linear model , the clowfs will fail to measure pointing errors if the stellar psf misses the focal plane mask ( pointing error is as large or larger than the mask size ) .
a separate coarse pointing sensor is then necessary to measure the pointing offset and then bring the stellar psf within the clowfs range .
the coronagraph image may be used for this purpose , provided that detector saturation does not prevent pointing error measurement . in this section
, we explore how the clowfs performance is affected by the choice of @xmath15 ( which sets the amount of light sent to the clowfs ) and the amount of @xmath43 introduced in the image of the focal plane mask . for each choice of the parameters @xmath15 and @xmath43 , the linear algorithm described in [ ssec : wfca ]
is used here to measure the clowfs sensitivity to pointing ( modes m1 , m2 ) , focus ( mode m3 ) , astigmatism ( modes m4 , m5 ) , differential tip - tilt ( modes m6,m7 ) and differential focus ( mode m8 ) .
the linear algorithm allows rapid evaluation of a large number of designs : 2-d sensitivity maps on a @xmath15 - @xmath43 plane can therefore be created , as shown in figure [ fig : map3d ] . under the ideal conditions used in this simulation ( perfect detector ) ,
figure [ fig : map3d ] shows that all aberrations are best measured if @xmath15 is close to zero : ideally , the focal plane mask should be fully reflective .
figure [ fig : map3d ] however shows that there is relatively little loss in sensitivity if @xmath15 is increased to up to @xmath44 . with
only 10% of the starlight reflected to the clowfs ( @xmath45 ) there is almost no loss in clowfs sensitivity to pointing errors .
for the pointing error to double , @xmath15 must be increased to 0.44 , leaving only 0.87% of the starlight for the clowfs .
if the same 0.87% were applied uniformly over the psf ( instead of masking the center of the psf ) , pointing errors would increase more than tenfold .
a similar behavior is observed for all other modes shown in figure [ fig : map3d ] : for example , the focus measurement error is doubled for @xmath46 , leaving only 0.06% of the starlight for the clowfs .
these results confirm the `` relative signal amplification '' effect previously claimed : the central dark part of the mask removes most of the clowfs flux but only mildly reduces the signal .
the clowfs design adopted in this paper ( @xmath10 , @xmath43 = 20 rad pv ) uses only 1.8% of starlight , at the expense of almost doubling tip - tilt and focus measurement errors .
the amount of @xmath43 has little effect on the clowfs sensitivity to pointing , unless it is sufficiently large ( @xmath47 rad ) to `` blend '' together light from opposite sides of the reflective focal plane ring .
some @xmath43 is however essential to allow good focus estimation with the clowfs , and helps with astigmatism as well . for focus
, there is however no gain beyond @xmath48 rad , at which point the loss in pointing sensitivity is still very small . for all pre - piaa low order aberrations , figure [ fig :
map3d ] shows that the residual sensing errors with a clowfs are within a factor @xmath62 of the theoretically optimal sensitivity .
for example , the clowfs pointing error is @xmath6 1 rad per mode per photon , or @xmath6 1.4 rad for tip and tilt combined .
this corresponds to just under 1 @xmath3/d pointing error for a single photon , even if the central opaque zone of the focal plane mask removes most of the starlight .
similarly , focus and astigmatisms are recovered with @xmath6 2 rad rms error each for a single photon .
the clowfs therefore makes a very efficient use of a limited number of photons .
due to the strong similarity between pre - piaa and post - piaa modes , the clowfs sensitivity to differential modes such as m6 , m7 or m8 , is much weaker . beyond the global trends outlined above , figure [
fig : map3d ] also shows more subtle effects : diffraction effects produce oscillations of the sensitivity to focus and astigmatism in the @xmath15 - @xmath43 plane . computing 2d sensitivity maps such as the ones shown in figure [ fig : map3d ] is therefore essential for fine tuning the clowfs performance to low order wavefront control requirements .
l c c required pointing calibration accuracy ( @xmath49 contrast ) & + maximum rms pointing excursion ( @xmath4 contrast ) & + required sampling time & 5 s & 38 @xmath50s + maximum allowed uncalibrated pointing drift rate & 0.026 @xmath51 & 3.4 arcsec / s + we consider here a 1.4-m diameter telescope observing a @xmath52 star in a 0.2 @xmath50 m wide band centered at @xmath53 .
we assume a 50% system throughput , which corresponds to @xmath54 at the telescope entrance .
as described in [ sec : requ ] , pre - piaa low order aberrations can affect coronagraphic contrast at small angles ( close to the iwa of the piaa coronagraph ) much more easily than post - piaa low order aberrations .
we therefore only consider in this section pre - piaa pointing errors , which could be generated by a telescope / spacecraft pointing error .
we adopt the requirement defined in [ sec : requ ] : pointing error must be measured to a 0.0016 @xmath1 1-@xmath55 accuracy . without clowfs ,
a 0.0016 @xmath3/d pointing error would have to be measured from the corresponding @xmath56 total coronagraphic leak in the science focal plane , equal to approximately @xmath57 .
a 1-@xmath55 measurement of this leak can be achieved with @xmath6 1 photon , provided that this leak is interferometrically combined with a much brighter and well known `` reference(s ) '' ( in 2-d , this technique is referred to a focal plane wavefront sensing , where a deformable mirror is used to mix coherent starlight with the scattered light halo ) .
this `` reference(s ) '' is necessary to measure the sign of the aberration and also to bring the signal above the detector readout noise and/or incoherent background in the image . in this scheme ,
detection of a faint companion can not be done at the same time as pointing measurement , and time must be shared between the two tasks ( we assume here that 50% of the time is spent for each task ) .
a measurement of pointing error with a 0.0016 @xmath1 1-@xmath55 error would therefore require 5 seconds . as shown in figure [ fig : map3d ] , the tip sensitivity for the clowfs is 1.2 rad rms for one photon at the telescope entrance .
the 0.0016 @xmath1 tip measurement accuracy ( equal to 0.0025 rad rms ) therefore requires @xmath58 photons , which can be gathered in 38 @xmath50s .
the clowfs is therefore capable of measuring pointing errors about 130,000 times faster than the science focal plane .
the pointing stability requirement is derived in table [ tab : perfsumm ] from both the required pointing calibration accuracy ( equal to 0.0016 @xmath1 in this example ) and the sampling time necessary to measure pointing errors to this level of accuracy .
the ratio between these two quantities defines a maximum allowable pointing drift rate beyond which the stability / calibration requirement previously defined can not be met .
as shown in table [ tab : perfsumm ] , measuring pointing errors with the clowfs is several orders of magnitude quicker than if only the light in the science focal plane were used .
this measurement sensitivity gain yields a much more relaxed pointing drift rate requirement : @xmath59 with a clowfs as opposed to @xmath60 without .
the clowfs camera requires a modest number of pixels : the signals shown in the left of figure [ fig : lowfs8mif ] contain little high spatial frequencies and can be accurately measured with approximately 10 pixels across one defocused spot images .
a 20 by 20 pixel window ( 400 pixels ) is sufficient , and can be read rapidly ( @xmath6110 khz ) with current technology .
the clowfs measures pointing and other low order aberration by detecting changes in the defocused image it acquires .
temporal changes in the detector response must therefore be small compared to the expected signal .
the clowfs is not sensitive to static spatial variations in the detector response ( flat field ) , as the signal is extracted from a difference between two images acquired at different times .
the effect of spatially uniform flat field variations can be removed by scaling of the images prior to this subtraction .
a 0.0016 @xmath3/d pointing offset corresponds to a 2% change in the surface brightness on the bright ring images by each frame of the clowfs : one side of the ring is 2% brighter while the opposite side is 2% fainter .
a comparable variation in detector sensitivity between the two sides of the ring images is very unlikely in modern detectors ( visible ccds ) , even over the course of several hours .
thanks to the large number of photons collected by the clowfs , a moderate amount of readout noise ( few photo - electron ) will still allow operation of the clowfs at high sampling rate .
for example , at 10 khz sampling rate , the defocused image contains 10800 photons on a @xmath52 target . assuming that the `` ring '' occupies a 50 pixel surface area on the detector , photon noise ( 15 photo - electrons per pixel ) is likely to be larger than readout noise on modern visible detectors .
the clowfs has been implemented in the piaa coronagraph testbed at the subaru telescope . as shown in fig .
[ fig : labsetup ] , the clowfs is driving 5 piezo actuators to move the light source in x , y , z and move a post - piaa mirror in tip and tilt .
schematic representation of the clowfs implementation for the piaa coronagraph testbed at subaru telescope .
the clowfs is extracting light reflected by the focal plane mask and drives 5 actuators : pre - piaa tip / tilt / focus ( the light source is mounted on a 3 axis piezo stage ) and post - piaa tip / tilt . ]
the clowfs focal plane mask used in this experiment is shown in fig .
[ fig : labmask ] , and was manufactured by lithography techniques .
the mask is not `` ideal '' : the central portion is not perfectly opaque , and reflects a few percent of the light .
since most of the starlight falls on this central part of the mask , the clowfs frame contains a bright peak at the center of the defocused mask image .
thanks to the defocus introduced in the clowfs , the fainter light reflected by the reflective annulus interferes with this central peak , and also forms fainter outer rings visible as shown in fig .
[ fig : labplots ] . with a finite reflectivity of the central part of the focal plane mask
, a more optimal design would be to reduce the size of the central `` dark '' area to increase the amount of light at the transition between the central dark spot and the reflective annulus .
clowfs focal plane mask used in the piaa coronagraph laboratory testbed at subaru telescope ( fabricated by hta photomask ) .
the 100 micron radius mask center is opaque ( low reflectivity ) , and is surrounded by a 100 micron wide highly reflective annulus .
the science field , transmiting light to the science camera , extends from 200 micron to 550 micron radius . ]
the linear scheme implemented for calibration and processing of the clowfs data is very insensitive to the details of the clowfs design , and the clowfs is able to measure simultaneously both pre and post - piaa tip / tilt with little cross - talk ( see fig . [
fig : labplots ] , upper right ) . as shown in fig .
[ fig : labplots ] , we have achieved @xmath0 closed loop pointing stability for both the light source position and post - piaa tip / tilt . in our laboratory testbed ,
pre - piaa pointing errors and post - piaa tip / tilt are similar in amplitude , while in an actual system , we would expect pre - piaa pointing errors ( telescope pointing , primary and secondary mirrors tilts ) to be much larger and faster than post - piaa tip / tilt ( most likely due to slow thermal drifts on small optics ) . on our testbed
, we have therefore operated the control loop with similar temporal bandwidth for both pre and post - piaa aberrations , unlike the control scheme proposed in fig .
[ fig : lowfs8mif ] .
the performance we have achieved in the laboratory is limited by both system stability ( our testbed is in air and includes 75 mm diameter optics separated by more than a meter ) and clowfs loop speed . in our experiment ,
the clowfs sampling interval was limited to 25 s due to hardware limitations ( readout speed of the camera image and time necessary to transfer the image from the computer which controls the camera to a separate computer performing the clowfs image analysis ) . for
a space coronagraphic mission , a better controlled environment and a faster readout camera ( 10 khz is reasonable for the small number of pixels needed ) would allow higher performance . despite these limitations ,
our laboratory demonstration of the clowfs concept has exceeded both the rms pointing excursion and the pointing measurement accuracy required for achieving 1e10 contrast at 2 @xmath1 with a visible piaa coronagraph mission .
the clowfs design presented in this paper can efficiently measure low order aberrations `` for free '' , as it uses light that would otherwise be discarded by the coronagraph .
both the hardware configuration and software algorithms presented are easy to implement and their performance is robust against calibration errors , chromaticity , non - common path errors and small errors / aberrations in the optical components .
the clowfs pointing measurement can also lead to improved astrometric accuracy for the position of faint companions , as the star position on the image is usually difficult to measure in coronagraphic images @xcite . in this paper , we have studied a clowfs design on a low - iwa piaa coronagraph . although coronagraphs with larger iwa can tolerate larger amount of low order aberrations ( see for example kuchner & traub 2002 ; kuchner et al . 2005 ) , they require these aberration to be measured ahead of the coronagraphed beam because they are `` blind '' to aberrations until they are large enough to produce significant coronagraph leaks
. the clowfs would therefore be very useful to any high contrast coronagraph , and the optical design presented in this paper can readily be used on any coronagraph where a focal plane mask physically blocks starlight .
the clowfs can also be applied to phase mask coronagraphs . for such coronagraphs
( see for example roddier & roddier 1997 ; rouan et al .
2000 ; palacios 2005 ) , where starlight is diffracted outside the pupil by the focal plane mask , a modified lyot stop is placed on the pupil plane to reflect starlight to the clowfs .
in addition , using a pattern matching algorithm , the clowfs can estimate low - order wavefront aberrations accurately and quickly even in non - linear region .
these new clowfs designs and their performances will be presented in an upcoming paper .
belikov , r. , kasdin , n.j . , & vanderbei , r.j .
2006 , , 652 , 833 digby , a.p .
2006 , , 650 , 484 guyon , o. 2003 , , 404 , 379 guyon , o. , pluzhnik , e.a . ,
galicher , r. , martinache , f. , ridgway , s.t . , & woodruff , r.a .
2005 , , 622 , 744 guyon , o. , pluzhnik , e. a. , kuchner , m. j. , collins , b. , & ridgway , s. t. 2006 , , 167 , 81 kasdin , n.j . , vanderbei , r.j . ,
spergel , d.n . , & littman , m.g .
2003 , , 582 , 1147 kuchner , m.j . , & traub , w.a .
2002 , , 570 , 900 kuchner , m.j . ,
crepp , j. , & ge , j. 2005 , , 628 , 466 lloyd , j.p . , & sivaramakrishnan , a. 2005 , , 621 , 1153 oppenheimer , b. r. , et al .
2004 , , 5490 , 433 palacios , d.m .
2005 , , 5905 , 196 roddier , f. , & roddier , c. 1997 , , 109 , 815 rouan , d. , riaud , p. , boccaletti , a. , clnet , y. , & labeyrie , a. 2000 , , 112 , 1479 shaklan , s.b .
, & green , j.j .
2005 , , 628 , 474 sivaramakrishnan , a. , soummer , r. , sivaramakrishnan , a.v . ,
lloyd , j.p . , oppenheimer , b.r . , & makidon , r.b .
2005 , , 634 , 1416 | high contrast coronagraphic imaging of the immediate surrounding of stars requires exquisite control of low - order wavefront aberrations , such as tip - tilt ( pointing ) and focus .
we propose an accurate , efficient and easy to implement technique to measure such aberrations in coronagraphs which use a focal plane mask to block starlight .
the coronagraphic low order wavefront sensor ( clowfs ) produces a defocused image of a reflective focal plane ring to measure low order aberrations . even for small levels of wavefront aberration ,
the proposed scheme produces large intensity signals which can be easily measured , and therefore does not require highly accurate calibration of either the detector or optical elements .
the clowfs achieves nearly optimal sensitivity and is immune from non - common path errors .
this technique is especially well suited for high performance low inner working angle ( iwa ) coronagraphs . on phase - induced amplitude apodization ( piaa )
type coronagraphs , it can unambiguously recover aberrations which originate from either side of the beam shaping introduced by the piaa optics .
we show that the proposed clowfs can measure sub - milliarcsecond telescope pointing errors several orders of magnitude faster than would be possible in the coronagraphic science focal plane alone , and can also accurately calibrate residual coronagraphic leaks due to residual low order aberrations .
we have demonstrated @xmath0 pointing stability in a laboratory demonstration of the clowfs on a piaa type coronagraph . |
semiclassical physics emerged during the development of quantum mechanics , almost one century ago , to account for the transition from wave physics in the quantum regime to classical mechanics @xcite .
today , semiclassical physics plays an essential role in quantum chaos @xcite , and its methods are applied in virtually any field that features wave dynamics , including acoustics @xcite , electromagnetism @xcite , hydrodynamics @xcite , and loop quantum gravity @xcite .
we demonstrate the power of these methods when applied to optical microresonators by studying square - shaped polymer - based microlasers .
their emission properties are little understood for two reasons .
firstly , the square resonator with dirichlet boundary conditions is a standard example of a separable system , whereas the dielectric square resonator is nonseparable and hence nonintegrable due to the diffraction at the dielectric corners .
this remains an open problem in mathematical physics with tremendous impact on radar or telecommunication applications @xcite .
secondly , it was observed that these lasers emit very narrow lobes in only a few directions .
directional emission from microlasers has been intensely investigated due to possible applications @xcite and observed for , e.g. , limaon- and stadium - shaped microlasers .
the underlying mechanism is understood in terms of their chaotic ray dynamics @xcite .
however , these explanations can not be applied to polygonal resonators since they do not exhibit chaotic ray dynamics .
while square and hexagonal microresonators and have been extensively studied , their far - field distributions have been scarcely investigated @xcite .
this is partly because rounded corners significantly influence the resonance frequencies and field distributions @xcite , and it is technologically challenging to fabricate sharp corners . here
we investigate the far - field emission of microlasers with sharp corners ( i.e. , with a radius of curvature much smaller than the wavelength ) . a semiclassical model for the dielectric square @xcite
is used to predict the far - field distributions , which are in very good agreement with numerical calculations .
careful comparison of the measured far - field distributions with the model allows to identify the lasing modes and understand the mechanism for their high directionality .
the microlasers consisted of a pmma polymer matrix doped by @xmath0 wt% dcm laser dye that was deposited in a @xmath1 nm - thick layer on a si / sio@xmath2 ( @xmath3 m ) wafer by spin - coating .
square cavities with side length ranging from @xmath4 to @xmath5 m were engraved by electron - beam lithography @xcite , which makes it possible to obtain two - dimensional cavities with nanoscale precision [ see ( a ) ] .
the microlasers were pumped by a pulsed , frequency - doubled nd : yag laser ( @xmath6 nm , @xmath7 ns , @xmath8 hz ) .
the short pulses and low duty cycle help to avoid problems from heating and quenching due to dark states .
the pump beam impinged vertically and it covered the whole cavity uniformly the lasing emission was collected in the sample plane by a lens and transferred to a spectrometer by an optical fiber .
it was polarized parallel to the sample plane .
the far - field intensity distributions were measured as a function of the azimuthal angle @xmath9 by rotating the cavity .
a typical spectrum is shown in .
it features a sequence of equidistant peaks .
its fourier transform ( right inset of ) shows that the free spectral range ( fsr ) of the spectrum corresponds to the optical length of the diamond orbit @xcite .
therefore it is expected that the observed lasing modes are localized on trajectories with an angle of incidence close to @xmath10 .
a typical far - field intensity distribution of a single resonance is presented in ( a ) .
it exhibits four emission lobes in the directions parallel to the side walls .
the emission lobes are very narrow with a full width at half maximum of about @xmath11 [ see ( b ) ] .
figure [ fig : images](b ) shows an image of the lasing cavity taken by a camera with a high - magnification zoom lens .
it evidences that the lasing emission ( red ) is emitted from the side walls parallel to the emission direction , whereas the side walls perpendicular to it do not emit but only scatter the green pump laser .
since the emission is at a grazing angle , it must stem from rays impinging on the cavity side walls with an angle of incidence very close to the critical one , @xmath12 , where @xmath13 is the refractive index @xcite .
this is in agreement with the previous observation that the ray trajectories on which the lasing modes are based have an angle of incidence close to @xmath10 .
the far - field distributions are calculated using the semiclassical model introduced in ref .
@xcite to identify the observed lasing modes .
the model is based on the observation that the wave functions are localized on classical tori and are thus composed of eight plane waves @xmath14 .
their directions are related to each other by the symmetry operations of the @xmath15 point group .
the corresponding ray trajectories in are given by the momentum vectors @xmath16 and their rotations by 90@xmath17 , @xmath18 .
to quantize the spectrum , the simplest approach is to separate the system along the @xmath19 and @xmath20 directions , and write two phase loop conditions analogous to that of a fabry - prot cavity @xcite , @xmath21 the main difference from a fabry - prot cavity are the fresnel coefficients @xmath22 corresponding to an incidence with angles @xmath23 $ ] instead of @xmath24 .
the momentum vector components @xmath25 can thus be formally written as @xmath26 \ } / a \ , , \\
k_y & = & \{\pi m_y + i \ln[r(\alpha_y ) ] \ } / a \ , .
\end{array}\ ] ] the resonance wave number is @xmath27 where @xmath28 is the free - space wavelength . note that @xmath29 are in general complex - valued .
all wave functions @xmath30 are obtained by adding up the @xmath31 plane waves with their correct momentum vectors and signs . for transverse magnetic ( tm ) [ transverse electric ( te ) ] polarization ,
the electric ( magnetic ) field is parallel to the plane of the cavity , and @xmath32 corresponds to the @xmath33 component of the electric ( magnetic ) field .
the resonances [ i.e. , solutions of ] are labeled by their quantum numbers @xmath34 and their symmetries @xmath35 ( @xmath36 ) with respect to the diagonal @xmath37 ( @xmath38 ) as @xmath39 , where @xmath40 ( @xmath41 ) means a symmetric ( antisymmetric ) wave function .
there are six symmetry classes corresponding to the different combinations of @xmath42 and the quantum numbers that are labeled by the mulliken symbols @xmath43 , @xmath44 and @xmath45 ( see ) . [ cols="^,^,^,^,^ , < " , ] green s identity is used to infer the far - field distribution @xcite .
the derivation is detailed in the supplementary materials along with the comparison to numerical simulations ( see figs . [
fig : specnum ] and [ fig : ffnum ] ) . the full expression for the far - field distribution of an @xmath46 mode
is given in .
it consists of @xmath31 terms like @xmath47 \end{array}\ ] ] where @xmath48 ( @xmath49 ) for tm ( te ) polarization and @xmath50 .
each such term corresponds to emission in the two directions @xmath51 defined by the roots of the argument of the @xmath52 function .
if the angle of incidence of the corresponding rays inside the resonator is smaller than @xmath53 , then @xmath51 is simply given by snell s law ( see ) , and the emission lobe is called _
refractive_. then the @xmath52 term is in fact the diffraction pattern of a plane wave going through a slit with a width @xmath54 that is the projection of the side of the square on the emission direction ( see ) . when @xmath51 is complex , corresponding to rays with angle of incidence larger than @xmath53 , the lobe is called _ nonrefractive _ and is emitted parallel to the side wall .
the nonrefractive lobes are jagged and much broader .
they have a much smaller amplitude than the refractive lobes due to the finite imaginary part of @xmath51 .
while no energy is transmitted when a plane wave impinges on an infinite dielectric interface with an angle larger than the critical one ( i.e. , it is totally reflected ) , this is not the case for an interface of finite length as illustrated by the existence of the nonrefractive emission lobes @xcite , which leads to a modification of the reflection and transmission coefficients .
two different cases of far - field diagrams can appear for @xmath55 , depending on @xmath56 . if @xmath57 , the plane waves impinging on a side wall with angle @xmath58 escape refractively , whereas those impinging with @xmath59 are totally reflected .
hence @xmath31 refractive and @xmath31 nonrefractive lobes are observed , where the latter are typically negligible compared to the former . if @xmath60 , all plane waves are totally reflected , and @xmath61 nonrefractive lobes are observed , though their directions are four - fold degenerate .
we now compare the model with experiments . a spectrum generated by the model and corresponding to the experimental conditions
is shown in .
the resonances are arranged in branches .
each of them consists of modes with identical longitudinal quantum number @xmath62 and increasing transverse quantum number @xmath63 as @xmath58 decreases @xcite .
the horizontal distance between these branches agrees well with the experimentally observed fsr , corresponding to the length of the diamond orbit .
therefore we conclude that the observed lasing resonances belong to different branches , i.e. , each has a different @xmath64 .
we use the far - field intensity distributions to infer @xmath58 and thus their transverse quantum numbers .
the model far - field patterns of three representative modes are presented in .
they are labeled by ( a ) , ( b ) , and ( c ) in .
mode ( c ) is located just below the critical angle ( @xmath65 ) and features two refractive emission lobes between @xmath66 and @xmath67 , while its nonrefractive lobes are not visible due to their negligible amplitude . since the measured far - field patterns feature only @xmath68 lobes in the total range of @xmath69 , such modes with @xmath57 can be excluded .
mode ( b ) with @xmath70 is located just above the critical angle .
the refractive lobes have become nonrefractive and merged into four narrow lobes parallel to the sides .
this is precisely the kind of far - field pattern observed experimentally .
mode ( a ) finally is well above the critical angle ( @xmath71 ) .
the nonrefractive lobes have become very broad and jagged unlike what was observed experimentally .
also their amplitude has decreased significantly .
this development continues with increasing @xmath58 .
the qualitative comparison with the measured far - field intensity distribution shows that the experimental lasing modes are only consistent with mode ( b ) , i.e. , an angle of incidence just above the critical angle .
this is confirmed by the fits described in the following .
the experimental far - field distribution measured with an angular resolution of 1@xmath17 as in is compared to the envelope of the model far - field distributions .
this is detailed in the supplementary materials .
the only two independent parameters are the amplitude and @xmath58 , which is considered a continuous variable due to the high resonance density .
the fit of the lobe yields @xmath72 and is plotted in ( b ) .
the agreement is excellent .
similar results are obtained for the other lobes and the other resonances as well as squares with various sizes , yielding values of @xmath73 from @xmath74 to @xmath75 . as expected from the qualitative considerations in the previous paragraph and the photograph in ( b ) , the angle of incidence is slightly above the critical angle .
in fact , the model predicts far - field distributions with a very quickly oscillating substructure that stems from the terms in the square brackets in , namely the @xmath76 and @xmath77 functions with the argument @xmath78 .
the actual model far - field intensity distribution for the fitted parameters is shown as thin green line in ( b ) . to confirm this prediction , the angular resolution of the setup
was improved by putting a slit in front of the collection lens .
the far - field intensity distribution measured this way is shown in the inset of ( b ) .
it features oscillations with a period of @xmath79 , while the expected period is @xmath80 according to .
the agreement is excellent for different wavelengths and cavity sizes .
systematic measurements with increased angular resolution will allow for a better determination of the symmetry of the lasing resonances .
furthermore , the symmetry of the far - field distributions can be changed by the polarization of the pump laser due to the fluorescence anisotropy of the laser dye @xcite : when the square microlaser is pumped with linear polarization parallel to the @xmath19 axis [ instead of circular polarization as in ( a ) ] , the emission is almost bidirectional with strong emission lobes at @xmath81 and @xmath82 , whereas the emission lobes at @xmath83 and @xmath84 are strongly suppressed ( not shown ) .
this effect is a consequence of the lower degree of symmetry induced by the pump polarization .
it can be explained by the formation of coherent superpositions of degenerate mode pairs like the @xmath85 and @xmath86 modes .
it should be noted that the lasing modes and hence far - field emission of microlasers can also be controlled by localized or selective pumping @xcite , whereas controlling it only with the pump polarization is an experimentally simpler approach .
these topics are , however , beyond the scope of this article and will be explored in a future publication .
the far - field intensity distributions of square - shaped organic microlasers were investigated and revealed highly directional emission in the four directions parallel to the cavity sides .
sharp corners were essential for obtaining this high directionality because rounded corners significantly change the emission behavior @xcite .
analytical formulas for the far - field distributions based on a semiclassical model @xcite were developed and showed excellent agreement both with experimental data and with numerical simulations .
they could be easily adapted to other polygonal cavities like hexagon or rectangles @xcite and other refractive indices .
the comparison between the model and experiments evidences that the lasing modes are based on trajectories with an angle of incidence almost equal to the critical angle that hence leave the resonator at a grazing angle .
interestingly , these modes do not feature the highest quality factors ; these would be modes with @xmath58 larger than @xmath53 .
maybe these modes actually lase , but are not observed because of their lower amplitude in the far - field compared to other modes . or they may actually have higher losses due to diffraction at the substrate interface @xcite
. the semiclassical model does not take into account diffraction by the corners , whereas it is precisely this effect which renders the system nonintegrable .
while the model calculations agree very well with the observed spectrum and far - field distributions , there are some minor observations that can not be explained by them . for example
, some light is emitted into other directions than the four main emission lobes due to diffraction at the corners , but it is much less intense .
this means that diffraction by the corner plays a secondary , but not negligible role for the emission from polygonal microlasers .
studying these effects would allow more insight into the open physical - mathematical problem of diffraction by a dielectric wedge .
the numerical calculations are based on a code developed by c. schmit .
s. b. gratefully acknowledges funding from the european union seventh framework programme ( fp7/2007 - 2013 ) under grant
this work was supported by a public grant from the laboratoire dexcellence physics atom light matter ( labex palm ) overseen by the french national research agency ( anr ) as part of the investissements davenir program ( reference no .
anr-10-labx-0039 ) .
27ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
_ ( , , ) @noop _ _ ( , , ) \doibase
doi:10.1088/1751 - 8113/40/50/r01 [ * * , ( ) ] link:\doibase 10.1103/physreve.63.026208 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.011301 [ * * , ( ) ] link:\doibase 10.1109/tap.2014.2364305 [ * * , ( ) ] `` , '' ( , , ) chap . , pp . link:\doibase 10.1364/josab.21.000923 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.033901 [ * * , ( ) ] link:\doibase 10.1103/physreva.67.023807 [ * * , ( ) ] link:\doibase 10.1109/jqe.2007.897879 [ * * , ( ) ] link:\doibase 10.1364/ol.34.001810 [ * * , ( ) ] link:\doibase 10.1063/1.3276069 [ * * , ( ) ] link:\doibase 10.1109/jqe.2005.846696 [ * * , ( ) ] link:\doibase 10.1063/1.4757572 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.023807 [ * * , ( ) ] link:\doibase 10.1103/physreve.88.062906 [ * * , ( ) ] link:\doibase 10.1103/physreve.90.052909 [ * * , ( ) ] link:\doibase 10.1063/1.4720474 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.023830 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.036208 [ * * , ( ) ] link:\doibase 10.1103/physreva.86.043817 [ * * , ( ) ] \doibase doi:10.1063/1.1949708 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.033903 [ * * , ( ) ] link:\doibase 10.1063/1.4883637 [ * * , ( ) ] link:\doibase 10.1088/1464 - 4258/5/1/308 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.53.1711 [ * * , ( ) ]
we use green s identity , @xmath87 to infer the far - field distribution ( cf . ref .
@xcite ) , where @xmath88 is the derivative with respect to the surface normal of the domain @xmath89 and @xmath90 is the green s function in two dimensions with @xmath91 being the wave number .
we choose the boundary of the resonator as the integration path @xmath92 and assume continuity of @xmath32 and @xmath93 at the resonator boundaries where @xmath48 ( @xmath49 ) for tm ( te ) polarization @xcite .
then the wave function outside of the resonator , @xmath94 , is related to that inside , @xmath95 , by @xmath96 we start with a single plane wave @xmath97 inside the resonator and calculate the corresponding outside wave function @xmath98 in the far field .
the integral can be simplified in this case where @xmath99 and @xmath100 with @xmath101 by using @xmath102 for @xmath103 .
furthermore , in this limit @xmath104 . the integration yields @xmath105 \sin\left[(k_y - k \sin{\varphi } ) \frac{a}{2}\right ] \left [ \frac{\mu k_x + k \cos{\varphi}}{k_y - k \sin{\varphi } } + \frac{\mu k_y + k \sin{\varphi}}{k_x - k \cos{\varphi } } \right ] \ , .\ ] ] it is instructive to replace @xmath106 and @xmath107 and look at the maxima of the two terms in the large square brackets .
the first term stems from the integration along the side walls parallel to the @xmath20 axis and is maximal when its denominator vanishes , i.e. , for @xmath108 .
this is simply snell s law for a ray refracted at a side wall parallel to the @xmath20 axis , and analogously the maximum of the second term corresponds to a ray refracted at a side wall parallel to the @xmath19 axis .
the results for the full wave functions of the dielectric square are obtained by adding up the @xmath31 plane waves with their correct momentum vectors and relative amplitudes .
for the symmetry class @xmath46 , this yields @xmath109 with @xmath110 - { \mathrm{sinc}}[(-k_y - k \sin{\varphi } ) \frac{a}{2 } ] \ } \\ & & \times [ \mu k_x \frac{a}{2 } \cos(k_x \frac{a}{2 } ) \sin(k \frac{a}{2 } \cos{\varphi } ) - k \frac{a}{2 } \cos{\varphi } \cos(k \frac{a}{2 } \cos{\varphi } ) \sin(k_x \frac{a}{2 } ) ] \\ \\ & & + 2 \ { { \mathrm{sinc}}[(k_x - k \cos{\varphi } ) \frac{a}{2 } ] - { \mathrm{sinc}}[(-k_x - k \cos{\varphi } ) \frac{a}{2 } ] \ } \\ & & \times [ \mu k_y \frac{a}{2 } \cos(k_y \frac{a}{2 } ) \sin(k \frac{a}{2 } \sin{\varphi } ) - k \frac{a}{2 } \sin{\varphi } \cos(k \frac{a}{2 } \sin{\varphi } ) \sin(k_y \frac{a}{2 } ) ] \\ \\ & & + 2 \ { -{\mathrm{sinc}}[(k_x - k \sin{\varphi } ) \frac{a}{2 } ] + { \mathrm{sinc}}[(-k_x - k \sin{\varphi } ) \frac{a}{2 } ] \ } \\ & & \times [ \mu k_y \frac{a}{2 } \cos(k_y \frac{a}{2 } ) \sin(k \frac{a}{2 } \cos{\varphi } ) - k \frac{a}{2 } \cos{\varphi } \cos(k \frac{a}{2 } \cos{\varphi } ) \sin(k_y \frac{a}{2 } ) ] \\ \\ & & + 2 \ { -{\mathrm{sinc}}[(k_y - k \cos{\varphi } ) \frac{a}{2 } ] + { \mathrm{sinc}}[(-k_y - k \cos{\varphi } ) \frac{a}{2 } ] \ } \\ & & \times [ \mu k_x \frac{a}{2 } \cos(k_x \frac{a}{2 } ) \sin(k \frac{a}{2 } \sin{\varphi } ) - k \frac{a}{2 } \sin{\varphi } \cos(k \frac{a}{2 } \sin{\varphi } ) \sin(k_x \frac{a}{2 } ) ] \ , . \\ \\ \end{array}\ ] ]
to validate this semiclassical model , numerical simulations were performed using the boundary element method @xcite . the complete spectrum of tm modes as well as the near- and far - field distributions were calculated in the range of @xmath111@xmath112 for @xmath13 .
figure [ fig : specnum ] shows the numerically calculated spectrum . for each resonance
, we inferred the quantum numbers @xmath113 and @xmath114 , and then the corresponding @xmath58 , by comparing the near - field distributions to the model predictions .
the very regular structure of the spectrum and the possibility to quantize it through @xmath113 and @xmath114 confirm that the dielectric square resonator is nearly an integrable system @xcite .
figure [ fig : ffnum ] presents typical far - field distributions , one with @xmath58 well below @xmath53 [ panel ( a ) ] , and the other with @xmath58 just above @xmath53 [ panel ( b ) ] . in fig .
s[sfig : ffnuma ] , the mode exhibits at total of @xmath31 narrow refractive emission lobes .
their directions agree very well with those expected from ray optics indicated by the arrows .
the analytically calculated far - field distribution ( red line ) agrees almost perfectly with the numerical calculation ( black line ) .
the @xmath31 nonrefractive emission lobes can not be recognized since their amplitudes are negligible compared to those of the refractive lobes .
it should be noted that the lobes of this example are much broader than those of the corresponding example in fig .
6(c ) of the main text for larger @xmath115 since the ratio between the effective slit size @xmath54 and the wavelength is much smaller here . in fig .
s[sfig : ffnumb ] , the directions of the various lobes are very well reproduced by the model although it sometimes fails to predict their amplitude correctly . in general , the ray - based model is less accurate for modes with @xmath116 since the fresnel reflection coefficients for an infinite interface that are used in the quantization condition are no longer a good approximation close to and above the critical angle .
this is evidenced by the prediction of zero losses for such modes by the model , whereas all resonances naturally have a finite @xmath117 in reality .
the nonrefractive lobes are no longer narrow but broad and jagged .
the model has only been tested for moderately sized cavities [ @xmath118 , and not for @xmath119@xmath120 like those investigated experimentally , because simulations in this size region are extremely cumbersome .
however , thanks to the semiclassical nature of the model , it seems reasonable that it works well for very large cavities .
the fast oscillations of the far - field intensity distribution are not visible with an angular resolution of @xmath121 , and the measured far - field patterns are hence compared with the envelope of the model far - field distributions .
the terms in the square brackets in , for example the first term of the equation , @xmath122 feature these very fast oscillations due to the @xmath76 and @xmath77 functions with the argument @xmath78 .
the envelope of @xmath123 is calculated by considering @xmath124 as an independent variable and replacing it by the value @xmath125 for which @xmath123 becomes extremal , yielding @xmath126 analogous formulas are obtained for the other terms .
the function that is fitted to the experimental data can be further simplified since for the parameter range of interest only one of the @xmath52 terms contributes significantly to each emission lobe .
so for example the lobe around @xmath127 is fitted by the formula @xmath128 h_{a_2 , 1}^\mathrm{(env)}(\varphi ) \right|^2\ ] ] where the only two fit parameters are the amplitude @xmath129 and the angle of incidence @xmath58 which is related to the momentum vector components via @xmath106 and @xmath107 .
each lobe of the measured far - field intensity distribution was fitted separately in the range @xmath130 $ ] where @xmath131 by or a corresponding expression . te polarization was used .
the envelopes of the far - field distributions are almost identical for the different symmetry classes and hence fitting the formula for another symmetry class to the experimental data yields nearly the same parameters and an equally good agreement . the effective refractive index of @xmath132 was assumed to be a constant since the shape of the envelope depends very little on it , and the fits are insofar independent of @xmath133 as the fit yields practically the same value for @xmath134 for a wide range of values of @xmath133 . | the emission from open cavities with non - integrable features remains a challenging problem of practical as well as fundamental relevance .
square - shaped dielectric microcavities provide a favorable case study with generic implications for other polygonal resonators .
we report on a joint experimental and theoretical study of square - shaped organic microlasers exhibiting a far - field emission that is strongly concentrated in the directions parallel to the side walls of the cavity .
a semiclassical model for the far - field distributions is developed that is in agreement with even fine features of the experimental findings .
comparison of the model calculations with the experimental data allows the precise identification of the lasing modes and their emission mechanisms , providing strong support for a physically intuitive ray - dynamical interpretation .
special attention is paid to the role of diffraction and the finite side length . |
one - way multitape finite automata were introduced in the seminal 1959 paper of rabin and scott @xcite . such automata ( under various restrictions )
are also commonly known as transducers see elgot and mezei @xcite for an early reference .
a multitape automaton with @xmath0 tapes accepts a @xmath0-ary relation on words .
the class of relations recognised by deterministic automata coincides with the class of @xmath0-ary rational relations @xcite .
two multitape automata are said to be equivalent if they accept the same relation .
undecidability of equivalence of non - deterministic automata is relatively straightforward @xcite .
however the deterministic case remained open for many years , until it was shown decidable by harju and karhumki @xcite .
their solution made crucial use of results about ordered groups specifically that a free group can be endowed with a compatible order @xcite and that the ring of formal power series over an ordered group with coefficients in a division ring and with well - ordered support is itself a division ring ( due independently to malcev @xcite and neumann @xcite ) . using these results
@xcite established the decidability of _ multiplicity equivalence _ of non - deterministic multitape automata , i.e. , whether two non - deterministic multitape automata have the same number of accepting computations on each input .
decidability in the deterministic case ( and , more generally , the unambiguous case ) follows immediately .
we refer the reader to @xcite for a self - contained account of the proof , including the underlying group theory .
harju and karhumki did not address questions of complexity in @xcite . however the existence of a * co - np * _ guess - and - check _
procedure for deciding equivalence of deterministic multitape automata follows directly from ( * ? ? ?
* theorem 8) .
this theorem states that two inequivalent automata are guaranteed to differ on a tuple of words whose total length is at most the total number of states of the two automata .
such a tuple can be guessed , and it can be checked in polynomial time whether the tuple is accepted by one automaton and rejected by the other . in the special case of two - tape deterministic automata ,
a polynomial - time algorithm was given in @xcite , before decidability was shown in the general case .
a * co - np * upper bound also holds for multiplicity equivalence of @xmath0-tape automata for each fixed @xmath0 .
however , as we observe below , if the number of tapes is not fixed , computing the number of accepting computations of a given non - deterministic multitape automata on a tuple of input words is # * p*-hard .
thus the guess - and - check method does not yield a * co - np * procedure for multiplicity equivalence in general .
it is well - known that the equivalence problem for single - tape weighted automata with rational transition weights is solvable in polynomial time @xcite . now
the decision procedure in @xcite reduces multiplicity equivalence of multitape automata to equivalence of single - tape automata with transition weights in a division ring of power series over an ordered group
. however the complexity of arithmetic in this ring seems to preclude an application of the polynomial - time procedures of @xcite .
leaving aside issues of representing infinite power series , even the operation of multiplying a family of polynomials in two non - commuting variables yields a result with exponentially many monomials in the length of its input .
in this paper we give an alternative proof that multiplicity equivalence of multitape automata is decidable , which also yields new complexity bounds on the problem .
we use the same basic idea as @xcite reduce to the single - tape case by enriching the set of transition weights .
however we replace their use of power series on ordered groups with results about matrix algebras and polynomial identity rings ( see remark [ rem : difference ] for a more technical comparison ) . in particular
, we use the amitsur - levitzki theorem concerning polynomial identities in matrix algebras .
our use of the latter is inspired by the work of @xcite on non - commutative polynomial identity testing , and our starting point is a simple generalisation of the approach of @xcite to what we call _ partially commutative _ polynomial identity testing .
our construction for establishing decidability immediately yields a simple randomised algorithm for checking multiplicity equivalence of multitape automata ( and hence also equivalence of deterministic automata ) .
the algorithm involves only matrix exponentiation , and runs in polynomial time for each fixed number of tapes .
let @xmath1 be an infinite field . recall that an @xmath1-algebra is a vector space over @xmath1 equipped with an associative bilinear product that has an identity . write @xmath2 for the free @xmath1-algebra over a set @xmath3 .
the elements of @xmath4 can be viewed as polynomials over a set of non - commuting variables @xmath3 with coefficients in @xmath1 .
each such polynomial is an @xmath1-linear combination of monomials , where each monomial is an element of @xmath5 .
the degree of a polynomial is the maximum of the lengths of its monomials .
let @xmath6 be an @xmath1-algebra and @xmath7 .
if @xmath8 evaluates to @xmath9 for all valuations of its variables in @xmath6 then we say that @xmath6 satisfies the _ polynomial identity _ @xmath10 .
for example , an algebra satisfies the polynomial identity @xmath11 if and only if it is commutative . note that since the variables @xmath12 and @xmath13 do not commute , the polynomial @xmath14 is not identically zero .
we denote by @xmath15 the @xmath1-algebra of @xmath16 matrices with coefficients in @xmath1 .
the amitsur - levitzki theorem @xcite is a fundamental result about polynomial identities in matrix algebras .
the algebra @xmath15 satisfies the polynomial identity @xmath17 where the sum is over the @xmath18 elements of the symmetric group @xmath19 .
moreover @xmath15 satisfies no identity of degree less than @xmath20 .
[ thm : al ] given a finite set @xmath3 of non - commuting variables , the _
generic @xmath16 matrix algebra _
@xmath21 is defined as follows . for each variable @xmath22
we introduce a family of commuting indeterminates @xmath23 and define @xmath21 to be the @xmath1-algebra of @xmath16 matrices generated by the matrices @xmath24 for each @xmath25
. then @xmath21 has the following universal property : any homomorphism from @xmath26 to a matrix algebra @xmath27 , with @xmath28 an @xmath1-algebra , factors uniquely through the map @xmath29 given by @xmath30 .
related to the map @xmath31 we also define an @xmath1-algebra homomorphism @xmath32 by @xmath33 where the matrix on the right has zero entries everywhere but along the superdiagonal . in this section
we introduce a notion of _ partially commutative polynomial identity_. we first establish notation and recall some relevant facts about tensor products of algebras .
write @xmath34 for the tensor product of @xmath1-algebras @xmath6 and @xmath35 , and write @xmath36 for the @xmath0-fold tensor power of @xmath6 .
if @xmath6 is an algebra of @xmath37 matrices and @xmath35 an algebra of @xmath38 matrices , then we identify the tensor product @xmath39 with the algebra of @xmath40 matrices spanned by the matrices @xmath41 , @xmath42 and @xmath43 , where @xmath44 in particular we have @xmath45 .
a _ partially commuting set of variables _ is a tuple @xmath46 , where the @xmath47 are disjoint sets .
write @xmath48 for the tensor product @xmath49 .
we think of @xmath48 as a set of polynomials in partially commuting variables .
intuitively two variables @xmath50 do not commute , whereas @xmath51 commutes with @xmath52 if @xmath53 .
note that if each @xmath47 is a singleton @xmath54 then @xmath55 is the familiar ring of polynomials in commuting variables @xmath56 . at the other extreme ,
if @xmath57 then we recover the non - commutative case .
an arbitrary element @xmath58 can be written uniquely as a finite sum of distinct _
monomials _ , where each monomial is a tensor product of elements of @xmath59 , @xmath60 , , and @xmath61 .
formally , we can write @xmath62 where @xmath63 and @xmath64 for each @xmath65 and @xmath66 . thus we can identify @xmath67 with the free @xmath1-algebra over the product monoid @xmath68 .
define the _ degree _ of a monomial @xmath69 to be the total length @xmath70 of its constituent words .
the degree of a polynomial is the maximum of the degrees of its constituent monomials .
let @xmath71 be a @xmath0-tuple of @xmath1-algebras .
a _ valuation _ of @xmath48 in @xmath72 is a tuple of functions @xmath73 , where @xmath74 .
each @xmath75 extends uniquely to an @xmath1-algebra homomorphism @xmath76 , and we define the map @xmath77 by @xmath78
. often we will abuse terminology slightly and speak of a valuation of @xmath48 in @xmath79 .
given @xmath80 , we say that @xmath72 satisfies the partially commutative identity @xmath10 if @xmath81 for all valuations @xmath82 .
next we introduce two valuations that will play an important role in the subsequent development .
recall that given a set of non - commuting variables @xmath3 , we have a map @xmath83 from the free @xmath1-algebra to the generic @xmath84-dimensional matrix algebra .
we now define a valuation @xmath85 by @xmath86 .
likewise we define @xmath87 by @xmath88 .
we will usually elide the superscript @xmath89 from @xmath90 and @xmath91 when it is clear from the context .
the following result generalises ( part of ) the amitsur - levitzki theorem , by giving a lower bound on the degrees of partially polynomial identities holding in tensor products of matrix algebras .
let @xmath80 and let @xmath92 be a field extending @xmath1 .
then the following are equivalent : ( i ) the partially commutative identity @xmath10 holds in @xmath93 ; ( ii ) @xmath94 . moreover , if @xmath8 has degree strictly less than @xmath84 then ( i ) and ( ii ) are both equivalent to ( iii ) @xmath95 ; and ( iv ) @xmath8 is identically @xmath9 in @xmath96 .
[ prop : sc - al ] the implication ( ii ) @xmath97 ( i ) follows from the fact that any valuation from @xmath48 to @xmath98 factors through @xmath99 . to see that ( i ) @xmath97 ( ii ) , observe that @xmath100 is an @xmath101 matrix in which each entry is a polynomial in the commuting variables @xmath102 . condition ( i ) implies in particular that each such polynomial evaluates to @xmath9 for all valuations of its variables in @xmath1 .
since @xmath1 is an infinite field , it must be that each such polynomial is identically zero , i.e. , ( ii ) holds . the implications ( ii ) @xmath97 ( iii ) and ( iv ) @xmath97 ( i ) are both straightforward , even without the degree restriction on @xmath8 .
finally we show that ( iii ) @xmath97 ( iv ) .
let @xmath103 be a monomial in @xmath67 , where @xmath104 has length @xmath105 .
then @xmath106 is an @xmath107 matrix whose first row has a single non - zero entry , which is the monomial @xmath108 at index @xmath109 .
it follows that @xmath110 maps the set of monomials in @xmath96 of degree less than @xmath84 injectively into a linearly independent set of matrices .
condition ( iv ) immediately follows .
the hypothesis that @xmath8 have degree less than @xmath84 in proposition [ prop : sc - al ] can be weakened somewhat , but is sufficient for our purposes .
a ring @xmath28 with no zero divisors is a _
domain_. if moreover each non - zero element of @xmath28 has a two - sided multiplicative inverse , then we say that @xmath28 is a _ division ring _
( also called a _ skew field _ ) .
a domain @xmath28 is a _ ( right ) ore domain _ if for all @xmath111 , @xmath112 .
the significance of this notion is that an ore domain can be embedded in a division ring of fractions ( * ? ? ?
* corollary 7.1.6 ) , something that need not hold for an arbitrary domain .
if the ore condition fails then it can easily be shown that the subalgebra of @xmath28 generated by @xmath113 and @xmath114 is free on @xmath113 and @xmath114 .
it follows that a domain @xmath28 that satisfies some polynomial identity is an ore domain ( * ? ? ?
* corollary 7.5.2 ) .
the tensor product of generic matrix algebras @xmath115 is an ore domain for each @xmath116 .
[ prop : ore ] we give a proof sketch here , deferring the details to appendix [ sec : appendix ] . by the amitsur - levitzki theorem , @xmath117 satisfies a polynomial identity .
thus it suffices to show that @xmath118 is a domain for each @xmath84 .
now it is shown in ( * ? ? ?
* proposition 7.7.2 ) that @xmath21 is a domain for each @xmath84 and set of variables @xmath3 .
while the tensor product of domains need not be a domain ( e.g. , @xmath119 ) , the proof in @xcite can be adapted _ mutatis mutandis _ to show that @xmath120 is also a domain . to prove the latter
, it suffices to find central simple @xmath1-algebras @xmath121 , each of degree @xmath84 , such that the @xmath0-fold tensor product @xmath122 is a domain .
such an example can be found , e.g. , in ( * ? ? ?
* proposition 1.1 ) .
then , using the fact that @xmath123 for any algebraically closed extension field of @xmath1 , one can infer that @xmath124 is also a domain .
let @xmath125 be a tuple of finite alphabets .
we denote by @xmath126 the product monoid @xmath127 .
define the length of @xmath128 to be @xmath129 and write @xmath130 for the set of elements of @xmath126 of length @xmath131 .
multitape automaton _ is a tuple @xmath132 , where @xmath133 is a set of _ states _ , @xmath134 is a set of _ edges _ , @xmath135 is a set of _ initial states _ , and @xmath136 is a set of _ final states_. a _ run _ of @xmath6 from state @xmath137 to state @xmath138 is a finite sequence of edges @xmath139 such that @xmath140 .
the _ label _ of @xmath141 is the product @xmath142 .
define the _ multiplicity _ @xmath143 of an input @xmath144 to be the number of runs with label @xmath145 such that @xmath146 and @xmath147 .
an automaton is _ deterministic _
if each state reads letters from a single tape and has a single transition for every input letter .
thus a deterministic automaton has a single run on each input @xmath148 .
we say that two automata @xmath6 and @xmath35 over the same alphabet are _ multiplicity equivalent _ if @xmath149 for all @xmath144 .
the following result implies that multiplicity equivalence of multitape automata is decidable .
given automata @xmath6 and @xmath35 with @xmath84 states in total , @xmath6 and @xmath35 are equivalent if and only if @xmath149 for all @xmath148 of length at most @xmath150 .
[ thm : harju ] theorem [ thm : harju ] immediately yields a * co - np * bound for checking language equivalence of deterministic multitape automata . given two inequivalent automata @xmath6 and @xmath35 , a distinguishing input @xmath145 can be guessed , and it can be verified in polynomial time that only one of @xmath6 and @xmath35 accepts @xmath145 .
a similar idea also gives a * co - np * bound for multiplicity equivalence in case the number of tapes is fixed .
in general we note however that the _ evaluation problem_given an automaton @xmath6 and input @xmath145 , compute @xmath143is @xmath151-complete .
thus it is not clear that the * co - np * upper bound applies to the multiplicity equivalence problem without bounding the number of tapes .
the evaluation problem for multitape automata is @xmath152-complete .
membership in @xmath151 follows from the observation that a non - deterministic polynomial - time algorithm can enumerate all possible runs of an automaton @xmath6 on an input @xmath148 .
the proof of @xmath151-hardness is by reduction from # sat , the problem of counting the number of satisfying assignments of a propositional formula .
consider such a formula @xmath153 with @xmath0 variables , each with fewer than @xmath84 occurrences .
we define a @xmath0-tape automaton @xmath6 , with each tape having alphabet @xmath154 , and consider as input the @xmath0-tuple @xmath155 .
the automaton @xmath6 is constructed such that its runs on input @xmath145 are in one - to - one correspondence with satisfying assignments of @xmath153 .
each run starts with the automaton reading the symbol @xmath9 from a non - deterministically chosen subset of its tapes , corresponding to the set of false variables .
thereafter it evaluates the formula @xmath153 by repeatedly guessing truth values of the propositional variables .
if the @xmath156-th variable is guessed to be true then the automaton reads @xmath157 from the @xmath156-th tape ; otherwise it reads @xmath158 from the @xmath156-th tape .
the final step is to read the symbol @xmath159 from a non - deterministically chosen subset of the input tapes
again corresponding to the set of false variables .
the consistency of the guesses is ensured by the requirement that the automaton have read @xmath145 by the end of the computation .
we start by recalling from @xcite an equivalence - respecting transformation from multitape automata to single - tape weighted automata .
recall that a single - tape automaton on a unary alphabet with transition weights in a ring @xmath28 consists of a set of _ states _
@xmath160 , _ initial states _
@xmath161 , _ final states _
@xmath162 , and _ transition matrix _ @xmath163 .
given such an automaton , define the _ initial - state vector _ @xmath164 and _ final - state vector _
@xmath165 respectively by @xmath166 then @xmath167 is the weight of the ( unique ) input word of length @xmath131 . consider a @xmath0-tape automaton @xmath168 , where @xmath169 , and write @xmath170 .
recall the ring of polynomials @xmath171 as defined in section [ sec : semi ] .
recall also that we can identify the monoid @xmath126 with the set of monomials in @xmath172 , where @xmath173 corresponds to @xmath174indeed @xmath172 is the free @xmath1-algebra on @xmath126 .
we derive from @xmath6 an @xmath175-weighted automaton @xmath176 ( with a single tape and unary input alphabet ) that has the same sets of states , initial states , and final states as @xmath6 .
we define the transition matrix @xmath177 of @xmath176 by combining the different transitions of @xmath6 into a single matrix with entries in @xmath175 . to this end , suppose that the set of states of @xmath6 is @xmath160 .
define the matrix @xmath178 by @xmath179 for @xmath180 .
let @xmath181 and @xmath182 be the respective initial- and final - state vectors of @xmath176
. then the following proposition is straightforward .
intuitively it says that the weight of the unary word of length @xmath131 in @xmath176 represents the language of all length-@xmath131 tuples accepted by @xmath6 . for all @xmath183 we have @xmath184 .
[ prop : matrix ] now consider two @xmath0-tape automata @xmath6 and @xmath35 .
let the weighted single - tape automata derived from @xmath6 and @xmath35 have respective transition matrices @xmath185 and @xmath186 , initial - state vectors @xmath187 and @xmath188 , and final - state vectors @xmath189 and @xmath190 .
we combine the latter into a single weighted automaton with transition matrix @xmath177 , initial - state vector @xmath181 , and final - state vector @xmath182 , respectively defined by : @xmath191 automata @xmath6 and @xmath35 are multiplicity equivalent if and only if @xmath192 for @xmath193 , where @xmath84 is the total number of states of the two automata .
[ prop : main ] since @xmath126 is a linearly independent subset of @xmath172 , from proposition [ prop : matrix ] it follows that @xmath6 and @xmath35 are multiplicity equivalent just in case @xmath194 for all @xmath195 .
the latter is clearly equivalent to @xmath196 for all @xmath183 .
it remains to show that we can check equivalence by looking only at exponents @xmath131 in the range @xmath197 .
suppose that @xmath198 for @xmath199 .
we show that @xmath192 for an arbitrary @xmath200 .
consider the map @xmath201 , as defined in ( [ eq : psi ] ) .
observe that @xmath167 is a polynomial expression in @xmath175 of degree at most @xmath131 .
therefore by proposition [ prop : sc - al ] ( ( ii ) @xmath202 ( iv ) ) , to show that @xmath192 it suffices to show that @xmath203 let us write @xmath204 for the pointwise application of @xmath205 to the matrix @xmath177 , so that @xmath204 is an @xmath16 matrix , each of whose entries is an @xmath101 matrix belonging to @xmath206 .
since @xmath205 is a homomorphism and @xmath181 and @xmath182 are integer vectors , ( [ eq : goal4 ] ) is equivalent to : @xmath207 recall from proposition [ prop : ore ] that the tensor product of generic matrix algebras @xmath206 is an ore domain and hence can be embedded in a division ring .
now a standard result about single - tape weighted automata with transition weights in a division ring is that such an automaton with @xmath84 states is equivalent to the zero automaton if and only if it assigns zero weight to all words of length @xmath84 ( see ( * ? ? ?
* pp143145 ) and @xcite ) . applying this result to the unary weighted automaton defined by @xmath181 , @xmath177 , and @xmath182
, we see that ( [ eq : goal2 ] ) is implied by @xmath208 but , since @xmath205 is a homomorphism , ( [ eq : goal3 ] ) is implied by @xmath209 this concludes the proof .
theorem [ thm : harju ] immediately follows from proposition [ prop : main ] .
the difference between our proof of theorem [ thm : harju ] and the proof in @xcite is that we consider a family of homomorphisms of @xmath175 into ore domains of matrices the maps @xmath205rather than a single `` global '' embedding of @xmath175 into a division ring of power series over a product of free groups . none of the maps @xmath205 is an embedding , but it suffices to use the lower bound on the degrees of polynomial identities in proposition [ prop : sc - al ] in lieu of injectivity . on the other hand , the fact that @xmath210 satisfies a polynomial identity makes it relatively straightforward to exhibit an embedding of the latter into a division ring .
as we now show , this approach leads directly to a very simple randomised polynomial - time algorithm for solving the equivalence problem .
[ rem : difference ] proposition [ prop : main ] reduces the problem of checking multiplicity equivalence of multitape automata @xmath6 and @xmath35 to checking the partially commutative identities @xmath192 , @xmath211 in @xmath175 .
since each identity has degree less than @xmath84 , applying proposition [ prop : sc - al ] ( ( iii ) @xmath202 ( iv ) ) we see that @xmath6 and @xmath35 are equivalent if and only if @xmath212 each equation @xmath213 in ( [ eq : pit ] ) asserts the zeroness of an @xmath101 matrix of polynomials in the commuting variables @xmath102 , with each polynomial having degree less than @xmath84 .
suppose that @xmath214 for some @xmath131say the matrix entry with index @xmath215 contains a monomial with non - zero coefficient . by ( [ eq : psi2 ] ) such a monomial
determines a term @xmath216 with non - zero coefficient in @xmath167 , and by proposition [ prop : matrix ] we have @xmath217 .
we can verify each polynomial identity in ( [ eq : pit ] ) , outputting a monomial of any non - zero polynomial , using a classical identity testing procedure based on the _ isolation lemma _ of @xcite .
there is a randomised polynomial - time algorithm that inputs a multilinear polynomial @xmath218 , represented as an algebraic circuit , and either outputs a monomial of @xmath8 or that @xmath8 is zero .
moreover the algorithm is always correct if @xmath8 is the zero polynomial and is correct with probability at least @xmath219 if @xmath8 is non - zero .
[ lem : pit ] the idea behind the algorithm described in lemma [ lem : pit ] is to choose a weight @xmath220 for each variable @xmath221 of @xmath8 independently and uniformly at random . defining the weight of a monomial @xmath222 to be @xmath223 , then with probability at least @xmath219
there is a unique minimum - weight monomial .
the existence of a minimum - weight monomial can be detected by computing the polynomial @xmath224 , since a monomial with weight @xmath225 in @xmath8 yields a monomial of degree @xmath225 in @xmath226 . using similar ideas one can moreover determine the composition of a minimum - weight monomial in @xmath8 .
applying lemma [ lem : pit ] we obtain our main result : let @xmath0 be fixed
. then multiplicity equivalence of @xmath0-tape automata can be decided in randomised polynomial time .
moreover there is a randomised polynomial algorithm for the function problem of computing a distinguishing input given two inequivalent automata .
[ thm : main ] the reason for the requirement that @xmath0 be fixed is because the dimension of the entries of the transition matrix @xmath177 , and thus the number of polynomials to be checked for equality , depends exponentially on @xmath0 . the above use of the isolation technique generalises @xcite , where it is used to generate counterexample words of weighted single - tape automata .
a very similar application in @xcite occurs in the context of identity testing for non - commutative algebraic branching programs .
we have given a simple randomised algorithm for deciding language equivalence of deterministic multitape automata and multiplicity equivalence of nondeterministic automata .
the algorithm arises directly from algebraic constructions used to establish decidability of the problem , and runs in polynomial time for each fixed number of tapes .
we leave open the question of whether there is a deterministic polynomial - time algorithm for deciding the equivalence of deterministic and weighted multitape automata with a fixed number of tapes .
( recall that the 2-tape case is already known to be in polynomial time @xcite . )
we also leave open whether there is a deterministic or randomised polynomial time algorithm for solving the problem in case the number of tapes is not fixed .
we first recall a construction of a _ crossed product division algebra _ from ( * ? ? ?
* proposition 1.1 ) .
let @xmath227 be commuting indeterminates and write @xmath228 for the field of rational functions obtained by adjoining @xmath229 to @xmath230 .
furthermore , let @xmath231 be a field extension whose galois group is generated by commuting automorphisms @xmath232 , each of order @xmath84 , which has fixed field @xmath1 .
( such an extension can easily be constructed by adjoining extra indeterminates to @xmath1 , and having the @xmath233 be suitable permutations of the new indeterminates . ) for each @xmath156 , @xmath234 , write @xmath235 for the subfield of @xmath236 that is fixed by each @xmath237 for @xmath238 ; then define @xmath239 to be the @xmath1-algebra generated by @xmath235 and @xmath240 such that @xmath241 for all @xmath242 .
then each @xmath239 is a simple algebra of dimension @xmath243 over its centre @xmath1 .
it is shown in ( * ? ? ?
* proposition 1.1 ) that the tensor product @xmath244 can be characterised as the localisation of an iterated skew polynomial ring and is therefore a domain . noting that the @xmath239 are all isomorphic as @xmath1-algebras , let @xmath248 be a basis of each @xmath239 over its centre @xmath1 . for each variable @xmath12 appearing in @xmath8 , introduce commuting indeterminates @xmath249 , @xmath250 , and write @xmath251 .
then we can express @xmath8 in the form @xmath252 where @xmath253 . by assumption
, each @xmath254 evaluates to @xmath9 for all values of the @xmath249 in @xmath1 .
since @xmath1 is an infinite field it follows that each @xmath254 must be identically zero .
now we can also regard @xmath248 as a basis for @xmath255 over @xmath92 . then by ( [ eq : formula ] ) , @xmath10 also on @xmath256 . recall that if @xmath92 is an algebraically closed field extension of @xmath1 , then we have @xmath257 for each @xmath156 . by proposition [ prop :
extend ] it follows that an identity @xmath10 holds in @xmath246 if and only if it holds in @xmath258 .
but by proposition [ prop : sc - al ] the latter holds if and only if @xmath100 is identically zero . to prove the proposition it will suffice to show that the image of @xmath99 contains no zero divisors , since the latter is a surjective map . now given @xmath259 with @xmath260
, we have that @xmath261 satisfies the identity @xmath262 .
since @xmath261 is a domain , it follows that it satisfies the identity @xmath263 for any @xmath264 in @xmath265 .
but now @xmath266 satisfies the identity @xmath263 for any @xmath264 .
since @xmath264 can take the value of an arbitrary matrix ( in particular , any matrix unit ) it follows that @xmath267 satisfies either the identity @xmath10 or @xmath268 , and so , by proposition [ prop : sc - al ] again , either @xmath269 or @xmath270 .
v. arvind and p. mukhopadhyay . derandomizing the isolation lemma and lower bounds for circuit size . in _ approx - random
_ , volume 5171 of _ lecture notes in computer science _ , pages 276289 .
springer , 2008 . | the decidability of determining equivalence of deterministic multitape automata ( or transducers ) was a longstanding open problem until it was resolved by harju and karhumki in the early 1990s .
their proof of decidability yields a * co - np * upper bound , but apparently not much more is known about the complexity of the problem . in this paper
we give an alternative proof of decidability , which follows the basic strategy of harju and karhumki but replaces their use of group theory with results on matrix algebras . from our proof
we obtain a simple randomised algorithm for deciding language equivalence of deterministic multitape automata and , more generally , multiplicity equivalence of nondeterministic multitape automata .
the algorithm involves only matrix exponentiation and runs in polynomial time for each fixed number of tapes . if the two input automata are inequivalent then the algorithm outputs a word on which they differ . |
interaction cross sections as well as reaction cross sections are intimately related to the size of nuclei @xcite . using this property ,
the halo structure has been found in some light neutron - rich nuclei .
this is a spatially extended density distribution of valence neutrons , and has been first recognized in @xmath4li by tanihata _
the root - mean - square radius diverges for @xmath5 and @xmath3 waves as the single - particle energy approaches to zero @xcite , and the halo structure has been ascribed to an occupation of an @xmath6 or @xmath7 orbit by the valence neutron @xcite .
@xmath4be @xcite and @xmath8c @xcite have been regarded as @xmath5-wave halo nuclei , and @xmath9he @xcite is an example of a @xmath3-wave halo nucleus .
a large interaction cross section for @xmath0ne was recently observed by takechi _
this observation suggests the extended density distribution for @xmath0ne , that is , the halo structure , being consistent also with a large coulomb breakup cross section measured by nakamura _
et al._@xcite .
et al . _ have analyzed the data using single particle levels in a deformed potential and argued that @xmath0ne is an @xmath5- or @xmath3- wave halo nucleus@xcite .
the ground state properties for @xmath0ne have not been known well .
for example , the one neutron separation energy s@xmath10 = 0.29 @xmath11 1.64 mev @xcite has a large uncertainty and the spin and parity have not yet been determined . in the core nucleus @xmath1ne , the candidates for the first excited 2@xmath12 and 4@xmath12 states
have been experimentally observed at excitation energies of 0.801 mev and 2.24 mev , respectively @xcite .
the energy ratio @xmath13=2.80 suggests this nucleus to be a transitional one in comparison with the ratio 3.33 for well - deformed nuclei .
hamamoto has carried out the nilsson model calculation with a deformed woods - saxon potential and argued that the [ 330 1/2 ] , [ 321 3/2 ] , and [ 200 1/2 ] nilsson levels occupied by the valence neutron can hold the halo structure @xcite .
the [ 330 1/2 ] and [ 321 3/2 ] configurations lead to @xmath14 , while the [ 200 1/2 ] leads to @xmath15 for the spin and parity of the ground state of @xmath0ne in the laboratory frame . in the previous publication
, we used a particle - rotor model ( prm ) @xcite to analyze the experimental data for the coulomb breakup cross section and discussed the ground state configuration for the @xmath0ne nucleus @xcite .
notice that the nilsson model corresponds to the adiabatic limit of prm .
we have shown that the ground state configuration corresponding to the [ 321 3/2 ] nilsson orbit can be excluded if the finite rotational excitation energy of the core nucleus is taken into account @xcite . in this paper
, we apply the same model to the reaction cross section of @xmath0ne .
the effect of deformation on the reaction cross section of the @xmath0ne nucleus has been discussed recently by minomo _
_ using the microscopic optical potential model @xcite , and has been shown to play an important role .
it would thus be of interest to discuss the role of deformation in the reaction cross section of @xmath0ne using the prm as an alternative approach , which has been successful in reproducing the coulomb breakup cross section .
notice that reaction cross sections with deformed projectiles have been evaluated by christley and tostevin with the optical limit glauber theory @xcite .
we will extend it to a system of deformed core nucleus plus a valence neutron , based on the formalism given in ref .
@xcite for single - nucleon knockout reactions .
the paper is organized as follows . in sec .
ii , we summarize the framework of prm and the calculation procedure for the reaction cross section . in sec .
iii , we present the results of the reaction cross section for @xmath16ne .
we discuss the effect of the finite rotational excitation energy on the reaction cross section .
we investigate also the dependence of the reaction cross section on the deformation , the ground state configuration , and the rms radius of @xmath0ne . in sec .
iv , we summarize the paper .
[ figure_prm ] ne plus a neutron .
@xmath17 denotes the symmetry axis for the deformed core nucleus @xmath1ne .
@xmath18 and @xmath19 are the coordinates for the valence neutron and the direction of the symmetry axis of the core nucleus in the laboratory frame , respectively .
@xmath20 is the angle between @xmath18 and @xmath19 .
, title="fig : " ] in order to compute the reaction cross section of the @xmath0ne nucleus , we assume that it consists of the statically deformed core nucleus @xmath1ne and one valence neutron as shown in figure 1 .
the relevant coordinate systems are also shown in the figure . in this model , the single particle motion of the valence neutron is coupled to the rotation of the deformed core nucleus . for simplicity , we assume the axially symmetric deformation for the core nucleus with the quadrupole deformation parameter @xmath21 .
we consider the same hamiltonian for this system as in ref .
@xcite , @xmath22 where @xmath23 is the reduced mass of the valence neutron , with @xmath24=30 and @xmath25 being the mass number of the core nucleus and the nucleon mass , respectively .
@xmath26 is the rotational hamiltonian for the core nucleus .
@xmath27 is the single - particle potential for the valence neutron interacting with the deformed core . @xmath18 and @xmath19 are the coordinates of the valence neutron and the direction of the symmetry axis of the core nucleus in the laboratory frame , respectively ( see fig .
1 ) . we use a deformed woods - saxon potential for @xmath28 and expand it up to the linear order of the deformation parameter @xmath21 as , v(,_c)~v_0(r)+v_def(r,_cn ) , where @xmath29 is a spherical woods - saxon potential together with the spin - orbit ( @xmath30 ) force , and @xmath31 is the deformed part of the potential given by , v_def(r,_cn ) = -r_0_2y_20(_cn ) , where @xmath32 is the central part of the spherical woods - saxon potential , @xmath33 , and @xmath20 is the angle between @xmath18 and @xmath19 .
the deformation of the @xmath30 potential is neglected for simplicity .
we have checked the validity of the expansion up to the linear order of @xmath21 by comparing to the calculation with the higher order terms , and have confirmed that it works well . since the calculation of the reaction cross section needs only the ground state wave function of @xmath0ne , it is sufficient to expand the wave function on the basis _
e.g. , _ the eigen - functions of the spherical part @xmath29 of the potential , @xmath34 , where @xmath35 is the radial wave function and @xmath36 is the spin - angular wave function .
the continuum spectrum can be discretized within a large box .
together with the rotational wave function @xmath37 , the total wave function for the @xmath2+@xmath1ne system is expanded as , @xmath38^{(im ) } , \label{wf}\ ] ] where @xmath39 is the spin of @xmath0ne and @xmath40 is its @xmath41-component .
the expansion coefficients @xmath42 as well as the corresponding eigen - energies for the @xmath0ne nucleus are obtained by numerically diagonalizing the hamiltonian @xmath43 .
we identify the ground state configuration in the same manner as in the previous work @xcite .
that is , we first solve the hamiltonian in the adiabatic limit by setting @xmath44 in eq .
( [ hprot ] ) , that is , by assuming that all the members of the ground rotational band are degenerate in energy . in this case , the @xmath45 quantum number , that is , the projection of the total angular momentum onto the @xmath41-axis in the body - fixed frame , is conserved , and several states with different @xmath39 , having the same value of @xmath45 , are degenerate in energy when the maximum value of @xmath46 included in the calculation is sufficiently large . the wave function in this limit is related to the wave function in the nilsson model , @xmath47 , in the sense that it is a transformation of the nilsson wave function from the body - fixed frame to the laboratory frame .
the eigen - energies so obtained thus form the single - particle nilsson levels .
the probability for the ( @xmath48 ) component in the nilsson wave function is represented in the following relation : @xmath49 which is independent of @xmath39 . in order to construct the ground state ,
we put two neutrons to each nilsson orbit from the bottom of the potential well , and seek the nilsson orbit which is occupied by the last unpaired neutron .
we then gradually increase the value of the 2@xmath50 energy of the core nucleus up to the physical value , @xmath51=0.801 mev , and monitor how the nilsson orbit for the valence neutron evolves . for a finite value of @xmath51
, the @xmath45 quantum number is not conserved any more due to the coriolis coupling , and the degeneracy with respect to @xmath39 is resolved .
we select the lowest energy state among several @xmath39 at @xmath51=0.801 mev as the ground state of @xmath0ne . in this way
, we take into account the pauli principle between the valence neutron and the neutrons in the core nucleus .
we consider two configurations with the spin - parity of @xmath52 in the laboratory frame as candidates for the ground state of @xmath0ne .
one with the deformation parameter @xmath53 corresponds to the nilsson level [ 330 1/2 ] , and the other with @xmath54 corresponds to the nilsson level [ 321 3/2 ] in the adiabatic limit where the rotational energy of the core nucleus is neglected . following ref .
@xcite , we use the same woods - saxon potential parameters as those given in table i of ref .
the depth of the woods - saxon potential is varied to reproduce the one neutron separation energy .
we use a similar value for the energy cut - off for the single particle basis and a similar size of the box to discretize the continuum spectrum as in ref .
@xcite . in this paper
, we discuss the reaction cross sections for @xmath16ne on the carbon target . for simplicity , we neglect the effect of the coulomb force . to verify this approximation ,
we have calculated the coulomb breakup cross sections of @xmath0ne for the configuration with @xmath55=0.2 at the separation energy @xmath56=0.2mev on the carbon target using the method given by ref .
the calculated coulomb breakup cross section , 0.0033b , is indeed small as compared to the total reaction cross section , suggesting that the nuclear force dominantly contributes to reaction cross sections for @xmath0ne on the carbon target . in order to compute the reaction cross sections
, we use the glauber theory where the eikonal approximation and the adiabatic approximation are adopted @xcite .
we closely follow the formalism in ref .
@xcite , in which the prm has been used to evaluate single - nucleon knockout reactions of a deformed odd - a nucleus based on the glauber theory . in the eikonal approximation ,
the final state @xmath57 after the collision is described with the initial state wave function @xmath58 as , _
i= _ i , where @xmath59 is the phase shift function .
the reaction cross section of @xmath0ne , defined as the difference between the total cross section and the elastic scattering cross section , is given with the ground state wave function of @xmath0ne , @xmath60 , as , & & _
r(^31ne)=d(1-_m + & & | _
im| s_cs_v|_im|^2 ) , [ sgm31 ] where @xmath61 is the impact parameter of the center of mass of the projectile nucleus @xmath0ne colliding with the target nucleus .
the @xmath62-matrix for the two - body projectile nucleus can be written by @xmath63 in the glauber approximation @xcite . here
, @xmath64 and @xmath65 are @xmath62 matrices for the core nucleus and the valence neutron , respectively .
notice that , since the directions of @xmath18 and @xmath66 are integrated in the whole space in eq .
( [ sgm31 ] ) before the integration over @xmath61 is carried out , the integrand does not depend upon the direction of @xmath61 @xcite .
the reaction cross section @xmath67 thus reads & & _
r(^31ne)=2bdb(1-_m + & & |dd_c s_vs_cf(,_c)|^2 ) , [ sgm31_1 ] where f(,_c ) & = & _ njli_c_njli_c^*_njli_c_njli_c r^*_njl(r)r_njl(r ) + & & _
m_jm_im_jm_i jm_ji_cm_i|imjm_ji_cm_i|im + & & y_jlm_j^*()_i_c^*(_c ) y_jlm_j()_i_c(_c ) . using the formula for the product of two spherical harmonics with the same angles ,
the function @xmath68 is transformed to @xcite f(,_c ) = _ l , m_l , l , m_l y_lm_l()y_lm_l(_c)u_lm_llm_l(r ) , where u_lm_llm_l(r ) & = & _ njli_c_njli_c_km_k ^*_njli_c_njli_c + & & r^*_njl(r)r_njl(r ) + & & ( -)^2j-j+i_c+m_k- + & & l0l0|l0i_c0i_c0|l0 + & & imk - m_k|imlm_llm_l|km_k + & & w(ljlj;l ) i_c & i_c & l + j & j & l + i & i & k , with @xmath69 and @xmath70 being the racah coefficients . in order to evaluate the @xmath62-matrices , we employ the optical limit approximation for simplicity . using the zero range interaction for the effective nucleon - nucleon interaction ,
we evaluate the @xmath62 matrices by folding the densities of the projectile and the target nuclei as , & s_c&= , [ sc ] + & s_v&= , [ sn ] where & _ c(_c,_c ) & = dz_cd_c( , _ c ) _
t(|+| ) , [ chic ] + & _ n(b_n)&=dz_n_t(r_n ) .
[ chin ] here , @xmath71 and @xmath72 are the coordinates of the center of mass of the core nucleus and the valence neutron from the target nucleus , respectively .
@xmath73 and @xmath74 are the densities of the core and the target nuclei , respectively .
we construct the density of the core nucleus @xmath1ne with the nilsson model . to this end
, we use the original values for the potential parameters given in table i of ref .
for the density distribution for the target nucleus @xmath75c , we use a one - range gaussian function whose width parameter is determined so as to reproduce the experimental root mean square radius .
@xmath76 in eqs .
( [ sc ] ) and ( [ sn ] ) is the average value of the total cross sections of the nucleon - nucleon scattering @xcite .
@xmath77 is also the average value of the ratio of the real to the imaginary part of the nucleon - nucleon scattering amplitudes .
we use the experimental values for @xmath78 , @xmath79 , @xmath80 and @xmath81 for the incident energy 240 mev / nucleon listed in ref .
@xcite .
notice that the reaction cross section of the core nucleus @xmath1ne is simply given by _
r(^30ne)=dd ( 1-| s_c| ^2 ) , [ sgm30 ] with @xmath64 given in eqs .
( [ sc ] ) and ( [ chic ] ) @xcite .
[ conf_exp ] ne obtained with the particle - rotor model as a function of the one neutron separation energy for @xmath0ne , @xmath82 . the upper and the lower panels show the results for the configuration with @xmath21=0.2 and 0.55 for @xmath0ne , respectively .
the dashed lines are the calculations in the adiabatic limit , while the solid lines are the calculations with the finite rotational energy .
the dotted lines denote the reaction cross sections for @xmath1ne .
the upper and the lower shaded regions indicate the experimental interaction cross sections for @xmath0ne and @xmath1ne , respectively .
, title="fig : " ] ne as a function of the one neutron separation energy .
the meaning of each line is the same as in fig .
we now numerically evaluate the reaction cross sections for @xmath16ne .
the upper and the lower panels of fig .
2 show the results for the configurations with @xmath53 and 0.55 , respectively .
since the measured one - neutron separation energy @xmath82 of @xmath0ne has a large error bar , s@xmath10 = 0.29 @xmath11 1.64 mev @xcite , we show the calculated reaction cross sections as a function of @xmath82 . the upper and
the lower shaded regions in each panel indicate the experimental interaction cross sections for @xmath0ne and @xmath1ne@xcite , respectively .
the dashed lines are the calculations in the adiabatic limit , while the solid lines take into account the finite rotational energy of the core nucleus .
we show also the reaction cross sections for @xmath1ne with the dotted lines .
as one can see , the results of prm are similar to those in the adiabatic limit for both the configurations with @xmath21=0.2 and 0.55 . in the adiabatic limit , since each component of ( @xmath83 ) with different values of @xmath84 has the same radial wave function , it is the total @xmath85 probability , summed with different @xmath84 values , that is relevant to the halo structure . on the other hand , due to the non - adiabatic effect
, the wave function [ @xmath86 is spatially most extended in the prm@xcite .
for the configuration with @xmath53 and @xmath87 mev , the probability for the total @xmath85 component is 54.9@xmath88 in the adiabatic limit , which is almost equal to the probability for the [ @xmath89 component in the prm , that is , 54.2@xmath88 .
the halo structure therefore retains even when the finite excitation energy is taken into account in the prm . for the configuration with @xmath54 ,
on the other hand , the total @xmath85 probability in the adiabatic limit is 25.7% and the probability for the [ @xmath89 component in the prm is 2.1 @xmath88 .
therefore , the halo structure disappears for this configuration when the finite rotational energy is taken into account@xcite .
the small difference between the solid and the dashed curves in the lower panel of fig.2 reflects this fact .
nevertheless , the halo contribution to the reaction cross section does not seem large in this mass region , and the adiabatic approximation still works for the reaction cross sections . in order to see the relation between the halo structure and the reaction cross section more clearly , fig .
3 shows the rms radii for those configurations as a function of @xmath82 .
the behaviors of the rms radii are qualitatively the same as the reaction cross sections shown in fig .
the rms radius is almost constant for @xmath55=0.55 when the finite rotational energy is taken into account , that is consistent with the disappearance of the halo structure .
the rms radii increase as the one - neutron separation energy , @xmath56 , decreases for the other cases , indicating the halo structure .
notice that in contrast to the rms radii and the coulomb breakup cross section , the reaction cross section is less sensitive to the extended density distribution , since the inner part of the density distribution also contributes to the cross section .
the calculated density distribution of @xmath0ne for the configuration with @xmath55=0.2 and @xmath56=0.2 mev is shown in figure 4 .
the dashed and the solid lines are the results for @xmath1ne and @xmath0ne , respectively . since the effects of the finite excitation of the core nucleus on the reaction cross sections and rms radii are small for this configuration , we calculate the density for @xmath0ne in the adiabatic limit , that is , nilsson model , which is defined in the body - fixed frame . the upper and the lower panels show the density distributions in the direction of the symmetry axis of the core nucleus @xmath1ne and in the direction perpendicular to the symmetry axis , respectively .
the density distribution has an exponentially extended tail , indicating the halo structure for this nucleus .
notice that the density distribution is proportional to @xmath90 for the pure @xmath85 state with @xmath91 , and it is extended more in the direction of the symmetry axis compared to the direction perpendicular to it .
ne with @xmath55=0.2 and @xmath56=0.2 mev obtained with the adiabatic limit of the particle - rotor model , that is , the nilsson model .
@xmath92 is the angle from the symmetry axis of the core nucleus .
the upper and the lower panels show the results in the direction of the symmetry axis of the core nucleus @xmath1ne and in the direction perpendicular to the symmetry axis , respectively . ] in figure 2 , the calculated reaction cross sections appear to reproduce the experimental data for both the configurations with @xmath55=0.2 and 0.55 .
however , the increase of the calculated reaction cross section from @xmath1ne to @xmath0ne is much smaller for the configuration with @xmath54 than with @xmath53 .
this is because the probability of the @xmath85 component is much larger at @xmath53 than at @xmath54 .
notice that these different @xmath85 probabilities stem from the nilsson levels [ 330 1/2 ] at @xmath53 and [ 321 3/2 ] at @xmath54 in the adiabatic limit , respectively .
the difference between the reaction cross sections for @xmath0ne and @xmath1ne may be approximately identified as the one - neutron removal cross section for @xmath0ne@xcite , _
-1n(^31ne)~_r(^31ne)-_r(^30ne ) .
[ rem ] figure 5 shows the one - neutron removal cross sections for @xmath0ne on the carbon target so obtained as a function of the one - neutron separation energy for @xmath0ne .
the shaded region indicates the experimental data with the incident energy of 230 mev / nucleon @xcite .
the thick and the thin lines are the one - neutron removal cross sections for the configuration with @xmath55=0.2 and 0.55 , respectively . the dashed and the solid lines are the results in the adiabatic limit and with the finite rotational excitation , respectively .
one can clearly see that the results with the configuration with @xmath55=0.2 reproduces the experimental data , while the configuration with @xmath55=0.55 is inconsistent with the experimental one - neutron removal cross section .
we conclude that the configuration with @xmath53 is a very promising candidate for the ground state of the deformed halo nucleus @xmath0ne , which is consistent with the analysis of the coulomb dissociation cross section of @xmath0ne with the prm @xcite .
ne defined as a difference between the reaction cross sections for @xmath0ne and @xmath1ne as a function of the one neutron separation energy for @xmath0ne , @xmath82 .
the thick and thin lines are the results for the configuration with @xmath55=0.2 and 0.55 for @xmath0ne , respectively .
the dashed lines are the calculations in the adiabatic limit of the particle - rotor model , while the solid lines are the calculations with the finite rotational energy .
the shaded region indicates the experimental one - neutron removal cross section for @xmath0ne , taken from ref .
] ne as a function of the deformation parameter @xmath21 . for @xmath0ne ,
the one neutron separation energy is assumed to be @xmath87 mev , with the configuration corresponding to that in the upper panel of fig .
the meaning of each line is the same as in fig .
we next investigate the deformation dependence of the reaction cross sections of @xmath16ne for the configuration which reproduces the experimental data of the reaction and the one - neutron removal cross section at @xmath53 . with
the method explained in section iia , this configuration remains the ground state in the range of the deformation parameter , [email protected] .
given the uncertainties of the potential parameters , this configuration may be the ground state even at around @xmath94 , as suggested by the anti - symmetrized molecular dynamics ( amd ) calculation for @xmath95ne @xcite , see table ii in ref .
figure 6 shows the reaction cross sections for @xmath16ne for the separation energy of @xmath87 mev , as a function of the deformation parameter in the region of @xmath96 .
the dashed line is the result in the adiabatic limit , while the solid line is the result with the finite rotational energy .
the result for @xmath1ne is shown with the dotted line .
the reaction cross sections for @xmath16ne smoothly increase only by about 0.01 b from @xmath97 to @xmath98 due to the deformation of the core density .
this may be understood in terms of the deformation dependence of the rms radius , see _
e.g. , _ eq . ( 1 ) in ref .
the total @xmath85 probability in the adiabatic limit varies from 54.9% to 57.0% as the deformation parameter changes from @xmath97 to @xmath98 .
consequently , the deformation dependence of the reaction cross section of @xmath0ne is small , as far as the same configuration is concerned , i.e. , the nilsson level [ 330 1/2 ] in the adiabatic limit .
the experimental reaction cross section can thus be reproduced within the region of @xmath99 .
we have discussed the reaction cross sections for @xmath16ne with the particle - rotor model . assuming the system with the deformed core nucleus and one valence neutron for @xmath0ne , the finite rotational excitation energy of the core nucleus @xmath1ne is taken into account . in order to calculate the reaction cross section on the carbon target
, we have used the optical limit approximation of the glauber theory .
we have considered two configurations with the spin - parity of @xmath52 at @xmath100 0.2 and 0.55 as candidates for the ground state of @xmath0ne , corresponding to the nilsson levels [ 330 1/2 ] and [ 321 3/2 ] in the adiabatic limit , respectively .
the effect of the finite rotational energy changes the probability of each component in the wave function , especially the proportion of the [ @xmath89 component as well as the [ @xmath101 component . we have found that the non - adiabatic effects on the reaction cross sections for these two configurations are small , and
it is concluded that the nilsson model works reasonably well for the reaction cross section for @xmath0ne .
we have also found that the difference of the reaction cross sections between @xmath0ne and @xmath1ne is much larger for the configuration with @xmath53 than for the configuration with @xmath54 , leading to a consistent description for one - neutron removal cross section for @xmath53 .
interaction cross sections of ne isotopes have been measured from @xmath102ne to @xmath103ne by takechi _
et al._@xcite .
the data show a large odd - even staggering for @xmath104ne , which has been understood in terms of the pairing anti - halo effect @xcite . as we have found , the adiabatic approximation works well for the reaction cross sections for neutron - rich ne isotopes
. it would thus be interesting to describe the deformed nucleus @xmath103ne with _
e.g. , _ the hartree - fock - bogoliubov ( hfb ) method taking into account the pairing interaction and then evaluate the interaction cross section in the adiabatic approximation of the glauber theory . a work toward this direction
is now in progress .
this work was supported by the global coe program `` weaving science web beyond particle - matter hierarchy '' at tohoku university , and by the japanese ministry of education , culture , sports , science and technology by grant - in - aid for scientific research under the program numbers ( c ) 22540262 and 20540277 . | using the glauber theory , we calculate reaction cross sections for the deformed halo nucleus @xmath0ne . to this end , we assume that the @xmath0ne nucleus takes the @xmath1ne + @xmath2 structure . in order to take into account the rotational excitation of the core nucleus @xmath1ne , we employ the particle - rotor model ( prm ) . we compare the results to those in the adiabatic limit of prm , that is , the nilsson model , and
show that the nilsson model works reasonably well for the reaction cross sections of @xmath0ne .
we also investigate the dependence of the reaction cross sections on the ground state properties of @xmath0ne , such as the deformation parameter and the @xmath3-wave component in the ground state wave function . |
in condensed matter physics the interplay among strain , strong correlation , and spin - orbit coupling ( soc ) has been becoming extremely important.@xcite this leads to abundant unconventional phenomena such as magnetic ordering , interface superconductivity , metal - insulator transitions , and more recently topological insulators.@xcite in a mott transition , in particular , the ratio @xmath8 of correlation strength @xmath9 to bandwidth @xmath10 determines charge- , spin- , and orbital - orderings .
@xmath10 can be controlled by external factors such as doping , pressure , and strain by a proper choice of substrate .
the strain has been of great interest for applications of band engineering.@xcite recently , another type of the mott insulator induced by correlation effects which are assisted by spin - orbit coupling ( soc ) , named relativistic mott insulator , has been observed in cubic double perovskite ba@xmath11naoso@xmath12 and a few ir - based oxides , which show substantially strong soc.@xcite the relativistic mott transition was also proposed to occur even in the @xmath13 li@xmath11rho@xmath0 system.@xcite in this transition , soc leads to the removal of degeneracy in partially filled bands or narrowing of the bandwidth , and then strong correlation selects a filled orbital .
however , pure correlation effects often lead to a metal - insulator transition in @xmath7 systems , where the strength of soc is negligible compared to that of correlation . here
, we will address a strain and soc - driven mott transition , accompanied by orbital - ordering ( oo ) , in the @xmath1 bacro@xmath0 thin film synthesized recently.@xcite in this system , strength of soc is indeed tiny , but crucial to open a gap .
this has been rarely observed in @xmath7 systems .
tetravalent @xmath14 chromate perovskites have recently drawn attention due to their atypical and controversial properties.@xcite although the orthorhombic antiferromagnet cacro@xmath0 is metallic , it was suggested that correlation effects play an important role.@xcite another metallic antiferromagnet srcro@xmath0 has a tetragonally distorted structure due to partial oo of @xmath15.@xcite through correlated first principles calculations , gupta _ et al .
_ proposed that lattice distortions in an ultrathin srcro@xmath0 film lead to oo , which in turn induces ferroelectricity.@xcite recently , zhu _ et al .
_ synthesized the missing member , bacro@xmath0 , of the tetravalent chromate perovskites , on srtio@xmath0 ( 001 ) surfaces.@xcite as observed in the isostructural srcro@xmath0,@xcite the distortion factor of @xmath16 is as small as approximately 0.5 % , and this prevents the cro@xmath12 octahedron from rotating and tilting ( see below ) .
the magnetic measurements show hysteresis with a small coercive field and small saturated moment of 0.028 @xmath17 with curie temperature @xmath18 k , indicating a weak ferromagnet ( fm ) or canted antiferromagnet ( afm ) .
the resistivity data follow an insulating behavior with a kink at @xmath19 k and indicate an activation energy gap of @xmath20 0.38 ev .
they also performed correlated first principles calculations using the local spin density approximation plus hubbard @xmath9 ( lsda+u ) approach , but failed to obtain the observed insulating state.@xcite motivated by the failure of lsda+u , giovannetti _ et al .
_ carried out dynamical mean field calculations and concluded that a jahn - teller distortion in the plane of the tetragonal cell with two different cr - o bond lengths would lead to an insulating state.@xcite the nature of the insulating state of bacro@xmath0 can be unveiled from first principles calculations , including @xmath9 and ( relativistic ) soc effects , using an accurate all - electron full - potential electronic method of wien2k.@xcite considering only @xmath9 and soc simultaneously throughout the lsda+soc+u approach , we obtained the experimentally observed insulating state,@xcite indicating a crucial role of soc . in contrast to the jahn - teller distortion incorporated scenario,@xcite our results show that this insulating state is induced by the small structure distortion and soc activated by narrowing @xmath10 , resulting from the enlarged volume due to the strain of the substrate .
the accompanying oo stabilizes the c - type afm ( c - afm ) , which is antiparallel in - plane and parallel along the @xmath21-axis , and leads to an unusually large orbital moment @xmath6 equal to half of the spin moment @xmath22 in this system .
interestingly , as shown below , this feature also drives a considerable @xmath6 in the apical oxygens .
according to the experiment by zhu _
, the tetragonal perovskite phase appears to be unstable in the bulk.@xcite we optimized the crystal structure to investigate stability of the tetragonal structure . during the optimization in the tetragonal phase ,
the ratio @xmath16 was fixed at the value obtained in the experiment.@xcite the structural parameters were optimized until forces were smaller than 2 mry / a.u . as shown in fig .
[ str](a ) , our optimized volume is about 18% less than that of the experiment . considering that the l(s)da usually underestimates an experiment volume by at most several percents , as observed for the similar compound srcro@xmath0,@xcite this difference is significant
this implies that the srtio@xmath0 substrate significantly expands the volume of bacro@xmath0 , consistent with the instability of the bulk state.@xcite remarkably , near the optimized volume the cubic and tetragonal phases are nearly degenerate .
we will return to this issue below . when @xmath21-axis lattice parameter in perovskites decreases , the in - plane lattice parameter tends to increase to prevent the volume from varying.@xcite this leads to the distortion from the cubic to the tetragonal structures .
for the experiment volume , the tetragonal phase has an energy 57 mev lower than the cubic phase , consistent with the experiment .
we also investigated a possible tilting and rotating of the cro@xmath12 octahedron .
however , our trials to obtain this gadolinium orthoferrite gdfeo@xmath0-type structure always converged to the tetragonal phase observed in the experiment,@xcite indicating that the gdfeo@xmath0-type distortion is not energetically favored in this system . from hereon ,
unless stated otherwise , we use the experiment lattice parameters @xmath23 and @xmath24 .@xcite in wien2k , the basis size was determined by r@xmath25k@xmath26=7 and augmented - plane - wave sphere radii ( 2.5 for ba , 2.01 for cr , and 1.82 for o ) .
the brillouin zone was sampled with up to a dense @xmath27-mesh of @xmath28 .
additionally , linear response full - phonon calculations were performed with quantum espresso.@xcite these calculations were carried out with a 2@xmath292@xmath292 @xmath30-mesh , a 14@xmath2914@xmath2914 @xmath27-mesh , an energy cutoff of 40 ry , and a charge cutoff of 400 ry . .
individual spin @xmath22 , orbital @xmath6 , and net @xmath31 moments of the c - afm ( in units of @xmath17 ) , obtained from lsda+soc and lsda+soc+u at @xmath32 ev calculations .
the moments of the planar oxygens are zero by symmetry . [
cols="^,^,^,^,^,^,^,^ " , ] [ table1 ]
the results of the generalized gradient approximation plus hubbard @xmath9 ( gga+u ) , performed by giovannetti _
et al._,@xcite show that the in - plane jahn - teller distortion is stable .
this is consistent with our lsda+u calculations , as shown in fig .
[ str](b ) .
however , the addition of soc to the lsda+u approach relieves the jahn - teller distortion in both metallic and insulating states .
this rules out the scenario of mott transition driven by the jahn - teller distortion in this system , which was proposed by giovannetti _
et al_. in the cro@xmath12 octahedron , the tetragonal distortion , shortening of @xmath21 lattice parameter , lowers the center of @xmath33 and results in the splitting of the @xmath34 manifold into the singlet @xmath33 and doublet @xmath35 .
thus , for this @xmath14 configuration , the doublet is half - filled in a spin ordered state .
initially , we applied the on - site coulomb repulsion @xmath9 to the cr ions to obtain the observed insulating state , using the lsda+u approach , in which the effective on - site coulomb repulsion @xmath36 was used where @xmath37 is the hund s exchange integral .
as observed in the previous calculation,@xcite inclusion of only @xmath38 can not produce an insulating state even for large values such as @xmath39 ev , which is already beyond a reasonable value for this system .
therefore , we add to the lsda+u approach the effects of soc in the magnetization direction along the @xmath21-axis , which is the high symmetry direction in the crystal structure . in the @xmath34 manifold , soc transforms \{@xmath33 , @xmath40 , @xmath41 } into \{@xmath33 , @xmath42}.@xcite in the lsda+soc+u calculations , an energy gap appears at the critical value @xmath43 ev,@xcite and finally reaches to the experimentally observed gap of @xmath44 ev at @xmath32 ev .
figure [ str](c ) shows the change in the energy gap as a function of @xmath38 in the c - afm .
the @xmath45 is somewhat larger than the value obtained from the constrained random - phase approximation,@xcite since @xmath9 acts on the mixture of cr @xmath7 and o @xmath46 due to the strong @xmath47 hybridization , which can be measured in the dos of fig .
it is noteworthy that there is a regime around @xmath45 in which both metallic and insulating solutions are obtained , as is extensively discussed in literature.@xcite however , for this system an insulating state is always energetically favored over a metallic state at the same @xmath38 in this regime . to disclose the nature of the insulating state
, we analyze the electronic and magnetic structures in detail .
figure [ str](c ) shows that the c - afm is energetically favored over fm , which has a much lower energy than the nonmagnetic state ( by 216 mev / f.u . in lsda+soc ) , regardless of the strength of @xmath38 . as @xmath38 increases
, the difference in energy ( @xmath48 ) per formula unit between these two states monotonically decreases with a slope of @xmath49 . using a simple heisenberg model @xmath50 , for this spin @xmath51 configuration the superexchange parameter @xmath37 is estimated by @xmath52 mev at @xmath32 ev .
the c - afm alignment has been observed in both cacro@xmath0 and srcro@xmath0.@xcite now , we focus on the c - afm alignment in both the metallic and insulating phases , where the latter is obtained from the lsda+soc+u calculations at @xmath32 ev . the total and atom - projected doss for both the metallic and insulating phases are displayed in fig .
the corresponding band structures with the emphasized fatbands of the @xmath33 and the @xmath53 orbitals are given in fig .
in fact , the electronic structures of lsda+soc are nearly identical to those of lsda ( not shown here ) , since the strength of soc is only several mev near the fermi energy . in the metallic state
, the partially filled @xmath34 manifold is separated from the unfilled @xmath54 manifold in most regimes , but the @xmath33 band touches the bottom of the conduction band with a @xmath55 character at the @xmath56 point .
as shown in the dos , @xmath10 of the @xmath57-filled @xmath34 manifold consistent with the @xmath14 configuration , is about 1.3 ev .
this leads to an estimated value of the nearest neighbor hopping parameter @xmath58 ev .
this value is a half of that for the nonmagnetic state , indicating a considerable narrowing due to the afm alignment .
our calculated moments are listed in table [ table1 ] .
the spin moment @xmath22 for cr is 1.68 @xmath17 , substantially lower than the formal value of 2 @xmath17 due to the strong @xmath47 hybridization .
these features can be visualized in a spin - density isocountour plot , as displayed in the top panel of fig .
[ spin ] . in the metallic phase ,
the minority of cr atoms has a @xmath59 character , while the majority of cr indicates an equally occupied @xmath34 manifold .
this clear visibility of the minority density indicates a strong @xmath60 hybridization . in the insulating state ,
the upper hubbard band is at the bottom of the conduction band , while the occupied @xmath33 and the lower hubbard band are mostly around 4 ev , as shown in the right panel of fig .
the corresponding spin - density plot is displayed in the middle panel of fig .
the cr majority character represents a mixture of @xmath61 . as given in table [ table1 ] , the cr orbital moment @xmath62 @xmath17 is antialigned to the corresponding @xmath22 , leading to the net moment @xmath63 @xmath17 .
this antialignment indicates that the lower hubbard band has mostly the @xmath64 character for the spin - up cr , but the @xmath65 character for the spin - down cr .
this zigzag oo in the @xmath66-plane is consistent with the c - afm alignment .
a similar oo was observed in the spinel znv@xmath11o@xmath67 , as including both @xmath9 and soc.@xcite although some @xmath7 systems show an unquenched @xmath6,@xcite this substantially large @xmath6 is rare in @xmath7 systems .
interestingly , the spin density of the apical oxygens clearly indicates the @xmath68 character , even though the @xmath22 of the apical oxygens is only 0.095 @xmath17 .
@xmath22 of the apical oxygen is nearly compensated by the considerable @xmath69 @xmath17 , which is unprecedented for an oxygen ion.@xcite however , recently this unusual feature has been observed in a @xmath70 system.@xcite
we also performed lsda+soc+u calculations for our optimized volume .
as expected from widening @xmath10 of the @xmath34 manifold by @xmath71% due to compression of the volume , in the optimized structure a gap appears at @xmath72 ev , about 50 % larger than in the experiment structure .
thus , if the optimized volume can be attained experimentally , this system would be metallic .
these results suggest that change in strain ( or applying a pressure ) leads to an insulator - metal transition in this system .
to investigate the stability of these structures , we carried out linear response full - phonon calculations for both the experiment volume and our optimized volume .
the results are displayed in fig .
[ phonon ] . in good agreement with the instability in the bulk,@xcite
there are several imaginary frequencies in the experiment volume .
however , our results show that in the optimized volume both the tetragonal and cubic perovskites are stable .
this suggests that these phases would be achieved by a proper choice of substrate or a high pressure technique , requiring further experiments .
considering the tiny difference in energy ( @xmath731 mev ) between the cubic and the tetragonal phases , our results imply that the cubic structure would be stabilized by quantum fluctuations.@xcite in summary , we have carried out _ ab initio _ calculations including both correlation and soc effects to investigate a tetragonally distorted perovskite bacro@xmath0 , whose the volume seems to be considerably enhanced by the strain of the srtio@xmath0 substrate .
inclusion of both soc and @xmath9 causes this @xmath14 system to change to insulating at @xmath74 ev , though the metallic state remains unchanged on applying @xmath9 only .
this indicates a crucial role of soc in this mott insulator , which is uncommon in @xmath7 systems . in the insulating state
, one electron occupies @xmath33 and the other alternatively occupies @xmath2 or @xmath3 orbital , resulting in an unquenched cr orbital moment @xmath6 which is antialigned to the cr spin moment @xmath5 .
this is consistent with the ground state of c - afm , implying that the observed small moment is due to a little canted spin .
remarkably , due to the spin and orbital orderings of the cr ions the spin density of the apical oxygen ions clearly shows the admixture of @xmath68 characters , leading to the considerable spin and orbital moments antialigned each other .
these may be directly probed by further experiments like x - ray resonant spectroscopy .
we acknowledged c. kim for useful discussions on a thin film growth , and w. e. pickett and d. kashinathan for useful communications on an oo , and hb rhee for a critical reading for the manuscript .
this research was supported by nrf-2013r1a1a2a10008946 .
a. c. komarek , s. v. streltsov , m. isobe , t. moller , m. hoelzel , a. senyshyn , d. trots , m. t. fernndez - daz , t. hansen , h. gotou , t. yagi , y. ueda , v. i. anisimov , m. gruninger , d. i. khomskii , and m. braden , phys .
* 101 * , 167204 ( 2008 ) .
p. giannozzi , s. baroni , n. bonini , m. calandra , r. car , c. cavazzoni , d. ceresoli , g. l. chiarotti , m. cococcioni , i. dabo , a. dal corso , s. fabris , g. fratesi , s. de gironcoli , r. gebauer , u. gerstmann , c. gougoussis , a. kokalj , m. lazzeri , l. martin - samos , n. marzari , f. mauri , r. mazzarello , s. paolini , a. pasquarello , l. paulatto , c. sbraccia , s. scandolo , g. sclauzero , a. p. seitsonen , a. smogunov , p. umari , and r. m. wentzcovitch , j. phys.:condens .
matter * 21 * , 395502 ( 2009 ) . as including @xmath9
, we used two popular double counting schemes of the so - called around mean field ( amf)@xcite and the fully - localized limit ( fll)@xcite .
however , in the lsda+soc+u calculations , only the amf scheme leads to open a gap above @xmath45 .
this is consistent with the fact that amf is often more suitable than fll for the moderately correlated systems , as discussed in detail in literature.@xcite thus , we will address the results from amf , here . | using _ ab initio _ calculations , we have investigated an insulating tetragonally distorted perovskite bacro@xmath0 with a formal @xmath1 configuration , the volume of which is apparently substantially enhanced by a strain due to srtio@xmath0 substrate .
inclusion of both correlation and spin - orbit coupling ( soc ) effects leads to a metal - insulator transition and in - plane zigzag orbital - ordering ( oo ) of alternating singly filled @xmath2 and @xmath3 orbitals , which results in a large orbital moment @xmath4 antialigned to the spin moment @xmath5 in cr ions .
remarkably , this ordering also induces a considerable @xmath6 for apical oxygens .
our findings show metal - insulator and oo transitions , driven by an interplay among strain , correlation , and soc , which is uncommon in @xmath7 systems . |
although casimir effect has attracted a lots of attention due to its important roles in various fields of physics @xcite , there have been relatively few works on casimir effect for quantum fields subject to robin boundary conditions .
robin boundary conditions arise naturally in the study of casimir effect for electromagnetic fields in perfectly conducting spheres @xcite .
however , in parallel plate configuration , perfectly conducting boundary conditions only lead to dirichlet boundary conditions and neumann boundary conditions .
nevertheless , it was shown that in some geometries , robin boundary conditions can be considered as extensions of perfectly conducting boundary conditions , where the robin coefficients are related to the skin - depth parameter describing the finite penetration of the field into the boundary @xcite .
robin type conditions are needed for conformally invariant field theories in the presence of boundaries , since robin boundary conditions can be made conformally invariant , but ordinary neumann boundary conditions can not .
the casimir effect for fields with robin boundary conditions was studied from different perspectives and under different context in @xcite for parallel plate geometry , in @xcite for a @xmath0-dimensional ball , in @xcite for spherical shell geometry , and in @xcite for the configuration of two spheres . on the other hand , quantum fields with nonlocal boundary conditions
can also be reduced to one with robin boundary conditions @xcite .
the importance of the robin boundary conditions to spacetime models and quantum gravity has been emphasized in @xcite . for randall - sundrum spacetime model @xcite ,
robin boundary conditions are natural for scalar fields and fermion fields , where the robin coefficients are related to the curvature scale and the boundary mass terms of the fields @xcite .
casimir effect in braneworld models with robin boundary conditions on the branes was discussed extensively in @xcite . despite the importance of the robin boundary conditions in field theory , the research on casimir effect with robin boundary conditions
has been limited to a small community , and very few of the works has taken into account the influence of finite temperature .
the goal of this paper is to consider the thermal casimir effect for the parallel plate configuration .
since spacetimes with extra dimensions have become prevalent in physics , we are going to discuss along this line .
our setup is the same as in @xcite , i.e. , we consider a pair of codimension one parallel plates in the background spacetime @xmath1 , where @xmath2 is the @xmath3-dimensional minskowski spacetime , and @xmath4 is a compact internal space .
in fact , casimir effect on parallel plates in a background spacetime of this form has been investigated in @xcite , when the boundary conditions are either dirichlet boundary conditions or neumann boundary conditions in the case of scalar fields , and perfect conductor boundary conditions in the case of electromagnetic fields . in the cases considered in @xcite ,
the dimension of the macroscopic space is @xmath5 and the internal space is assumed to be the simplest one - dimensional compact space @xmath6 or the more general @xmath7-dimensional toroidal manifold @xmath8 . in @xcite , the internal space is generalized to arbitrary compact manifold and the parallel plates are allowed to be finite with arbitrary cross section .
the temperature corrections to the casimir effect in the scenario of @xcite were considered in @xcite .
it has been shown that for massless scalar field with dirichlet boundary conditions on both plates or neumann boundary conditions on both plates , the casimir force is always attractive , at any temperature and for any spacetime geometry .
for massless scalar field with dirichlet boundary condition on one plate and neumann boundary condition on another plate , the casimir force is always repulsive .
the present work can be considered as a generalization of @xcite to take into account the finite temperature correction , or a generalization of @xcite where the boundary conditions are the more general robin boundary conditions instead of the dirichlet boundary conditions or neumann boundary conditions .
we are mainly interested in studying the dependence of the casimir force on the interplay between the spacetime geometry , temperature and boundary conditions . as in @xcite
, we first consider the casimir force acting on a piston which moves freely inside a closed cylinder with arbitrary cross section ( see fig .
[ f1 ] ) . in section
[ ce ] , the casimir energy in the two regions divided by piston is computed using cut - off method . in section [ cf ] , the casimir force acting on the piston is derived . the limit where one end of the cylinder is moved to infinity is equivalent to two parallel plates embedded orthogonally inside an infinitely long cylinder . letting the cross section of the plates become infinitely large give the usual infinite parallel plate configuration .
this piston scenario introduced by cavalcanti @xcite has the advantage that the contribution to the casimir force due to the vacuum fluctuations of the quantum fields in the region between the plates and the region outside the plates are both taken into account , and the divergences of the casimir forces from the two regions will cancel . in section [ acf ] , we study in detail the sign of the casimir force at small and large plate separations . in this article , we use the units where @xmath9 .
as in @xcite , we consider a massive scalar field @xmath10 in a background spacetime of the form @xmath11 , where @xmath2 is the @xmath3-dimensional minkowski spacetime and @xmath4 is an @xmath7-dimensional compact internal manifold .
let @xmath12 be the spacetime metric , where @xmath13 , @xmath14 for @xmath15 and @xmath16 is a riemannian metric on @xmath4 , and @xmath17 .
the equation of motion for a massive scalar field @xmath10 is @xmath18 we will first consider the casimir force acting on a movable piston inside a closed cylinder of arbitrary cross section as shown in fig .
let the cylinder be the region @xmath19\times \omega\times \mathcal{n}^n$ ] in @xmath20 , where @xmath21 is the cross section of the piston , which we assume to be a simply connected region in @xmath22 .
the position of the piston is denoted by @xmath23 .
if the right chamber of the cylinder is infinitely long , i.e. , @xmath24 , we obtain the configuration equivalent to two codimension one parallel plates located at @xmath25 and @xmath23 embedded in an infinitely long cylinder . for the boundary conditions
, we assume that the field @xmath10 satisfies dirichlet boundary conditions on the curved surface @xmath19\times{\partial}\omega\times\mathcal{n}^n$ ] of the cylinder . the interesting part is the boundary conditions on the three plates perpendicular to the @xmath26-direction , which we take to be robin boundary conditions : @xmath27 where @xmath28 are nonnegative dimensionless constants , and @xmath29 are constants with dimension of length . for @xmath30 , @xmath31 and @xmath32
can not be zero simultaneously .
@xmath33 corresponds to the dirichlet boundary conditions , @xmath34 corresponds to the neumann boundary conditions , and the general robin boundary conditions only depend on the ratio @xmath35 .
we would like to remark that our @xmath35 is equal to the @xmath36 in @xcite . to find the casimir force acting on the piston , one does not have to calculate the casimir energy of the region outside the cylinder , since it is independent of the position of the piston @xcite .
therefore , we only have to compute the casimir energies inside the left chamber and the right chamber .
it is easy to see that the casimir energy inside the right chamber can be obtained from the casimir energy inside the left chamber by replacing @xmath37 with @xmath38 , @xmath39 with @xmath40 and @xmath41 with @xmath42 . using separation of variables
, one can show that the eigenmodes of the field @xmath10 confined in the left chamber satisfying the boundary conditions specified above are given by @xmath43for @xmath44 , @xmath45 is an eigenfunction with eigenvalue @xmath46 for the laplace operator with dirichlet boundary conditions on @xmath47 . for @xmath48
, @xmath49 is an eigenfunction with eigenvalue @xmath50 for the laplace operator on @xmath4 . by convention ,
when @xmath51 , @xmath52 is a constant function with eigenvalue @xmath53 .
the boundary conditions imposed on @xmath25 and @xmath23 imply that @xmath54this system gives a nontrivial solution for @xmath55 if and only if @xmath56 satisfies the following condition : @xmath57the zeros of @xmath58 are either real or purely imaginary .
as is shown in @xcite , these zeros are all simple . in the following ,
we assume that @xmath58 does not have imaginary zeros .
this is the case if @xcite : @xmath59 notice that @xmath60 is always a solution of .
however , it is easy to check that @xmath61 gives to a nontrivial function @xmath62 satisfying if and only if @xmath63 , i.e. , in the case of neumann boundary conditions on both plates .
we let @xmath64 be all the positive solutions of .
then the set of eigenfrequencies @xmath65 for the field @xmath10 is given by @xmath66where @xmath67 in the case @xmath63 , we have to let @xmath68 starts from @xmath69 instead of @xmath70 , and @xmath71 .
the finite temperature casimir energy inside the left chamber is given by @xmath72where @xmath73 is a cut - off parameter .
as was shown in @xcite , up to the term of order @xmath74 , @xmath75-\psi(1)-\log 2 + 1}{2\sqrt{\pi}}c_{\text{cyl},d+1}-\frac{t}{2}\left(\zeta_{\text{cyl } , t}'(0)+\log[\mu^2 ] \zeta_{\text{cyl } , t}(0)\right),\end{split}\]]where @xmath76 is a normalization constant with dimension length@xmath77 , @xmath78 is the finite temperature zeta function defined by @xmath79 ^ 2\right)^{-s } , \end{split}\ ] ] and @xmath80 , @xmath81 , are heat kernel coefficients defined by @xmath82it can be shown that ( see appendix [ a3 ] ) these coefficients are linear functions of @xmath37 .
moreover , the coefficient of @xmath37 in @xmath80 is independent of the robin coefficients @xmath83 , @xmath84 . using a more general form of the generalized abel - plana formula @xcite ( see appendix [ a1 ] )
, we compute @xmath85 and @xmath86 in appendix [ a2 ] .
the results substituted into give the casimir energy in the left chamber : @xmath87 ^ 2}-\alpha_1\right)\left ( \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2}-\alpha_2\right)}{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_1\right)\left ( \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2\right)}e^{-2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}\right ) , \end{split}\]]where @xmath88 and @xmath89 are terms that are independent of @xmath37 , and @xmath89 is also independent of the robin coefficients . in fact
, @xmath89 can be interpreted as the vacuum energy per unit length in the @xmath26 direction that would present in the region between the plates if the plates are absent .
our assumption that @xmath58 does not have imaginary zeros will make each term under the logarithm in nonzero .
however , if @xmath90 or @xmath91 , we may get terms that are infinite . to avoid this kind of complications , from now on we only consider the case where @xmath92 and @xmath93 .
for the piston system shown in fig .
[ f1 ] , the casimir force acting on the piston is given by @xmath94upon differentiation with respect to @xmath37 , the terms @xmath88 and @xmath95 that are independent of @xmath37 will be killed . on the other hand , for the term proportional to @xmath37 in @xmath96
, we find that it is equal to @xmath97 , which is independent of @xmath37 .
therefore , this term will also be killed after differentiation with respect to @xmath37 . consequently , all the terms that would diverge when @xmath98 will not contribute to the casimir force . therefore , we can set @xmath99 after taking derivative with respect to @xmath37 to obtain @xmath100 where @xmath101 ^ 2}}{\frac{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_1\right)\left ( \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2\right)}{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi p t]^2}-\alpha_1\right)\left ( \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2}-\alpha_2\right)}e^{2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}-1},\]]and @xmath102 notice that even though different robin conditions are imposed on the walls @xmath25 and @xmath103 , the divergent terms of the casimir energy will not contribute to the casimir force acting on the piston . therefore the casimir force acting on the piston is independent of the regularization procedure , as in the piston system for scalar fields with dirichlet or neumann boundary conditions .
a crucial reason why this is so is that the term @xmath89 is independent of the robin coefficients .
it is easy to see that @xmath104 vanishes as @xmath24 . in other words
, @xmath105 can be interpreted as the casimir force acting between two parallel plates embedded orthogonally inside an infinitely long cylinder . in the following , we will drop the superscript @xmath106 when we discuss the casimir force between parallel plates . setting @xmath107 or @xmath63 or @xmath108 in
, one obtains respectively the casimir force acting on a pair of parallel plates with dirichlet - dirichlet , neumann - neumann or dirichlet - neumann boundary conditions .
the results coincide with the results derived in @xcite .
notice that the casimir force between parallel plates @xmath109 comes from the derivative with respect to the third term @xmath110 ^ 2}-\alpha_1\right)\left ( \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2}-\alpha_2\right)}{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_1\right)\left ( \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2\right)}e^{-2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}\right)\]]in , which was called the interaction term @xcite . usually this is also regarded as the renormalized casimir energy between the plates .
it has the property that it vanishes when the plate separation @xmath37 goes to infinity .
the expression for the casimir force between parallel plates shows that the casimir force is always attractive ( negative ) if the same boundary conditions ( i.e. , @xmath111 ) are imposed on both the plates .
this is a special case of the theorem @xcite which states that the casimir force between two bodies with the same property is attractive . from
, we also find that the casimir force decays exponentially as the separation of the plates @xmath37 becomes large . as a function of the temperature @xmath112 , shows that the high temperature leading term of the casimir force is linear in @xmath112 , given by the sum of the terms with @xmath113 . on the other hand ,
let @xmath114 and @xmath115 be the size of the cross section @xmath21 and the size of the internal space @xmath4 respectively .
recall that @xmath116 . since @xmath117 and @xmath118 , shows that the casimir force @xmath109 between a pair of parallel plates goes to zero when the size @xmath119 of the cross section goes to zero or the mass @xmath120 goes to infinity .
when the size @xmath121 of the internal space goes to zero , all the terms with @xmath122 go to zero , and the limit of the casimir force is the casimir force in the @xmath3-dimensional minkowski spacetime given by the sum of the terms with @xmath51 . for the low temperature behavior , we use the abel - plana summation formula with @xmath123 ^ 2}}{\frac{\left ( \beta_1 \sqrt{m_{j ,
l}^2+[2\pi tz]^2}+\alpha_1\right)\left ( \beta_2 \sqrt{m_{j , l}^2+[2\pi tz]^2}+\alpha_2\right)}{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi tz]^2}-\alpha_1\right)\left ( \beta_2\sqrt{m_{j , l}^2+[2\pi tz]^2}-\alpha_2\right)}e^{2a\sqrt{m_{j , l}^2+[2\pi t z]^2}}-1}.\ ] ] the term @xmath124 on the right - hand side of gives the zero temperature casimir force acting on the parallel plates : @xmath125the term @xmath126gives @xmath127this term is independent of @xmath37 and the robin coefficients @xmath128 , @xmath84 . finally , it is easy to check that the poles of @xmath129 are exactly at @xmath130with residues @xmath131therefore , @xmath132from these
, we find that the temperature correction to the casimir force is @xmath133 in the case of @xmath5 with dirichlet - dirichlet boundary conditions , @xmath107 and @xmath134 , we find that this formula agrees with the formula ( 3.7 ) we derived in @xcite . from , it is obvious that the temperature correction goes to zero exponentially fast when the temperature @xmath112 approaches zero . when @xmath135 , since @xmath136 , the leading term of the temperature correction is the first term on the right hand side of that is independent of @xmath37 .
the next thing we would like to investigate is the limit of the casimir force when the cross section of the plates is infinitely large , i.e. @xmath137 . in this case
, we have to consider the casimir force density acting on the plates , which is defined as @xmath138this limit is given by replacing @xmath139 by @xmath140 and turning the summation over @xmath141 into an integral : @xmath142after a change of variables , we find that the finite temperature casimir force density acting on a pair of infinite parallel plates is given by @xmath143 ^ 2}}^{\infty}\frac{\left(x^2- m_l^2-[2\pi p t]^2\right)^{\frac{d_1 - 3}{2}}x^2dx}{\frac{(\beta_1 x+\alpha_1)(\beta_2x+\alpha_2)}{(\beta_1 x-\alpha_1)(\beta_2x-\alpha_2)}e^{2ax}-1 } , \end{split}\]]where @xmath144the corresponding interaction term of the casimir energy density is @xmath145 ^ 2}}^{\infty } \left(x^2- m_l^2-[2\pi p t]^2\right)^{\frac{d_1 - 3}{2}}x\\&\hspace{3cm}\times\log\left\{1-\frac{(\beta_1 x-\alpha_1)(\beta_2x-\alpha_2)}{(\beta_1 x+\alpha_1)(\beta_2x+\alpha_2)}e^{-2ax}\right\}dx . \end{split}\ ] ] one can check that in the case of dirichlet - dirichlet , neumann - neumann or dirichlet - neumann boundary conditions , agrees with the results derived in @xcite .
applying the prescription to the zero temperature casimir force , we find that the zero temperature casimir force density acting on a pair of infinite parallel plates is @xmath146this agrees with the result of @xcite .
the finite temperature casimir force density is the sum of the zero temperature casimir force density and the thermal correction @xmath147 given by @xmath148 in the massless case , the sum of the terms with @xmath51 in the first term on the right hand side of has to be replaced by @xmath149we observe that for massive case , the temperature correction term to the casimir force density acting on a pair of infinite parallel plates is exponentially suppressed .
however , for the massless case , the leading term of the thermal correction is of order @xmath150 , and this leading term is independent of the boundary conditions imposed on the plates . the low temperature expansion of the interaction term of the casimir energy density can be obtained by directly integrating and .
it is given by @xmath151 the leading behavior of the finite temperature casimir force acting on a pair of parallel plates when @xmath152 ( i.e. , the plate separation is much larger than the size of the extra dimensions , but much smaller than the size of the cross section ) can be obtained from the corresponding leading behavior of the casimir force density acting on infinite parallel plates .
we need to consider the case of high temperature and the case of low temperature separately . in the high temperature regime , @xmath153 .
the leading term of the casimir force when @xmath152 and @xmath154 is obtained from the @xmath155 term in : @xmath156if @xmath157 and @xmath158 ( both non - dirichlet conditions ) , then in the limit @xmath159 and @xmath160 , we find that the leading term is @xmath161this leading term is the same as for the case of @xmath107 ( dirichlet - dirichlet boundary conditions ) .
if @xmath162 and @xmath158 ( one dirichlet and one non - dirichlet ) , then in the limit @xmath160 , the leading term is @xmath163notice that if both plates have non - dirichlet boundary conditions , the leading behavior is the same as both plates having dirichlet boundary conditions . in this case , the casimir force is attractive when @xmath152 . on the other hand ,
if one plate assumes dirichlet boundary condition and the other assumes non - dirichlet boundary condition , then the leading behavior is the same as the dirichlet - neumann case , i.e. , the casimir force is repulsive in the limit @xmath152 . in the low temperature regime ,
i.e. @xmath164 , the leading behavior of the casimir force when @xmath152 can be obtained analogously from .
we find that when both plates assume non - dirichlet boundary conditions or when both plates assume dirichlet boundary conditions , the leading term is @xmath165when one plate assumes dirichlet boundary condition and the other plate assumes non - dirichlet boundary condition , the leading term is @xmath166 these zero temperature asymptotic behaviors have been observed in @xcite .
the asymptotics , , and are derived under the assumption that @xmath167 , i.e. the size of the extra dimensions are much smaller than the plate separation . in the case @xmath168 , one has to take into account the correction terms from the @xmath122 terms in and . in the other extreme
where @xmath169 , the extra dimensions play the same role as the cross section in the macroscopic spacetime .
therefore , when @xmath169 , the asymptotics of the casimir force are obtained from , , and by replacing @xmath170 with @xmath17 .
we see that the sign of the casimir force is not changed when we pass from @xmath167 to @xmath169 , as long as @xmath171 . at first sight
, it might be quite surprising to find that when the plate separation is small , the leading behavior of the casimir force when non - dirichlet boundary conditions are imposed on both plates is the same as when dirichlet boundary conditions are imposed on both plates ; but the leading behavior of the casimir force when dirichlet boundary condition is imposed on one plate and non - dirichlet boundary condition is imposed on the other plate is the same as when dirichlet boundary condition is imposed on one plate and neumann boundary condition is imposed on the other . in fact , this can be explained as follows .
the behavior of the casimir force with respect to @xmath37 is governed by the solutions @xmath56 of the function @xmath58 , i.e. , @xmath56 satisfies @xmath172when @xmath173 , @xmath174 . therefore
if @xmath157 and @xmath158 , @xmath175consequently , we find that when @xmath68 is large enough , @xmath176is close to the corresponding @xmath56 for the case where @xmath107 . on the other hand , if @xmath162 and @xmath158 , then when @xmath68 is large enough , @xmath177this implies that @xmath178is close to the corresponding @xmath56 for the case of dirichlet - neumann boundary conditions .
in this section , we analyze in more detail the sign of the casimir force .
the cases of dirichlet - dirichlet ( @xmath107 ) , neumann - neumann ( @xmath63 ) or dirichlet - neumann ( @xmath179 ) boundary conditions have been studied in @xcite , where it was proved that for dirichlet - dirichlet or neumann - neumann case , the casimir force is always attractive . for dirichlet - neumann case ,
the casimir force is always repulsive . in the following
, we will not consider these cases .
first we consider the case @xmath162 , @xmath180 and @xmath158 where dirichlet boundary condition is imposed on one of the plates , and generic robin boundary conditions on the other . in this case , the finite temperature casimir force between the plates is @xmath181 ^ 2 } } { \frac { \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2 } { \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2 } -\alpha_2}e^{2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}+1},\end{split}\]]and the zero temperature casimir force is @xmath182 = 0.49 = 0.49 = 0.49 = 0.49 if @xmath183 ^ 2}\right\}= \sqrt{\omega_{\omega,1}^2+m^2 } , \end{split}\]]then each term in the sum of and is positive .
we find that the casimir force between the plates is always repulsive and the casimir force is a monotonically decreasing function of @xmath37 .
moreover , the presence of the extra dimensions enhances the casimir force .
the condition means that the ratio @xmath184 can not be too large .
in other words , this is a close - to - neumann condition .
therefore , it is reasonable that the casimir force is repulsive as in the dirichlet - neumann case . in general , we have shown in section [ cf ] that for any @xmath158 , the casimir force is repulsive when @xmath171 .
when @xmath153 , the leading term of the casimir force is determined by the @xmath185 term in , which is repulsive if is satisfied , and attractive if is not satisfied . at zero temperature , also shows that if is not satisfied , then when @xmath186 or @xmath187 , the casimir force will eventually become attractive .
therefore we see that if is not satisfied , the casimir force will change from repulsive to attractive at any temperature . in the massless case
, the right hand side of goes to zero when the size of the cross section @xmath119 goes to infinity .
therefore , for a pair of infinite parallel plates with dirichlet boundary condition on one plate and non - dirichlet and non - neumann boundary condition on the other plate , the casimir force always change from repulsive to attractive when @xmath37 increases from 0 to @xmath188 .
= 0.49 = 0.49 = 0.49 = 0.49 an example of the first case is shown in fig .
[ f2 ] and fig .
[ f2_1 ] , where the graphs for the casimir energy and casimir force acting on a pair of parallel plates embedded in an infinitely long rectangular cylinder with cross section @xmath189 due to massless scalar field is shown .
the internal manifold is a circle with radius @xmath121 . in these figures , we show the variation of the casimir force with respect to the plate separation @xmath37 at @xmath190 and @xmath191 ( @xmath192k ) , in the case that the internal manifold has radius @xmath193 m , @xmath194 m and @xmath195 m and in the case without internal manifold @xmath196 .
one of the plates assumes dirichlet boundary condition , i.e. , @xmath197 , and the other assumes robin boundary condition with @xmath198 .
notice that @xmath199 .
these graphs show that the casimir force is repulsive at small @xmath37 and becomes attractive for @xmath200 .
the critical point @xmath201 is a minimum point of the energy .
therefore , it is a stable equilibrium point .
the position of the minimum point can be affected by the size of the extra dimension and temperature .
the temperature dependence of the casimir force when @xmath202 m and @xmath203 m are shown in fig .
they verify the linear dependence of the casimir force when @xmath204 .
notice that with plate separation @xmath203 m , the energy between the plates is of order @xmath205tev at @xmath190 and of order @xmath206tev at @xmath207tev .
the casimir energy increases by a factor of @xmath208 from @xmath190 to @xmath207tev .
therefore the temperature correction is important in the high temperature regime .
experiments on casimir effect have not been able to reach this high energy regime .
however , with the advent of technology , one can expect that future experiments will be able to explore the high temperature regime which might bring forward some new applications of casimir effect in technology .
return to fig .
[ f2_2 ] , we also notice that when the size of the extra dimension change from 0.2@xmath76 m to @xmath209 m , the casimir force change from attractive to repulsive . as we have discussed in the previous section
, this can not happen for dirichlet - dirichlet , neumann - neumann or dirichlet - neumann boundary conditions . in fig .
[ f2_3 ] , we show the dependence of the casimir force on plate separation when the internal manifold is @xmath210 , @xmath211 and @xmath212 respectively , where @xmath8 is the product of @xmath7-circles each with radius @xmath193 m .
this figure shows that contrary to the dirichlet - dirichlet , neumann - neumann or dirichlet - neumann cases , the increase in the number of extra dimensions can reduce the magnitude of the casimir force . the second case where @xmath213 , @xmath180 and @xmath158 , i.e. , neumann boundary condition is imposed on one of the plates , and generic robin boundary condition on the other is exactly the opposite of the first case .
more precisely , in this case the finite temperature casimir force and the zero temperature casimir force are given respectively by @xmath214 ^ 2 } } { \frac { \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2 } { \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2 } -\alpha_2}e^{2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}-1},\end{split}\]]and @xmath215therefore , we conclude immediately that when @xmath184 is small enough to satisfy , the casimir force is always attractive as in the neumann - neumann case . however , if @xmath184 does not satisfy , the casimir force is attractive at small plate separation , but will eventually turn to repulsive when @xmath216 or @xmath217 or @xmath218 is large enough . in the case is satisfied , we can say more . as in the dirichlet - dirichlet or neumann - neumann case
, implies that the magnitude of the casimir force is a monotonically decreasing function of the plate separation .
moreover , since each term in the sum of and is positive , and the sum of the @xmath51 terms in and corresponds to the casimir forces without extra dimensions , we find that when is satisfied , the presence of extra dimensions enhances the magnitude of the casimir force .
= 0.49=0.49 = 0.49 = 0.49 finally , we consider the generic case where @xmath219 and @xmath220 . without loss of generality , assume that @xmath221 .
as is already observed in section [ cf ] , when the robin conditions on the two plates are the same , i.e. , @xmath111 , then the casimir force is always attractive .
if @xmath184 satisfies , so does @xmath222 .
the casimir force is then always attractive and its magnitude is enhanced by the presence of extra dimensions . for any @xmath223 and @xmath224 ,
we have shown in section [ cf ] that the casimir force is always attractive when the plate separation is small enough .
in the other extreme where @xmath187 , if the temperature @xmath112 is not zero , the dominating term of the casimir force is given by the term with @xmath225 and @xmath226 in .
we conclude that when @xmath37 is large enough , then the casimir force is repulsive if @xmath227and is attractive otherwise .
if the temperature @xmath112 is zero , the same conclusion can be derived from .
[ f3 ] and fig .
[ f3_1 ] show the dependence of the casimir force on plate separation when @xmath228 . in fig .
[ f3 ] , the internal manifold is @xmath6 with different radius @xmath121 . in fig .
[ f3_1 ] , the internal manifold is @xmath229 respectively .
the robin coefficients in these graphs are @xmath230 and @xmath198 .
notice that @xmath231 .
the graphs show that the casimir force is attractive at small @xmath37 and becomes repulsive at large @xmath37 .
there is only one equilibrium point which is unstable .
another interesting case is shown in fig .
[ f4 ] , with @xmath232 and @xmath198 . in this case ,
the graph shows that the casimir force is attractive at small and at large @xmath37 as dictated by our analysis above .
however , the casimir force can become repulsive at some intermediate values of @xmath37 .
this implies that there are two equilibrium points @xmath234 and @xmath235 , one unstable and one stable .
if the initial separation of the two plates is in the range of the two equilibrium points , i.e. @xmath236 , then the two plates would tend to repulse each other until they settle at the distance @xmath237 .
= 0.49=0.49 in the figures shown in this section , the size of the internal manifold is chosen to be between @xmath238 m to @xmath209 m so as to demonstrate significant difference with the casimir force without extra dimensions .
extra dimensions of this size is not physically interesting .
one can show that if the size of the internal manifold @xmath121 is ten times smaller than the plate separation , there is no significant difference between the casimir force with or without the internal manifold .
therefore for physically interesting size of extra dimensions of order @xmath239 nm , it will not be detectable by the present experiments in casimir effect which measures casimir force between objects that are @xmath240 nm @xmath241 nm apart . before ending this section
, we would like to remark on the original setup we consider a piston moving freely inside a closed cylinder . as is shown by , in this case , the casimir force is the difference of the casimir force between the left end of the cylinder and the piston , and the casimir force between the right end of the cylinder and the piston .
since the magnitude of the casimir force is very large when the plate separation is very small , and is very small when the plate separation is very large , we can deduce that when the piston is close to one end , whether it is attracted to or pushed away from that end only depends on the robin coefficients on the piston and on that end .
however , when the piston is away from both ends , then which side it will move to depends on the robin coefficients on the piston and on the two ends .
there are some combinations of robin coefficients that will make the piston stay in the middle region .
in this article , we studied the interplay between geometry , temperature and boundary conditions on the sign and magnitude of the casimir force acting on parallel plates .
we first derived the finite temperature casimir force acting on a piston moving freely inside a closed cylinder due to a scalar field with robin boundary conditions .
it is shown that even the robin coefficients on the two ends of the cylinder are different , the casimir force acting on the piston is independent of the regularization procedure , since the divergent terms of the casimir forces from the two chambers of the cylinder divided by the piston are independent of the robin coefficients and cancel each other . by moving one end of the cylinder to infinity , we obtain the casimir force acting on a pair of parallel plates . in the high temperature regime ,
the leading term of the casimir force is linear in temperature , which shows that the casimir force has a classical limit .
a modified abel - plana summation formula is used to rewrite the casimir force which is suitable for the analysis of its low temperature behavior .
when the parallel plates has finite size , the casimir force decays exponentially when the temperature tends to zero .
in case of infinite parallel plates , the temperature correction is of order @xmath150 , where @xmath170 is the dimension of the macroscopic space .
interestingly , these behaviors are independent of the values of the robin parameters .
the sign of the casimir force when the plate separation is small and when the plate separation is large is analyzed in detail .
it is found that if dirichlet condition is imposed on one plate and non - dirichlet condition is imposed on the other plate , then the casimir force is repulsive when the plate separation is small enough .
if non - dirichlet conditions are imposed on both plates , the casimir force is attractive for small enough plate separations .
we give an explanation for these behaviors by the asymptotics of the frequencies .
when the separation between the plates becomes sufficiently large , the sign of the casimir force depends not only on the boundary conditions , but also on the geometry of the transversal dimensions .
we show that for a wide range of robin coefficients , the casimir force can change from attractive to repulsive or repulsive to attractive , giving rise to unstable equilibrium and stable equilibrium respectively .
this can be applied in nanotechnology if robin conditions is used to model the skin depths of real materials . as mentioned in the introduction ,
robin boundary conditions arise naturally in randall - sundrum spacetime model .
the results in this article is not readily transferred to the randall - sundrum model except for massless scalar field that couples conformally to scalar curvature .
there is a brief discussion of this in the zero temperature case in @xcite
. it will be interesting to consider the finite temperature casimir effect due to a bulk massive scalar field with general curvature coupling , and with general robin boundary conditions on the branes .
a special case has been considered in @xcite where the thermodynamic energy was shown to have a minimum that might give rise to brane stabilization mechanism .
here we present a more general abel - plana summation formula , which is a direct generalization of those presented in @xcite .
if @xmath242 and @xmath243 are meromorphic functions , and @xmath244 then @xmath245@xmath246here @xmath247 and @xmath248 are weight functions defined by@xmath249 this formula can be proved in the same way as in @xcite .
it is a direct consequence of the residue theorem . to recover the original abel - plana summation formula ,
let @xmath129 be a meromorphic function , and define @xmath250if for all @xmath251 , @xmath252we can apply the formula , which gives @xmath253here we assume that @xmath129 does not have poles at @xmath254
. if @xmath129 is analytic in the right - half plane , then the last three terms that contain the residues of @xmath255 in the right - half plane are identically zero . in this case
, we obtain the original abel - plana summation formula .
in this section , we want to compute the zeta function @xmath79 ^ 2\right)^{-s } , \end{split}\]]and its derivative at @xmath257 . by definition ,
@xmath258 are the zeros of @xmath259 on the right half - plane .
we can rewrite @xmath58 as @xmath260let @xmath261 , @xmath262and define @xmath263 ^ 2\right)^{-s}\frac{d}{dz}\log f_i(z),\hspace{1cm}i=0,1,2 .
\end{split}\]]it is easy to verify that the conditions are satisfied .
notice that @xmath264 has zeros at @xmath265 , @xmath266 , and @xmath267 has zeros at @xmath268 , @xmath266 . for @xmath269 ,
since @xmath270 is a holomorphic function with simple zeros , the poles of @xmath271 coincide with the zeros of @xmath270 . applying the generalized abel - plana summation formula
, we find that @xmath256 can be written as a sum of three terms : @xmath272the term @xmath273 is independent of @xmath37 : @xmath274 ^ 2\geq \left[\frac{\alpha_i}{\beta_i}\right]^2}}\left(-\left[\frac{\alpha_i}{\beta_i}\right]^2+m_{j , l}^2+[2\pi pt]^2\right)^{-s}\\&+\sum_{i=1}^2 w_{i } \sum_{\substack{k , j\in\mathbb{n } , l\in\mathbb{n}_0 , p\in{{\mathbb{z}}}\\ m_{j , l}^2+[2\pi pt]^2\leq \left[\frac{\alpha_i}{\beta_i}\right]^2}}\cos(\pi s)\left(\left[\frac{\alpha_i}{\beta_i}\right]^2-m_{j , l}^2-[2\pi
pt]^2\right)^{-s}\\ & + \frac{1}{\pi}\int_0^{\infty } \sum_{j=1}^{\infty}\sum_{l=0}^{\infty}\sum_{p=-\infty}^{\infty}\left(x^2+m_{j , l}^2+[2\pi pt]^2\right)^{-s}\sum_{i=1}^2\frac{\alpha_i\beta_i}{\alpha_i^2+\beta_i^2x^2 } dx . \end{split}\]]the first term @xmath275 ^ 2\right)^{-s}\ ] ] comes from the zero of @xmath276 at @xmath60 . if @xmath63 , we have to change the sign of this term to positive .
the second and third terms in come from the zeros of @xmath264 and @xmath267 .
the weights @xmath277 are defined so that if @xmath34 , then @xmath278 ; if @xmath279 , @xmath280 , then @xmath281 and if @xmath282 , @xmath283 , then @xmath284 . the last term in comes from the @xmath37-independent part of @xmath285 the term @xmath286 is proportional to @xmath37 , coming from the @xmath37-dependent part of : @xmath287 ^ 2\right)^{-s}dx\\ = & \frac{a}{2\sqrt{\pi}}\frac{\gamma\left(s-\frac{1}{2}\right)}{\gamma(s)}\zeta_{\omega\times\mathcal{n},t}\left(s-\frac{1}{2}\right ) .
\end{split}\]]it is interesting to note that this term is independent of the robin coefficients @xmath83 , @xmath84 .
finally the first two terms on the right hand side of give @xmath288 : @xmath289 ^ 2}}^{\infty}\sin ( \pi s ) \left(z^2-m_{j , l}^2-[2\pi p t]^2\right)^{-s } \frac{d}{dz}\log\left(1-\frac{(\beta_1
z-\alpha_1 ) ( \beta_2z-\alpha_2 ) } { ( \beta_1 z+\alpha_1 ) ( \beta_2 z+\alpha_2)}e^{-2az}\right)dz .
\end{split}\]]this term goes to zero as @xmath290 . from
, we find that @xmath291 .
therefore @xmath292 is linear in @xmath37 .
moreover , the coefficient of @xmath37 is independent of the robin coefficients .
the derivative of @xmath288 at @xmath257 can be easily computed from and we find that @xmath293 ^ 2}-\alpha_1\right)\left ( \beta_2\sqrt{m_{j , l}^2+[2\pi p t]^2}-\alpha_2\right)}{\left ( \beta_1 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_1\right)\left ( \beta_2 \sqrt{m_{j , l}^2+[2\pi p t]^2}+\alpha_2\right)}e^{-2a\sqrt{m_{j , l}^2+[2\pi p t]^2}}\right ) .
\end{split}\ ] ]
in this section , we show that the heat kernel coefficients @xmath80 , @xmath81 , are linear functions of @xmath37 . moreover , the coefficients of @xmath37 is independent of the robin coefficients @xmath128 , @xmath84 . from the theory of elliptic operators
, we have @xmath294 using inverse mellin transform , we find that@xmath295where @xmath296as in appendix [ a2 ] , we find that @xmath297where @xmath298 ^ 2+m^2\right)^{-s}\delta\left(m^2-\left[\frac{\alpha_i}{\beta_i}\right]^2\right ) + \sum_{i=1}^2 w_{i } \cos(\pi s)\left(\left[\frac{\alpha_i}{\beta_i}\right]^2-m^2\right)^{-s}\delta\left(\left[\frac{\alpha_i}{\beta_i}\right]^2-m^2\right)\\ & + \frac{1}{\pi}\int_0^{\infty } \left(x^2+m^2\right)^{-s}\sum_{i=1}^2\frac{\alpha_i\beta_i}{\alpha_i^2+\beta_i^2x^2 } dx .
\end{split}\]]is independent of @xmath37 , @xmath299is proportional to @xmath37 and independent of the robin coefficients , and @xmath300is an analytic function of @xmath301 on the complex plane . therefore , all the poles of the function @xmath302 come from @xmath303 .
it is easy to see that the poles of @xmath304 are at @xmath305 and all of them are simple poles . for @xmath306 , the poles can only comes from the term @xmath307now , @xmath308this shows that @xmath309has an asymptotic expansion of the form @xmath310 as @xmath311 .
it is then standard to show that the function @xmath312 has simple poles at @xmath313 .
it follows that the function @xmath314 only has simple poles at @xmath315 .
applying residue theorem to , we find that @xmath316using the fact that @xmath317we conclude from that the heat kernel coefficients @xmath80 are linear functions of @xmath37 .
moreover , the coefficient of @xmath37 in @xmath80 does not depend on the robin coefficients @xmath83 , @xmath84 since @xmath318 does not .
this project is funded by ministry of science , technology and innovation , malaysia under e - science fund 06 - 02 - 01-sf0080 .
we are grateful to a. flachi and the anonymous referee for the helpful comments and suggestions .
mostepanenko and n.n .
trunov , sov .
. phys . * 42 * ( 1985 ) , 812 .
lebedev , sov .
jetp * 83 * ( 1996 ) , 423 .
lebedev , phys . atom .
* 64 * ( 2001 ) , 1337 .
setare and r. monsouri , class .
quantum grav . * 18 * ( 2001 ) , 2659 . l.c .
de albuquerque and r. m. cavalcanti , j. phys .
a * 37 * ( 2004 ) , 7039 .
z. bajnok , l. palla and g. tak@xmath319cs , phys .
d * 73 * ( 2006 ) , 065001 .
setare , int . j. mod
. phys . a * 22 * ( 2007 ) , 1771 .
d. dean , phys .
e * 79 * ( 2008 ) , 011108 .
e. elizalde , s.d .
odintsov and a. a. saharian , phys .
d * 79 * ( 2009 ) , 065023 .
m. bordag , e. elizalde and k. kirsten , j. math . phys .
* 37 * ( 1996 ) , 895 .
saharian and m.r .
setare , class .
quantum grav .
* 20 * ( 2003 ) , 3765 .
t. emig , n. graham , r. l. jaffe , and m. kardar , phys .
d * 77 * ( 2008 ) , 025005 .
saharian and g. esposito , j. phys .
a * 39 * ( 2006 ) , 5233 .
moss , class .
quantum grav .
* 6 * ( 1989 ) , 759 .
g. esposito and a.yu kamenshchik , class .
quantum grav .
* 12 * ( 1995 ) , 2715 .
saharian , phys .
d * 73 * ( 2006 ) , 064019 .
setare , phys .
b * 637 * ( 2006 ) , 1 .
saharian and m.r .
setare , phys .
b * 637 * ( 2006 ) , 5 .
saharian , phys .
d * 74 * ( 2006 ) , 124009 .
saharian and m.r .
setare , jhep * 0702 * ( 2007 ) , 089 . | in this article , we study the finite temperature casimir effect for scalar field with robin boundary conditions on two parallel plates in a background spacetime that has a compact internal manifold with arbitrary geometry .
the finite temperature casimir force acting on the plates is derived using the piston approach , with which we show that the divergences of the casimir forces from the region between the plates and the region outside the plates cancel each other .
the sign and asymptotic behaviors of the casimir force at different limits such as small and large plate separation , low and high temperature , are studied in detail .
the range of robin coefficients where the casimir force are always attractive or always repulsive are determined .
there are also some range of the robin coefficients where the casimir force is shown to be attractive at small plate separation and repulsive at large plate separation , or vise versa . |
the standard cosmological scenario assumes that before big bang nucleosynthesis ( bbn ) , the universe was dominated by radiation with the hubble parameter evolving like @xmath0 , where @xmath1 is the energy density of radiation and @xmath2 is the scale factor of the universe .
however , if we imagine to add a significant fraction of scalar energy density @xmath3 to the background radiation at some time in the past , and if @xmath4 , then @xmath3 would decay more rapidly than @xmath1 , but temporarily increase the global expansion rate .
a measurable effect of this modification to the standard scenario is the anticipation of the freeze - out time of neutralinos and the consequent enhancement of their relic abundance , as explicitly calculated in @xcite . in a flat universe , a scalar field with potential
@xmath5 obeys the equations @xmath6 for any given time during the cosmological evolution , the relative importance of the scalar energy density w.r.t . to matter and radiation in the total energy density
@xmath7 depends on the initial conditions , and is constrained by the available cosmological data on the expansion rate and large scale structure .
if the potential @xmath5 is of the runaway type , the initial stage of the scalar evolution is typically characterized by a period of so called ` kination ' @xcite during which the scalar energy density @xmath8 is dominated by the kinetic contribution @xmath9 , giving @xmath10 .
after this initial phase , the field comes to a stop and remains nearly constant for some time ( ` freezing ' phase ) , until it eventually reaches an attractor solution @xcite .
then , if we modify the standard picture according to which only radiation plays a role in the post - inflationary era and suppose that at some time @xmath11 the scalar contribution was small but non negligible w.r.t .
radiation , then at that time the expansion rate @xmath12 should be correspondingly modified . during the kination phase the scalar to radiation energy density ratio evolves like @xmath13 , and so the scalar contribution would rapidly fall off and leave room to radiation . in this way , we can respect the bbn bounds and at the same time keep a significant scalar contribution to the total energy density just few red - shifts before at bbn ( @xmath14 ) would imply @xmath15 at @xmath16 and @xmath17 at @xmath18 , if the scalar field is undergoing a kination phase . ] .
the increase in the expansion rate @xmath19 due to the additional scalar contribution would anticipate the decoupling of particle species and result in a net increase of the corresponding relic densities .
as shown in @xcite , a scalar to radiation energy density ratio @xmath20 at bbn would give an enhancement of the neutralino codensity of roughly three orders of magnitude .
the enhancement of the relic density of neutralinos requires that at some early time the scalar energy density was dominating the universe .
this fact raises a problem if we want to identify the scalar contribution responsible for this phenomenon with the quintessence field @xcite .
indeed , the initial conditions must be such that the scalar energy density is sub - dominant at the beginning , if we want the quintessence field to reach the cosmological attractor in time to be responsible for the presently observed acceleration of the expansion @xcite . for initial conditions
@xmath21 we obtain instead an ` overshooting ' behaviour : the scalar field rapidly rolls down the potential and after the kination stage remains frozen at an energy density much smaller than the critical one .
however , as shown in @xcite , more complicated dynamics are possible if we relax the hypothesis of considering a single uncoupled scalar .
the presence of several scalars and/or of a small coupling with the dark matter fields could modify the dynamics in such a way that the attractor is reached in time even if we started , for example , in the overshooting region
. * more fields . *
consider a potential of the form @xmath22 , with m a constant of dimension mass . in this case , as discussed first in @xcite , the two fields dynamics enlarges the range of possible initial conditions for obtaining a quintessential behaviour today .
this is due to the fact that the presence of more scalars allows to play with the initial conditions in the fields values , while maintaining the total initial scalar energy density fixed .
doing so , it is possible to obtain a situation in which for a fixed @xmath23 in the overshooting region , if we keep initially @xmath24 we actually produce an overshooting behaviour , while if we choose to start with @xmath25 ( and _ the same _ @xmath23 ) it is possible to reach the attractor in time .
this different behaviour emerges from the fact that , if at the beginning @xmath26 then , in the example at hand , @xmath27 and so @xmath28 ( the smaller field ) will run away more rapidly and tend to overshoot the attractor , while @xmath29 ( the larger field ) will move more slowly , join the attractor trajectory well before the present epoch and drive the total scalar energy density towards the required value
. * interaction . *
suppose now that the quintessence scalar is not completely decoupled from the rest of the universe . among the possible interactions ,
two interesting cases are the following : @xmath30 if we add @xmath31 or @xmath32 to @xmath33 , the potential will acquire a ( time - dependent ) minimum and the scalar field will be prevented from running freely to infinity . in this way , the long freezing phase that characterises the evolution of a scalar field with initial conditions in the overshooting region can be avoided .
effective interaction terms like @xmath31 in eq .
( [ inter ] ) are discussed in refs .
@xcite , where it is shown that supersymmetry breaking effects in the early universe can induce mass corrections to the scalar lagrangian of order @xmath34 . the second type of interaction ( @xmath32 in eq .
( [ inter ] ) ) can emerge in the context of scalar - tensor theories of gravity ( see , for example ref .
@xcite ) , in which a metric coupling exists between matter fields and massless scalars .
the scalar field equation in this context is modified by the presence of an additional source term @xmath35 where @xmath36 is a generic function of @xmath37 and @xmath38 is the trace of the matter energy - momentum tensor @xmath39 .
the case @xmath40 corresponds to a scalar field decoupled from the rest of the world .
p. j. steinhardt , l. m. wang and i. zlatev , 1999 , phys .
d * 59 * , 123504 ; s. c. ng , n. j. nunes and f. rosati , 2001 , phys .
d * 64 * , 083510 ; a. de la macorra and g. piccinelli , 2000 , phys .
d * 61 * , 123503 . | results of a recent study @xcite on how the quintessence scalar could affect the relic abundance of dark matter particles are presented . |
accretion powers a wide range of energetic objects , systems ranging in size from low - mass x - ray binary systems to the disks surrounding supermassive black holes in active galaxies and quasars .
theoretical models of these systems must be , necessarily , highly simplified .
one such simplification that is often employed is to assume a time stationary disk .
but the rapid time - variability seen in space- and ground - based observations emphasizes that accretion disks are highly dynamic systems .
these observations highlight the need to solve the fully time - dependent equations describing accretion disks , and that , in turn , requires numerical techniques .
recently , time - dependent numerical simulations have been applied to the problem of accretion disk transport .
the discovery of the magnetorotational instability ( mri ) in accretion disks ( balbus & hawley 1991 ) has elucidated the physical basis for this transport : magnetohydrodynamic ( mhd ) turbulence . at the same time , the development of practical three - dimensional mhd codes brought the study of disk mhd turbulence , transport , and evolution into the computational domain . to date , most of the numerical simulations of the mri have been carried out in a local disk approximation known as the shearing box ( hawley , gammie & balbus 1995 , hereafter hgb ) . in the shearing box model , one considers a frame corotating with the angular velocity at a fiducial radius @xmath2 . by restricting the computational domain to small excursions from this fiducial radius
, one can reduce the geometry ( but not the dynamics ) to a simple cartesian system while retaining tidal and coriolis forces .
local shearing box simulations ( e.g. , hgb ; hawley , gammie , & balbus 1996 , hereafter hgb2 ; brandenburg et al . 1995 ; matsumoto & tajima 1995 ; stone et al .
1996 ) have demonstrated that mhd turbulence is the nonlinear outcome of the mri , and that outward angular momentum transport is its natural consequence .
the total stress @xmath3 is dominated by the maxwell , or magnetic , component rather than the kinematic , or reynolds stress . in disks
the amplitude of this stress is often parameterized with the shakura - sunyaev ( 1973 ) alpha formulation , @xmath4 , where @xmath5 is the pressure . in the local simulations
typical stress values range from @xmath6 to @xmath7 .
these simulations are more fully reviewed in balbus & hawley ( 1998 ) .
an important step in increasing the realism of numerical disk studies is to move from this local approximation to fully global disk simulations .
the difficulties of three dimensional global simulations and , in particular , the demands such simulations place on computer hardware , make this step challenging .
in general , the time- and length - scale differences between the extended radial scales of an accretion disk and the scale of the mhd turbulence are too great to be fully resolved
. nevertheless , global simulations are now possible for somewhat restricted ranges in radius , height , and angle .
several global mhd disk simulations have already been done .
one of the first , by matusmoto & shibata ( 1997 ) , followed the evolution of a thick torus embedded in a weak vertical magnetic field .
what transpires is highly dynamic : the outer layers of the torus slough inward in a process referred to as avalanche accretion .
this avalanche accretion is the global consequence of the radial streaming motion ( the `` channel solution '' ) which is the nonlinear manifestation of the vertical field instability in the local limit ( hawley & balbus 1992 ; goodman & xu 1994 ) .
the vertical field winds up as it is twisted by the torus rotation , and material is accelerated upward and outward along the field lines . in subsequent work , matsumoto ( 1999 ) evolved constant angular momentum tori containing an initially toroidal field .
the nonaxisymmetric mri rapidly leads to turbulence .
magnetic energy is amplified and approaches a value of @xmath8 . again
the onset and appearance of the nonlinear stage of the instability are understandable in light of what has been seen in local simulations . in this simulation , however , the mri has a global consequence : the resulting mhd stress transports angular momentum outward , rapidly changing the angular momentum distribution from constant with radius to nearly keplerian .
recently , armitage ( 1998 ) performed a global calculation of a cylindrical ( no vertical structure ) keplerian disk using the zeus code ( stone & norman 1992a , b ) .
the initial magnetic field was @xmath9 , with the sine function chosen to vary once over the central part of the disk .
this initial state is unstable , and turbulence rapidly develops along with significant angular momentum transport ; @xmath10 values are on order 0.1 . as in local simulations ,
the maxwell stress is several times the reynolds stress .
the magnetic field in the turbulent disk is dominated by the toroidal component , and the turbulence itself is characterized by structures with small azimuthal wavenumbers @xmath1 .
hawley & balbus ( 1999 ) similarly computed cylindrical models of keplerian disks and constant angular momentum tori .
they found in all cases that the mri rapidly developed leading directly to turbulence and significant angular momentum transport .
global simulations such as these have become possible only recently with advances in available computer speed and memory .
as computer capacities continue to increase , ever more ambitious simulations will become possible .
the aim of this paper is to begin a systematic effort to develop such three - dimensional global mhd disk simulations . to a large extent
, this paper will focus on the technical aspects of such simulations , and the gross dynamics of the mri in global disks .
what type of simulations are currently practical ?
what sort of diagnostics are useful ?
what are the advantages and limitations of simplifications such as axisymmetry or the cylindrical approximation ?
as these questions are addressed , however , several significant conclusions about accretion torus dynamics will be obtained .
the plan of this paper is as follows : 2 discusses technical aspects of the global simulations , including the equations , the numerical algorithm , test problems , initial conditions , and diagnostic procedures . in
3 the results from a range of simulations are presented .
two important simplifications are investigated : two - dimensional axisymmetry , and a cylindrical gravitational potential .
these approximations reduce the computational complexity of global simulations , and a comparison with the full three - dimensional treatment allows their limitations to be assessed .
simulations of fully global three dimensional tori follow . finally
4 summarizes the conclusions from this work .
the dynamical evolution of an accretion disk is governed by the equations of ideal mhd , i.e. , @xmath11 @xmath12 @xmath13 @xmath14 where @xmath15 is the mass density , @xmath16 is the specific internal energy , @xmath17 is the fluid velocity , @xmath5 is the pressure , @xmath18 is the gravitational potential , and @xmath19 is the magnetic field vector . in the present study
we adopt an adiabatic equation of state , @xmath20 , and ignore radiation transport and losses . since there is no explicit resistivity or physical viscosity , the gas can heat only through adiabatic compression , or in shocks through explicit artificial viscosity , @xmath21 .
artificial viscosity appears as an additional pressure term in equation ( [ ene ] ) , i.e. , @xmath22 ( stone & norman 1992a ) .
the absence of physical shear viscosity in these equations is worth noting .
this is the actual state of the plasma : the molecular viscosity is utterly negligible .
accretion disks evolve due to transport resulting from what is normally referred to as anomalous viscosity .
but this is not an additional term in the equations ; transport results from the dynamics contained within the above equations . in some simulations below
, the code employs the usual newtonian gravitational potential , @xmath23 , where @xmath24 is the spherical radius . to deal with the inner boundary in a more realistic manner
, however , the code also makes use of the pseudo - newtonian potential ( paczysky & wiita 1980 ) to model the relativistic effects associated with a schwarzschild metric ( e.g. , minimum stable orbit ) .
this inner boundary permits supersonic ( and super - alfvnic ) accretion off of the inner radial grid , reducing the likelihood of unphysical influences from the boundary conditions .
the pseudo - newtonian potential has the form @xmath25 where @xmath26 is the `` gravitational radius '' ( akin to the black hole horizon ) , here set equal to 1 . for this potential ,
the keplerian specific angular momentum is @xmath27 and @xmath28 .
here we set @xmath29 , and , for both gravitational potentials , @xmath30 .
this determines the units of time in the simulation with @xmath31 at @xmath32 for a newtonian gravitational potential . all times given
will be reported in these units ; the corresponding orbital periods at locations in the tori will also be given where appropriate .
equations ( [ mass])-([ind ] ) are solved using time - explicit eulerian finite differencing .
the global disk code is written in in cylindrical coordinates , @xmath33 .
the center of the coordinate system is excised , i.e. , the radial coordinate begins at a nonzero minimum @xmath34 ; this avoids the coordinate singularities associated with the axis . in some simulations
, the gravitational potential is assumed to be cylindrical ; this removes the vertical component of gravity , and reduces the number of @xmath35 grid zones required .
when the vertical component of gravity is included , the @xmath35 coordinate is centered on the equatorial plane .
the angle @xmath36 is periodic and covers the full @xmath37 or some integer fraction of @xmath37 . in some simulations ,
only the two - dimensional axisymmetric system @xmath38 will be computed .
a schematic of the computational domain is shown in figure 1 .
most of the three dimensional global simulations are run on parallel computer systems with a version of the global code that uses the message passing interface ( mpi ) library for interprocess communication .
the full computational domain is divided into overlapping subgrids , one for each processor .
boundary zone values are passed between processors using explicit mpi routines . in the description of
what follows , however , this subdivision will be ignored ; the physical domain size and the number of grid zones discussed will be the totals .
the mhd algorithm now has a long history of use .
the magnetic field is evolved with the constrained transport ( ct ) approach of evans and hawley ( 1988 ) that preserves the constraint @xmath39 .
the ct scheme was designed to work in any general curvilinear coordinate system , and here the addition of appropriate geometric terms easily adapts the procedure to cylindrical coordinates .
the algorithm uses information propagated along alfvn characteristics to solve a restricted set of characteristic equations for time - advanced fields and electromotive forces within the ct framework .
this approach is known as the method of characteristics constrained transport ( mocct ) algorithm , and its first implementation is described in detail in stone & norman ( 1992b ) ; the current version is described in hawley & stone ( 1995 ) .
considerable effort has gone into determining the most satisfactory scheme for the individual numerical terms in the present application .
local conservation of angular momentum is particularly important , and a number tests using hydrodynamic keplerian disks on a variety of grids suggest that some numerical approaches are better than others . specifically , it is best to evolve @xmath40 , where @xmath41 is the specific angular momentum , as a fundamental variable ( as opposed to angular velocity , or specific angular momentum alone ) and to use consistent advection ( norman , wilson & barton 1980 ) where the advection of all variables is tied to the advection of the density . to minimize errors associated with the coordinate singularity located at @xmath42
we make use of a regularized form of the operator @xmath43 ( evans 1986 ) .
a variety of test simulations were used to verify code performance .
the hydrodynamic implementation was tested with keplerian disks and a series of simulations of constant angular momentum tori similar to those studied by hawley ( 1991 ) .
constant and near - constant angular momentum tori are subject to particularly vigorous growth of the papaloizou - pringle instability ( papaloizou & pringle 1984 ) , especially when such tori are `` slender , '' meaning that their cross section is small compared to the radius of the pressure maximum .
slender tori are unstable to the principal mode of the papaloizou - pringle instability , with growth rates near the maximum value .
these tests gave results that were consistent with past simulations .
for an mhd test , the grid was moved out to large radius and an isolated vertical magnetic field flux tube was embedded in a keplerian flow .
this test is a three - dimensional version of the axisymmetric simulations described in hawley & balbus ( 1991 ) .
again the results were fully consistent with the earlier work .
most of the simulations discussed below consider the problem of the nonlinear evolution of the magnetorotational instability in accretion tori .
tori have significant internal pressure gradients that balance the gravitational and centrifugal accelerations in hydrostatic equilibrium .
this pressure gives the torus a vertical thickness @xmath44 that can be comparable to its radius @xmath45 .
significant departures from keplerian angular momentum distributions are possible as the torus becomes ever thicker ; the limit is the constant angular momentum torus .
a particularly useful feature of these tori for numerical simulations is that they are well - defined equilibria that can be completely contained on the finite domain of a computational simulation .
this allows their evolution to proceed ( at least initially ) independent of the grid boundary conditions . for an adiabatic equation of state , a gravitational potential @xmath18 , and an assumed rotation law @xmath46 , the density in the torus
is determined by @xmath47 where @xmath48 is the keplerian angular momentum at the pressure maximum and @xmath49 is a constant of integration that establishes the zero pressure surface of the torus .
a given torus is specified by the angular velocity gradient @xmath50 ( with @xmath51 corresponding to constant angular momentum ) , the radial location of the pressure maximum @xmath52 , and the radial location of the inner edge of the torus , @xmath53 ( which determines @xmath49 ) . outside of the torus
the grid is filled with a cold gas with low density .
a floor value of the density and internal energy is applied to ensure that these zones do not become evacuated , or develop extremely high sound speeds .
the floor value of density should be set low enough so that its contribution to the total mass or energy is not significant at any place on the grid . here
the floor value is @xmath54 and the typical torus initial density maximum is @xmath55 .
tori are known to be subject to the purely hydrodynamic global instability discovered by papaloizou & pringle ( 1984 ) .
the papaloizou - pringle instability has been extensively investigated both with perturbation analyses , and with several numerical simulations .
the mechanism of the instability was elucidated by goldreich , goodman & narayan ( 1986 ) and involves the establishment of a global wave that gains energy through the exchange of angular momentum between the inner and outer region of the disk .
amplification requires wave reflection at at least one disk boundary .
the nonlinear evolution of the papaloizou - pringle instability has been followed for both slender ( hawley 1987 ) , and relatively wide ( blaes & hawley 1988 ; hawley 1991 ) tori .
in the slender torus case , the instability saturates as a highly nonaxisymmetric orbiting fluid ellipse ( the `` planet '' solution ; hawley 1987 ) which itself proves to be an equilibrium solution ( goodman , narayan , goldreich 1987 ) .
wider tori saturate in spiral pressure waves ( blaes & hawley 1988 ; hawley 1991 ) .
torus linear stability and , as we shall see , nonlinear dynamics are quite different when a magnetic field is present .
the largest growth rate of the mri greatly exceeds those typical of the hydrodynamic papaloizou - pringle instability .
linear stability analyses bear this out .
curry & pudritz ( 1996 ) and ogilvie & pringle ( 1996 ) have done global linear mhd stability analyses of cylindrical systems ( @xmath56 gravitational potential ) with a variety of initial angular momentum distributions .
they find rapidly growing unstable modes for a wide range of equilibria and field strengths .
although global linear analyses are most appropriate for strong magnetic fields whose most unstable wavelengths are comparable to the size of the torus , weak fields have unstable wavelengths that are much smaller than the torus dimensions .
the essential physics of the mri in these circumstances is local , and the nonlinear evolution should be similar to that seen in the local shearing box simulations .
this is , of course , a testable proposition . in any case
, global simulations will provide information regarding the ultimate consequences of the mri for the evolution , and , indeed , existence of thick tori .
obviously , the initial equilibrium torus can not persist in the face of vigorous instability and the resulting angular momentum transport .
its overall global evolution , however , may present novel features , and the end state of its evolution may represent a more realistic accretion disk structure .
a global simulation will involve a grid with a million to a billion zones , run for possibly hundreds of thousand of timesteps .
a major task is to develop useful diagnostics that concisely and adequately characterize the essential behavior of an evolving disk in such a simulation .
the simplest type of diagnostic is the volume integral of a quantity over the entire grid .
for example , the total mass is @xmath57 time histories are computed of this quantity and others such as total kinetic and magnetic energies ( by component ) , and the radial mass flux @xmath58 through both the inner and outer boundaries .
standard steady - state thin disk models are often expressed in terms of vertically averaged quantities . here
such values can be computed by integrating density - weighted quantities over @xmath35 and @xmath36 .
the vertically- and azimuthally - averaged mass density is @xmath59 the averaged mass flux , @xmath60 , is defined by ( [ mdot ] ) , and one can construct similar averages for other quantities . for non - keplerian disks the radial dependence of the average specific angular momentum will be of interest , and this can be computed from @xmath61 the stress @xmath62 consists of a magnetic component ( the maxwell stress ) @xmath63 and a kinematic component ( the reynolds stress ) @xmath64 the value of the perturbed angular velocity @xmath65 is defined for a keplerian disk by @xmath66 . for other angular momentum distributions , including average angular momentum distributions that change with time , we adopt an alternative definition for the perturbed orbital velocity @xmath65 in terms of the difference between the total instantaneous angular momentum flux , and the mass flux times the average angular momentum .
the difference then represents the excess or deficit angular momentum transport due to orbital velocity fluctuations compared to the mean .
specifically , @xmath67 together the maxwell and reynolds stress make up the total stress , which , in the shakura - sunyaev ( 1973 ) parameterization , is set equal to @xmath68 . the stresses observed in the simulations
can be similarly scaled using the vertically averaged pressure to derive an @xmath10 as a function of @xmath45 . in the simulations
, @xmath10 varies with both time and space .
its utility is mainly to provide a familiar point of reference for characterizing the observed stress .
the local shearing box simulations ( hgb ; hgb2 ) found that the maxwell stress , which dominates the total transport , is highly correlated with the total magnetic pressure .
this suggests an alternate stress parameter , the `` magnetic alpha value '' , defined as @xmath69 . in a steady state disk , the total stress , maxwell plus reynolds ,
would be equal to radial angular momentum flux carried inward by the net radial drift velocity , minus the angular momentum flux at the inner boundary of the disk where the stress is presumed to vanish .
the latter term is generally ignored for radii large compared to the inner disk boundary . in the simulations none of these assumptions
hold _ a priori _ : the typical torus radius is comparable to the inner disk boundary , the stress does not vanish at the inner disk boundary , and the disk will not be in steady state .
in fact , the simulations will prove to be highly dynamic with rapid variations in both time and space .
this is best appreciated from time - dependent animations which also have been used as diagnostics for these global simulations .
while reporting time- and space - averaged quantities tends to obscure this feature , such averages are more practical for a written summary .
in this section the results from a range of three dimensional simulations are presented .
the simulations are listed in the summary table which gives a model designation , the grid resolution used , the gravitational potential , the initial torus angular momentum distribution @xmath50 , the initial magnetic field topology , and the time to which the simulation was run .
the first set of simulations make use of a cylindrical gravitational potential @xmath70 ( the cylindrical disk limit ) .
this is a significant simplification , as it permits the use of periodic boundary conditions in the vertical direction .
further , with this potential there is only one important vertical lengthscale , namely that of the mri ; this reduces the number of @xmath35 grid zones required .
finally , in the cylindrical limit one can study the evolution of models with vertical fields without the stringent courant limits due to the high alfvn speeds associated with strong fields passing through the low density region above the disk .
the first two cylindrical simulations assume a keplerian initial disk .
keplerian simulations provide an immediate comparison with the local shearing box results which also , for the most part , assumed keplerian flow .
consider first the run labeled ck1 in the summary table ; this is a cylindrical computational domain @xmath33 running from 1 to 4 in @xmath45 , 0 to 1 in @xmath35 , and 0 to @xmath71 in @xmath36 .
the grid has @xmath72 zones .
an outflow boundary condition is used at both the outer and inner radial boundaries , and periodic boundary conditions are used for @xmath36 and @xmath35 .
the initial disk has a constant mass density from @xmath73 to the outer boundary .
the adiabatic sound speed is constant and equal to 5% of the orbital velocity at the inner edge of the disk .
the initial magnetic field is vertical and proportional to @xmath74 between @xmath75 and 3.5 , with a maximum strength corresponding to @xmath76 .
the strength of the field , and hence the alfvn speed @xmath77 , was chosen so that the characteristic wavelength of the mri was @xmath78 at the location of the field strength maximum .
this ensures that the @xmath35 domain size and the vertical grid resolution will be adequate to resolve the fastest growing modes .
random nonaxisymmetric pressure fluctuations are added to create a full range of low amplitude initial perturbations .
as the evolution proceeds , the mri sets in , field energy is amplified , and soon the characteristic radial streaming structures ( referred to as the channel solution ) of the vertical field instability appear , much as they do in the local shearing box models . in the present simulation ,
these structures develop first at the inner part of the disk where the rotation frequency is the highest .
the amplitude of the mri becomes nonlinear by 3 orbits at the center of the grid @xmath79 ) , and filaments of strong magnetic field are carried inward and outward by fluid elements well out of keplerian balance .
these reach the outer part of the disk even before the local mri in that region becomes fully nonlinear .
thus there are two immediate global effects not seen in local simulations : linear growth rates that vary strongly with radius ( @xmath80 ) , and extended radial motions of significantly non - keplerian fluid .
in local simulations with initial vertical fields that vary sinusoidally with radius , the initial phase of the instability promptly breaks down into mhd turbulence .
this happens in the global simulation as well .
after the onset of turbulence , the disk displays many tightly wrapped ( i.e. , large radial wavenumber ) trailing spiral features with low azimuthal wavenumber , @xmath1 . the magnetic field energy saturates at about @xmath81 , with the toroidal component dominant : @xmath82 and @xmath83 .
similar ratios were found in the shearing box simulations with zero net initial magnetic field ( hgb2 ) .
the mhd turbulence produces rapid angular momentum transport and mass accretion , with @xmath10 peaking at 0.21 at @xmath84 , and declining beyond this point , dropping to 0.06 at the end of the simulation .
the maxwell stress displays a high correlation with the total magnetic pressure . after an early peak at 0.68 , @xmath85 declines with time ; at the end of the simulation @xmath86 . the ratio maxwell to reynolds stress is about 3 at the time of the largest total stress and rises to about 9 at the end of the simulation .
the overall angular momentum distribution remains nearly keplerian throughout the simulation , becoming slightly sub - keplerian outside of the original inner disk edge , and slightly super - keplerian in the region between this point and the inner grid boundary .
the mass flux , @xmath87 , varies both in space and time .
after 8 orbits of evolution , over half of the initial disk mass has been lost through the boundaries , particularly the outer boundary . since the initial keplerian disk abuts the outer boundary , and the mass increases with radius for a constant initial @xmath15 , it is relatively easy for a considerable amount of the mass to leave the grid , driven in part by a pressure gradient created by a rarefaction wave .
this simulation is similar to that of armitage ( 1998 ) except that it uses a smaller domain in the angular direction , and a constant initial density in the disk .
qualitatively , the two runs exhibit similar behavior , and both are consistent with results from local simulations .
one difference noted by armitage , is that @xmath6 in the global simulation ; this is larger than the value of @xmath88 seen in local simulations with zero net vertical field .
the larger @xmath10 is , however , consistent with what is seen in local simulations with net vertical field .
the global simulations have much larger radial and azimuthal wavelengths available to the mri .
further , the radial scale over which the vertical field sums to zero in the global simulation is much larger than in the local shearing box . here
@xmath89 initially varies between @xmath75 and 3.5 ; this range has a large variation in @xmath90 and hence a large range in growth rates of the mri .
finally , the presence of outflow radial boundaries means the net field need not remain zero over the simulation . in the local simulation ,
periodic boundary ensure that the net field can not change , and by definition , the vertical field sums to zero over a @xmath91 which is small .
reconnection in the local box should be much more efficient , reducing the total magnetic pressure and hence the total stress .
the next simulation , labeled ck2 , begins with the same keplerian disk , but with an initial toroidal field of constant strength , @xmath81 , lying between @xmath92 and 3.5 . generally speaking
, toroidal field simulations are more challenging than vertical field simulations .
although the presence of the linear instability is independent of the orientation of the background magnetic field , the fastest growing toroidal field modes are associated with large vertical and radial wavenumbers .
further , the critical azimuthal wavenumber is large , unless the magnetic energy is comparable to the rotation energy . here @xmath93 .
since there are 80 azimuthal grid zones over @xmath71 in angle , the critical wavenumbers of the instability will be resolved for this choice of field strength .
but it is clear that it takes a relatively strong toroidal field to do so for a reasonable number of grid zones .
this illustrates one of the difficulties with global models : it is necessary to reconcile the practical limitations of grid resolution with the length scales associated with the weakest , and hence most interesting , magnetic fields . with a toroidal initial field
, the disk evolves at a slower rate that ck1 .
the instability grows over the first few orbits , with the fastest rate of growth associated with the innermost radius .
the growing modes of the instability have the same appearance as seen in the local simulations ; high @xmath1 structures appear first , building to lower @xmath1 with time .
the vertical and radial structure also features high wavenumbers .
the rapid growth phase of the instability ceases after about 8 orbits at the grid center .
at this point the field exhibits a low @xmath1 , tightly wrapped structure , with rapid variations in @xmath45 and @xmath35 .
this behavior is consistent with the linear analysis ( balbus & hawley 1992 ) .
turbulence and accretion begin , and by 12 orbits over half of the disk mass has been lost from the grid ( much of it through the outer boundary ) ; what remains is piled up near @xmath94 and is slowly accreting .
the magnetic field is subthermal and dominated by the toroidal component .
the component magnetic energies at late time are quite similar to those seen in the vertical field run ck1 .
angular momentum transport is again mainly due to the maxwell stress .
compared with ck1 , the toroidal field model has lower overall stress .
this is partly because the vertical field model has a very vigorous initial phase associated with the saturation of the linear instability . at the end of both ck1 and ck2 , @xmath95 .
the mass accretion rate , magnetic field strength , and @xmath10 fluctuate considerably both in time and in space .
the total @xmath10 value rises to 0.12 and then declines to 0.06 by the end of ck2 .
the average ratio of the maxwell to reynolds stress is 6 at late times .
the value of @xmath10 never rises as high as in local vertical field models , including ck1 .
it is comparable , however , to the typical value found for weak field toroidal field models in the local shearing box simulations that began with comparable initial toroidal field strengths ( hgb ) .
these keplerian disk models suffer from significant mass loss through the outer boundary .
in contrast , tori can be completely and self - consistently contained initially within the computational domain .
we next turn to cylindrical models of constant angular momentum ( @xmath51 ) tori , beginning with a torus with @xmath96 and @xmath97 .
the torus outer boundary is at @xmath98 .
two simulations are done , model ct1 with an initial vertical field , and ct2 with an initial toroidal magnetic field ; both use the same computational grid as above .
model ct1 begins with a vertical field with constant @xmath99 from @xmath100 to 3.1 .
this gives @xmath101 at the pressure maximum .
as the evolution proceeds , the magnetorotational instability rapidly develops , again with the characteristic channel .
early on , the beginnings of the papaloizou - pringle principal mode can also be seen as @xmath102 oscillations at the edges of the torus . but
long before this global instability can develop , the torus is dramatically altered by the local mri .
the torus does not endure as a torus : it expands rapidly outward as the angular momentum distribution shifts from constant toward keplerian .
after 4 orbits at the initial torus pressure maximum the system has evolved to a nearly keplerian disk that fills the computational domain .
the magnetic field is amplified to an overall average value of @xmath103 .
toroidal field dominates : @xmath104 , and @xmath105 .
as always , the maxwell stress exceeds the reynolds stress by a factor of several .
the value of @xmath85 peaks at 0.7 and declines to 0.4 at the end of the simulation ( @xmath106 ) .
the overall @xmath10 value varies throughout the disk ; at @xmath107 it varies between 0.3 and 0.4 .
next take the same torus and apply an initial toroidal field with @xmath108 ( model ct2 ) . with this strength field ,
the critical azimuthal wavenumber at the pressure maximum is @xmath109 , so the fastest growing modes are underresolved on this grid . as with the keplerian simulation ck2 ,
the initial toroidal field model becomes turbulent at a later time compared with an initial vertical field model .
the total poloidal magnetic field amplification is also considerably smaller than seen in the vertical field model . despite this ,
the qualitative outcome of the instability for the torus is largely the same .
the slower onset of the mri allows the principal mode of the papaloizou - pringle instability to appear , but soon the transport of angular momentum brings this to a halt .
the disk spreads outward with the bulk of the mass slowly moving inward .
the overall angular momentum distribution changes from constant to keplerian from the inner boundary to @xmath110 , and sub - keplerian but increasing beyond this point . as in the previous simulations ,
the disk exhibits large local fluctuations in density , stress and field strength , and many low azimuthal , high radial wavenumber features . similar time and space variations
were observed in the local models as well ; they are characteristic of the mhd turbulence . at late time
the magnetic energy has risen to @xmath111 , with the toroidal field dominant by a large factor : @xmath112 .
the ratio of the vertical to radial field energy is 0.16 .
the value of @xmath85 rises to @xmath113 , close to , but below , the value in the keplerian toroidal field simulation above .
the total stress corresponds to an @xmath114 - 0.03 at late time . again
the maxwell stress exceeds the reynolds component by a factor of 3 to 4 .
these properties are consistent with toroidal field shearing box simulations ( hgb )
. finally , consider a constant angular momentum torus , ct3 , with an initial field constructed by setting the azimuthal component of the vector potential equal to the density in the torus , @xmath115 .
the resulting field is normalized to an average @xmath116 of 100 . in the cylindrical limit
, the density depends only on radius @xmath45 and the resulting field is vertical with zero net value integrated over the disk .
this model also adopts the pseudo - newtonian potential @xmath117 with @xmath118 .
the computation domain runs from @xmath75 to 13.5 , @xmath119 to @xmath71 and @xmath120 to 2 .
the grid resolution is @xmath121 .
the torus has an inner edge at @xmath122 and a pressure maximum at @xmath123 where @xmath124 ; the orbital period at the outer boundary is 289 .
a new feature in this model is the presence of the marginally stable orbit at @xmath125 . inside of this radius
, matter will rapidly accelerate inward .
the mri grows rapidly in the inner regions of the disk , again with the characteristic radial channel appearance .
accretion through the inner boundary begins at about @xmath126 .
the magnetic energy rises to peak at @xmath127 at @xmath128 .
the magnetic energy grows more slowly after that point ; additional small peaks are observed due to the growth of the mri in the outer parts of the disk .
mass loss through the outer boundary begins at @xmath129 , after which the initial linear growth phase is over and the disk is fully turbulent . between @xmath129 and the end of the simulation
, @xmath116 decreases from 10 to 6 . as always , the toroidal field exceeds the radial field , with @xmath130 .
the radial field is , in turn , greater than the vertical field , @xmath131 .
the overall maxwell stress in the mhd turbulence reaches a peak at @xmath128 , drops off , and then slowly climbs again . at late time @xmath132 , and
the total stress parameter @xmath10 varies with radius inside the disk , from 0.1 near the initial pressure maximum , to 0.2 at the outer regions .
the angular momentum distribution rapidly evolves away from constant , toward keplerian .
by @xmath133 it is nearly keplerian from the pressure maximum inward , and sub - keplerian out beyond this point . inside of @xmath125 the maxwell and
reynolds stresses decline with radius , although not as fast as the pressure .
this makes @xmath10 rise sharply . by the end of the simulation ( @xmath133 ) over one quarter of the disk mass
has been lost , and 70% of it has gone in past the marginally stable orbit .
an examination of the various vertically- and azimuthally - averaged velocities at the end of the run ( fig .
2 ) reveals that the inflow velocity @xmath134 accelerates rapidly inside of @xmath135 . by the time the inner grid boundary is encountered , the inflow speed strongly supersonic and super - alfvnic . inside the disk
the radial speed an order of magnitude smaller than the sound speed , and smaller than the toroidal or poloidal alfvn speeds , which themselves remain subthermal .
going inward from the marginally stable orbit the toroidal alfvn speed rises more slowly that the poloidal alfvn speed .
the radial field strength is roughly constant , but the fluid density drops inside of @xmath136 as the flow accelerates inward .
the specific angular momentum drops 3% from @xmath137 to the inner boundary at @xmath75 indicating that there is still some net stress inside @xmath136 .
the cylindrical disk limit represents a useful way of simplifying the full global problem .
another potentially useful simplification is the axisymmetric limit . in this series of simulations , the torus evolution problem is considered in the axisymmetric @xmath38 plane .
these simulations use a pseudo - newtonian potential , and begin with a pressure - supported torus that is fully contained on the grid .
the initial magnetic field is chosen to satisfy two requirements : it must have a poloidal component to allow for the development of the mri , and , as a practical matter , it should be contained completely within the torus to avoid the courant limitations caused by high alfvn speeds due to strong fields in low density regions . a suitable initial configuration consists of magnetic field loops lying along equidensity surfaces in the torus .
this initial setup will develop strong toroidal field due to shearing of the initial radial field .
however , experience has shown that strong toroidal fields are the outcome of all initial field choices , so this should not represent too great an idealization .
all of the two - dimensional simulations considered here will also be run in three dimensions .
the first simulation has a radial grid running from @xmath75 to an outer boundary at @xmath138 and a vertical grid running from @xmath139 to 5 .
the grid resolution is @xmath140 .
a constant-@xmath41 torus is placed on the grid with an inner boundary at the marginally stable orbit , @xmath141 and a pressure maximum at @xmath142 .
the orbital period at the pressure maximum is 50 .
the initial field is obtained from an azimuthal vector potential @xmath143 , and is normalized to @xmath108 using the total integrated magnetic and thermal energies .
density grayscale plots from this simulation are presented in figure 3 .
the initial period of evolution is dominated by the growth of toroidal magnetic field due to shear .
as the magnetic pressure increases , the torus expands , particularly at the inner edge . at first
, low density material is driven outward perpendicular to the torus surface ; subsequently , it flows radially out and around the torus . because of the initial reflection symmetry across the equator , the toroidal field changes sign there and a strong current sheet forms
this current sheet proves to be unstable , and oscillates around the equator .
this is an important part of the evolution , and indicates that it is necessary to simulate the full @xmath38 plane rather than adopt the equator as an explicit boundary . as the toroidal field pressure increases
, it drives inflow through the marginally stable orbit ; this initial accretion flow peaks at @xmath144 then begins to decline . in the meantime
, the poloidal field mri grows within the torus , and begins to manifest itself visibly in the typical form of radial channels by @xmath145 ( 3.6 orbits ) .
there follows a period of violent readjustment within the disk , featuring strong mass inflow punctuated by episodic accretion events .
this phase lasts until @xmath146 ( 7 orbits ) , beyond which the poloidal magnetic energy declines , and with it the level of turbulence in the disk , and the accretion rate . at the end of the simulation ( @xmath147 , @xmath148 orbits ) about
60% of the initial disk mass has been lost .
most of this mass is lost by @xmath149 ; after this time the inflow accretion rate is very small .
thus there are three distinct phases to the two - dimensional torus evolution : expansion due to the shear - amplified toroidal magnetic pressure , strong nonlinear evolution of the poloidal field mri , and finally a more quiescent turbulent state with declining poloidal magnetic field strengths .
angular momentum transport occurs in all three phases at different rates . in the first phase there is a growing maxwell stress from the amplified @xmath150 field , mainly in the inner region of the disk where the orbital frequency is highest .
the reynolds stress is negligible .
the initially constant angular momentum distribution is unchanged within the disk except in this inner region . with the onset of the mri ,
angular momentum transport occurs everywhere , and the specific angular momentum begins to increase with radius . during the middle of this second phase , angular momentum transport peaks , with @xmath10 ranging between 0.2 and 0.5 through the disk .
as the turbulence subsides , the rate of transport declines with @xmath10 varying between .02 and .1 within the main part of the disk . since this is an axisymmetric simulation , the strength of the poloidal magnetic field
is limited topologically by the antidynamo theorem , and the mhd turbulence must eventually die out .
the next model is a torus located initially at a larger radius , which increases the amount of time that it can evolve prior to reaching the marginally stable orbit .
the grid runs from @xmath75 to @xmath151 and from -4.5 to 4.5 in @xmath35 .
the grid resolution is @xmath152 .
the constant-@xmath41 torus has an inner edge at @xmath153 and a pressure maximum at @xmath154 .
the orbital period at the pressure maximum is 88 .
again , magnetic field loops are placed along the equidensity surfaces within the torus , with a total magnetic energy corresponding to @xmath99 . as before , this torus undergoes three phases of evolution : toroidal field amplification , development of the nonlinear poloidal mri , and subsequent turbulence .
the mri saturates around @xmath155 ( 3.4 orbits ) with total magnetic energy @xmath156 .
the phase of strong mri turbulence is over by @xmath157 ( 5.1 orbits ) when the average @xmath116 rises to 6 ; fairly steady accretion follows for the remainder of the simulation which runs to time @xmath158 ( 9.5 orbits ) . in the third simulation
, the initial torus has an angular momentum distribution closer to keplerian , specifically @xmath159 .
the torus inner boundary is at 4 and its pressure maximum at @xmath160 ( orbital period @xmath161 ) .
the computational domain runs from @xmath162 to 21.5 , and from @xmath163 to 10 in @xmath35 ; the grid resolution is @xmath140 .
although it began with a much different initial angular momentum distribution , the evolution is quite similar to the two previous cases . in the early stage , toroidal field
is amplified by shear .
the poloidal mri soon comes into play , producing the characteristic radial streams .
the total magnetic energy peaks at about @xmath164 ( 4.5 orbits ) .
after this the poloidal field energy declines steadily , as does the maxwell stress and the mass accretion through the inner radial boundary .
toroidal field dominates , with the ratio @xmath165 . beyond @xmath166 ,
the toroidal field energy is essentially constant with a @xmath167 .
angular momentum transport begins almost immediately .
initially it is confined to the inner regions of the torus where shear is strongest , and dominated by the maxwell stress associated with the growing toroidal field
. however , the growth of the mri produces stress throughout the torus .
by @xmath168 the averaged @xmath10 value is between 0.1 and 0.2 . at the end of the simulation at @xmath169 , @xmath10 ranges between 0.01 and 0.04 within the torus .
the ratio of the maxwell stress to the magnetic pressure , @xmath85 , begins small , rises to a value of 0.1 during the initial saturation of the mri , then steadily declines to 0.03 by the end of the simulation as poloidal field is preferentially destroyed .
since this torus began with an angular momentum distribution similar to the end state of the initially @xmath51 tori , there is only a slight change in the slope of the average angular momentum .
once the torus expands , this slope is extended in range all the way from @xmath170 to the outer radial boundary .
now we turn to the evolution of fully three dimensional tori using a pseudo - newtonian potential .
the simulations described in this section are three - dimensional versions of the tori considered above in the axisymmetric limit .
the first , model gt1 , is the constant angular momentum torus with @xmath171 and @xmath172 .
the orbital period at the pressure maximum is @xmath173 .
the computational domain runs from 1.5 to 11.5 in radius , from @xmath174 to @xmath175 in @xmath35 , and from @xmath176 to @xmath71 .
the initial magnetic field is constructed by setting the toroidal component of the vector potential equal to the density inside the disk , @xmath177 , for all @xmath15 greater than a minimum value .
the total magnetic energy is then normalized to a value of @xmath108 , using the total integrated gas pressure of the torus .
different regions within the torus will , of course , have larger or smaller values of @xmath116 .
the strongest initial fields are found between the pressure maximum and the inner edge of the torus .
the simulation is run to time @xmath178 ( 15.6 orbits ) .
the evolution of this torus is illustrated with a series of grayscale plots in @xmath179 ( fig .
4 ) of vertical @xmath38 and equatorial @xmath180 slices . as with the axisymmetric torus ( fig .
3 ) , the initial phase of evolution is controlled by the shear amplification of the toroidal field . however , in three - dimensions this toroidal field is itself unstable to the mri .
the characteristic large - wavenumber structures associated with the instability quickly appear in the inner regions of the disk ( by @xmath181 ) .
this creates turbulence , and momentarily increases the mass accretion rate over that seen in the axisymmetric simulation ( fig .
5 ) . by @xmath129 the disk
contains tightly wrapped , low @xmath1 trailing spiral waves .
the presence of this nonaxisymmetric structure affects the development of the poloidal field instability , eliminating the organized radial streaming flows seen in axisymmetry .
the flow is more turbulent and the mass accretion rate is steadier , without the large impulsive spikes in @xmath182 associated with the radial streaming in two dimensions .
this overall turbulence declines after saturation of the mri , but continues through the end of the run .
this is the stage when the accretion rate in the two - dimensional simulation drops to very small levels . while in three dimensions one might hope that the disk could achieve a nearly steady state mass accretion rate ,
there is , of course , only a finite amount of disk mass available to accrete . at the end of the run 85% of the initial total mass has been lost from the grid , most of it through the innermost stable orbit .
the amount of mass ejected off the outer boundary is about 30% of that accreted through the inner boundary .
although the total volume - averaged magnetic energy drops with time beyond @xmath183 , most of this decline is due to loss from accretion .
average @xmath116 after @xmath129 is @xmath184 and remains fairly constant thereafter .
the magnetic energy is dominated by the toroidal component ; the poloidal field @xmath116 value has an average value of @xmath185 , and also remains nearly constant after @xmath129 . in the axisymmetric simulation
the poloidal @xmath116 value attains a value of @xmath186 at @xmath187 , but rises steadily to @xmath188 by @xmath164 , indicating a poloidal magnetic energy loss rate exceeding that due to accretion alone .
angular momentum transport results in rapid restructuring of the disk .
the angular momentum distribution in the inner half of the torus steepens to a nearly keplerian value by @xmath144 ( fig .
the angular momentum in the outer part of the torus increases over the remainder of the simulation . at the end time
the angular velocity @xmath189 has a slope near @xmath190 .
this is essentially the same as in the axisymmetric run above ( represented by the short dashed line in fig .
this slope adequately characterizes the disk since the angular momentum , and hence angular velocity , is nearly constant on cylinders .
after turbulence has developed , the average @xmath10 values in the main part of the disk are between 0.1 and 0.2 .
the ratio of the maxwell to reynolds stress varies , but lies between 1 and 4 throughout the main part of the disk . the time- and space - averaged value of @xmath85 after @xmath129 is 0.40 .
figure 6 shows that the specific angular momentum continues to decline even inside the marginally stable orbit .
this indicates that even here there remains a significant net stress .
in fact , inside the marginally stable orbit @xmath10 rapidly increases because the gas pressure drops while the maxwell stress remains roughly constant .
the presence of this stress means there is no maximum in the epicyclic frequency , @xmath191 . in the pseudo - newtonian potential ( and , of course in the relativistic potential
that the pseudo - newtonian potential was designed to imitate ) , the keplerian value of @xmath192 has a maximum that occurs at about @xmath193 .
more generally , the epicyclic frequency will go to zero at a stress - free inner edge of a disk where @xmath194 , ensuring a @xmath195 maximum somewhere in the disk . here , however , the inner boundary of the disk is not characterized by a zero - stress condition , the slope of @xmath196 is nearly constant , and the epicyclic frequency does not turn over but continues to rise .
it has been proposed that potentially observable oscillatory modes might be trapped near the disk inner edge where the epicyclic frequency turns over ( e.g. , nowak & wagoner 1993 ) .
the present result must be regarded as preliminary , but if significant stress inside the marginally stable orbit proves to be a generic property of magnetic turbulence , trapped disk oscillations may not be present . in some respects ,
the two- and three - dimensional simulations are similar .
they both have shear amplification of toroidal field , they both evolve due to the resulting increase in toroidal magnetic pressure , and they both develop turbulence due to the poloidal field mri . both rapidly evolve from constant to nearly keplerian specific angular momentum distributions ( fig .
the three dimensional simulation , however , permits the development of the nonaxisymmetric mri which increases and sustains turbulence and mass accretion . in two dimensions the organized poloidal field channel solution produces an impulsive accretion rate greater than that seen in three dimensions during the initial saturation of the poloidal mri .
but axisymmetry causes the two - dimensional poloidal field to decline , and with it the maxwell stress .
the contrast between @xmath85 in the three- and the two - dimensional simulations is instructive ( fig .
7 ) . in three dimensions
@xmath85 remains relatively constant and near the value typically seen in the local shearing box simulations . in two dimensions
@xmath85 declines steadily with time following the saturation of the mri .
maintenance of the poloidal field through dynamo action is possible only in three dimensions .
although two - dimensional simulations can not capture these essential features of global evolution , they do have one clear advantage : they are considerably easier to compute .
two dimensional simulations are useful for searching a wide range of initial conditions in support of the more challenging three - dimensional models .
to test the idea of using an evolved two - dimensional simulation as an initial condition , consider next a constant-@xmath41 torus with an initial inner edge at @xmath197 and a pressure maximum at @xmath198 where @xmath124 .
this same initial torus was run both in the cylindrical limit , and in the axisymmetric @xmath38 limit above . here
the computational domain runs from @xmath75 to 13.5 , @xmath119 to @xmath71 and @xmath199 to 4.5 .
the grid resolution is @xmath200 .
the three dimensional torus gt2 is initialized by taking the output from the axisymmetric simulation at time @xmath201 and expanding it out in the @xmath36 direction .
this time corresponds to shortly before the nonlinear aspects of the mri first manifest themselves noticeably in the density plots .
random pressure perturbations are added to break the axisymmetry .
while this initialization procedure reduces the total computational time required , it does nt allow the toroidal field instability the opportunity to grow during the initial phase .
this means that the strong coherent poloidal field instability develops as it does in two dimensions .
nevertheless , significant nonaxisymmetric structure appears by @xmath183 , and the overall turbulence is increased over that seen in axisymmetry .
accretion inflow at the inner boundary is about twice what it is in two dimensions after this point in time .
after @xmath183 the average magnetic field strength is @xmath202 , and the poloidal field strength @xmath203 . at late time
@xmath204 , and @xmath205 , a level similar to that seen in gt1 . in two dimensions
the maxwell stress drops and the poloidal @xmath116 increase .
a the end of the simulation , @xmath206 and @xmath207 .
the toroidal field @xmath116 is fairly constant with time in both two- and three - dimensions runs after @xmath157 .
this run can also be compared with the cylindrical run ct3 .
figure 8 is a plot of the radial dependence of the azimuthally and vertically averaged speeds .
these curves are very similar to those in figure 2 from ct3 .
gt2 shows a stronger initial field amplification phase and a stronger mri saturation ( created by initializing from the axisymmetric run ) .
however the magnetic field amplitudes and stress levels are comparable at late times in both simulations .
these similarities indicate that the cylindrical limit provides a useful approximation for investigating the nonlinear evolution of mhd turbulence near the midplane of a disk . to summarize
, gt2 shows that fully three dimensional turbulence can be rapidly produced and sustained in a simulation initialized from the output of an axisymmetric simulation .
the late - time properties of such a simulation , essentially one with complicated nonlinear initial perturbations , are quite similar to those obtained from a simulation evolved from a formal equilibrium and linear perturbations .
the two - dimensional simulations , therefore , serve as just another type of initial condition , albeit more complicated than usually adopted .
one characteristic of all these simulations is that at late times the magnetic field is predominantly toroidal .
the next simulation considers a torus that begins with only a toroidal field .
model gt3 is of a @xmath51 torus with an inner boundary at 4.5 and a pressure maximum at @xmath208 ( orbital period @xmath209 ) .
the initial toroidal field has @xmath81 everywhere within the torus .
the mri that will result is nonaxisymmetric ; an axisymmetric @xmath38 version of this model does not develop turbulence or transport angular momentum .
the computational domain runs from @xmath162 to 13.5 , @xmath35 from @xmath210 to 6 , and over the full @xmath37 in angle ; the grid resolution is @xmath211 .
no attempt is made to keep the initial torus in pressure balance ; the initial magnetic field simply adds to the equilibrium hydrodynamic pressure . as a consequence
, the torus undergoes an axisymmetric readjustment due to this additional magnetic pressure .
the toroidal field mri develops rapidly at the inner edge of the disk where @xmath90 is largest , before spreading throughout the disk .
the poloidal field energy grows steadily with time until about @xmath212 when it reaches a value of @xmath213 .
after @xmath183 , @xmath214 increases with time to 65 at the end of the run .
grayscale images of gt3 are given in figure 9 .
models gt2 and gt3 began with the same hydrodynamic torus ; they differ in their initial magnetic fields , in the size of the @xmath36 domain , and in resolution . despite these differences ,
the late time evolution in these two runs is very similar .
they have comparable mass accretion rates through @xmath170 ( fig . 5 ) .
both have @xmath215 at @xmath216 .
figure 10 illustrates the redistribution of mass and angular momentum as a function of time .
the slope of the angular momentum distribution at late time is the same in gt3 as in both gt1 and gt2 .
figure 11 shows the radial run of density @xmath217 , velocity @xmath218 , and mass flux @xmath219 at time @xmath216 .
density is normalized by the maximum @xmath217 at @xmath220 . within the torus
the radial drift velocity is small compared to the orbital velocity ; @xmath221 reverses outside of @xmath222 beyond which there is net outflow .
inside @xmath136 the density decreases rapidly as the inflow accelerates inward .
the accretion rate increases slightly from @xmath223 inward .
figure 12 shows gas and magnetic pressures at this same end time ( normalized by the initial pressure maximum value ) , as well as the maxwell stress .
the magnetic energy remains subthermal throughout , although it drops less rapidly than the gas pressure inside of @xmath136 .
the ratio of the maxwell stress to the gas pressure has a minimum of 0.03 between @xmath135 and 5 , rising sharply inside @xmath136 and more slowly toward the outer boundary .
the final global simulation , gt4 , is the @xmath159 torus with an inner boundary at 4 and a pressure maximum at @xmath160 .
this angular momentum distribution yields a torus that extends over a large radial distance without becoming too thick in the vertical direction .
the computational domain runs from @xmath162 to 21.5 , from @xmath163 to 10 in @xmath35 , and over the full @xmath37 in angle ; the grid resolution is @xmath224 .
the simulation is run to time @xmath225 .
this is longer than the other simulations , although it amounts to fewer orbits at the pressure maximum : @xmath225 is 7.2 orbits at @xmath52 .
this evolution time is sufficient to observe all of the stages seen in the other three - dimensional simulations , although not long enough for the torus to have settled into a quasi - steady state .
density plots from gt4 are presented in figure 13 . at the beginning
, the toroidal field grows by shearing out the radial field , but as it does so , the toroidal mri sets in .
this soon leads to turbulence .
the poloidal mri develops rapidly in the inner regions of the disk and subsequently spreads throughout .
the inner edge of the disk moves slowly inward until it reaches the marginally stable orbit at @xmath226 after which gas plunges inward .
initially the inflowing fluid is confined largely to the equatorial plane .
as time passes , however , this region fills with gas and becomes thicker . as with the previous global simulations ,
the mass accretion rate is larger in three dimensions compared to two . near
the end of the simulation , @xmath87 at the inner radial boundary is approximately 2.5 times that seen in axisymmetry .
the value of @xmath85 is rising with time indicating that the poloidal field strength is still increasing at the end of the simulation .
figure 14 shows the radial mass and angular momentum distributions at the initial and final times in gt4 .
also shown are curves from two times in the equivalent axisymmetric calculation .
the average angular velocity parameter @xmath50 decreases by a very small amount over the course of the evolution .
the torus mass , on the other hand , has been substantially redistributed .
in this paper , we have carried out global mhd simulations of accretion tori .
we have also made use of several different limits or approximations : two dimensional axisymmetry , and three dimensional cylindrical gravitational potential , in addition to the full three dimensional global model .
in addition to the intrinsic interest in the results , these simulations begin to map out what is currently possible with the present hardware and algorithms .
first , how useful are the two - dimensional and cylindrical approximations ?
while neither should be relied on exclusively , they both have appropriate applications , and they greatly reduce the computational difficulty of the simulations .
two - dimensional axisymmetric torus simulations demonstrate effects from the generation of toroidal field due to shear , and from the development of the poloidal field mri .
the latter leads to turbulence , rapid angular momentum transport and evolution toward a nearly keplerian angular momentum distribution .
a significant limitation of the axisymmetric system , however , is embodied in cowling s antidynamo theorem .
the @xmath227 component of the vector potential is conserved , except for losses due to dissipative processes .
poloidal field can grow from axisymmetric stretching and folding , but this is ultimately limited . in the simulations , after the nonlinear saturation of the mri , the poloidal magnetic field energies decline and the turbulence begins to die out . despite this
, considerable similarity is seen between the two- and three - dimensional simulations .
this is due to the dominance in the initial stages of the torus evolution by what is largely axisymmetric dynamics : increase in toroidal field pressure due to shear , the development of the `` channel solution , '' and the rapid redistribution of angular momentum by large stresses . during phases of a disk s evolution when such effects are most important , two - dimensional simulations are a good approximation
. their utility must be limited , however , since genuine nonaxisymmetric effects , including the development of the toroidal field mri , dynamo amplification and maintenance of poloidal fields , and nonaxisymmetric spiral waves , are generally of dominant importance over the long term .
cylindrical disks are a natural extension of the shearing box model .
the cylindrical disk , like the shearing box simulations of hgb and hgb2 , does not include the effects of vertical gravity .
cylindrical keplerian disk simulations , initialized with vertical fields or toroidal fields show rapid development of the mri consistent with the shearing box results . field amplification and stress levels are comparable between the two types of simulation as well . with the cylindrical disk , however , we are able to observe the direct consequences of the stress : redistribution of angular momentum toward keplerian , and mass accretion .
cylindrical disk simulations can provide detailed information about the radial dependence of the growth and saturation of the mri , the resulting mhd turbulence , the transport of angular momentum , and the net accretion flow .
simulations of cylindrical constant angular momentum tori with vertical or toroidal initial fields illustrate many aspects of a full global evolution .
regardless of the initial field configuration , the initially constant angular momentum tori evolve toward radial angular momentum distributions that are nearly , but not quite , keplerian .
the final outcome is largely the same for either initial vertical or toroidal fields , although the early growth of the mri is more rapid for vertical fields , and dominated by small scale structure for toroidal fields .
this is entirely consistent with the local linear analysis .
of course , we want to simulate disks fully globally and with as few approximations or geometric constraints as possible .
indeed , full three dimensionality is essential .
cylindrical simulations can provide little , if any , information about the vertical structure of a disk , energy transport , or the possible formation of winds or jets . as with the cylindrical disk simulations , the overall evolution of fully global tori are consistent with the intuitions developed from shearing box simulations .
the mri grows rapidly , and produces mhd turbulence that transports angular momentum . in all cases , toroidal field dominates , followed by radial and then vertical field .
one difference is that the shearing box calculations with zero net field ( e.g. , hgb2 ) typically have a total stress value of @xmath88 , whereas here @xmath6 in the heart of the turbulent disks .
this is more a matter of the magnetic pressure that is sustained in the torus versus the shearing box , rather than some qualitative difference in the behavior of the mri . in both the global and local simulations , the stress is directly proportional to @xmath228 : @xmath229 .
all of the present global simulations began with relatively strong field , either in the form of poloidal field loops which immediately generated strong toroidal field through shear , or from the presence of an initially strong toroidal field .
this field strength is by and large sustained , and thus the observed @xmath10 values are consistent with @xmath230 .
it is suggestive that matsumoto ( 1999 ) also obtained @xmath231 for his toroidal field simulations , even one that began with @xmath232 .
global ( rather than local ) length scales make it natural to have larger magnetic field strengths in the saturated state .
computational problems associated with the inner boundary are greatly reduced through the use of the pseudo - newtonian potential .
this provides a physical inner radius for the disks , and ensures that gas and field will flow smoothly off the inner radial grid .
the accretion flow rapidly accelerates inward near @xmath233 , with the radial speed quickly exceeding the sound speed .
interestingly , the angular momentum continues to drop inside of @xmath233 , indicating the continuing presence of maxwell stress .
the absence of a stress - free inner disk boundary is one of the ways in which these disks differ from standard analytic models .
one immediate conclusion from these simulations is that the constant , or nearly - constant angular momentum torus is a remarkably unstable structure in the presence of a weak magnetic field .
the mhd turbulence resulting from the mri simply transports angular momentum too efficiently . within a few orbits the angular momentum distribution changes from @xmath51 to @xmath234 .
the tori considered here did not remain maintain their initial constant angular momentum distribution long enough to develop the coherent structures associated with the nonlinear outcome of the papaloizou - pringle instability .
this does not , however , necessarily imply that accretion disks must be thin .
moderately thick configurations , as seen in the grayscale plots , are still possible even for nearly keplerian angular momentum distributions , if there is significant internal pressure .
averaged properties of the torus such as density , pressure or angular momentum distribution fail to convey the impression of disorder seen in animations of the evolution .
low density , high magnetic field filaments entwine themselves throughout the torus
. regions of strong field develop and rise through the torus into the surrounding low - density atmosphere .
low density material flows outward around the bulk of the torus .
since angular momentum transport is by mhd turbulence , it follows that the disk should be highly dynamic .
the effect of this on observed properties of disks must be considered in subsequent , more sophisticated simulations .
it is not premature , however , to question the relevance of the traditional image of a quiescent steady - state disk .
there remain several limitations to the present simulations to be addressed in subsequent work .
first , greater grid resolution is always welcome , particularly where the accretion inflow is squeezed into a narrow equatorial flow , or for following the growth of the mri from substantially weaker initial field strengths .
because the disks develop oscillations across the equatorial plane , particularly in the simulations that begin with poloidal field loops , the full @xmath38 plane must be simulated in global models , without applying explicit ( e.g. , reflecting ) equatorial boundary conditions .
this doubles the number of grid zones required , but appears to be necessary .
higher resolution simulations are feasible with the addition of more processors to the parallel computation .
the simulations described here used a maximum of 64 .
indeed , a great many issues can be profitably investigated using simulations with more grid zones .
for example , the global simulations presented here featured tori that began close to the marginally stable orbit .
little evolution is required to accrete through @xmath233 and into the central hole .
simulating a greater dynamic range in disk radii , for longer times and with larger numbers of radial grid zones , is an immediate next goal .
the present simulations focus only on the dynamical properties of mhd in tori .
as the thermodynamic properties of the disk are of obvious importance in establishing an overall global disk structure , further algorithm development is desirable . in the present simulations
the simplified equation of state and lack of explicit resistivity mean that the only sources of gas heating are artificial shock viscosity and adiabatic compression .
some amount of energy is necessarily lost through the numerical reconnection of magnetic field .
further , neither radiative transport nor simple radiative losses were included .
these and other issues of global disk evolution must be deferred to subsequent work . this work is supported in part by nasa grants nag5 - 3058 and nag5 - 7500 .
three dimensional simulations were run on the cray t90 and t3e systems operated by the san diego supercomputer center with resources provided through a national resource allocation grant from the national science foundation .
the three - dimensional torus simulation gt4 was run by greg lindahl on a cluster of 64 alpha processors , part of the centurion system developed by the legion project at the university of virginia .
a portion of this work was completed while attending the program on black hole astrophysics at the institute for theoretical physics , supported in part by the nsf under grant no .
phy94 - 07194 .
i thank steve balbus , charles gammie , james stone , and wayne winters , and the members of the itp black hole astrophysics program for helpful comments and discussions .
armitage , p. 1998 , apj , 501 , l189 balbus , s. a. , & hawley , j. f. 1991 , apj , 376 , 214 balbus , s. a. , & hawley , j. f. 1992 , apj , 400 , 610 balbus , s. a. , & hawley , j. f. 1998 , rev mod phys , 70 , 1 blaes , o. m. , & hawley , j. f. 1988 , apj , 326 , 277 brandenburg , a. , nordlund , , stein , r. f. , & torkelsson , u. 1995 , apj , 446 , 741 curry , c. , & pudritz , r. 1996 , mnras , 281 , 119 evans , c. r. 1986 , in dynamical spacetimes and numerical relativity , ed . j. m. centrella ( new york : cambridge univ . press ) , 3 evans , c. r. , & hawley , j. f. 1988 , apj , 332 , 659 goldreich , p. , goodman , j. , & narayan , r. 1986 , mnras , 221 , 339 goodman , j. , narayan , r. , & goldreich , p. 1987
, mnras , 225 , 695 goodman , j. , & xu , g. 1994 , apj , 432 , 213 hawley , j. f. 1987 , mnras , 225 , 677 hawley , j. f. 1991 , apj , 381 , 496 hawley , j. f. & balbus , s. a. 1991 , apj , 376 , 223 hawley , j. f. & balbus , s. a. 1992 , apj , 400 , 595 hawley , j. f. & balbus , s. a. 1999 , in astrophysical discs , eds .
sellwood , j. a. & goodman , j. ( cambridge univ .
press : cambridge ) , 108 hawley , j. f. , gammie , c. f. , balbus , s. a. 1995 , apj , 440 , 742 ( hgb ) hawley , j. f. , gammie , c. f. , balbus , s. a. 1996 , apj , 464 , 690 ( hgb2 ) hawley , j. f. , & stone , j. m. 1995 , comp phys comm , 89 , 127 matsumoto , r. 1999 , in numerical astrophysics , eds .
miyama , s. , tomisaka , k. & hanawa , t ( kluwer : dordrecht ) , 195 matsumoto , r. , & shibata , k. 1997 , in accretion phenomena and related outflows , eds .
d. wickramsinghe , l. ferrario , & g. bicknell ( san francisco : pasp ) matsumoto , r. , & tajima , t. 1995 , apj , 445 , 767 norman , m. l. , wilson , j. r. , & barton , r. t. 1980 , apj , 239 , 968 nowak , m. a. , & wagoner , r. v. 1993 , apj , 418 , 187 ogilvie , c. , & pringle , j. e. 1996 , mnras 279 , 152 paczysky , b. , & wiita , p. j. 1980 , a&a , 88 , 23 papaloizou , j. c. b. & pringle , j. e. 1984 , mnras , 208 , 721 shakura , n. i. , & sunyaev , r. a. 1973 , a&a , 24 , 337 stone , j. m. , hawley , j. f. , gammie , c. f. , & balbus , s. a. 1996 , apj , 463 , 656 stone , j. m. , & norman , m. l. 1992a , apjs , 80 , 753 stone , j. m. , & norman , m. l. 1992b , apjs , 80 , 791 ck1 & @xmath72 & @xmath56 & 1.5 & @xmath235 & 188 + ck2&@xmath72 & @xmath56 & 1.5&@xmath236 & 285 + ct1&@xmath72 & @xmath56 & 2&@xmath237 & 102 + ct2&@xmath72 & @xmath56 & 2&@xmath236 & 403 + ct3&@xmath121 & @xmath238 & 2 & @xmath239 & 420 + gt1&@xmath240 & @xmath241 & 2 & @xmath242 & 780 + gt2&@xmath200&@xmath241 & 2 & @xmath242 & 727 + gt3&@xmath224&@xmath241&2&@xmath243 & 727 + gt4&@xmath224&@xmath241&1.68&@xmath244 & 1283 + | global time - dependent simulations provide a means to investigate time - dependent dynamic evolution in accretion disks .
this paper seeks to extend previous local simulations by beginning a systematic effort to develop fully global three - dimensional simulations .
the nonlinear development of the magnetorotational instability is investigated using a time - explicit finite difference code written in cylindrical coordinates .
the equations of ideal magnetohydrodynamics are solved with the assumption of an adiabatic equation of state .
both a newtonian , and a pseudo - newtonian potential are used .
two simplifications are also explored : a cylindrical gravitational potential ( the `` cylindrical disk '' ) , and axisymmetry . the results from those simulations are compared with fully three dimensional global simulations .
the global simulations begin with equilibrium pressure supported accretion tori .
two different initial field geometries are investigated : poloidal fields that are constant along initial equidensity surfaces , and toroidal fields with a constant ratio of gas to magnetic pressure . in both cases the magnetorotational instability rapidly develops , and the torus becomes turbulent .
the resulting turbulence transports angular momentum , and the torus develops an angular momentum distribution that is near keplerian .
a comparison with axisymmetric simulations shows that in three dimensions the magnetorotational instability can act as a dynamo and regenerate poloidal field thereby sustaining the turbulence .
as previously observed in local simulations , the stress is dominated by the maxwell component .
the total stress in the interior of the disk is @xmath0 - 0.2 the thermal pressure . at late time
the disks are characterized by relatively thick configurations , with rapid time - dependence , and tightly - wrapped , low-@xmath1 spiral structures . |
the statistical properties of primordial fluctuations provide crucial information on the physics of the very early universe @xcite ( see @xcite for a review ) . in the simplest single field inflation models where the scalar field has a canonical kinetic term and quantum fluctuations are generated from the standard bunch - davis vacuum , non - gaussianity of the fluctuations is too small to be observed even with future experiments @xcite .
thus the detection of non - negligible departures from gaussinaity of primordial fluctuations will have a huge impact on the models of early universe .
so far , most of the studies have focused on the leading order non - gaussianity measured by the three - point function of cosmic microwave background ( cmb ) anisotropies , i.e. the bispectrum @xcite . especially , the optical method of extracting the bispectrum from the cmb data has been sufficiently developed @xcite ( for a more general approach , see @xcite ) .
however , future experiments like planck @xcite can also prove the higher order statistics such as the trispectrum @xcite .
the trispectrum gives information that can not be obtained from the bispectrum @xcite .
in addition , it is possible that the trispectrum can be the leading order non - gaussianity , that is , even if we do not detect the bispectrum , this does not mean that the primordial fluctuations are confirmed to be gaussian . in the local type non - gaussian models @xcite ,
the primordial curvature perturbation @xmath6 is modeled as @xmath7 where @xmath8 is a gaussian variable @xcite . in this model
, the bispectrum has maximum amplitude for the squeezed configurations in the fourie space where one of wavenumbers is small compared with others . in these models
, the trispectrum indeed gives very interesting tests on multi - field inflation models where there are several gaussian variables @xmath9 using the consistency relation between the amplitudes of the bispectrum and trispectrum @xcite .
also the trispectrum can constrain the cubic - order non - linearities in primordial curvature perturbation , @xmath10 that can not be constrained by the bispectrum measurements . estimators to measure the trispectrum in the local - type non - gaussianity have been developed and the kurtosis based estimator @xcite have been used to obtain constraints on the amplitude of the trispectrum , @xmath11 at @xmath12 confidence level from the wmap 5-year data @xcite .
there are another class of non - gaussian models .
a typical example is dirac - born - infled ( dbi ) inflation @xcite whose non - gaussian property was extensively studied by @xcite ( see also @xcite for reviews ) . in this model , like k - inflation @xcite , the inflaton field has non - canonical kinetic term and non - linear derivative interactions can give rise to large non - gaussianity of quantum fluctuations . for current observational constraints on dbi inflation
see @xcite . in these models ,
the amplitude of the bispectrum has a peak typically for the equilateral configuration in the fourie space .
the shape of the trispectrum is more complicated .
for example , for the bispectrum , the equilateral condition @xmath13 completely specifies the shape of the bisepctrum , but this is not the case for the trispectrum .
the shape of the trispectrum has been analyzed in several inflationary models such as single field dbi inflation @xcite , multi - field dbi inflation @xcite and the models motivated by effective theory of inflation @xcite .
regardless of these efforts , since the form of the trispectrum is generally very complicated , estimators for the trispectrum in this class of non - gaussian models have not been implemented yet so far .
it was suggested that the form of trispectrum given by @xmath14 represents the shape of the trispectrum in equilateral non - gaussian models very well . here
the trispectrum of the curvature perturbation is defined as @xmath15 where @xmath16 is the power spectrum given by @xmath17 .
this trispectrum ( [ fiducial ] ) appears in dbi inflation as a contribution from the fourth order interacting hamiltonian ( the contact interaction " ) @xcite . in the effective theory of inflation ,
it was shown that the trispectrum of this shape can have the amplitude that is not constrained by the bispectrum measurements @xcite . in ref .
@xcite , it was suggested that this trispectrum can be used to construct an estimator because by introducing the integral @xmath18 , this function is factorisable ( see appendix a ) . therefore , in this paper , we compare the shapes of trispectra in single field and multi - field dbi inflation with eq .
( [ fiducial ] ) based on a shape correlator introduced by regan et.al @xcite and investigate whether the estimator constructed from the trispectrum ( [ fiducial ] ) represents the shapes of trispetrum in these models or not .
this paper is organized as follows . in section
ii , we review the shape correlator introduced by regan et al . @xcite . in section iii
, we study the overlap between the shape given by eq .
( [ fiducial ] ) and trispectra in single field and multi - field dbi inflation models . in section iv , we give theoretical predictions for @xmath2 in some concrete theoretical models .
we conclude in section v. in appendix a , we present the optimal estimator using eq .
( [ fiducial ] ) explicitly . in appendix b , we summarise the shape function of the reduced trispectra appeared in general single field k - inflation models and give the shape correlations among the representative shapes . in appendix c
, we check the validity of our method to relate the amplitude of the estimator to the theoretical predictions using the bispectrum .
in this section , we review the shape correlator introduced by regan et al . @xcite .
first we exploit the symmetry of the trispectrum to define the _ reduced _ trispectrum as follows @xcite .
we rewrite the definition of the trispectrum as @xmath19\,.\end{aligned}\ ] ] then we need to consider only the reduced trispectrum @xmath20 from one particular arrangement of the vectors , such as @xmath21 with @xmath22 and form the other contributions by considering permutations . here , for the later convenience
, we use the symmetrised reduced trispectrum @xmath23\ , , \label{sym_reducedtrispectrum}\end{aligned}\ ] ] and from now on we omit the superscript @xmath24 for simplicity .
the reduced trispectrum is a function of six variables .
we can choose them to be @xmath25 where @xmath26 represents the deviation of the quadrilateral from planarity which is specified by the triangle @xmath27 .
we find that in terms of these variables @xmath28 is expressed as @xmath29 which implies that the valid range of @xmath30 is constrained by @xmath31 motivated by the relation between the cmb trispectrum and the trispectrum for @xmath6 , the shape function for the reduced trispectrum is defined as @xmath32 regan et al .
@xcite proposed to define an overlap between two different shape functions @xmath33 and @xmath34 as @xmath35 where @xmath36 is an appropriate weight function .
the weight function should be chosen such that @xmath37 in @xmath38 space produces the same scaling as the estimator in @xmath39 space and we adopt the one used in ref .
@xcite , @xmath40 with this choice of weight , the shape correlator is defined as @xmath41 first to parameterise the magnitude of the momenta , we use the semiperimeter of the triangle formed by the vectors @xmath42 , @xmath43 from the scaling behaviour , the form of the shape function on a constant-@xmath44 cross section becomes independent of @xmath44 and we can write @xmath45 where @xmath46 and @xmath47 . since we are restricted to the region where the momenta @xmath27 and @xmath48 form triangles by momentum conservation , we will reparameterise the allowed region to separate out the overall scale @xmath44 from the behaviour on a constant @xmath44 cross - sectional slice .
this five - dimensional slice is spanned by the remaining coordinates . for triangle
@xmath27 we have @xmath49 while for triangle @xmath48 @xmath50 where @xmath51 parameterises the ratio of the perimeters of the two triangles , i.e. @xmath52 .
we do not lose the generality to consider @xmath53 .
the different expressions for @xmath54 imply that @xmath55 from which @xmath56 is eliminated to give @xmath57 the conditions for triangle @xmath27 that @xmath58 imply that @xmath59 and @xmath60 , while the condition for triangle @xmath48 that @xmath61 imply that @xmath62 .
furthermore , in terms of these variables , the condition ( [ const_costheta4 ] ) is expressed as @xmath63 in summary , we have the following domains , @xmath64 together with eq .
( [ const_costheta4_ito_abce ] ) . in practice
, we introduce a cutoff for the integration of @xmath51 as @xmath65 is decreasing with @xmath51 asymptotically after integrating out the dependence of @xmath66 , @xmath67 and @xmath68 in the overlap integral .
we set the cut - off to be @xmath69 but the dependence on this cut - off is very weak . also we should emphasize that the cmb measurements will never prove the parameter region where @xmath70 . making use of this parameterisation , the shape function ( [ pritrishapefunction ] ) , the weight function ( [ trispectrum_weight ] ) and the volume element can be rewritten as @xmath71 it is worth noting that although the integration @xmath72 is five - dimensional in eq . ( [ pritrispectrum_correlator_part ] ) , for scale - invariant shape functions with constant @xmath73 , it is enough to evaluate shape correlations only for the four dimensional slices with constant @xmath44 .
in this section , we study the overlap between eq .
( [ fiducial ] ) and the trispectra in single field and multi - field dbi inflation .
it is worth mentioning that from the definition of the shape correlator ( [ pritrispectrum_correlator ] ) , the shape correlations are independent of the normalisations of shape functions .
we will discuss the normalisation of the trispectrum in the next section .
first , we find that the shape function for the trispectrum ( [ fiducial ] ) is given by @xmath74 where @xmath75 is given by eq .
( [ shape_c1 ] ) . we will assume the scale independence of the spectrum @xmath76 in the rest of the paper . in k - inflation model
, this class of models are characterised by @xmath77 for the action given by eq .
( [ action ] ) .
it was also shown that , in the context of the effective theory of inflation , it is possible to construct consistent inflationary models where the trispectrum is characterised by this shape function and its amplitude is not constrained by the bispectrum @xcite .
as this shape function depends on @xmath66 , @xmath68 , @xmath67 , @xmath51 , first , we clarify the @xmath51 dependence of the signal which is given by @xmath78 .
we find that after integrating out the dependence of @xmath66 , @xmath68 and @xmath67 in the overlap integration , the amplitude of the signal is proportional to @xmath79 asymptotically .
this shows that the dominant contribution to the signal for this shape is coming from @xmath80 . in fig . 1
, we show the @xmath51 dependence of @xmath78 .
next , we examine the @xmath66 , @xmath68 , @xmath67 dependence of @xmath81 . for this purpose ,
we plot @xmath82 evaluated at @xmath83 for given @xmath68 in fig .
[ c1_alpha_gamma_beta ] where @xmath84 is symmetric under the exchange of @xmath66 and @xmath67 for the configurations with @xmath85 and the physical region is given by @xmath86 .
it was suggested in @xcite that the shape function corresponding to the reduced trispectrum of single field dbi inflation at leading order in the slow - roll expansion is given by @xmath87\ , , \label{dbisigma_temp}\\ n^{dbi(\sigma)}&= & \frac{h^{12}}{\dot{\phi}^6 c_s^4}\ , , \label{norm_sdbi}\end{aligned}\ ] ] where @xmath88 , @xmath89 and @xmath90 are given by eqs .
( [ shape_s1 ] ) , ( [ def_shape_s2 ] ) and ( [ def_shape_s3 ] ) , respectively . similar to the case of the equilateral shape , first we examine the @xmath51 dependence . in fig . 1
, we plot @xmath91 .
we find that after integrating out the dependence of @xmath66 , @xmath68 and @xmath67 in the overlap integration , the amplitude of the signal is proportional to @xmath92 asymptotically .
this shows that the dominant contribution to the overlap between the single field dbi model and the equilateral shape is coming from @xmath80 .
the difference of the asymptotic @xmath51 dependence between the trispectra corresponding to the differences of the shapes between @xmath75 , which is coming from the contact interaction and @xmath93 , @xmath94 , @xmath95 , which arise from the scalar exchanges .
this was pointed out in ref .
@xcite by considering the double squeezed limit ( @xmath96 ) .
therefore , it is natural that the asymptotic @xmath51 dependence between @xmath84 and @xmath97 is different , as @xmath97 is obtained by a linear combination of @xmath93 , @xmath94 and @xmath95
. next , we examine the @xmath66 , @xmath68 , @xmath67 dependence of @xmath97 . for this purpose , we plot @xmath98 evaluated at @xmath83 for given @xmath68 in fig .
[ sdbi_alpha_gamma_beta ] where @xmath97 is symmetric under the exchange of @xmath66 and @xmath67 for the configurations with @xmath85 and the physical region is given by @xmath86 .
except for the region with very small value of @xmath68 ( from @xmath99 to @xmath100 ) the shapes are very similar to @xmath84 .
this explains that the overlap between @xmath97 and @xmath84 is sufficiently large for the configurations with @xmath83 .
table [ table_summary ] provides a summary of correlations between @xmath81 and @xmath97 .
in addition to the correlation considering full configurations dealing with five dimensional parameter space , for comparisons , we also consider the configurations limited with @xmath83 and equilateral configurations ( @xmath83 , @xmath101 ) .
the overlap decrease if we include the non - equilateral configurations keeping @xmath83 .
this is due to the difference of the shapes for small @xmath68 .
also the different asymptotic behaviours with respect to @xmath51 further reduces the overlap if we integrate over @xmath51 .
however , even after we perform the all integration , the overlap remains high at around 87@xmath102 level .
it was suggested in @xcite that the reduced trispectrum of multi - field dbi inflation dominated by originally purely entropic perturbations at leading order in the slow - roll expansion is given by @xmath103\ , , \label{dbis_temp}\\ n^{dbi(s ) } & = & \frac{h^{12}}{\dot{\phi}^6 c_s^4 } t_{\mathcal{r } s}^4\ , , \label{norm_mdbi}\end{aligned}\ ] ] where @xmath104 and @xmath105 are given by eqs .
( [ def_shape_s2_tilde ] ) and ( [ def_shape_s3_tilde ] ) , respectively .
again , we first examine the @xmath51 dependence of the overlap ( in fig . 1 ) .
we find that after integrating out the dependence of @xmath66 , @xmath68 and @xmath67 , the amplitude of the overlap , @xmath106 is proportional to @xmath92 asymptotically as in the single field case .
this shows that the dominant contribution to the signal for this shape is coming from @xmath80 .
the asymptotic @xmath107 dependence of @xmath108 is the same as that of @xmath97 .
we find that asymptotically @xmath109 , @xmath110 and @xmath111 given by eqs .
( [ shape_s2b ] ) , ( [ shape_s2d ] ) and ( [ shape_s3b ] ) give the dominant contribution to both @xmath97 and @xmath108 , which characterises the asymptotic @xmath51 dependence .
however , as is shown in fig .
[ epsilon_dep2 ] , @xmath112 become negative above some critical value of @xmath51 .
this reduces the final overlap once we integrate over @xmath51 .
next , we examine the @xmath66 , @xmath68 , @xmath67 dependence of @xmath108 . for this purpose ,
again we plot @xmath113 evaluated at @xmath83 for given @xmath68 in fig .
[ mdbi_alpha_gamma_beta ] where @xmath108 is symmetric under the exchange of @xmath66 and @xmath67 for the configurations with @xmath85 and the physical region is given by @xmath86 .
except for the region with very small value of @xmath68 ( from @xmath99 to @xmath100 ) the shape is different from @xmath84 .
this explains that the overlap between @xmath108 and @xmath84 is not so large even for the configurations with @xmath83 .
table [ table_summary ] provides a summary of correlations between @xmath81 and @xmath108 . as in the single field case ,
in addition to the correlation considering full configurations dealing with five dimensional parameter space , for comparisons , we also consider the configurations limited with @xmath83 and equilateral configurations ( @xmath83 , @xmath101 ) .
table [ table_summary ] shows that the overlap becomes smaller once we include the non - equilateral configurations with @xmath83 .
this is clear from the shape difference for @xmath114 .
in addition the shape correlation for full configurations becomes further smaller once we integrate over @xmath51 . as explained
before , this is due to the fact that while @xmath115 is always positive , @xmath116 become negative above some critical value of @xmath51 .
this confirms the fact that the shape dependence of trispectrum can in principle distinguish multi - field dbi inflation models form single field dbi inflation models shown by refs .
@xcite . in practice ,
the overlap still remains at 33@xmath102 level after integrating over all the shape parameters and the equilateral shape could still be used to get a reasonable estimation for the constraints on multi - field dbi inflation model .
in this section , making use of the shape correlations investigated in the previous section , we give theoretical predictions for @xmath2 . as a consistency check
, we applied the same method to estimate the amplitude of the bispectrum in dbi inflation in appendix [ sec_bispectrum_estimation ] . in k - inflation models , as is shown in appendix [ parts ] by setting @xmath117 , the shape function
is given by @xmath118 then , by comparing eqs .
( [ rel_shape_equil_shape_c1 ] ) with ( [ shape_func_equi_effect ] ) , we find @xmath2 is obtained as @xmath119 where we have used @xmath120 for single field k - inflation . in order to express the amplitude of trispectrum in single field dbi inflation in terms of @xmath2 ,
we rewrite eq .
( [ dbisigma_temp ] ) in the following form : @xmath121\ , , \label{shape_func_sdbi_est}\end{aligned}\ ] ] where the numerical factor in eq .
( [ shape_func_sdbi_est ] ) is chosen so that when @xmath122 and @xmath123 , the following conditions are satisfied , @xmath124 of course , @xmath123 is not true in reality and this factor will enhance the amplitude of the signal for a given @xmath2 .
this term is necessary because when we use the estimator related with @xmath125 for the signal whose shape is characterised by @xmath126 , the observed signal is suppressed by @xmath127 and it is necessary to compensate this .
then , by comparing eqs .
( [ dbisigma_temp ] ) with ( [ def_sdbi_uni ] ) , we can relate @xmath2 with the sound speed @xmath128 as @xmath129 where we have used @xmath130 for single field dbi inflation .
similarly , in order to express the amplitude of trispectrum in multi - field dbi inflation in terms of @xmath2 , we rewrite eq .
( [ dbis_temp ] ) in the following form : @xmath131\ , , \label{shape_func_mdbi_est}\end{aligned}\ ] ] where the numerical factor in eq .
( [ shape_func_mdbi_est ] ) is again chosen so that when @xmath122 and @xmath132 , the following conditions are satisfied , @xmath133 then , by comparing eqs .
( [ dbis_temp ] ) with ( [ def_mdbi_uni ] ) , we can relate @xmath2 with the sound speed @xmath128 and the transfer coefficient that relate the amplitude of original entropy perturbations to the final curvature perturbation @xmath134 as @xmath135 where we have used @xmath136 for multi field dbi inflation . in table
[ table_summary ] , we summarise theoretical predictions for @xmath137 for the models discussed in this section .
[ h ! ]
|c|c|c|c|c|c| & overlap - full & @xmath83 & equilateral & theoretical prediction for @xmath137 & @xmath138 + equilateral shape & @xmath139 & @xmath139 & @xmath139 & @xmath140 & @xmath138 + single dbi & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 + multi dbi & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 + it is instructive to compare the values ( [ gnl_sdbi ] ) and ( [ gnl_mdbi ] ) with previous results of @xmath151 @xcite based on the matching of the amplitude at a specific equilateral configuration .
we define the non - linear parameter @xmath152},\ ] ] and evaluated it for the configuration specified by @xmath153 , @xmath154 . for single field and multi - field dbi inflation models ,
we obtain @xmath155 using the same procedure , we get @xmath156 . by comparing this to ( [ tausdbi ] ) and ( [ taumdbi ] ) we can estimate @xmath5 as @xmath157 which underestimates the amplitude by factor @xmath158 compared with the results of eqs .
( [ gnl_sdbi ] ) and ( [ gnl_mdbi ] ) .
this demonstrates that unlike the bispectrum case where the matching of the amplitude at the equilateral configuration gives a reasonable estimation for @xmath159 , it is necessary to calculate the overlap between the shapes in full five - dimensional parameter space to extract the amplitude of the trispectrum @xmath5 .
it is well known that there are many interesting early universe models that predict equilateral type primordial non - gaussianity motivated by string theory and effective field theory . taking into account the fact that future experiments such as planck can prove even next order statistics , it is important to study the primordial trispectrum in these models .
for example , we had shown previously that the trispectrum can in principle distinguish multi - field dbi inflation models from single field dbi inflation models from the shape dependence @xcite . on the other hand , at the practical level ,
since the form of the trispectrum is too complicated , the estimator in this class of models had not been implemented explicitly .
therefore , in this work we have presented a method to estimate primordial trispectrum in equilateral type non - gaussian models such as k - inflation model whose action is given by eq .
( [ action ] ) , single field dbi inflation and multi - field dbi inflation .
our method is based on the following two facts .
one is that the equilateral shape given by eq .
( [ fiducial ] ) becomes factorisable by introducing the integral @xmath18 as was suggested in ref .
the other is that in terms of the shape correlation proposed by ref .
@xcite , we can relate the amplitudes of trispectra with different shapes .
after reviewing the shape correlator , we have calculated the overlaps between the equilateral shape and the shapes of trispectra in single field dbi inflation and multi - field dbi inflation .
we have shown that the shape is @xmath0 correlated with the one in single field dbi inflation , while it is @xmath1 correlated with that in multi - field dbi inflation when the curvature perturbation is originated from purely entropic perturbations during inflation .
we have summarised the overlaps including the configurations restricted to @xmath83 and equilateral configurations ( @xmath83 , @xmath101 ) in table [ table_summary ] .
we found that the main difference between the the equilateral shape and the shape in single field dbi inflation comes from the configurations with @xmath160 , which can be seen even in the equilateral configurations ( @xmath161 ) .
for the shape in multi - field dbi inflation , as the behaviour of the shape function is different from the one in single field dbi inflation @xcite , the overlap becomes smaller .
regardless of this , when we take into account of the fact that this overlap is calculated in the five - dimensional parameter space , the @xmath1 correlation is not necessarily small .
for example , the overlap between equilateral shape and local shapes , which depend on cutoffs in the integration due to divergences in various limits , is less than @xmath162 .
then , we have given theoretical predictions for @xmath2 , which enables us to constrain this type of non - gaussian models from future experiments .
for the model with equilateral shape motivated by k - inflation , we obtained @xmath163 , while for single field dbi inflation and multi - field dbi inflation , @xmath164 and @xmath165 , respectively . to obtain this value , instead of matching the amplitudes of the shape functions at a specific point in the parameter space ,
we have adopted an overlap function , @xmath166 , defined in eq .
( [ pritrispectrum_correlator_part ] ) , which involves integration over five - parameters . before closing , let us comment on the detectability of the trispectrum in future experiments . according to the estimation in refs .
@xcite , the observational errors on @xmath167 and @xmath168 scales as @xmath169 where @xmath170 represents the number of data points of the experiment . therefore ,
current limit on @xmath168 by wmap is of order @xmath171 , while the future experiments like planck @xcite and 21-cm line experiments @xcite are expected to produce a limit @xmath172 and @xmath173 , respectively . for the models like single field dbi inflation where
non - gaussian parameters are given by @xmath174 and @xmath175 , the trispectrum is not detectable even by the planck satellite since there is already a constraint like @xmath176 from @xmath177 from the bispectrum measurement @xcite .
however , for multi - field dbi inflation models where non - gaussian parameters are given by @xmath178 and @xmath179 , it might be possible to detect the trispectrum if there is a large transfer from the entropy mode .
for example , it is detectable by planck if @xmath180 . in this context , to construct a concrete theoretical model which gives a large transfer coefficient @xmath181 is important
. the model with the equilateral shape motivated by effective theory of inflation @xcite can give @xmath168 much larger than @xmath182 which is detectable by planck while keeping the value of @xmath167 to be just of order one .
finally , although we have not studied in this paper , it is known that the ghost inflation also gives equilateral type non - gaussianity @xcite and recently the shape dependence of the trispectrum was also calculated @xcite .
it might be interesting to express the amplitude of the trispectrum in terms of the estimator proposed in this paper .
we would like to thank rob crittenden and dominic galliano for useful discussions .
kk thanks the yukawa institute for theoretical physics , kyoto university and the royal society for two workshops , `` non - linear cosmological perturbations '' ( yitp - w-09 - 01 ) and `` the non - gaussian universe '' ( yitp - t-09 - 05 ) where he is benefitted from many stimulating discussions .
he is also grateful to the organizers of the workshop
the almost non - gaussian universe " held at the institut de physique theorique de saclay and thank leonard senatore and sebastien renaux - petel for useful discussions .
sm is supported by jsps .
kk is supported by european research council , research councils uk and stfc .
in this section , we present the optimal estimator to detect the trispectrum in equilateral type non - gaussian models . as was shown in the main text , eq .
( [ fiducial ] ) is a representative form of the trispectrum which has sufficiently large overlaps between trispectra in physically motivated models such as single field and multi - field dbi inflation .
moreover , this trispectrum can be written in a factorisable form as @xmath183 where @xmath184 then the connected part of the trispectrum of the cmb temperature anisotropies is calculated as @xmath185 where @xmath186 @xmath187 is the radiative transfer function , @xmath188 is the spherical bessel function and @xmath189 is an experimental window function .
the optimal estimator is given by @xcite @xmath190 where @xmath191,\end{aligned}\ ] ] and @xmath192 here @xmath193 using the expression for the trispectrum ( [ fiducialint ] ) , the estimator for the equilateral trispectrum can be written as @xmath194\ ] ] where @xmath195 the ensemble average can be evaluated using the monte carlo simulation .
the fisher error bound of @xmath5 is given by @xmath196 where @xmath197 where we assume the covariant matrix is diagonal and the reduced trispectrum is given by @xmath198 here @xmath199 is given by @xmath200
here , based on our previous work @xcite , we summarise the shape functions for the reduced trispectra in general single field k - inflation described by the following action : @xmath201,\label{action}\ ] ] where @xmath202 is the inflaton field , @xmath203 is the ricci scalar and @xmath204 , where @xmath205 is the metric tensor . for this class of models , the third and the fourth order interaction hamiltonian of the field perturbation @xmath206 in the flat gauge at leading order in the slow - roll expansion
are given by @xmath207\ , , \label{k_inf_third_hamiltonian}\\ h_i^{(4)}(\eta)&=&\int d^3x\left[\beta_1\delta\phi'^4 + \beta_2\delta\phi'^2\left(\partial\delta\phi\right)^2 + \beta_3\left(\partial\delta\phi\right)^4\right]\ , , \label{k_inf_fourth_hamiltonian}\end{aligned}\ ] ] where prime denotes derivative with respect to conformal time @xmath208 and coefficients @xmath209 , @xmath210 , @xmath211 , @xmath212 and @xmath213 are given by @xmath214 @xmath215 then , the shape function @xmath216 is composed of two parts @xmath217 where @xmath218 denotes the contribution from the contact interaction and @xmath219 denotes that from the scalar exchange interaction , respectively . @xmath218 is given by @xmath220 here @xmath221 , @xmath222 and @xmath223 are the following shape functions : @xmath224 + 3\;\ ; { \rm perms.}\ , , \label{shape_c2}\\ s^{c_{3 } } _ { \mathcal{t } } & = & \frac{k_{12 } ( \mathbf{k_1}\cdot\mathbf{k_2 } ) ( \mathbf{k_3}\cdot\mathbf{k_4 } ) } { \sum_{i=1}^4k_i\,\pi_{i=1}^4k_i } \left(1+\frac{\sum_{i < j}k_ik_j}{\left(\sum_{i=1}^4k_i\right)^2}+3\frac{\pi_{i=1}^4k_i}{\left(\sum_{i=1}^4k_i\right)^3}\sum_{i=1}^4\frac{1}{k_i}+12\frac{\pi_{i=1}^4k_i}{\left(\sum_{i=1}^4k_i\right)^4}\right)+3\;\ ; { \rm perms.}\ , , \label{shape_c3}\end{aligned}\ ] ] where
@xmath225 " denotes the permutations @xmath226 , @xmath227 and @xmath228 . in eq .
( [ shape_general_k_cont ] ) , @xmath229 .
similarly , @xmath219 is given by @xmath230 here @xmath88 , @xmath89 and @xmath90 are the following shape functions : @xmath231 + 3\;\ ; { \rm perms.}\ , , \label{shape_s1}\end{aligned}\ ] ] @xmath232 + 3\;\ ; { \rm perms.}\ , , \label{shape_s2a } \\ s^{s_{2b } } _ { \mathcal{t } } & = & -2 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_1}\cdot\mathbf{k_2 } ) ( \mathbf{k_{12}}\cdot\mathbf{k_4 } ) \biggl [ \tilde{\mathcal{f}}_2 ( -k_{12},k_1,k_2,k_3,k_4,k_{12 } ) - \tilde{\mathcal{f}}_2(-k_{12},-k_1,-k_2,k_3,k_4,k_{12 } ) \nonumber\\ & & + \tilde{\mathcal{f}}_2 ( k_3,k_4,-k_{12},k_{12},k_1,k_2 ) - \tilde{\mathcal{f}}_2(-k_3,-k_4,-k_{12},k_{12},k_1,k_2 ) \biggr]+3\;\ ; { \rm perms.}\ , , \label{shape_s2b } \\
s^{s_{2c } } _ { \mathcal{t } } & = & 2 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_{12}}\cdot\mathbf{k_2 } ) ( \mathbf{k_3}\cdot\mathbf{k_4 } ) \biggl [ \tilde{\mathcal{f}}_2 ( k_1,k_2,-k_{12},k_{12},k_3,k_4 ) - \tilde{\mathcal{f}}_2(-k_1,-k_2,-k_{12},k_{12},k_3,k_4 ) \nonumber\\ & & + \tilde{\mathcal{f}}_2 ( -k_{12},k_3,k_4,k_1,k_2,k_{12 } ) - \tilde{\mathcal{f}}_2(-k_{12},-k_3,-k_4,k_1,k_2,k_{12 } ) \biggr]+3\;\ ; { \rm perms.}\ , , \label{shape_s2c } \\ s^{s_{2d } } _ { \mathcal{t } } & = & 4 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_{12}}\cdot\mathbf{k_2 } ) ( \mathbf{k_{12}}\cdot\mathbf{k_4 } ) \biggl [ \tilde{\mathcal{f}}_2 ( k_1,k_2,-k_{12},k_3,k_4,k_{12 } ) - \tilde{\mathcal{f}}_2(-k_1,-k_2,-k_{12},k_3,k_4,k_{12 } ) \nonumber\\ & & + \tilde{\mathcal{f}}_2 ( k_3,k_4,-k_{12},k_1,k_2,k_{12 } ) - \tilde{\mathcal{f}}_2(-k_3,-k_4,-k_{12},k_1,k_2,k_{12 } ) \biggr]+3\;\ ; { \rm perms.}\ , , \label{shape_s2d}\end{aligned}\ ] ] @xmath233 + 3\;\ ; { \rm perms.}\ , , \label{shape_s3a } \\ s^{s_{3b } } _ { \mathcal{t } } & = & 6 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_{12}}\cdot\mathbf{k_4 } ) \biggl [ \tilde{\mathcal{f}}_3 ( k_1,k_2,-k_{12},k_3,k_4,k_{12 } ) - \tilde{\mathcal{f}}_3(-k_1,-k_2,-k_{12},k_3,k_4,k_{12 } ) \nonumber\\ & & + \tilde{\mathcal{f}}_4 ( k_3,k_4,-k_{12},k_1,k_2,k_{12 } ) - \tilde{\mathcal{f}}_4(-k_3,-k_4,-k_{12},k_1,k_2,k_{12 } ) \biggr]+3\;\ ; { \rm perms.}\ , , \label{shape_s3b}\\ s^{s_{3c } } _ { \mathcal{t } } & = & 3 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_1}\cdot\mathbf{k_2 } ) \biggl [ \tilde{\mathcal{f}}_4 ( -k_{12},k_1,k_2,k_3,k_4,k_{12 } ) - \tilde{\mathcal{f}}_4(-k_{12},-k_1,-k_2,k_3,k_4,k_{12 } ) \nonumber\\ & & + \tilde{\mathcal{f}}_3 ( k_3,k_4,-k_{12},k_{12},k_1,k_2 ) - \tilde{\mathcal{f}}_3(-k_3,-k_4,-k_{12},k_{12},k_1,k_2 ) \biggr]+3\;\ ; { \rm perms.}\ , , \label{shape_s3c}\\ s^{s_{3d } } _ { \mathcal{t } } & = & -6 k_{12 } ( k_1 k_2 k_3 k_4)^{1/2 } ( \mathbf{k_{12}}\cdot\mathbf{k_2 } ) \biggl [ \tilde{\mathcal{f}}_4 ( k_1,k_2,k_{-12},k_3,k_4,k_{12 } ) - \tilde{\mathcal{f}}_4(-k_1,-k_2,k_{-12},k_3,k_4,k_{12 } ) \nonumber\\ & & + \tilde{\mathcal{f}}_3 ( k_3,k_4,-k_{12},k_1,k_2,k_{12 } ) - \tilde{\mathcal{f}}_3(-k_3,-k_4,-k_{12},k_1,k_2,k_{12 } ) \biggr]+3\;\ ; { \rm perms . } \
, , \label{shape_s3d}\end{aligned}\ ] ] where again
@xmath225 " denotes the permutations @xmath226 , @xmath227 and @xmath228 .
here we have defined four @xmath234 functions ( with @xmath235 ) as follows ; @xmath236\,,\\ \tilde{\mathcal{f}}_3(k_1,k_2,k_3,k_4,k_5,k_6 ) & = & 2\frac{|k_1k_2k_3k_4|^\frac{1}{2}}{|k_5k_6|^\frac{3}{2}}\frac{1}{\mathcal{a}\mathcal{c}^3 } \left [ 1+\frac{k_5+k_6}{\mathcal{a}}+2\frac{k_5k_6}{\mathcal{a}^2 } + \frac{3}{\mathcal{c}}\left(k_5+k_6 + 2\frac{k_5k_6}{\mathcal{a}}\right ) + 12\frac{k_5k_6}{\mathcal{c}^2 } \right]\,,\\ \tilde{\mathcal{f}}_4(k_1,k_2,k_3,k_4,k_5,k_6 ) & = & 2\frac{|k_1k_4k_5k_6|^\frac{1}{2}}{|k_2k_3|^\frac{3}{2}}\frac{1}{\mathcal{a}^3\mathcal{c } } \bigg[1+\frac{\mathcal{a}}{\mathcal{c}}+\frac{\mathcal{a}^2}{\mathcal{c}^2}+\frac{k_2+k_3}{\mathcal{c}}+2\frac{\mathcal{a}\left(k_2+k_3\right)+k_2k_3}{\mathcal{c}^2 } \nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\quad + 3\frac{\mathcal{a}}{\mathcal{c}^3}\left(\mathcal{a}\left(k_2+k_3\right)+2k_2k_3\right)+12k_2k_3\frac{\mathcal{a}^2}{\mathcal{c}^4}\bigg]\,,\end{aligned}\ ] ] where @xmath237 is defined by the sum of the last three arguments of the @xmath234 functions as @xmath238 and @xmath239 is defined by the sum of all the arguments as @xmath240 . in tables
[ table_shape_correlation_trispectrum ] , we summarise the correlations among these shape functions for the full configurations , the configurations with @xmath83 , the equilateral configurations ( @xmath83 , @xmath101 ) , respectively . [ h ! ] |c|c|c|c|c|c|c|c|c| & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + overlap - full & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255&@xmath256 + @xmath83 & @xmath249 & @xmath257 & @xmath258 & @xmath249 & @xmath259 & @xmath260 & @xmath261&@xmath262 + @xmath263 & @xmath249 & @xmath264 & @xmath265 & @xmath252 & @xmath142 & @xmath266 & @xmath267&@xmath268 + it is worth noting that the following properties @xmath269 hold for the shape function given by @xmath270 where @xmath271 and @xmath272 s are corresponding coefficients . by combining the shape correlations obtained in table [ table_shape_correlation_trispectrum ] and the properties shown above
, we can calculate correlations of the shape functions in any general single field k - inflation models against the equilateral shape ( [ shape_c1 ] ) , once the action ( [ action ] ) is specified .
especially , in the case of single field dbi inflation , the coefficients in the hamiltonians ( [ k_inf_third_hamiltonian ] ) and ( [ k_inf_fourth_hamiltonian ] ) are given by @xmath273 in multi - field dbi inflation model @xcite , in addition to the shape functions @xmath221 , @xmath222 , @xmath223 , @xmath88 @xmath89 , @xmath90 , we find it convenient to define the following shape functions @xmath104 and @xmath105 given by @xmath274 the table [ table_shape_correlation_trispectrum ] shows the shape correlations between these shapes and the equilateral shape .
it is clear that these shapes have significantly low correlations which explain the reason why the final correlation between the shapes in the equilateral model and the multi - field dbi models is lower than the single field model .
in this section , we explain our method to compare the amplitudes of the bispectrum using the overlap integration .
we introduce the shape function @xmath275 @xmath276 for the bispectrum of the curvature perturbation @xmath277 defined by @xmath278 in @xcite , it was shown that at leading order in the slow roll expansion and small sound speed limit , shape function for the bispectrum in the single field dbi inflation is given by @xmath279 where @xmath280 .
the shape ( [ shape_func_b_dbi_theory ] ) is not factorisable and it is not easy to perform an optimal analysis using cmb observations .
thus a factorisable shape which approximates ( [ shape_func_b_dbi_theory ] ) was proposed by @xcite which is given by @xmath281 where the permutations act only on the last term in parentheses . one way to relate @xmath159 to the prediction of dbi
inflation is to compare the amplitude of the shape function for the equilateral configuration @xmath13 .
we get @xmath282 in fact , in @xcite , these two shapes are shown to be @xmath283 correlated to each other based on the following primordial shape correlator for two different shape functions @xmath275 and @xmath284 @xmath285 which is constructed from @xmath286 in eq .
( [ pribispectrum_correlator_part ] ) the integration is performed for the region where the triangle condition for @xmath287 holds and weight function @xmath288 is given by @xmath289 for these two slightly different shapes , it would be enough to match the amplitudes of the bispectra evaluated at the equilateral configuration ( @xmath13 ) where the amplitude have a peak by setting the relation ( [ fnl_dbi_conv ] ) .
however , it is more appropriate to match the amplitude and shape of the bispectra by taking into account all possible configurations . in this context
, we use a different way to estimate the amplitude of the bispectrum in single field dbi inflation using the information of all possible configurations .
for this purpose , we rewrite eq .
( [ shape_func_b_dbi_theory ] ) in the following form : @xmath290 where the numerical factor in eq .
( [ shape_func_b_dbi_est ] ) is chosen so that when @xmath291 and @xmath292 , the following conditions are satisfied , @xmath293 of course , @xmath292 is not true in reality and this factor will enhance the amplitude of the signal for a given @xmath294 .
then , by comparing eqs .
( [ shape_func_b_dbi_theory ] ) with ( [ shape_func_b_dbi_est ] ) , we can relate @xmath294 with @xmath128 as @xmath295 where we have used @xmath296 for single field dbi inflation .
as is expected , this gives almost the same value as the one given by eq .
( [ fnl_dbi_ours ] ) , due to the fact that there is a large overlap between the two shapes .
however , for the trispectrum the difference between the two approaches tends to be large as the trispectrum has five parameters even assuming the scale invariance and thus matching the amplitude at a specific point in the five - dimensional parameter space is generally not enough to ensure that we get the same signal . in this case
, it is more appropriate to use all the shape information using the overlap integration to compare the shape and amplitude of trispectra .
l. senatore , k. m. smith and m. zaldarriaga , jcap * 1001 * ( 2010 ) 028 [ arxiv:0905.3746 [ astro-ph.co ] ] .
j. r. fergusson , m. liguori and e. p. s. shellard , arxiv:1006.1642 [ astro-ph.co ] .
n. bartolo , e. komatsu , s. matarrese and a. riotto , phys .
rept . * 402 * , 103 ( 2004 ) [ arxiv : astro - ph/0406398 ] .
v. acquaviva , n. bartolo , s. matarrese and a. riotto , nucl .
b * 667 * ( 2003 ) 119 [ arxiv : astro - ph/0209156 ] .
j. m. maldacena , jhep * 0305 * ( 2003 ) 013 [ arxiv : astro - ph/0210603 ] .
d. seery and j. e. lidsey , jcap * 0506 * ( 2005 ) 003 [ arxiv : astro - ph/0503692 ] .
l. verde , l. m. wang , a. heavens and m. kamionkowski , mon . not .
* 313 * ( 2000 ) l141 [ arxiv : astro - ph/9906301 ] .
l. m. wang and m. kamionkowski , phys . rev .
d * 61 * ( 2000 ) 063504 [ arxiv : astro - ph/9907431 ] .
e. komatsu and d. n. spergel , phys .
d * 63 * , 063002 ( 2001 ) [ arxiv : astro - ph/0005036 ] .
e. komatsu , d. n. spergel and b. d. wandelt , astrophys .
j. * 634 * ( 2005 ) 14 [ arxiv : astro - ph/0305189 ] .
d. babich , p. creminelli and m. zaldarriaga , jcap * 0408 * , 009 ( 2004 ) [ arxiv : astro - ph/0405356 ] .
s. yadav , e. komatsu and b. d. wandelt , astrophys .
j. * 664 * ( 2007 ) 680 [ arxiv : astro - ph/0701921 ] .
s. yadav , e. komatsu , b. d. wandelt , m. liguori , f. k. hansen and s. matarrese , astrophys .
j. * 678 * ( 2008 ) 578 [ arxiv:0711.4933 [ astro - ph ] ] . j. r. fergusson and e. p. s. shellard , phys .
d * 76 * ( 2007 ) 083523 [ arxiv : astro - ph/0612713 ] .
j. r. fergusson and e. p.
s. shellard , phys .
d * 80 * ( 2009 ) 043510 [ arxiv:0812.3413 [ astro - ph ] ] .
j. r. fergusson , m. liguori and e. p. s. shellard , arxiv:0912.5516 [ astro-ph.co ] . w. hu , phys .
d * 64 * , 083005 ( 2001 ) [ arxiv : astro - ph/0105117 ] .
t. okamoto and w. hu , phys . rev .
d * 66 * , 063008 ( 2002 ) [ arxiv : astro - ph/0206155 ] .
n. kogo and e. komatsu , phys .
d * 73 * , 083007 ( 2006 ) [ arxiv : astro - ph/0602099 ] .
x. chen , phys .
d * 72 * ( 2005 ) 123518 [ arxiv : astro - ph/0507053 ] .
s. e. shandera and s. h. tye , jcap * 0605 * ( 2006 ) 007 [ arxiv : hep - th/0601099 ] .
x. chen , m. x. huang , s. kachru and g. shiu , jcap * 0701 * ( 2007 ) 002 [ arxiv : hep - th/0605045 ] .
d. langlois , s. renaux - petel , d. a. steer and t. tanaka , phys . rev .
* 101 * , 061301 ( 2008 ) [ arxiv:0804.3139 [ hep - th ] ] .
d. langlois , s. renaux - petel , d. a. steer and t. tanaka , phys . rev .
d * 78 * , 063523 ( 2008 ) [ arxiv:0806.0336 [ hep - th ] ] .
f. arroja , s. mizuno and k. koyama , jcap * 0808 * , 015 ( 2008 ) [ arxiv:0806.0619 [ astro - ph ] ] .
j. khoury and f. piazza , jcap * 0907 * ( 2009 ) 026 [ arxiv:0811.3633 [ hep - th ] ] .
d. langlois , s. renaux - petel and d. a. steer , jcap * 0904 * ( 2009 ) 021 [ arxiv:0902.2941 [ hep - th ] ] .
y. f. cai and h. y. xia , phys .
b * 677 * , 226 ( 2009 ) [ arxiv:0904.0062 [ hep - th ] ] .
s. kecskemeti , j. maiden , g. shiu and b. underwood , jhep * 0609 * , 076 ( 2006 ) [ arxiv : hep - th/0605189 ] . j. e. lidsey and d. seery , phys .
rev . * d75 * , 043505 ( 2007 ) , astro - ph/0610398 .
d. baumann and l. mcallister , phys .
d * 75 * , 123508 ( 2007 ) [ arxiv : hep - th/0610285 ] .
r. bean , s. e. shandera , s. h. henry tye and j. xu , jcap * 0705 * , 004 ( 2007 ) [ arxiv : hep - th/0702107 ]
. j. e. lidsey and i. huston , jcap * 0707 * , 002 ( 2007 ) [ arxiv:0705.0240 [ hep - th ] ] .
h. v. peiris , d. baumann , b. friedman , and a. cooray , phys .
rev . * d76 * , 103517 ( 2007 ) , 0706.1240 .
t. kobayashi , s. mukohyama and s. kinoshita , jcap * 0801 * , 028 ( 2008 ) [ arxiv:0708.4285 [ hep - th ] ] .
d. a. easson , r. gregory , d. f. mota , g. tasinato and i. zavala , jcap * 0802 * ( 2008 ) 010 [ arxiv:0709.2666 [ hep - th ] ] .
l. lorenz , j. martin , and c. ringeval , jcap * 0804 * , 001 ( 2008 ) , 0709.3758 .
r. bean , x. chen , h. peiris and j. xu , phys .
d * 77 * , 023527 ( 2008 ) [ arxiv:0710.1812 [ hep - th ] ] .
s. bird , h. v. peiris and d. baumann , phys .
d * 80 * ( 2009 ) 023534 [ arxiv:0905.2412 [ hep - th ] ] .
d. bessada , w. h. kinney and k. tzirakis , jcap * 0909 * ( 2009 ) 031 [ arxiv:0907.1311 [ gr - qc ] ] .
e. j. copeland , s. mizuno and m. shaeri , phys .
d * 81 * ( 2010 ) 123501 [ arxiv:1003.2881 [ hep - th ] ] .
x. chen , m. x. huang and g. shiu , phys .
d * 74 * ( 2006 ) 121301 [ arxiv : hep - th/0610235 ] .
n. arkani - hamed , p. creminelli , s. mukohyama and m. zaldarriaga , jcap * 0404 * ( 2004 ) 001 [ arxiv : hep - th/0312100 ] .
l. senatore , phys .
d * 71 * , 043512 ( 2005 ) [ arxiv : astro - ph/0406187 ] .
k. izumi and s. mukohyama , jcap * 1006 * , 016 ( 2010 ) [ arxiv:1004.1776 [ hep - th ] ] .
q. g. huang , arxiv:1004.0808 [ astro-ph.co ] . | we investigate an estimator to measure the primordial trispectrum in equilateral type non - gaussian models such as k - inflation , single field dbi inflation and multi - field dbi inflation models from cosmic microwave background ( cmb ) anisotropies .
the shape of the trispectrum whose amplitude is not constrained by the bispectrum in the context of effective theory of inflation and k - inflation is known to admit a separable form of the estimator for cmb anisotropies .
we show that this shape is @xmath0 correlated with the full quantum trispectrum in single field dbi inflation , while it is @xmath1 correlated with the one in multi - field dbi inflation when curvature perturbation is originated from purely entropic contribution .
this suggests that @xmath2 , the amplitude of this particular shape , provides a reasonable measure of the non - gaussianity from the trispectrum in equilateral non - gaussian models .
we relate model parameters such as the sound speed , @xmath3 and the transfer coefficient from entropy perturbations to the curvature perturbation , @xmath4 with @xmath2 , which enables us to constrain model parameters in these models once @xmath5 is measured in wmap and planck . |
the joint efforts of physicists in the last four decades has resulted in solid evidence , not only at astrophysical scales , but also cosmological , which leave no doubt that our universe is mainly populated by a still undetected non - interactive type of matter , the so - called dark matter , which nature is still unknown . amongst the numerous theories devised to solve this problem , the picture of a weakly interactive massive particle ( wimp ) arises as the most favourable , since the dark matter abundance inferred today from the cosmic microwave background ( cmb ) matches the abundance of a relic particle with an annihilation cross - section of the order of the weak scale .
furthermore , new particle physics theories motivated by different reasons , provide natural candidates for this type of matter , making this a highly interdisciplinary field of investigation .
+ in the picture of particle dark matter , wimps that populate the milky way can be gravitationally captured by the sun @xcite . since wimps are mandatorily stable , they will accumulate inside sun and annihilate to become standard model particles .
this annihilation will produce a distinctive neutrino signal that can be detected in current and projected neutrino detectors , providing an excellent indirect survey to dark matter properties , which has been extensively studied @xcite .
+ however , most indirect dark matter searches focus on simple models where wimps are majorana particles and annihilate through an @xmath1-wave , velocity independent thermally averaged cross - section . in these models
, it is usual to fix @xmath2 , in order to respect the dark matter density @xcite , which is precisely determined from cmb measurements @xcite .
theoretically , the dark matter abundance is defined by the time of wimp freeze out , which happens when the universe s temperature drops below the wimps mass , resulting in the decoupling of dark matter particles from the primordial universe thermal bath .
this will result in a constant number of wimps per co - moving volume which corresponds to the one measured today .
+ despite the fact that the usual approach in literature is to use a constant thermally averaged cross - section , there are a large number of well - motivated models which have a @xmath0-wave contribution , ( i.e. a dependence in the relative velocity between wimps ) to the annihilation cross - section which can be dominant .
models where dark matter is a majorana particle , such as the _ neutralino _ , a natural candidate which arises in the minimal super symmetric standard model ( mssm ) , @xmath1-wave annihilation to fermion anti - fermion pairs is helicity suppressed by a factor of @xmath3 , where @xmath4 is the wimps mass @xcite .
furthermore , due to @xmath5 conservation , final states with @xmath5=+1 , are only accessible through @xmath0-wave annihilations ( @xmath1-wave states for two identical majorana fermions are @xmath5=-1 ) . hence , neutralino annihilation to @xmath6 or any combination of the vectorial bosons @xmath7 and @xmath8 , can only occur through @xmath0-wave annihilation @xcite .
another well known example is the case of parity conserving minimal extensions to the standard model ( sm ) with a fermionic dark matter candidate - a gauge singlet dirac fermion - in which annihilation to scalar states with even parity , such as @xmath9 , does not receive a contribution from the @xmath1-wave annihilations @xcite .
+ in this article we obtain new limits on the dark matter scattering cross - section from the upper limits on the neutrino fluxes measured by the super - kamiokande and icecube neutrino telescopes , using a general model where @xmath0-wave annihilation is the leading contribution to the total annihilation cross - section , i.e. , where we focus in the second term in the thermally averaged annihilation cross - section expansion , + @xmath10 with @xmath11 , and where we also assumed that the @xmath1-wave contribution to the annihilation , as well as higher order terms with @xmath12 , are negligible .
the coefficient @xmath13 in [ eq : anncs ] is assumed constant and obtained taking into account the dark matter density at the time of freeze - out , which is roughly the same as today ( see sec . [
sec : dmabundance ] ) .
it is important to note that , since we are mainly interested in the epochs from the moment of wimp freeze - out , the expansion in [ eq : anncs ] is only accurate if wimps freeze out at non - relativistic velocities , i.e. if @xmath14 , where @xmath15 is the temperature of freeze - out .
however , in most cases this happens at @xmath16 @xcite , which means that the thermally averaged annihilation cross - section in [ eq : anncs ] is a reliable approximation for our analysis .
+ in section [ sec : dmabundance ] we obtain the coefficient @xmath13 by solving the boltzmann equation for a relic particle with a velocity dependent annihilation cross - section . in sec .
[ sec : dmsun ] we present the formalism needed to compute the neutrino flux from dark matter annihilation in this picture , as well as the stellar evolution code used to compute the dark matter capture and annihilation to sm particles . in sec .
[ sec : dmresults ] we present our results followed by some final remarks .
the dark matter particle co - moving number density @xmath17 is governed the boltzmann equation @xcite @xmath18 where @xmath19 is the hubble parameter , and @xmath20 is the wimp s number density when in equilibrium with the thermal plasma
. equation [ eq : boltzmann ] takes into account the universe s expansion rate as well as the annihilation and production of wimps from the thermal plasma ( first and second terms in the r.h.s . of eq .
[ eq : boltzmann ] ) . to simplify equation [ eq : boltzmann ]
it is usual to use the law of entropy conservation defining the number density @xmath21 in terms of the universe s total entropy @xmath22 , and changing the independent variable from @xmath23 to @xmath24 , where @xmath25 is the photon temperature . using the annihilation cross - section given in [ eq : anncs ] , where @xmath26 , the boltzmann equation eq . [ eq : boltzmann ] can be written as @xmath27 with @xmath28 where @xmath29 is the planck s mass .
the effective number of relativistic degrees of freedom in the universe contributing to the energy density , @xmath30 , can be approximated by a step - function @xcite and in the epoch of wimp freeze - out is essentially the same as the effective number of relativistic degrees of freedom contributing to the total entropy , @xmath31 .
for this reason , hereafter we will use @xmath32 . in this article
we are interested in the standard case where wimps fully thermalize with the thermal plasma in the early universe . in this case , in the early universe ( for low @xmath33 ) , @xmath34 tracks the equilibrium number density @xmath35 since wimps are constantly annihilating and being produced .
when the annihilation rate drops below the expansion rate , wimps freeze - out , i.e. they fall out of thermodynamic equilibrium and their abundance is fixed .
again , assuming that wimps usually freeze - out at temperatures below their mass @xmath36 , one can use a non - relativistic maxwell - boltzmann distribution for @xmath35 for a wimp with @xmath37 degrees of freedom , @xmath38 numerically solving equation [ eq : simpboltz ] with [ eq : lambda ] and [ eq : yeq ] , using the condition that @xmath39 at @xmath40 we obtained the wimp number density today , @xmath41 ( fig .
[ fig : boltzmann ] ) . solved numerically for different wimp masses . for the sake of simplicity it is only shown the equilibrium abundance for @xmath42 .
the number density corresponding to the wimp relic density ( eq . [ eq : relicdens ] ) inferred from cmb measurements is also shown ( dotted line ) . ] finally , we can compute the wimp relic density using @xmath43 from where we obtained the value @xmath44 which best reproduces the relic abundance measured from the cmb , @xmath45 @xcite , for a wimp of @xmath46 gev is @xmath47 it is important to stress that this result is mainly independent from the mass , as we can see in fig . [
fig : boltzmann ] , except for logarithmic corrections .
we can also see by fig .
[ fig : boltzmann ] that wimps freeze - out at a temperature of roughly @xmath48 which is consistent with the hypothesis that freeze - out occurs when wimps are non - relativistic .
dark matter particles in the milky way halo can be gravitationally trapped inside the sun .
this capture process will continue until equilibrium with annihilation is attained , which will fix the number of wimps inside the sun .
the hot plasma will have impact , not only in the dark matter distribution inside the sun , but also on the neutrino flux resulting from its annihilation .
the number of wimps in the sun is governed by three distinct processes : capture , annihilation and evaporation ( which is the inverse reaction of capture ) .
when the evaporation process is negligible , which is our case since we are studying wimps with @xmath49 gev @xcite , the number of wimps is given by @xmath50 where @xmath51 is the capture rate , and @xmath52 ^ 2}\ ] ] is the annihilation coefficient , which is integrated over the sun s volume taking into account the wimps distribution , @xmath53 .
equation [ eq : wimpnumber ] has a straightforward solution , which yields @xmath54 defining the equilibrium time - scale @xmath55 , which represents the time needed to achieve equilibrium between wimp capture and annihilation , the number of wimps in the sun today is in equilibrium if @xmath56 , where @xmath57 gyr . in that case ,
equation [ eq : wimpnumber2 ] simplifies to @xmath58 and the annihilation rate is simply given by @xmath59 as we can see by equation [ eq : annihilationrate ] , if the wimp number equilibrium has been attained , the annihilation rate , @xmath60 , and consequently the neutrino flux , will depend exclusively on the capture rate , which means that no conclusions can be made regarding the nature of the dark matter annihilation cross - section using data from neutrino experiments .
+ in the usual approach , where wimps annihilate through @xmath1-wave , the number of wimps in the sun is in the equilibrium regime for most of the wimp parameter region of interest , and the last equality of eq .
[ eq : annihilationrate ] is an excellent approximation .
however , in our case , where the thermally averaged annihilation cross - section is given by [ eq : anncs ] , the annihilation coefficient is @xmath61 which will generally yield @xmath62 , rendering the approximation in [ eq : annihilationrate ] unreasonable .
for example , simple order of magnitude estimations yield that wimps with @xmath63 gev will mainly populate the inner 6 % of the solar radius ( see eq
. 2.3 of * ? ? ?
* ) . , where the temperature is approximately @xmath64 k , which is at least 4 orders of magnitude lower than the freeze - out temperature for a wimp with the same characteristics .
this means that @xmath65 for wimps annihilating in the sun will be dramatically lower than @xmath66 at freeze - out , which will in turn ( eq . [ eq : annint ] ) increase the equilibrium time - scale .
this will be further explained in sec .
[ sec : resultsequilibrium ] . to compute the annihilation coefficient in eq .
[ eq : annint ] one needs to know the properties of dark matter inside the sun , namely its distribution and temperature .
the wimps distribution in the sun is governed by the knudsen number , @xmath67 which is the ratio between the wimps mean free path in the center of the sun , and the length - scale of its distribution @xcite , @xmath68 where @xmath69 and @xmath70 are the solar central temperature and density respectively . if @xmath71 , wimps are in local thermodynamic equilibrium ( lte ) with the solar plasma and reflect its temperature , i.e. @xmath72 @xcite .
if @xmath73 , wimps will transport energy non - locally and will have an isothermal distribution characterized by a unique temperature , @xmath74 @xcite . to obtain the wimp distribution for arbitrary @xmath75 , it is usual to interpolate between the lte and isothermal distributions .
the interpolation is computed using a suppression factor , function of @xmath75 , which accounts for the departure of the lte regime .
wimps inside the sun will also contribute as an effective mechanism of energy transport @xcite , which efficiency will depend on the knudsen number , @xmath75 , and will be different for the two regimes discussed above , lte and non - local .
this process can have impact in the sun s structure @xcite , however , in our case this impact is negligible , hence we will not pursue this subject further . in our work we used a stellar evolution code which is based on cesam @xcite that has been further developed to include not only dark matter capture and annihilation , but also its energy transport and distribution inside the sun @xcite , as described in sec .
[ sec : dmsunnumber ] and [ sec : dmsundist ] . the reference model without dark matter is calibrated to achieve a standard solar model ( ssm ) @xcite in full agreement with the predominant solar models used in the literature @xcite .
a brief remark regarding the theoretical uncertainty in current solar models , which is mainly caused by the uncertainty on the heavy element abundances @xcite , should be made at this point .
recent independent determinations of the photospheric heavy element abundances @xcite have led to lower metallicity to hydrogen ratio when compared with older estimations by @xcite . despite using complex and improved solar atmosphere models , these downward revisions of the photospheric heavy element abundances result in solar models hard to reconcile with helioseismological and solar neutrino data @xcite , in opposition to the older estimation by @xcite which is in agreement with observations . the solution to this discrepancy , known as the solar abundance problem , is yet to be found , as well as the consensus on which heavy element mixture best reproduces the actual sun ( see * ? ? ? * and ref . therein ) .
for this reason we decided to use both the heavy element mixture by ( * ? ?
* agss09 ) with a metallicity to hydrogen ratio of @xmath76 , and the older mixture by ( * ? ? ?
* gs98 ) with a lower metallicity , @xmath77 , in order to study the impact of the solar chemical composition uncertainty in our results .
the sun is evolved from the zero age main sequence ( zams ) in a galactic halo with a constant dark matter energy density until its present age . for each series of wimp parameters ,
the model is calibrated by automatic adjustment of the convection mixing length parameter and helium abundance , until it reaches a precision of @xmath78 of the present luminosity , radius and mass .
the calibration process takes an average of 10 iterated full - runs to obtain the desired precision .
dark matter capture , thermal annihilation ( eq . [ eq : annint ] ) and energy transport are computed at each time step , in order to consistently achieve the values expected today .
in this section we present the results obtained for thermally annihilating dark matter in the sun for different wimp masses , spin - independent ( si ) and spin - dependent ( sd ) scattering cross - sections .
we used a dark matter density of @xmath79 @xcite .
we assume the standard value for the local orbital speed of the sun @xmath80 @xcite , which results in a velocity dispersion for the standard maxwellian dark matter halo of @xmath81 . to study how wimps are distributed in the solar medium , we computed the knudsen number ( eq . [ eq : knudsen ] ) at the present age , @xmath82 , in the wimp parameter region of interest [ fig : knudsen ] . due to the low scattering cross - sections
, wimps will orbit the sun several times before scattering with nuclei , which means that in this region , @xmath83 is always larger than 1 . taking this into account , and the fact that wimps have thermalized in the sun @xcite
, we can assume that the wimp population in the sun reflects a global temperature @xmath84 , and which distribution is simply given by a boltzmann distribution with state density defined by the gravitational potential @xmath85 , @xmath86 taking into account that wimps tend to strongly cluster on the sun s core ( for example , @xmath87 for @xmath88 gev and @xmath89 for @xmath90 gev ) it is accurate to approximate the potential @xmath85 for @xmath91 by assuming a constant solar central density @xmath92 , which yields @xmath93 , resulting in a much simpler expression for the wimp distribution , @xmath94 in equation [ eq : dmdistsimp ] we also used eq .
[ eq : dmlenght ] assuming that @xmath95 , which is accurate for higher @xmath4 . for lower masses , @xmath96 gev ,
the error in this approximation is always lower than 7% regarding the temperature , which is somewhat mitigated by the fact that wimps with lower masses will be close to equilibrium ( see sec .
[ sec : resultsequilibrium ] ) . using equation [ eq : dmdistsimp ]
, we can solve the integral for annihilation coefficient ( eq . [ eq : annint ] ) , which yields , @xmath97 \label{eq : annana}\ ] ] where @xmath98 cm , is the sun radius .
taking into account that in our case @xmath99 , we can simplify eq .
[ eq : annana ] to the final annihilation coefficient expression , @xmath100 which is independent of the sun radius , as expected .
temperatures inside the sun range from @xmath101 k , which means that the thermally averaged cross - section ( eq . [ eq : annint ] ) inside the sun will be dramatically lower comparatively to the freeze - out epoch , that , as we saw in sec .
[ sec : dmabundance ] , happens for much higher temperatures .
a lower annihilation cross - section , will allow further wimp accretion , resulting in a larger number of wimps in the sun relatively to the @xmath1-wave annihilation case .
if the number of wimps in the sun is in equilibrium , the total annihilation rate will be the same as the @xmath1-wave case , since in this regime , the annihilation rate depends exclusively on the capture rate ( see eq .
[ eq : annihilationrate ] ) .
however , a lower annihilation cross - section can cause the number of wimps to fall out of equilibrium , and in this case , the annihilation rate will be different from the case with standard @xmath1-wave annihilations . + in figure [
fig : annihilation ] we computed the annihilation rate , for wimps isothermally distributed ( eq .
[ eq : dmdistsimp ] ) using the velocity dependent annihilation cross - section in eq .
[ eq : annint ] for different wimp masses and scattering cross - sections . for comparison
, the annihilation rate for the @xmath1-wave case is also shown .
as we can see , for larger scattering cross sections ( @xmath102 and @xmath103 ) there is almost no difference between the two cases , since the number of wimps has achieved equilibrium in the two cases .
however , as we increase the mass and decrease the scattering cross - section , there is a large difference between the @xmath1-wave and @xmath0-wave case .
this is also visible in figure [ fig : equilibrium ] , where we plotted the hyperbolic tangent in eq .
[ eq : wimpnumber ] , @xmath104 which measures the departure from wimp number equilibrium .
regions where the number of wimps is far from equilibrium , i.e. where @xmath105 , the annihilation rate ( fig .
[ fig : annihilation ] ) will be lower in the @xmath0-wave case .
in fact , for smaller scattering cross - sections ( @xmath106 and @xmath107 ) and higher masses , the difference between the two annihilation cases can be of several orders of magnitude , which will result in more loose constraints , as we will show in the next section .
+ wimps in the sun will annihilate to sm particles , from which only neutrinos will be able to escape the sun without loosing all of their energy , due to their weakly interacting nature .
the resulting neutrino flux from wimp annihilation in the sun is given by @xmath108 where the sum is done over the @xmath109 possible annihilation channels , with branching ratio @xmath110 and spectra @xmath111 .
note that the branching ratios for each annihilation channel are not known for a general wimp model .
however , most wimps annihilate predominantly to one particular state , such as @xmath112 , @xmath113 and @xmath114 .
the neutrino signal can be detected on earth using large cerenkov detectors , such as the icecube , in the south pole , and the super - kamiokande , in the _ kamioka mine _
, japan . to compute the neutrino flux measured at the earth , we used wimpsim @xcite ( integrated with our stellar evolution code ) , which uses an event - based framework .
wimpsim computes wimp annihilation to sm particles , and their subsequent hadronization or decay to neutrinos using pythia 6.400 @xcite .
after that , it propagates the neutrino signal through the solar medium and vacuum , using a three - flavour oscillation framework and taking into account neutrino charged currents ( cc ) and neutral currents ( nc ) with the solar nuclei , which will have impact in the neutrino spectra for higher energies . + [ cols="<,<,^,^,^,^,<,^,^ " , ] [ tab : ic ] the upper - limits on the neutrino flux from wimp annihilation in the icecube and super - kamiokande were converted to annihilation rates using wimpsim ( see tables [ tab : sk ] and [ tab : ic ] ) .
the annihilation rates , @xmath60 , where converted into upper limits on the scattering cross - section ( sd and si ) using our stellar evolution code to compute dark matter capture and @xmath0-wave annihilation
. we also computed the upper limits on the scattering cross - section for wimps annihilation through @xmath1-wave for comparison reasons . in fig .
[ fig : constraints ] we plotted the upper limits for @xmath1 and @xmath0-wave annihilation , as well as limits from different direct detection experiments ( see figure caption for references ) . + as expected , for regions where @xmath115 , the limits on the scattering cross - section for wimp models with dominant @xmath0-wave annihilation are coincident with the limits for the standard @xmath1-wave case . however , as we increase @xmath4 and decrease the scattering cross - section , the number of wimps will fall out of equilibrium resulting in smaller annihilation rates , as shown in sec .
[ sec : resultsequilibrium ] , which will convert to upper limits on the sd and si scattering cross - sections above the standard case , specially for higher masses .
in sec . [ sec : limits ] we obtained the limits on the scattering cross - section for p - wave annihilating wimps , placed by the super - kamiokande and icecube neutrino telescopes .
we also computed the limits for the standard case of s - wave annihilation , which are in fair agreement with the results obtained by previous analysis , including the experimental collaborations @xcite .
it should be noted that our constraints for the s - wave case are slightly below than those obtained by the experimental collaborations , which is simply due to the fact that we used a newly determined dark matter local halo density of @xmath116@xcite , while in the other analysis the traditional density @xmath117 is used . despite being presented in a separate fashion , the constraints presented in sec .
[ sec : limits ] can be interpreted as the limiting cases of wimps annihilating with both s and p - wave contributions .
a previous analysis by @xcite also included the computation of the limits on the scattering cross - section for light wimps ( @xmath118 gev for si and @xmath119 gev for sd ) with pure p - wave annihilation . albeit not including the most recent super - kamiokande run ( sk iv , 2008 - 2012 )
, the analysis by @xcite already showed that the p - wave limits are considerably less stringent than those from pure s - wave annihilations , which stems from the fact that the wimp annihilation cross - sections within the sun are smaller than those during the freeze - out epoch . in this work
, we extended and improved this analysis by employing the most recent limits to this date by the super - kamiokande and icecube experiments to a wider range of wimp masses , which are currently responsible for the strongest bounds on the spin - dependent cross - section for wimps with dominant s - wave annihilation . similarly to @xcite , in the si case , we found that the favoured regions by the cdms @xcite , dama / libra @xcite and cresstii @xcite remain viable for wimps annihilating mainly through the @xmath113 or @xmath112 annihilation channel with pure dominant p - wave annihilation . on the other hand , for the sd scenario ,
the dama / libra favoured region is ruled out for both s and p - wave annihilations .
the result obtained also shows that as experiments are built to further constrain the neutrino signal from wimp annihilation , it will be continuously harder to survey lower scattering cross - sections for models with pure @xmath0-wave annihilation , since we will enter the parameter region where the number of wimps is far from equilibrium ( see fig .
[ fig : equilibrium ] ) . in this limit , where @xmath120 , the number of wimps in the sun can be approximated by @xmath121 resulting in @xmath122 which means that if we are able to lower the upper limit in the neutrino flux by a factor of 10 , it will result in lowering the upper limit on the scattering cross - section by a factor of approximately 3 , for models with pure @xmath0-wave annihilation .
+ the limits on the scattering cross - section presented in [ sec : dmresults ] are susceptible to various uncertainties , mainly stemming from the current uncertainty on the capture rate of wimps from the sun , which in turn is sensible to the uncertainty on the local dark matter density @xmath123 and the velocity distribution of wimps in the milky way . regarding the former , independent determinations have shown that the local dark matter can be at least a factor of two larger than the value used here @xcite .
this uncertainty can be simply accounted , since the capture rate is directly proportional to @xmath124 .
the same however is not true for the uncertainty associated with the velocity distribution of wimps in the milky way , which contributes non - trivially for the capture rate .
motivated by the growing tensions between direct detection results , recent n - body simulations have shown that the assumption of a maxwellian velocity distribution can overestimate the number of wimps in the high velocity tail of the velocity distribution @xcite , on which the direct detection rates depend strongly .
it is important to note these results stem from dark matter only simulations , and more recent distributions inferred from simulations including baryons somewhat reduce the discrepancy with the standard maxwellian distribution @xcite .
moreover , the capture rate is only sensible to the low velocity tail of the wimp velocity distribution , given that low velocity wimps are more prone to be gravitationally captured by the sun .
@xcite have studied the impact of the wimp velocity distribution on the capture rate , and they found that for light wimps ( @xmath125 gev ) there is an uncertainty on the capture rate of 24% ( 17% ) for sd ( si ) interactions , while for heavier wimps ( @xmath126 gev ) , the impact is more dramatic , with uncertainties up to 50 % .
these results include both the uncertainty on the wimp velocity distribution and on the parameters @xmath127 and @xmath128
. the error induced by the capture rate uncertainty on the limits from neutrino telescopes will be different for the s and p - wave cases .
for the s - wave case , since @xmath129 , there is a direct correspondence between uncertainties ( see * ? ? ?
* for a review on the uncertainties of the s - wave limits from neutrino telescopes ) . for the p - wave scenario however , as shown in sec
[ sec : resultsequilibrium ] , the number of wimps will generally not be in equilibrium . in the extreme case @xmath130 , and thus we can assume as a conservative estimation for the error of the scattering cross - section limits in the p - wave case as @xmath131 we also studied how the current uncertainty on the overall solar models affects our results , by repeating the computation of the p - wave limits from neutrino telescopes for meticulously chosen benchmark values of @xmath132 , @xmath133 and @xmath134 using the older mixture of abundances by @xcite ( gs98 ) .
again , we found that the largest source of uncertainty is the capture rate @xmath51 , which for sd ( only scattering with hydrogen is taken into account ) scattering is larger for agss09 ( @xmath135 independent of @xmath4 ) , and for si ( scattering with all the elements ) scattering is larger ( @xmath136 for @xmath137 gev and @xmath138 for @xmath139 gev ) for gs98 .
this difference , which is a direct consequence of the difference in @xmath140 between models ( see the discussion in [ sec : dmmodel ] ) , can have impact in the upper limits for @xmath141 since the annihilation rate @xmath60 depends on the number of wimps @xmath142 , which in turn depends on @xmath51 .
the difference in the central temperature between solar models evolved with the mixture by gs98 and agss09 , ( @xmath143 higher @xmath144 for gs98 ) will produce an uncertainty in the annihilation coefficient ( eq . [ eq : annsimp ] ) which is sub - dominant compared to the uncertainty in the capture rate . in the overall picture , for solar models with gs98 , upper limits on the sd cross - section will be @xmath135 more relaxed , while upper limits in the si cross - section will be tighter ( @xmath145 for annihilation to @xmath146 and @xmath147 for @xmath148 ) .
dark matter particles trapped in the sun will annihilate and create a neutrino signal that can be used to survey its properties . in this article we studied the neutrino emission for a simple model in which the main contribution for the annihilation comes from the @xmath0-wave channel , i.e. a velocity dependent cross - section . to obtain the annihilation cross - section we numerically solved the boltzmann equation for the density of a relic particle satisfying the dark matter abundance as measured today from the cmb . to convert the upper limits on the neutrino flux from the icecube and superkamiokande detectors to upper limits on the wimp scattering cross - section we used a robust stellar evolution code to model the sun including dark matter capture and annihilation . assuming that wimps distribute isothermally in the sun s core , we derived an analytical expression for the annihilation coefficient which is directly proportional to the solar central temperature . + differently from the usual case adopted in literature , where wimps annihilate through a velocity independent constant annihilation cross - section ( @xmath1-wave channel ) , the neutrino signal will be directionally proportional to the annihilation coefficient , resulting in upper limits on the scattering cross - section of at least one order of magnitude above the @xmath1-wave case , which reduces the tension with results from other detection experiments .
+ we also studied the impact of the current uncertainty on solar models ( mainly due to the imposition of different solar heavy elements mixtures ) in our results .
we found out that models with a higher metallicity to hydrogen ratio have an enhancement of @xmath147 on the capture rate for spin independent scattering , while models with lower @xmath140 capture @xmath145 more dark matter for spin dependent scattering .
this variations can reflect a maximum of @xmath147 uncertainty on the upper limits for the scattering cross - section for both @xmath1 and @xmath0-wave annihilations .
+ we would like to acknowledge the authors of the numerical packages used in this work , namely morel and lebreton for the stellar evolution code , cesam ; eugnio and casanellas for including all the dark matter framework in the code ; and finally blennow , edsj and ohlsson for wimpsim .
we would also like to thanks the anonymous referee for the insightful comments that helped improve this article .
aartsen , m.g .
( icecube collaboration ) 2013 , , 110 , 131302 ade , p.a.r et .
al ( planck collaboration ) 2015 , e - print : arxiv:1502.01589 .
agnese , r. et al .
( cdms collaboration ) 2013 , , 111 , 251301 akerib , d.s .
et al . ( lux collaboration ) 2014 , , 112 , 091303 angloher , g. , bauer , m. , bavykina , i. , bento , a. , bucci , c. et al .
2012 , eur .
j. , c72 , 1971 aprile , e. et al .
( xenon100 collaboration ) 2012 , , 109 , 181301 asplund , m. , grevesse , n. , sauval , a.j . & scott , p. 2009 , annual review of astronomy and astrophysics bahcall , j.n . & pinsonneault , m.h .
2004 , , 92 , 121301 bahcall , j.n . ,
serenelli , a.m. & basu , s. 2005 , , 621 , l85 bahcall , j.n . ,
serenelli , a.m. & basu , s. 2006 , , 165 , 400 basu , s. & antia , h.m .
2004 , , 606 , l85 bernabei , r. et al .
( dama / libra collaboration ) 2008 , eur .
j. , c56 , 333 bertone , g. , hooper , d. & silk , j. 2005 , physics reports , 405 , 279 busoni , g. , de simone , a. , huang , w. 2013 , jcap , 07 , 010 blennow , m. , edsj , j. & ohlsson , t. 2008 , jcap , 01 , 021 caffau , e. , ludwig , h.g . , steffen , m. , freytag , b. & bonifacio , p. 2011 , sol .
, 268 , 255 catena , r. & ullio , p. 2010 , jcap , 08 , 004 choi , k. , rott , c. & itow , y. 2014 , jcap , 05 , 049 choi , k. et .
al ( super - kamiokande collaboration ) 2015 , , 114 , 141301 danninger , m. & rott , c. 2014 , physics of the dark universe , 5 , 35 dent , j.b . , dutta , s. , scherrer , r.j .
2010 , physics letters b , 687 , 275 drees , m. , & nojiri , m. m. 1992 , 1992nrdm.book.....d gaisser , t.k .
, steigman , g. , tilav , s. 1986 , , 34 , 2206 garbari , s. , read , j.i . & lake , g. 2011 , mon . not .
r. astron .
, 416 , 2318 garbari , s. , liu , c. , read , j.i . & lake , g. 2012 , mon . not .
r. astron .
soc . , 425 , 1445 goldberg , h. 1983 , , 50 , 1419 gould , a. & raffelt , g. 1990 , , 352 , 654 grevesse , n. & sauval , a.j .
1998 , space sci .
, 85 , 161 griest , k. , seckel , d. 1987 , nuclear physics b , 283 , 681 haxton , w.c . ,
hamish robertson , r.g .
& serenelli , a.m. 2013 , annu .
, 51 , 21 jungman , g. , kamionkowski , m. , & griest , k. 1996 , physics reports , 267 , 195 kappl , r. & winkler , m.w .
2011 , nucl .
b , 850 , 505 kelso , c. , savage , c. , valluri , m. , freese , k. , stinson , g.s .
& bailin , j .
2016 , e - print : arxiv:1601.04725 kerr , f.j . & lynden - bell , d. 1986 , mon .
r. astron .
soc . , 221 , 1023 kim , h.g . , lee , k.y . 2007 , , 75 , 115012 kolb , e.w .
& turner , m.s .
, _ the early universe _ , addison - wesley , redwood city , usa 1989 lopes , i. & silk , j. 2010 , science , 330 , 462 lopes , i. , & silk , j. 2010 , , 722 , l95 lopes , i. , casanellas , j. & eugnio , d. 2011 , , 83 , 63521 lopes , i. , panci , p. , & silk , j. 2014 , , 795 , 162 mao , y.y .
, strigari , l.e . ,
wechsler , r.h .
, wu , h.y . &
oliver , h. 2013 , , 764 , 35 mao , y.y . ,
strigar , l.e . & wechsler , r. h. 2014 , , 89 , 063513 morel , p. 1997
, a & a supplement series , 124 , 59 pato , m. , agertz , o. , bertone , g. , moore , b. & teyssier , r. 2010 , 82 , 023531 salucci , p. , nesti , f. , gentile , g. & martins , c.f .
2010 , a&a , 523 , a83 serenelli , a.m. , basu , s. , ferguson , j.w . & asplund , m. 2009 , the astrophysical journal letters , 705 , l123 shelton , j. , shapiro , s. , & fields , b. 2015 , , 115 , 23 sheldon , c. , dutta , b. , & komatsu , e. 2010 , , 82 , 9 silk , j. , olive , k. , srednicki , m. 1985 , , 55 , 257 sjstrand , t. , mrenna , s. & skands , p. 2006
, j. high energy phys . , 05 , 026 sloane , j.d .
, buckley , m.r . ,
brooks , a.m. & governato , f. 2016 , e - print : arxiv:1601.05402 spergel , d.n . & press , w.h .
1985 , , 294 , 663 steigman , g. , sarazin , c.l . ,
quintana , h. , faulkner , j. 1978 , astronomical journal , 83 , 1050 lopes , i. & turck - chieze , s. 1993 , , 408 , 347 turck - chieze , s. , couvidat , s. , piau , l. , ferguson , j. , lambert , p. , ballot , j. , garcia , r.a . &
nghiem , p.a.p . 2004 , , 93 , 211102 wilkstrom , g. , edsjo , j. 2009 , jcap , 04 , 009 | we used a consistent and robust solar model to obtain upper limits placed by neutrino telescopes , such as icecube and super - kamiokande , on the dark matter - nucleon scattering cross - section , for a general model of dark matter with a velocity dependent ( @xmath0-wave ) thermally averaged cross - section . in this picture , the boltzmann equation for the dark matter abundance is numerically solved satisfying the dark matter density measured from the cosmic microwave background ( cmb ) .
we show that for lower cross - sections and higher masses , the dark matter annihilation rate drops sharply , resulting in upper bounds on the scattering cross - section one order of magnitude above those derived from a velocity independent ( @xmath1-wave ) annihilation cross - section .
our results show that upper limits on the scattering cross - section obtained from dark matter annihilating in the sun are sensible to the uncertainty in current standard solar models , fluctuating a maximum of 20 % depending on the annihilation channel . |
large data sets of crystalline structures are nowadays available in two major contexts . on one hand , databases of materials have been created containing structural information of both experimental and theoretical compounds from high - throughput calculations , which are the basis for data - mining techniques in materials discovery projects @xcite . on the other hand ,
ab initio structure predictions @xcite can produce a huge number of new structures that have either not yet been found experimentally or are metastable @xcite . in both cases
it is essential to quantify similarities and dissimilarities between structures in the data sets , requiring a configurational distance that satisfies the properties of a metric . databases frequently contain duplicates and insufficiently characterized structures which need to be identified and filtered . in experimental data ,
the representation of identical structures as obtained from different experiments will always slightly differ due to noise in the measurements , such that the configurational distance is never exactly zero .
noise is also present in theoretical calculations where a geometry relaxation is for instance stopped once a certain , possibly insufficient convergence threshold is reached . in ab initio structure prediction
schemes it is typically necessary to maintain some structural diversity which can be quantified as a certain minimal configurational distance .
all these examples clearly show the need for a metric that allows to measure configurational distances and local structures in a reliable and efficient way .
crystalline structures are typically given in a dual representation .
the first part specifies the cell and the second part the atomic positions within the cell .
the former can for instance be given by the three lattice vectors @xmath0 , @xmath1 and @xmath2 , or by their lengths @xmath3 , @xmath4 and @xmath5 , and the intermediate angles @xmath6 , @xmath7 and @xmath8 . the atomic positions can either be specified by cartesian coordinates or the reduced coordinates with respect to the lattice vectors .
however , such representations are not unique , since any choice of lattice points can serve as cell vectors of the same crystalline structure .
unique and preferably standardized cell parameters are required for comparison and analysis of different crystals @xcite .
algorithms to transform unit cells to a reduced form are frequently used in crystallography , such as the niggli - reduction @xcite which produces cells with shortest possible vectors ( @xmath9 ) .
unfortunately , in the presence of noisy lattice vectors , cells can change discontinuously within the niggli - reduction algorithm .
symmetry analysis and the corresponding classification in the 230 crystallographic space groups are another tool to compare crystal structures
. however , the outcome of a symmetry analysis algorithm strongly depends on a tolerance parameter such that the introduction of some noise can change the resulting space group in a discontinuous manner . because of the above described problems it is difficult to quantify similarities based on dual representations . within the structure prediction community fingerprints that are not based on such a dual representation have been proposed .
oganov et al .
@xcite introduced element resolved radial distribution functions as a crystal fingerprint . for a crystal containing one element only a single function is obtained for the entire system .
the difference between the radial distribution functions of two crystals is then taken as the configurational distance . by definition
the radial distribution function contains only radial information , but no information about the angular distribution of the atoms .
such angular information has been added in the bond characterization matrix ( bcm ) fingerprint @xcite . in this fingerprint spherical harmonic and exponential functions
are used to set up modified bond - orientational order metrics @xcite of the entire configuration .
the distance between two configurations can be measured by the euclidean distance between their bcms .
atomic environment descriptors are also needed in the context of machine learning schemes for force fields @xcite , bonding pattern recognition @xcite , or to compare vacancy , interstitial and intercalation sites @xcite .
these descriptors could also be used to measure similarities between structures .
even though they have never been used in this context we will present a comparison with such a descriptor . when humans decide by visual inspection whether two structures are similar they proceed typically in a different way .
they try to find matching atoms which have the same structural environment .
if all the atoms in one structure can be matched with the atoms of the other structure , the two structures are considered to be identical .
such a matching approach based on the hungarian algorithm @xcite has already turned out to be useful for the distinction of clusters @xcite . in this paper
we will present a fingerprint for crystalline structures which is based on such a matching approach .
the environment of each atom is described by an atomic fingerprint which is calculated in real space for an infinite crystal and represents some kind of environmental scattering properties observed from the central atom .
therefore , all the ambiguities of a dual representation do not enter into the fingerprint , allowing an efficient and precise comparison of structures .
recently we have proposed an configurational fingerprint for clusters @xcite . in this approach
an overlap matrix is calculated for an atom centered gaussian basis set . the vector formed by the eigenvalues of this
matrix forms a global fingerprint that characterizes the entire structure .
the euclidian norm of the difference vector between two structures is the configurational distance between them and satisfies the properties of a metric .
since there is no unique representation of a crystal by a group of atoms ( e.g. the atoms in some unit cell ) we will use atomic fingerprints instead of global fingerprints in the crystalline case .
however , this atomic fingerprint is closely related to our global fingerprint for non - periodic systems . for each atom @xmath10 in a crystal located at @xmath11 we obtain a cluster of atoms by considering only those contained in a sphere centered at @xmath11 . for this cluster
we calculate the overlap matrix elements @xmath12 as described in reference for a non - periodic system , i.e we put on each atom one or several gaussian type orbitals and calculate the resulting overlap integral .
the orbitals are indexed by the letters @xmath13 and @xmath14 and the index @xmath15 gives the index of the atom on which the gaussian @xmath16 is centered , i.e. @xmath17 in this first step , the amplitudes of the gaussians @xmath18 are chosen such that the gaussians are normalized to one . to avoid that the eigenvalues have discontinuities when an atom enters into or leaves the sphere we construct in a second step another matrix @xmath19 such that @xmath20 the cutoff function @xmath21 smoothly goes to zero on the surface of the sphere with radius @xmath22 @xmath23 in the limit where @xmath24 tends to infinity the cutoff function converges to a gaussian of width @xmath25 .
the characteristic length scale @xmath25 is typically chosen to be the sum of the two largest covalent radii in the system .
the value @xmath24 determines how many derivatives of the cutoff function are continuous on the surface of the sphere , and @xmath26 was used in the following .
one can consider the modified matrix @xmath19 to be the overlap matrix of the cluster where the amplitude of the gaussian at atom @xmath13 is determined by @xmath27 . in this way
atoms close to the surface of the sphere give rise to very small eigenvalues of @xmath19 and are thus weighted less than the atoms closer to the center .
the eigenvalues of this matrix @xmath19 are sorted in descending order and form the atomic fingerprint vector @xmath28 .
since we can not predict exactly how many atoms will be in the sphere we estimate a maximum length for the atomic fingerprint vector .
if the number of atoms is too small to generate enough eigenvalues to fill up the entire vector , the entries at the end of the fingerprint vector are filled up with zeros .
this also guarantees that the fingerprint is a continuous function with respect to the motion of the atoms when atoms might enter or leave the sphere . if an atom enters into the sphere some zeros towards the end of the fingerprint vector are transformed in a continuous way into some very small entries which only contribute little to the overall fingerprint .
the euclidean norm @xmath29 measures the dissimilarity between the atomic environments of atoms @xmath10 and @xmath30
. the atomic fingerprints @xmath31 and @xmath32 of all the @xmath33 atoms in two crystalline configurations @xmath34 and @xmath35 can now be used to define a configurational distance @xmath36 between the two crystals : @xmath37 where @xmath38 is a permutation function which matches a certain atom @xmath10 in crystal @xmath34 with atom @xmath39 in crystal @xmath35 . the optimal permutation function which minimizes @xmath36
can be found with the hungarian algorithm @xcite in polynomial time .
if the two crystals @xmath34 and @xmath35 are identical the hungarian algorithm will in this way assign corresponding atoms to each other .
the hungarian algorithm needs as its input only the cost matrix @xmath40 given by @xmath41 in the following it will be shown that @xmath36 satisfies the properties of a metric , namely * positiveness : @xmath42 * symmetry : @xmath43 * coincidence axiom : @xmath44 if and only if @xmath45 * triangle inequality : @xmath46 . from the definition ( eq . [ def ] )
it is obvious that the positiveness and symmetry conditions are fulfilled .
the coincidence theorem is satisfied if the individual atomic fingerprints are unique , i.e if there are not two different atomic environments that give rise to identical atomic fingerprints . in our work on fingerprints for clusters
we have shown that the fingerprints can be considered to be unique if they have a length larger or equal to 3 per atom .
the triangle inequality can be established in this way : @xmath47 where @xmath38 , @xmath48 and @xmath49 are assumed to be the permutations that minimize respectively the euclidean vector norms associated to @xmath50 , @xmath51 and @xmath36 .
since the @xmath11-centered spheres contain typically about 50 atoms , an atomic fingerprint has at least length 50 if only @xmath52-type gaussian orbitals or length 200 if both @xmath52 and @xmath34 orbitals are used . since a configuration is characterized by the ensemble of all the atomic fingerprints of all the atoms in the cell , the amount of data needed to characterize a structure is quite large even though it is certainly manageable for crystals with a small number of atoms per unit cell .
storage requirements might however become too high in certain cases such as large molecular crystals .
we will , therefore , introduce contraction schemes that allow to considerably reduce the amount of data necessary to characterize a crystalline structure .
two such schemes will briefly be discussed below .
let us introduce a function @xmath53 that designates a certain property of the gaussian orbital @xmath13 and encodes it in form of a contiguous integer index . in case of a multicomponent crystal
it can indicate on which kind of chemical element the gaussians are centered and whether the orbital is of @xmath52 or @xmath34 type .
the principal vector is thus chopped into pieces whose elements all carry the same value @xmath53 . in the following presentation of numerical results
we have always considered the central atom to be special , independent of its true chemical type .
having @xmath54 atomic species in the unit cell and using atomic gaussian orbitals with a maximum angular momentum @xmath55 , @xmath53 runs from 1 to @xmath56 .
now we can construct a contracted matrix @xmath57 @xmath58 together with its metric tensor @xmath59 @xmath60 where @xmath61 is the principal vector of the matrix @xmath19 of eq . [ cutoff ] .
the eigenvalues @xmath62 of the generalized eigenvalue problem @xmath63 form again an atomic fingerprint of length @xmath56 which is much shorter than the non - contracted fingerprint @xmath28 .
the fingerprints described so far can in principle also be used for molecular crystals .
however , the amount of data needed to characterize such crystals can be quite large if the molecules forming the crystal contain many atoms . by creating molecular orbitals in analogy with standard methods in electronic structure calculations the required amount of data
can be considerably reduced .
the eigenvalues arising from the overlap matrix in this molecular basis set will then form a fingerprint for the molecular crystal .
the molecular orbitals can be obtained in the following way : for each molecule @xmath10 in our unit cell we cut out a cluster of molecules within a sphere of a certain radius . for each molecule
@xmath6 in this sphere we set up the overlap matrix by putting gaussian type orbitals on all its constituent atoms .
then we calculate for this matrix the eigenvalues and eigenvectors .
the principal vectors @xmath64 belonging to several of the largest eigenvalues @xmath65 are subsequently used for the contraction : @xmath66 no metric tensor is required since the set of vectors used for the contraction is orthogonal .
the molecular orbitals have characteristic patterns , such that the orbital corresponding to the first principal vector has no nodes , while the orbitals of the following principal vectors have increasing number of nodes .
they are therefore similar to the atomic orbitals of @xmath52 , @xmath34 and higher angular momentum character , which were used for the fingerprints in the ordinary crystals . in fig .
[ orbitals ] these orbitals are shown for the case of the paracetamol molecule . by multiplying @xmath67 with some cutoff function as in eq .
[ cutoff ] we can then obtain molecule centered overlap matrices in this molecular basis which is free of discontinuities with respect to the motion of the atoms . in the molecular case
the value of the cutoff function depends on some short range pseudo - interaction between the central and the surrounding molecules .
this interaction @xmath68 between the central molecule @xmath10 and another molecule @xmath6 is given by @xmath69 where the sum over @xmath13 runs over all the atoms in the central molecule @xmath10 and the sum over @xmath14 over all the atoms in the surrounding molecule @xmath6 .
@xmath70 is the distance between the atoms @xmath13 and @xmath14 and @xmath71 is the sum of the van der waals radii of the two atoms .
the interaction is taken to vanish beyond its first zero .
because of the short range of the interaction , molecules sharing a large surface will be coupled strongly .
the analytical form of the cutoff function is identical to the one for the atomic case ( eq . [ fcut ] ) .
however , since a cartesian distance between molecules is ill defined , the argument in eq .
[ fcut ] is modified .
the scaled distance @xmath72 between the atoms is replaced by the normalized interaction between the molecules @xmath73 the eigenvalues of this final overlap matrix form now a fingerprint describing the environment of this molecule @xmath10 with respect to the other molecules . to compare two structures this procedure
is done for all molecules contained in the corresponding unit cell .
a configurational distance is calculated then as in eq .
[ def ] by using the hungarian algorithm @xcite .
structural data found in various material databases is frequently obtained from measurements at different temperatures which results in thermal expansion .
similarly , measurements at different pressures or low quality x - ray diffraction patterns can lead to slight cell distortions .
obviously our fingerprint distances among such expanded or distorted but otherwise identical structures are different from zero . for these reasons
we have introduced a scheme where the six degrees of freedom associated to the cell are optimized while keeping the reduced atomic coordinates fixed such as to obtain the smallest possible distance to a reference configuration .
the gradient of our fingerprint distance with respect to the lattice vectors can be calculated analytically using the hellmann feynman theorem .
an application of the lattice optimization scheme was applied to a subset of zro@xmath74 structures taken from the open quantum materials database ( oqmd ) @xcite , as will be discussed in further detail later in the following section .
fig . [ big ] shows all possible pairwise configurational distances obtained with several fingerprints for various data sets .
different fingerprints are plotted along the x and y axis .
lfp stands for the uncontracted long fingerprint and in square parenthesis it is indicated whether only @xmath52 or both @xmath52 and @xmath34 orbitals were used to set up the overlap matrix , sfp[@xmath52 ] stands for the short contracted fingerprint with @xmath52 orbitals only where the properties used for the contraction are central atom and the element type of the neighboring atoms in the sphere . for materials that have only one type of element ( si in our case ) the atomic fingerprint has only length two and the coincidence theorem is not satisfied .
even though there are hyperplanes in the configurational space where different configurations have identical fingerprints , it is very unlikely that different local minima lie on such hyperplanes and the fingerprint can therefore nevertheless well distinguish between identical and distinct structures .
if both @xmath52 and @xmath34 orbitals are used ( sfp[@xmath76 ) the atomic fingerprint has at least length 4 and no problem with the coincidence theorem arise .
in addition we also show the configurational distances arising from the oganov @xcite and bcm @xcite fingerprints as well as from a fingerprint based on the amplitudes of symmetry functions @xcite .
all our data sets contain both the global minimum ( geometric ground state ) as well as local minima ( metastable ) structures , obtained from minima hopping runs @xcite .
energies and forces were calculated with the dftb+ @xcite method for sic and the molecular crystals , and the lenosky tight - binding scheme was used for si @xcite . for the cspbi@xmath77 perovskite and the transparent conductive oxide zn@xmath74sno@xmath78 plane wave density functional theory ( dft ) calculations were used as implemented in the quantum espresso code @xcite .
the first test set consists of clathrate like structures of low density silicon allotropes @xcite .
low density silicon gives rise to a larger number of low energy crystalline structures than silicon at densities of diamond silicon and thus poses an ideal benchmark system .
in the first line of the figure we show the results of a relatively sloppy local geometry optimization , where the relaxation is stopped once the forces are smaller than 5.e-2 ev / .
gaps separating identical from distinct structures are hardly visible for all fingerprints .
once a very accurate geometry optimization with a force threshold of 5.e-3 ev / @xmath79 is performed , gaps become visible for all the fingerprints . the second data set is silicon carbide , a material well known for its large number of polytypes .
our fingerprint gives rise to a small gap whereas the configurational distances based on all other fingerprints do not show any gap at all .
the opening of a gap can again be observed once the geometry optimization is done with high accuracy . for this case
all fingerprints result in a gap , but like for all test sets it is the least pronounced for the bcm fingerprint .
both the oganov and bcm fingerprints are global ones such that information is lost in the averaging process of these fingerprints as the system gets larger .
therefore , it is not surprising that the gap again disappears even for the high quality geometry optimization once one goes to large cells .
the next two test sets consist of an oxide material and a perovskite with their characteristic building blocks of octahedra and tetrahedra which can be arranged in a very large number of different ways .
all our fingerprints give rise to clear gaps separating identical from distinct structures .
the oganov fingerprint also gives rise to clear gaps whereas the bcm fingerprint only weakly indicates some gap .
the behler fingerprint gives a well pronounced gap for zn@xmath74sno@xmath78 but only a blurred gap for cspbi@xmath77 .
the last theoretical test system is a platinum surface . in this case
the energies were calculated with the morse potential @xcite .
the geometry optimization were done with high accuracy and therefore a big gap is visible in all cases . fig .
[ energy ] shows the correlation between the energy difference and the fingerprint distance for all the test cases of fig .
[ big ] . except for the very large 256 atoms system there exists always a clear energy gap if the geometry optimization was done with high accuracy .
even though there is of course the possibility of nearly degenerate structures , this seems to happen rarely in practice and energy is thus a rather good and simple descriptor for small unit cells . to test our molecular fingerprint ,
two test systems were employed , namely crystalline formaldhyde and paracetamol .
the formaldehyde system comprised 240 structures with 8 molecules per cell and the paracetamol system 300 structures with 4 molecules per cell .
the two top panels of fig . [ molecular ] show the molecular fingerprint distance versus the energy difference of different structures of paracetamol and formaldehyde , respectively .
the two bottom panels show the correlation of the standard fingerprint against the molecular fingerprint for both systems .
the existence of a gap in the pairwise distance distributions clearly indicates that identical and distinct structures can be identified by both fingerprints .
however , the molecular fingerprint vector is considerably shorter because only six principal vectors were used ( shown in fig .
[ orbitals ] ) . since six is the number of degrees of freedom of a rigid rotator it is expected that this fingerprint is long enough to satisfy the coincidence theorem .
top panels : correlation between the energy difference and the molecular fingerprint distance ( mfp ) for formaldehyde ( a ) and paracetamol ( b ) .
bottom panels : correlation between molecular fingerprint distance and standard fingerprint distance ( short contracted fingerprint with s orbitals only , sfp[s ] ) for formaldehyde ( c ) and paracetamol ( d ) . ] the nodal character of the first six principal vectors for the paracetamol molecule . the atoms are colored according to the sign of the elements of the first six principal vectors @xmath64 . a systematic colour pattern can be observed .
the first principal eigenvector never changes sign and has therefore no nodes ( a ) .
higher principal vectors exhibit more and more nodes ( b - f ) . ]
next we applied our fingerprint to zro@xmath74 structures contained in the oqmd .
115 different entries were available at this composition .
the structures were either based on experimental data retrieved from the inorganic crystal structure database ( icsd ) or on binary structural prototypes .
when the oqmd was initially created , duplicate entries were identified with the structure comparison algorithm as implemented in the materials interface ( mint ) software package @xcite which employs a 6-level test that includes cell reduction as well as an analysis of the lattice symmetry .
structures classified as identical to an existing entry in oqmd were mapped to that entry without performing a structural relaxation .
therefore , the structural data set contains both dft optimized and experimental structures , resulting in noise on the atomic and cell coordinates arising from the numerical calculations as well as from the different experiments and thermal effects . in fig .
[ zro2]a we show the ordinary and the lattice vector optimized fingerprint distances for all 115 structures from the database .
we can see that the fingerprint distance can be reduced down to about 1.e-7 for many structures .
for some of them the initial fingerprint distances were as large as 0.1 .
this allows to detect some identical structures whose initial large fingerprint distance was only due to thermal expansion .
however , even with lattice vector optimization it was not possible to decide for the whole data set in an inambiguous way which structures are identical and which were not .
therefore , local geometry optimizations were performed at the dft level for all structures using the vasp code @xcite .
a plane wave cutoff energy of 520 ev was used together with a dense @xmath10-point mesh .
both the atomic and cell variables were relaxed until the maximal force component was less than 2.e-3 ev / @xmath79 and the stress below 1.e-2 gpa .
panel ( b ) of fig .
[ zro2 ] shows the dft energy differences of the relaxed structures against the fingerprint distances , showing a clear gap that allows to distinguish between identical and different structures .
applying the lattice vector optimization scheme on these relaxed structures was not able to further lower the fingerprint distances of identical structures .
the coloring in fig .
[ zro2 ] indicates how the two structures belonging to a fingerprint distance were classified by mint . assuming that there are no different structures with degenerate dft energies
, one can conclude that mint was not able to extract from the non - relaxed data set the information whether structures are identical or not and has erroneously assigned numerous identical structures as distinct , and vice versa to a lesser extent .
since both oganov and bcm methods are global fingerprints that discard crucial information , they can fail to describe structural differences , a problem that becomes especially apparent when considering defect structures in complex materials . as an example
, a @xmath80 supercell was constructed of the cubic perovskite structure of laalo@xmath77 @xcite .
half of the al atoms on the b - sites were replaced by mn .
then , single oxygen vacancies were introduced on symmetrically inequivalent x - sites .
obviously , the structural symmetry was reduced from the initial space group @xmath81 of laalo@xmath77 to the orthorhombic space group @xmath82 of the supercell la(al , mn)o@xmath77 , and the oxygen vacancies resulted in structures with @xmath83 and @xmath84 symmetry .
both mint and our fingerprint confirm that the structures are clearly different , whereas the oganov and bcm fingerprint erronously classify both structures as identical .
atomic fingerprints that describe the scattering properties as obtained from an overlap matrix are well suited to characterize atomic environments .
an ensemble of atomic fingerprints forms a global fingerprint that allows to identify crystalline structures and to define configurational distances satisfying the properties of a metric .
the widely used oganov and bcm fingerprints do not have these properties and do also in practice not allow a reliable way to distinguish identical from distinct structures .
symmetry function based fingerprints are of similar quality as our scattering fingerprints .
however , they are much more costly to calculate .
both fingerprints have a cubic scaling with respect to the number of atoms within the cutoff range , but our prefactor of the matrix diagonalization is much smaller then the prefactor for the 3-body terms required for the calculation of the symmetry functions .
in contrast to ` true'`false ' schemes such as employed in mint which rely on a threshold and affirm that two structures are either identical or distinct , our fingerprint gives a distance between configurations .
the appearance of a gap in the distance distribution indicates that a reliable assignment of identical and distinct structure can be performed .
in addition , strong reductions in the fingerprint distances upon lattice vector optimization can detect and eliminate thermal noise on the data set , rendering our fingerprint ideal to scan for duplicates in large structural databases .
our scheme can easily be extended to molecular crystals by introducing quantities that are analogous to molecular orbitals .
furthermore , the new fingerprint can be used to accurately explore local environments to create atomic and structural attributes for machine learning techniques . in summary
, we have demonstrated that this approach allows to characterize crystalline structures by rather short fingerprint vectors and to decide more reliably whether structures are identical or not than previously proposed methods .
we thank vinay hegde and antoine emery for valuable expert discussions .
this work was done within the nccr marvel project .
ma gratefully acknowledges support from the novartis universitt basel excellence scholarship for life sciences and the swiss national science foundation .
computer resources were provided at the cscs under project s499 and the national energy research scientific computing center , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 .
46ifxundefined [ 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 10.1088/0957 - 0233/16/1/039 [ * * , ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.4812323 [ * * , ( ) ] link:\doibase 10.1038/sdata.2015.9 [ * * , ( ) ] link:\doibase 10.1016/j.commatsci.2015.02.050 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/a901227c [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.cpc.2011.11.007 [ * * , ] @noop _ _ , , vol .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) link:\doibase 10.1103/physrevb.90.054102 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.cpc.2013.10.027 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.014101 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop `` , '' @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | measuring similarities / dissimilarities between atomic structures is important for the exploration of potential energy landscapes .
however , the cell vectors together with the coordinates of the atoms , which are generally used to describe periodic systems , are quantities not suitable as fingerprints to distinguish structures . based on a characterization of the local environment of all atoms in a cell
we introduce crystal fingerprints that can be calculated easily and allow to define configurational distances between crystalline structures that satisfy the mathematical properties of a metric .
this distance between two configurations is a measure of their similarity / dissimilarity and it allows in particular to distinguish structures .
the new method is an useful tool within various energy landscape exploration schemes , such as minima hopping , random search , swarm intelligence algorithms and high - throughput screenings . |
ndal@xmath0 , upd@xmath0al@xmath1 and ybrh@xmath0si@xmath0 are all metallic compounds . the aim of this paper is to study crystal - field ( cef ) effects in these metallic compounds owing to the fact that still the concept and the use of the crystal - field approach is questioned , if applied to metallic compounds .
there are opinions that obtained agreements with many experimental data within the cef - based approaches are accidental ( for further details of these reproaches see arxiv / cond - mat/0504199 ) . for us
the many - electron cef approach is obvious and well founded in solid - state physics .
the many - electron cef approach captures the essential physics of transition - metal compounds .
we think that much of the localized / itinerant controversy is related to misunderstandings due to a long - lasting lack of the open scientific discussion that would allow for good , scientifically clear , formulation of the subject of the controversy .
the cef is a result of an inhomogeneous charge distribution in a solid .
so , it is everywhere .
another problem is its quantification . in a crystal due to the translational symmetry and a high point symmetry the cef potential can be quantified by cef coefficients that reflect multipolar charge moments of the surroundings . a transition - metal cation with the incomplete 4f/5f shell serves as a probe of this cef multipolar potential showing the discrete splitting of the localized 4f/5f states .
here we will present this discrete splitting of the localized 4f/5f states in ndal@xmath0 , upd@xmath0al@xmath1 and ybrh@xmath0si@xmath0 .
the shown transition - metal cations have in these compounds an odd number of 4f/5f electrons .
thus , they have charge - formed kramers doublet ground state . due to the presence of the orbital moment
all of these compounds can serve as example of the anisotropic spin lattice , here chosen in a sequence of the lowering ordering temperature .
ndal@xmath0 crystallizes in the cubic mgcu@xmath0 structure and orders ferromagnetically below 65 k [ 1 ] . among other papers in ref . 2
it has been shown that its magnetic and electronic properties are very well described as being predominantly determined by the nd@xmath5 ion with the 4f@xmath6 configuration . in ref .
2 the temperature dependence of the magnetic moment and of the heat capacity have been calculated .
1 , similar to that shown in ref .
[ 2 ] - here we use the same parameters , only we derive now more physical quantities , shows a discrete electronic structure of 10 lowest localized states originating from the lowest multiplet j=9/2 .
the 4f@xmath6 configuration has in total 120 localized states .
the localized f electrons coexist with itinerant electrons originating from outer 3 nd electrons and from outer al electrons .
itinerant electrons form a band and are responsible for the conductivity .
their contribution to the magnetic moment or to the heat capacity , for instance , is negligible in comparison to that of localized electrons .
the crystal - field splitting amounts to 160 k. magnetic interactions enlarge this splitting to 350 k. the calculated magnetic moment at 0 k of 2.62 @xmath7 is in good agreement with the experimental datum of [email protected] @xmath7 [ 1 ] .
the calculated total moment is composed from the spin moment m@xmath9 = 3.35 @xmath7 and orbital moment m@xmath10=-0.73 @xmath7 . for calculations we make use of a single - ion like hamiltonian for the ground multiplet j=9/2 [ 2 ] : + + h = @xmath11 + @xmath12 = @xmath13 in equation ( 1 ) the first term is the ( cubic ) crystal - field hamiltonian with b@xmath14= -11.48 mk and b@xmath15= + 0.464 mk .
these parameters yield the multipolar moment of the nd surroundings a@xmath16 = b@xmath17/(@xmath18 @xmath19 of a@xmath14=+13.5 ka@xmath20 and a@xmath15=-0.81 ka@xmath21 .
@xmath18 is a respective stevens factor of the ground multiplet and @xmath19 is the respective power of the 4f radius .
the second term takes into account intersite spin - dependent interactions ( n@xmath22 - molecular field coefficient , magnetic moment @xmath23 = -@xmath24j@xmath7 , @xmath24=8/11 - lande factor ) - it produces the magnetic order below t@xmath25 what is seen in fig . 1 as the appearance of the splitting of the kramers doublets and in experiment as the @xmath26-peak in the heat capacity at t@xmath27 .
upd@xmath0al@xmath1 exhibits supercoductivity below 2 k coexisting with antiferromagnetism that appears below 14 k [ 3,4 ] .
a controversy is related to description of f electrons , itinerant or localized in a number of 2 or 3 . in our refs .
[ 5 ] and [ 6 ] , in contrary to authors of refs [ 7,8 ] claiming the 5f@xmath30 configuration , we have interpreted the excitations observed by krimmel _
[ 9 ] as related to the energy level scheme : 0 , 7 mev ( 81 k ) , 10 mev ( 116 k ) and 23.4 mev ( 271 k ) of the @xmath31 ( u@xmath5 ) configuration .
the fine electronic structure of the @xmath31 ( u@xmath5 ) consists of five kramers doublets split by multipolar charge interactions ( cef interactions ) .
the fine electronic structure originates from the lowest multiplet @xmath32i@xmath33 , higher multiplets are at least 0.2 ev above and do not affect practically the ground - multiplet properties .
these 5 doublets are further split in the antiferromagnetic state , i.e. below t@xmath34 of 14 k as is shown in fig .
2 . for calculations we make use the same hamiltonian as for the nd ion in ndal@xmath0 . a derived set of cef parameters of the hexagonal symmetry : b@xmath35=+5.3 k , b@xmath36=+40 mk , b@xmath37=-0.02 mk and b@xmath38=+26 mk yields states at 81 k , 120 k ,
270 k and 460 k , energies of which are in perfect agreement with the experimentally observed excitations .
the highest state was not observed in ins experiment .
this electronic structure accounts also surprisingly well for the overall temperature dependence of the heat capacity , the substantial uranium magnetic moment and its direction [ 6 ] . in fig .
3 we present temperature dependence of the splitting energy between two conjugate kramers ground state .
surprisingly , both the value of the energy and its temperature dependence is in close agreement to a low - energy excitation of 1.7 mev at t=0 k observed by sato _
[ 8 ] which has been attributed by them to a magnetic exciton .
we are convinced that the 5f@xmath39(u@xmath5 ) scheme provides a physical explanation for the 1.7 mev excitation ( magnetic exciton ) - this excitation is associated to the removal of the kramers - doublet ground state degeneracy in the antiferromagnetic state .
ybrh@xmath0si@xmath0 is one of a hallmark heavy - fermion compound with a kondo temperature t@xmath43 of 25 - 30 k. it becomes recently of great scientific importance after the discovery in prof .
f. steglich group [ 10 ] of the well - defined electrons spin resonance ( esr ) signal at temperature so low as 1.5 k similar to the esr signal observed in conventional diluted yb alloys with the yb@xmath5 ion . according to the kondo model below t@xmath43
there should not be discrete states .
the derived @xmath24 tensor is very anisotropic : @xmath44 = 3.561 and @xmath45 = 0.17 at 5 k. at 2 k @xmath44 decreases to 3.50 . in ref .
[ 11 ] we have derived two sets of cef parameters that yield two kramers doublet ground - state eigenfunctions , @xmath46 or @xmath47 , that reproduce this g tensor at 5 k : @xmath46 = 0.944 @xmath48 + 0.322 @xmath49 - -0.052 @xmath50 - 0.047 @xmath51 @xmath52 = 0.803 @xmath53 + 0.595 @xmath54 - -0.027 @xmath55
- 0.009 @xmath56 in ref .
[ 12 ] we have proposed a physical experiment in order to distinguish these two ground states .
we have calculated temperature dependence of the quadrupolar interactions for these two states .
as it is very different for these two ground states the mossbauer experiment should easily distinguish them .
for perfect reproduction of the @xmath24 tensor within the fully - localized crystal - field model it was necessary to introduce the cef b@xmath57 term .
it indicates lowering of the tetragonal symmetry of ybrh@xmath0si@xmath0 at low temperatures to the orthorhombic symmetry .
the increase of this distortion with lowering temperature is , according to us , the reason for temperature variation of the @xmath24 tensor .
a success in description of ndal@xmath0 within the atomic - based cef approach is related to a fact that there is only one crystallographic nd site up to lowest temperatures . due to
the kramers doublet ground state being not susceptible to the charge surroundings and due to relatively large separation to the first excited state mechanisms leading to a lowering symmetry are not effective as by them the system can not gain energy .
ndal@xmath0 being ferromagnet assures conditions for all nd moments to have the same magnetic interactions .
it causes that all nd ions equally contribute to the macroscopic magnetization and thermodynamical properties . as a consequence
, the macroscopic molar values are obtained from the atomic - scale properties by simple multiplication by the avogadro number .
we advocate by years for the importance of the crystal - field approach in 4f/5f compounds , also in those exhibiting heavy - fermion phenomena .
of course , we do not claim that cef explains everything ( surely cef itself can not explain the formation of a magnetic ordering , for which in eq .
1 we add the second term ) , but surely the atomic - like 4f/5f/3d cation is the source of the magnetism of a whole solid .
the atomic - scale magnetic moment is determined by local effects known as crystal field and spin - orbit interactions as well as very strong electron correlations .
these strong electron correlations are predominantly of the intra - atomic origin and taking into account , in the first approximation , via on - site hund s rules .
these strong correlations lead to many - electron version of the cef approach .
finally , we use a name `` @xmath58 electrons '' , as often is used in literature .
of course , @xmath58 electrons are not special electrons but it means electrons in @xmath58 states .
@xmath58 states have well defined characteristics , like the orbital quantum number @xmath59 . by identifying states in a solid we can identified @xmath58 electrons , in particular their number .
owing to a debate on the formation and a role of localized crystal - field states in conducting magnetic materials we point out that in metallic compounds localized crystal - field states coexist with conduction electrons .
the former are predominantly responsible for magnetic and low - energy spectroscopic properties , the latter for the conduction . in ndal@xmath0 and upd@xmath0al@xmath1
we illustrate the physical realization of the anisotropic spin liquid ( lattice of kramers ions with an odd number of electrons ) , that is unstable with respect to spin fluctuations for t@xmath20 k. in these three compounds the magnetic ordering temperature decreases from 65 k via 14 k to 0.07 k only in case of ybrh@xmath0si@xmath0 .
the difficulty in the removal of the kramers degeneracy in case of ybrh@xmath0si@xmath0 is , according to us , the reason for the quantum critical point phenomena observed at ultra - low temperatures . in our atomic - start approach to 3d-/4f-/5f - atom containing compounds we take into account crystal field and spin - orbit interactions as well as very strong electron correlations , predominantly of the intra - atomic origin .
+ @xmath60 dedicated to the pope john paul ii , a man of freedom and truth in life and in science . | the existence of cef states and the magnetism in conducting compounds ndal@xmath0 , upd@xmath0al@xmath1 and ybrh@xmath0si@xmath0 is discussed .
we point out that in metallic compounds localized crystal - field states coexist with conduction electrons .
these compounds are physical realization of an anisotropic spin liquid , which in case of the atomic - like configuration with an odd number of f electrons , is unstable with respect to spin fluctuations for t@xmath20 k. ybrh@xmath0si@xmath0 is discussed as a compound with the kramers doublet ground state with a very low magnetic ordering temperature being a reason for the quantum critical point phenomena at low t. |
there has been growing observational evidence indicating that our universe has entered the epoch of accelerated expansion @xcite . within the framework of einstein gravity
, this acceleration can be attributed to the so called dark energy ( de ) with negative pressure , which dominates the content of universe at present .
the leading interpretation of such de is a cosmological constant with equation of state ( eos ) @xmath0 @xcite . though the cosmological constant is consistent with the observation data , at the fundamental level it fails to be convinced : the vacuum energy density falls far below the value predicted by any sensible quantum field theory by many orders of magnitude , and it unavoidably leads to the coincidence problem , i.e. ,
why are the vacuum and matter energy densities of precisely the same order today ? " .
more sophisticated models have been proposed to replace the cosmological constant by a dynamical dark energy in the conjectures relating the de either to a scalar field called quintessence with @xmath1 , or to an exotic field called phantom with @xmath2 .
it would be fair to say that there is no clear winner in sight to explain the nature of de at the moment .
on the other hand , the remarkable discovery of the cosmic acceleration might also be doubted as the result of a modification of standard einstein gravity at large distances .
this happens in @xmath3 theories @xcite and in braneworld models@xcite .
how to distinguish such modified gravity theory from that of de is an imperative task which will bring about not only a breakthrough in cosmology , but also in the field of high energy physics .
the de is usually not supposed to clump on the scales of the largest cosmic structures , and the most powerful way where its nature can be unveiled is to investigate the expansion history of universe@xcite .
however , the current observations of cosmic expansion history can not break the degeneracies among different approaches trying to explain the cosmic acceleration . considering that the rate of expansion as a function of cosmic time can in turn affect the process of structure formation ,
an alternative channel to study the clustering properties of cosmic structure has been used to detect the possible effects of de .
it is expected that the growth history of dark matter ( dm ) perturbations can provide a complementary probe to distinguish the modified gravity from de@xcite .
however , most of these studies have neglected the effect of de perturbations .
the de perturbations do not affect the background evolution , but they are crucial in determining the de clustering properties@xcite , which consequently have effects on the evolution of dm perturbations .
recently it has been found that for small sound speed and de eos prominently away from @xmath4 , the presence of de perturbations may leave a significant imprint on the growth function of dm perturbations@xcite . in the minimal picture , dm does not feel any significant interactions from de .
although this picture is consistent with current observations , considering that de accounts for a large fraction of the universe , it is natural , in the framework of field theory , to consider its interaction with the remaining fields in the universe .
the possibility that de and dm can interact has been widely discussed in @xcite-@xcite .
it has been shown that the coupling between de and dm can provide a mechanism to alleviate the coincidence problem @xcite .
in addition , it has been argued that an appropriate interaction between de and dm can influence the perturbation dynamics and affect the lowest multipoles of the cmb angular power spectrum @xcite .
furthermore it was suggested that the dynamical equilibrium of collapsed structures such as clusters would be modified due to the coupling between de and dm @xcite . in the presence of such coupling , there has been some concerns about the stability of the perturbations @xcite .
however , it was proved in @xcite that the stability of the curvature perturbation depends on the forms of coupling between dark sectors . since observational signatures on the dark sectors mutual interaction
have been found in the probes of the cosmic expansion history @xcite , it is interesting to ask whether the interaction can have an effect on the growth of structure , whether it can provide a more consistent check for the coupling between dark sectors . in this paper
we are going to study the effect of the interaction between de and dm on the growth function of dm perturbations .
recently there is an attempt on this study@xcite . here
we will incorporate the de perturbations in our study .
we will restrict our investigation to constant sound speed as well as constant eos of de . from our formalism
we will see that the growth of dm perturbations gets more modification due to the dark sectors interaction than that of the de perturbations .
this provides an opportunity to probe the interaction between de and dm through the growth history of dm .
the organization of the paper is as follows . in the following section we will provide our analytical framework for the perturbation equations at the linear level . in sec.iii
, we will present our numerical results and discuss the growth function of dm .
we will give our conclusions and discussions in the last section .
in this section we will derive the second order differential equations for the perturbations of dm and de with couplings between them in a spatially flat friedmann - robertson - walker ( frw ) background .
the perturbed space - time at first order reads , @xmath5,\label{perturbedspacetime}\ ] ] where @xmath6 is scalar metric perturbations , @xmath7 is the cosmic scale factor and @xmath8 .
we work with general stress - energy tensor @xmath9 for a two - component system consisting of de and dm .
each energy - mementum tensor satisfies the conservation law @xmath10 where @xmath11 denotes the interaction between different components and @xmath12 denotes either the dm or the de sector .
the perturbed energy - momentum tensor reads , @xmath13\psi+\delta \rho_{\lambda}'+(p_{\lambda}+\rho_{\lambda})\theta_{\lambda}+3\mathcal { h}(\delta p_{\lambda}+\delta \rho_{\lambda})+3(p_{\lambda}+\rho_{\lambda})\phi'\}\nonumber\\ & = & \delta q^0_{\lambda}\nonumber \label{perturbation } \\
\partial_i \delta \nabla_{\mu}t^{\mu i}_{(\lambda ) } & = & \frac{1}{a^2}\{[p'_{\lambda}+\mathcal { h}(p_{\lambda}+\rho_{\lambda})]\nabla^2b+[(p'_{\lambda}+\rho'_{\lambda})+4\mathcal{h}(p_{\lambda}+\rho_{\lambda})]\theta_{\lambda}\\\nonumber & & + ( p_{\lambda}+\rho_{\lambda})\nabla^2b'+\nabla^2\delta p_{\lambda}+(p_{\lambda}+\rho_{\lambda})\theta_{\lambda}'+(p_{\lambda}+\rho_{\lambda})\nabla^2\psi\ } = \partial_i \delta q^i_{(\lambda)}\end{aligned}\ ] ] where @xmath14 , @xmath15 is the potential of three velocity and the prime denotes the derivative with respect to the conformal time @xmath16 .
the perturbed einstein equations yield , in the linear order @xmath17 ^ 2e&=&-4\pi ga^2\delta \rho\nonumber \\ \mathcal{h}\nabla^2\psi-\nabla^2\phi'+2\mathcal{h}^2\nabla^2b-\frac{a^{''}}{a}\nabla^2b+\frac{1}{6}[\nabla^2]^2e'&=&-4\pi
g a^2(\rho+p)\theta\nonumber \label{pereinstein}\\ -\partial^{i}\partial_{j}\psi-\partial^{i}\partial_{j}\phi+\frac{1}{2}\partial^{i}\partial_{j}e^{''}+\mathcal{h}\partial^{i}\partial_{j}e'+\frac{1}{6}\partial^{i}\partial_{j}\nabla^2e-2\mathcal{h}\partial^{i}\partial_{j}b-\partial^{i}\partial_{j}b'&=&8\pi ga^2\pi^{i}_{j}\end{aligned}\ ] ] where @xmath18 is the total energy perturbation , @xmath19 and @xmath20 . considering an infinitesimal transformation on the coordinates@xcite , @xmath21 the perturbed quantities behave as , @xmath22 inserting eq([quantities])into the medium part of eq([perturbation ] ) , we obtain the behavior of @xmath23 on the right hand side of eq([perturbation ] ) @xmath24 where @xmath25 and @xmath26 is the potential of three vector @xmath27 .
this is consistent with the results got from lie derivatives , @xmath28 which shows that @xmath29 is covariant .
we expand the metric perturbations in fourier space by using scalar harmonics@xcite , @xmath30 and the perturbed conservation equations eq ( [ perturbation ] ) read @xmath31 the interaction @xmath32 can be decomposed into two parts , @xmath33 where @xmath34 is the total four - velocity as defined in@xcite@xcite , @xmath35 is a spacial vector and vanishes in back ground @xmath36@xcite , @xmath37 is the projection operator .
the perturbed quantities read @xmath38 and eq ( [ perturbed ] ) can go back to eq(20 ) and eq(21 ) in @xcite , if we neglect the anisotropic stress @xmath39 . constructing the gauge invariant quantities@xcite , @xmath40'\nonumber\\ \delta q_{p\lambda}^{i}&=&\delta q_{p\lambda } -q^0_{\lambda }
\frac{e'}{2k}.\nonumber\end{aligned}\ ] ] the perturbed einstein equations eq ( [ pereinstein ] ) become @xmath41 where we have assumed that the pressure perturbation of de is isotropic @xmath42 . using the gauge invariant quantities
eq ( [ gaugeinvariant ] ) , we can obtain the linear perturbation equations for dm from eq([perturbed ] ) , @xmath43'=-kv_m-3\phi'+2\psi\frac{a^2q_m^0}{\rho_m}-\delta_m\frac{a^2q_m^0}{\rho_m}-\frac{\rho_m'}{\rho_m}\frac{v_m}{k}\frac{a^2q_m^0}{\rho_m}+\frac{a^2\delta q^{0i}_m}{\rho_m}\nonumber\\ & & + \frac{a^2q_m^{0'}}{\rho_m\mathcal{h}}\phi-\frac{a^2q_m^0}{\rho_m}\left[\frac{\phi}{\mathcal{h}}\right]'\label{dm1}\\ v_m'&=&-\mathcal{h}v_m+k\psi-\frac{a^2q_m^0}{\rho_m}v_m+\frac{a^2\delta q_{pm}^{i}}{\rho_m}\label{dm2}.\end{aligned}\ ] ] considering the pressure perturbation of de @xcite @xmath44 where @xmath45 is the sound speed in de rest frame , @xmath46 is the adiabatic sound speed , and further noting that @xmath47 , we have the gauge invariant form of de perturbation equations @xmath48'+3\mathcal{h}c_e^2(\delta_d+\frac{\rho_d'}{\rho_d}\frac{v_d}{k})-3\mathcal{h}(c_e^2-c_a^2)\frac{\rho_d'}{\rho_d}\frac{v_d}{k}-3w \mathcal{h}(\delta_d+\frac{\rho_d}{\rho_d}\frac{v_d}{k})= \nonumber \\ & & -k(1+w)v_d-3(1+w)\phi'+2\psi\frac{a^2q_d^0}{\rho_d}-(\delta_d+\frac{\rho_d'}{\rho_d}\frac{v_d}{k})\frac{a^2q_d^0}{\rho_d}+\frac{a^2\delta q^{0i}_d}{\rho_d}+\frac{a^2q_d^{0'}}{\rho_d\mathcal{h}}\phi-\frac{a^2q_d^0}{\rho_d}\left[\frac{\phi}{\mathcal{h}}\right]'\\ & & v_d'+\mathcal{h}(1 - 3w)v_d=\frac{k}{1+w}\left[c_e^2(\delta_d+\frac{\rho_d'}{\rho_d}\frac{v_d}{k})-(c_e^2-c_a^2)\frac{\rho_d'}{\rho_d}\frac{v_d}{k}\right]-\frac{w'}{1+w}v_d+k\psi-\frac{a^2q_d^0}{\rho_d}v_d+\frac{a^2\delta q_{pd}^{i}}{(1+w)\rho_d}.\end{aligned}\ ] ] inserting eq.([dm2 ] ) into eq.([dm1])to eliminate @xmath49 , in the subhorizon approximation @xmath50 , we obtain the second order equation for the dm perturbation @xmath51 similarly for the de perturbation we have @xmath52\delta_d'-k(\frac{a^2\delta q_{pd}^{i}}{\rho_d})-k^2c_e^2\delta_d\nonumber\\ & & + 3(w'\mathcal{h}+w\mathcal{h}')\delta_d - k^2(1+w)\psi+ \frac{a^2\delta q_{d}^{0i}}{\rho_d}[\mathcal{h}(1 - 3w)-\frac{w}{1+w}\frac{\rho_d'}{\rho_d}+\frac{a^2q_d^0}{\rho_d}]+(\frac{a^2\delta q_{d}^{0i}}{\rho_d})'.\end{aligned}\ ] ] changing the conformal time into the cosmic proper time , we can rewrite the above second order equations as @xmath53\delta_m=\nonumber \\ & & ( h+\frac{aq_m^0}{\rho_m})\frac{a\delta q_m^{0i}}{\rho_m}+\frac{1}{a}\dot{(\frac{a^2\delta q_m^{0i}}{\rho_m})}-k(\frac{\delta q_{pm}^{i}}{\rho_m})-\frac{k^2}{a^2}\psi,\end{aligned}\ ] ] @xmath54\delta_d+\frac{a\delta q_{d}^{0i}}{\rho_d}\left\{h(1 - 3w)-\frac{w}{1+w}\frac{\dot{\rho}_d}{\rho_d}+\frac{aq_d^0}{\rho_d}\right\}+\frac{1}{a}\dot{(\frac{a^2\delta q_d^{0i}}{\rho_d})}.\nonumber\end{aligned}\ ] ] where the dot denotes the derivative with respect to the proper time @xmath55 . for the convenience in the following discussions , we can further rewrite the second order perturbation equations for de and dm into dimensionless form @xmath56\frac{dln\delta_m}{dlna}+\left(\frac{dln\delta_m}{dlna}\right)^2+\frac{1}{h^2}\left\{\frac{aq_m^0}{\rho_m}h+\left(\frac{aq^0_m}{\rho_m}\right)^2+\frac{h}{a}\frac{d}{dlna}\left(\frac{a^2q_m^0}{\rho_m}\right)\right\}\nonumber\\ & = & \left(h+\frac{aq_m^0}{\rho_m}\right)\frac{a\delta q_m^{0i}}{h^2\rho_m}e^{-ln\delta_m}-\frac{2}{h}\frac{aq_m^0}{\rho_m}\frac{dln\delta_m}{dlna}+\frac{1}{ah}\frac{d}{dlna}\left(\frac{a\delta q_m^{0i}}{\rho_m}\right)e^{-ln\delta_m}-\frac{k\delta q_{pm}^{i}}{h^2\rho_m}e^{-ln\delta_m}\nonumber\\ & & + \frac{3}{2}\left[\omega_m\delta_m+(1-\omega_m)\delta_d\right]e^{-ln\delta_m},\end{aligned}\ ] ] @xmath57\frac{dln\delta_d}{dlna}+\left(3c_e^2 - 6w+\frac{2aq_d^0}{h\rho_d}-\frac{w}{1+w}\frac{1}{\rho_d}\frac{d\rho_d}{dlna}\right)\frac{dln\delta_d}{dlna}\nonumber\\ & = & -3\left(\frac{1}{h}\frac{dh}{dlna}+1\right)c_e^2-\frac{1}{ah}\frac{d}{dlna}\left(\frac{a^2q_d^0}{\rho_d}\right)+\left\{1 - 3w-\frac{w}{1+w}\frac{1}{\rho_d}\frac{d\rho_d}{dlna}+\frac{a}{h}\frac{q_d^0}{\rho_d}\right\}\times\left\{-3c_e^2 + 3w-\frac{a}{h}\frac{q_d^0}{\rho_d}\right\}\nonumber\\ & & -\frac{k\delta q_{pd}^{i}}{\rho_dh^2}e^{-ln\delta_d}-\frac{k^2c_e^2}{a^2h^2}+(1+w)\frac{3}{2}\left\{\omega_m\delta_m+(1-\omega_m)\delta_d\right\}e^{-ln\delta_d}+3\left[\frac{dw}{dlna}+w\left(\frac{1}{h}\frac{dh}{dlna}+1\right)\right]\nonumber \\ & & + \frac{a\delta q_d^{0i}}{h\rho_d}\left\{1 - 3w-\frac{w}{1+w}\frac{1}{\rho_d}\frac{d\rho_d}{dlna}+\frac{aq_d^0}{h\rho_d}\right\}e^{-ln\delta_d}+\frac{1}{ah}\frac{d}{dlna}\left(\frac{a^2\delta q_d^{0i}}{\rho_d}\right)e^{-ln\delta_d}.\end{aligned}\ ] ] in the subhorizon approximation , from the perturbed einstein equations eq ( [ einstein ] ) we can get the poission equation " @xmath58 where we have used friedmann equation in the derivation
. this equation can be used to build the bridge between the matter perturbations to the metric perturbations .
since the nature of de and dm remains unknown , it will not be possible to derive the precise form of the interaction between them from the first principle .
one has to assume a specific coupling from the outset @xcite or determine it from phenomenological requirements @xcite . from eq([qtransform])and eq([lie ] ) , we know that @xmath29 is a covariant vector , which does not need to depend on the four velocity .
for the generality , we can assume the phenomenological description of the interaction between dark sectors in the comoving frame as @xmath59^{t}\nonumber\\ q_d^{\nu}=\left[-\frac{3\mathcal{h}}{a^2}(\delta_1\rho_m+\delta_2\rho_d),0,0,0\right]^{t},\end{aligned}\ ] ] where @xmath60 are small positive dimensionless constants and superindex @xmath61 is the transpose of the vector .
choosing positive sign in the interaction , one can ensure the direction of energy transfer from de to dm , which is required to alleviate the coincidence problem @xcite and avoid some unphysical problems such as negative de density etc @xcite .
in the subhorizon approximation @xmath50 , @xmath62 further noting that the gauge - invariant momentum transfer @xmath63 refers to the intrinsic momentum transfer between dark sectors and as explained in @xcite , such intrinsic momentum transfer is due to the collision of particles from different fluids . such collision can produce acoustics in dm fluid as well as pressure which may resist the squeeze of the attraction of gravity and hinder the growth of gravity fluctuations during tightly coupled photon baryon period , we therefore set @[email protected] is a choice of interaction and the result should not heavily depend on such setting . employing the above interaction form
, we can finally arrive at perturbation equations for dark sectors with constant eos of de @xmath65 \frac{dln \delta_m}{d lna } -(3\delta_1 + 6\frac{\delta_2}{r})\frac{dln \delta_m}{dlna}+3\frac{\delta_2}{r}\frac{dln \delta_d}{dlna } exp(ln \frac{\delta _ d } { \delta _ m})\nonumber\\ & & + \frac{3[exp ( ln \frac{\delta _ d } { \delta _ m})-1]}{r}\left(\delta_2 + 3\delta_1\delta_2 + 3\delta_2 ^
2/r+\delta_2(\frac{h'}{h}+1)- \delta_2\frac{r'}{r}\right)+\frac{3}{2}\left[\omega_m+(1 - \omega_m)exp ( ln \frac{\delta _ d } { \delta _ m})\right]\label{dmevolution}\end{aligned}\ ] ] @xmath66\frac{dln \delta_d}{d lna } + ( 1 + w)\frac{3}{2}\left[\omega_mexp(ln\frac{\delta_m}{\delta_d})+ ( 1 - \omega_m)\right]-\frac{k^2c_e^2}{a^2h^2}\nonumber \\ & + & \left[3\delta_2 + 6\delta_1r+6w-3c_a^2 + 3(c_e^2-c_a^2)\frac{\delta_1r+\delta_2}{1+w}+\frac{c_e^2}{1+w}\frac{\rho_d'}{\rho_d}\right]\frac{dln\delta_d}{dlna}\nonumber \\ & -&3\delta_1 r exp(ln\frac{\delta_m}{\delta_d})\frac{dln\delta_m}{dlna}+3(\frac{h'}{h}+1)(w - c_e^2 ) + 3\delta_1(\frac{h'}{h}r+r+r')- 3\delta_1\left[(\frac{h'}{h}+1)r+r'\right]exp(ln\frac{\delta_m}{\delta_d})\nonumber \\ & + & 3\left[w - c_e^2 + \delta_1r(1- exp(ln\frac{\delta_m}{\delta_d}))\right]\left[(1 - 3w)-3\frac{c_e^2-c_a^2}{1+w}(1+w+\delta_1r+\delta_2)-3(\delta_1r+\delta_2)-\frac{c_e^2}{1+w}\frac{\rho_d'}{\rho_d}\right],\label{deevolution}\end{aligned}\ ] ] where @xmath67 and the prime denotes @xmath68 .
in this section we present numerical results of solving the above perturbation equations .
we concentrate on the behavior of the evolution of the dm perturbation .
is the amplitude of curvature perturbations from wmap five - year results and @xmath69 is the density contrast of dm.,width=336,height=336 ] it was argued that the presence of the interaction between de and dm may give rise to dynamical instabilities in the growth of structure @xcite . in our general form of the phenomenological interaction , when we choose the coupling between dark sectors in proportion to the dm energy density by setting @xmath70 while keeping @xmath71 nonzero , we do find the consistent fast growth of the fluctuations of dm , as seeing in fig [ blowup ] .
however , when we choose the dark sectors mutual interaction in proportion to the energy density of de by taking @xmath72 and @xmath73 , we have the stable dm perturbation .
this result tells us that the stability in the growth of structure depends on the type of coupling between dark sectors , which is consistent with the findings in the curvature perturbation in @xcite .
[ cols="^,^ " , ] to see more clearly of the influence of different parameters on the growth history of the dm perturbation , we introduce the growth index @xmath74 with the definition @xcite @xmath75 the growth index is generically not constant which was first emphasized and investigated in terms of cosmological parameters in @xcite . the growth index has been argued as a useful way in principle to distinguish the modified gravity models from de models @xcite . in fig [ growthindex_without_coupling ] , we show that the influence of the de perturbation on the growth index for the lack of interaction between dark sectors .
the red lines in the figure mark the result without de perturbation . in fig [ growthindex_without_coupling ] a
, we see that for fixed de eos @xmath76 , when @xmath77 decreases , the growth index with the de perturbation deviates more from the result without de perturbation
. this deviation will be even more prominent when de eos is further away from @xmath4 as shown in fig .
[ growthindex_without_coupling ] b. we can see clearly that the difference between the growth index with and without de perturbation can be as big as @xmath78 .
our numerical result further supports what found in @xcite . in fig [ growthindex_without_coupling ] c , d
, we observed that in the subhorizon approximation the value of @xmath79 does not influence as much as the parameters @xmath80 on the result of the growth index . in fig [ couple_growth_index ]
, we present our numerical results when we incorporate the interaction between de and dm .
the solid lines are for the results with de perturbation , while the dotted lines are for the results without de perturbation .
it is clearly shown that the growth index got more influenced from the interaction between dark sectors than the de perturbation .
if we take the best fitting value of @xmath81 from observations of the expansion history of the universe , @xmath82 @xcite , its influence on the growth index will even overwhelm that of the de perturbation . although the enhancement of the growth index due to the de perturbation and the interaction shown in fig .
[ growthindex_without_coupling ] , [ couple_growth_index ] is clear , the available accuracy from the observations such as dune@xcite etc .
is calculated for the @xmath83cdm model and may not be true for our interacting cosmology , since the current values of @xmath84 may not typically be equal in the two cases as argued in @xcite
. however , this phenomenon is interesting , as it opens the possibility that the future measurement of the growth factor may be helpful to reveal the presence of the interaction between de and dm .
the cosmological observations have provided a firm evidence for significant physics beyond standard models .
it is clear now that the formation of structure in the universe demands dm and the accelerated expansion of our universe requires some kind of de or a significant infrared modification on einstein gravity .
de and dm are two major components of the cosmic energy budget .
it is reasonable to explore possible interactions between them in the framework of field theory . in this paper
we have concentrated on the time evolution of the dm perturbation .
we have derived general equations to describe the perturbations of dm and de which incorporates the interactions between them .
it was argued that the interaction between dark sectors might give rise to dynamical instabilities on the growth of structure @xcite .
however , we observed that this instability depends on the form of the coupling .
for example , if we choose the interaction in proportion to the energy density of de , we did observe the stable growth of structure .
this result is consistent with what is found in the curvature perturbation @xcite .
besides the interaction between dark sectors , we have also discussed the effects of the nonvanishing de perturbation on the evolution of dm perturbation .
usually in the discussion of the growth of structure , the de perturbation was neglected . for the minimal picture without the coupling between dark sectors ,
the influence of the de perturbation on the growth function of dm perturbation has been examined in @xcite . in our general formalism , by taking the appropriate coupling between dark sectors which enables the stable growth of structure , we have found that the effect of the interaction between dark sectors overwhelms that of the de perturbation on the growth function of dm perturbation .
when the de eos @xmath76 is in the vicinity of @xmath4 somewhere abound the best fitted value at the moment , the de perturbation is suppressed , however , when the interaction presents , the growth index can differ from the value without interaction by a big amount up to the observational sensibility .
this provides an interesting way to probe the interaction between dark sectors through the observations on the growth of structure in large scale @xcite .
it would be of great interest to confront our theoretical work to observations to constrain the interaction between dark sectors .
however , the data available on the growth of structure are still poor and there is still a long way to go before we can talk about precision cosmology in this respect .
we will leave our work in this direction in the future . on the other hand
, it is also very interesting to include the interaction between dark sectors to modify the code in studying the n - body cosmological simulations on the structure formation .
work in this direction is in progress .
99 s. j. perlmutter et al . ,
nature 391 , 51 ( 1998 ) ; a. g. riess et al .
j. 116 , 1009 ( 1998 ) ; s. j. perlmutter et al .
, astroph .
j. 517 , 565 ( 1999 ) ; j. l. tonry et al .
, astroph .
j. 594 , 1 , ( 2003 ) ; a. g. riess et al .
, astroph .
j. 607 , 665 ( 2005 ) ; p. astier et al .
447 , 31 ( 2005 ) ; a g. riess et al .
, astroph .
j. 659 , 98 ( 2007 ) .
g. r. dvali , g. gabadadze , and m. porrati , phys .
b485 , 208 ( 2000 ) ; a. sheykhi , b. wang , and n. riazi , phys .
d 75 , 123513 ( 2007 ) ; s. y. yin , b. wang , e. abdalla , c. y. lin , phys.rev.d76 , 124026 , ( 2007 ) .
d. n. spergel et al .
( wmap ) , astrophys .
148 , 175 ( 2003 ) , astro - ph/0302209 ; j. l. sievers et al .
( 2005 ) , astro - ph/0509203 ; d. n. spergel et al .
( wmap ) , astrophys .
170 , 377 ( 2007 ) , astro - ph/0603449 ; c .- l .
kuo et al . ( 2006 ) , astro - ph/0611198 ; c. l. reichardt et al . ( 2008 ) , 0801.1491 ; m. r. nolta et al .
( wmap ) ( 2008 ) , 0803.0593 ; s. cole et al .
( the 2dfgrs ) , mon . not .
362 , 505 ( 2005 ) , astro - ph/0501174 ; m. tegmark et al .
d74 , 123507 ( 2006 ) , astroph/ 0608632 ; w. j. percival et al .
, astrophys .
j. 657 , 645 ( 2007 ) , astroph/ 0608636 ; d. j. eisenstein et al .
( sdss ) , astrophys .
j. 633 , 560 ( 2005 ) , astro - ph/0501171 ; w. j. percival et al . , mon . not .
381 , 1053 ( 2007 ) , 0705.3323 ; a. g. riess et al .
( supernova search team ) , astron .
j. 116 , 1009 ( 1998 ) , astro - ph/9805201 ; s. perlmutter et al .
( supernova cosmology project ) , astrophys .
j. 517 , 565 ( 1999 ) , astro - ph/9812133 ; p. astier et al .
447 , 31 ( 2006 ) , astroph/ 0510447 ; a. g. riess et al .
, astrophys .
j. 659 , 98 ( 2006 ) , astroph/ 0611572 .
e. bertschinger and p. zukin , arxiv:0801.2431 ; s. tsujikawa , phys .
d76 , 023514 , ( 2007 ) ; b. jain and p. zhang , arxiv : 0709.2375 ; h. wei and s. n. zhang , arxiv : 0803.3292 ; y. g. gong , arxiv : 0808.1316 .
n. afshordi , m. zaldarriaga , and k. kohri , phys .
rev.d72 , 065024 ( 2005 ) , astro - ph/0506663 ; m. kaplinghat and a. rajaraman , phys .
d75 , 103504 ( 2007 ) , astro - ph/0601517;o .
e. bjaelde et al . , jcap 0801 , 026 ( 2008 ) , 0705.2018;r .
bean , e. e. flanagan , and m. trodden , new j. phys .
10 , 033006 ( 2008 ) , 0709.1124;r .
bean , e. e. flanagan , and m. trodden , phys .
d78 , 023009 ( 2008 ) , 0709.1128;l .
vergani , l. p. l. colombo , g. la vacca , and s. a. bonometto ( 2008 ) , 0804.0285 . | = 0.6 cm we present a general formalism to study the growth of dark matter perturbations when dark energy perturbations and interactions between dark sectors are present .
we show that the dynamical stability on the growth of structure depends on the form of coupling between dark sectors . by taking the appropriate coupling which enables the stable growth of structure , we find that the effect of the interaction between dark sectors overwhelms that of dark energy perturbation on the growth of dark matter perturbation . due to the influence of the interaction ,
the growth index can differ from the value without interaction by an amount up to the observational sensibility , which provides an opportunity to probe the interaction between dark sectors through future observations on the growth of structure .
= 0.8 cm |
there is growing evidence that the classification of galactic nuclei based on their optical spectra alone provides an incomplete and sometimes deceiving description of their true nature . indeed , some `` classical '' starburst galaxies show signatures of a hidden active galactic nucleus ( agn ) at non - optical wavelengths .
the most convincing evidence for obscured agn hidden in starburst nuclei comes from x - ray observations .
the prototype of this class of objects is ngc4945 which hosts a powerful nuclear starburst , identified through emission in the br@xmath1 and pa@xmath2 lines and through its mid - ir spectrum @xcite .
the observations indicate that a starburst superwind has created a large cavity where shocked clouds emit faint liner - type lines .
studies at optical and near - ir wavelengths reveal no evidence for an agn in this galaxy .
however , x - ray observations unambiguously prove the presence of a powerful and heavily obscured agn : the 2 - 10 kev spectrum shows a reflection - dominated component with a prominent fe line , while the intrinsic transmitted component is observed as a strong , variable excess in the range 10 - 100 kev @xcite .
similar cases are ngc6240 ( vignati et al . 1999 ) , which has an optical spectrum very similar to that of ngc4945 ( i.e. weak liner - type lines probably associated with the starburst superwind shocks ; * ? ? ?
* ) , and ngc3690 @xcite which has an hii - type optical spectrum . in all three galaxies ,
x - ray observations clearly indicate the presence of an obscured active nucleus , yet their optical spectra show no signs of the `` classical '' seyfert - type emission lines .
we define `` optically elusive agns '' , or simply `` elusive agns '' , those objects which do not show a seyfert - like spectrum in the optical , but which host a hard x - ray nuclear source whose intrinsic luminosity is in the seyfert range , i.e. @xmath3 .
therefore , elusive agn are generally hosted in nuclei optically classified either as hii or liner .
we emphasize that a liner spectrum does not necessarily imply the presence of an agn with seyfert - like luminosity .
for instance , most low - luminosity nuclei studied by @xcite with classical liner spectra are intrinsically much weaker than classical seyferts ( @xmath4 ) .
alternatively , liner spectra may originate in starburst - driven shocks @xcite and therefore be totally unrelated with the presence of an agn .
an object with a liner - like spectrum but @xmath3 will be therefore considered as an elusive agn .
we have started a program to search for elusive agn in starburst galaxies using infrared , radio and x - ray observations . in this paper , we present the results of a small set of chandra observations . in combination with literature x - ray data for other galaxies ,
we derive a first estimate of the fraction of elusive agns in the local universe and attempt to constrain their nature .
the parent sample consists of all non - seyfert galaxies ( typically starburst with hii or liner spectra ) which show a high brightness temperature ( @xmath5k ) in the vlbi observations obtained by @xcite and @xcite .
the selection based on the presence of a radio core was made to maximize the chance of finding an elusive agn .
indeed , most seyfert nuclei show a compact radio core with brightness temperature higher than @xmath6k .
however , brightness temperatures even in excess of @xmath7k can also be achieved by compact radio sne @xcite .
therefore , the detection of a radio core alone is not a proof for the presence of an agn .
@xcite selected nearby galaxies ( [email protected] ) from the iras catalog with luminosity @xmath9 ( most of these have @xmath10 ) , and warm infrared colours : @xmath11 _ and _ @xmath12 . the warm colours are expected to favor the selection of galaxies hosting agn .
the kewley et al .
sample contains 61 galaxies , 48 of which were observed with the parkes tidbinbilla interferometer ( pti ) .
pti detections imply @xmath13 k. @xcite selected the 40 most luminous galaxies in the iras bright galaxy sample ( @xmath14 ) and observed with the vlbi the subsample of the 31 galaxies with best vla compact fluxes .
detections with the vlbi generally identify higher brightness temperatures ( @xmath15 k in several cases ) .
since , by definition , elusive agn can not be identified through the classical diagnostic diagrams involving optical line ratios , we have to rely on x - ray based classification schemes .
we infer the presence of an agn if one of the following conditions is met .
i ) for compton - thin agn : an absorption - corrected x - ray _ nuclear _ luminosity of @xmath3 and a spectrum that can be described by an absorbed powerlaw with agn - like slope ( @xmath16 ) .
ii ) for compton - thick agn : a flat x - ray spectrum ( powerlaw with @xmath17 ) in the case of cold reflection was chosen because most cold - reflection dominated , compton - thick sources are in this range @xcite and also to discard ulxs , which have @xmath18 @xcite . ] , and/or an fe line at 6.4 - 6.7 kev with equivalent width ( ew ) in excess of 500 ev , and/or the direct detection of the agn emission at e@xmath1910kev ( e.g. ngc4945 ) .
this definition is not as well - defined as the one adopted for the optical identification of seyfert nuclei , and is subject to some ambiguities , as discussed below .
we have used chandra / acis - s to search for elusive agn in four galaxies of our parent sample , namely ugc2369 , ngc2623 , ngc4418 , and ngc4691 .
in addition , we include acis - s observations of ngc2993 which were serendipitously obtained during observations of the companion ngc2992 ( colbert et al . in prep ) .
the data were reduced using ciao v.2.3 , following standard procedures .
table [ tab_chandra ] gives a short observation log . in all cases ,
the nuclear x - ray source discussed here lies within 1@xmath20 of the position of the radio core or the near - ir peak , and thus can confidently be identified with the galaxy nucleus .
the x - ray fluxes from the nuclear sources are too weak to perform a detailed spectral analysis . for our analyses we used the c - statistics which is more appropriate in the case of low signal - to - noise data @xcite .
the weakest sources were fit with a simple power - law using a column density n@xmath21 fixed to the galactic value .
for the brighter sources , the fit was improved by including a thermal component and/or by varying the value of intrinsic n@xmath21 .
a summary of the best fit for each source is given in table [ tab_chandra ] .
ngc2623 is the only source which unambiguously meets the above criteria for the presence of an obscured agn , since its very flat hard x - ray spectrum can only be explained by the presence of a compton - thick , cold reflection dominated agn ( fig.1 ) .
our data constrain the ew of the fe line to an upper limit of 1 kev .
this value is somewhat lower than in typical compton - thick agn , but not unusual among other elusive agn such as ngc4945 ( bassani et al .
1999 ) or ngc3690 @xcite .
an absorbed powerlaw with @xmath22 frozen to 1.9 gives a much worse fit .
ngc4418 also shows evidence for a flat spectrum emission component which may imply the presence of a compton - thick agn , but the limited photon statistics make this identification somewhat tentative .
we note , however , that infrared observations also suggest the presence of a heavily obscured agn in this galaxy @xcite .
also in this case , an absorbed powerlaw with @xmath22 frozen to 1.9 gives a significantly worse fit .
lcccccc name & t@xmath23 & cts@xmath24 & f@xmath25@xmath26 & n@xmath27 & @xmath22 & kt + u2369 & 10 & 112 & 4.9 & @xmath28 & @xmath29 & @xmath30 + n2623 & 20 & 129 & 13.0 & @xmath31 & @xmath32 & @xmath33 + n2993 & 50 & 650 & 8.7 & @xmath81.2 & @xmath34 &
+ n4418 & 20 & 25 & 1.5 & 0.2@xmath35 & @xmath36 & + n4691 & 10 & 24 & 1.4 & 0.2@xmath35 & @xmath37 & + @xmath38 integration time in ksec . + @xmath24 background subtracted counts .
+ @xmath26 0.38 kev observed flux in @xmath39 .
+ @xmath40 absorbing column density in units of @xmath41 . + @xmath35 frozen to the galactic value . [ tab_chandra ]
we have collected additional hard x - ray data available in the literature for other galaxies of our parent sample .
the objects for which hard x - ray data are available are listed in table [ tab_tot ] , where we also report the optical classification of the nuclei : hii and liner are self - explanatory , while `` n4945 '' indicates a spectrum similar to that of ngc4945 , i.e. weak liner - type lines around h@xmath2 , while h@xmath42 and [ oiii]@xmath435007 are undetected . in principle , the absence of h@xmath42 and [ oiii]@xmath435007 does not allow one to distinguish between sy2 and liner .
however , in ngc4945 @xcite showed that the overall spectral properties are inconsistent with sy2 . for the other two objects with n4945-like spectra ( namely ngc2623 and ngc4418 )
the spectral coverage and quality of the optical spectra do not allow to reach similarly firm conclusions .
however , at least in the case of ngc4418 , for which @xcite present data with higher spectral resolution , the width of the lines is more consistent with liners than seyferts . in the third column of table [ tab_tot ] , we mark those objects that have x - ray spectra with evidence for an obscured agn . in these cases
, we provide the absorption - corrected 2 - 10kev luminosity and the column density of absorbing gas n@xmath21 . for compton - thick agn
where only the reflection component is observed , we estimate the intrinsic luminosity by assuming a reflection efficiency of 1% ( with the exception of ngc253 discussed below ) . some additional explanation is required for ngc253 and ngc1808 .
ngc253 was found to host a low - luminosity agn with observed @xmath44 which is obscured by @xmath45 @xcite .
however , as discussed by weaver et al . , the detection of a broad radio recombination line by @xcite strongly suggests that the nucleus is absorbed by compton - thick material and that the intrinsic x - ray luminosity is @xmath461000 times higher .
ngc1808 , on the other hand , appears to be variable : in 1990 , @xcite found evidence for a mildly absorbed agn with @xmath47 , but observations at later epochs @xcite demonstrated that the nuclear source was fainter by a factor of a few .
it is possible that the elusive nature of this agn is caused by fading rather than obscuration . lcccc + name & opt.@xmath38 & x@xmath24 & @xmath48@xmath26 & @xmath49@xmath40 + ugc2369 & hii@xmath50 & @xmath51 & & + ngc2623 & n4945@xmath50 & agn@xmath51 & @xmath52 42.9@xmath35 & @xmath1924 + ugc5101 & liner@xmath53 & agn@xmath54 & @xmath52 43.3@xmath35 & @xmath1924 + ngc3690 & hii@xmath50 & agn@xmath55 & 42.7 & 24.4 + i1525 + 36 & liner@xmath56 & @xmath57 & & + ngc6240 & liner@xmath58 & agn@xmath59 & 44.2 & 24.3 + + name & opt.@xmath38 & x@xmath24 & @xmath48@xmath26 & @xmath49@xmath40 + ngc253 & hii@xmath60 & agn(?)@xmath61 & 39 ( 42@xmath35 ) & 23.3 ( @xmath1924 ) + ngc1672 & hii@xmath62 & @xmath63 & & + ngc1808 & hii@xmath64 & agn@xmath65 & 41.3 & 22.5 + ngc2993 & hii@xmath66 & @xmath51 & & + ngc4418 & n4945@xmath67 & agn(?)@xmath51 & ( 41.2@xmath35 ) & ( @xmath1924 ) + ngc4691 & hii@xmath68 & @xmath51 & & + ngc4945 & n4945@xmath69 & agn@xmath70 & 42.8 & 24.6 + @xmath38 optical classification ( see text ) ; references and notes : @xmath71@xcite ; @xmath72@xcite ; @xmath73@xcite ; @xmath74@xcite ; @xmath75@xcite ; @xmath69@xcite ; @xmath76@xcite ; @xmath77@xcite ; @xmath78the integrated line ratios of ugc5101 are liner - like ( @xmath56 ) , but @xcite claim the detection of sy - like optical signatures after a multi - component fitting analysis ; @xmath79the integrated line ratios of ngc1672 are hii - like ( @xmath80 ) , but @xcite claim the detection of sy - like optical signatures after a multi - component fitting analysis ; @xmath81@xcite dismissed previous claims for the presence of sy - like signatures in the optical spectrum ; @xmath82@xcite ; @xmath83@xcite .
+ @xmath24 x - ray evidence for an agn is marked on this column .
references for the x - ray observations are : @xmath51 this work ; @xmath54 @xcite ; @xmath55 @xcite ; @xmath57 @xcite ; @xmath59 @xcite ; @xmath61 @xcite ; @xmath63 @xcite ; @xmath65 @xcite ; @xmath70 @xcite .
+ @xmath26 log of the absorption corrected 210 kev luminosity ( erg s@xmath84 ) .
+ @xmath40 log of absorbing @xmath85 in units of @xmath86 .
+ @xmath35 compton - thick objects for which the intrinsic x - ray luminosity is inferred by assuming a 1% reflection efficiency ( with the exception of ngc253 , see text ) .
[ tab_tot ]
despite the limited statistics , we can make a first attempt at estimating the fraction of elusive agn in the local universe . as discussed by various authors ( e.g. * ? ? ? * ; * ? ? ? * ) , the fraction of agn in galaxies increases with increasing infrared luminosity .
similarly to these studies , we divide our sample by infrared luminosity and , more specifically , into two bins : @xmath87 and @xmath88 . for consistency with veilleux et al .
, we adopt @xmath89 . the @xmath90 range . in this luminosity range ,
the evidence for elusive agn is more ambiguous than in the higher luminosity bin .
ngc4945 falls into this luminosity range : it is a clear case of an elusive agn and is listed in table [ tab_tot ] as a prototype for this class of objects .
however , it is not included in our parent sample because its infrared colours are cooler than the selection threshold of @xcite .
therefore , ngc4945 will not be counted in the statistics used to derive the local fraction of elusive agn . about 48% of the local galaxies with @xmath87 ( from the iras bright galaxy survey ,
* ) have infrared colours matching the selection criterion used by kewley et al . ( 2000 ) .
we conservatively assume that elusive agn are only hosted in galaxies matching the kewley et al .
ir colour criteria . in their sample ,
42% of the galaxies not classified as seyfert host a radio core . of these
, 6 galaxies have been observed in the hard x - rays ( table [ tab_tot ] ) and 3 of them appear to host an elusive agn , though 2 of these are dubious .
therefore , a rough estimate of the fraction of elusive agn in this luminosity range is 48%@xmath9142%@xmath913/6@xmath5210% ; if we exclude the two dubious cases the fraction estimate drops to @xmath463% .
we summarize this result in fig .
[ frac_el ] , where we include the fraction of local galaxies with optically identified agn derived by @xcite , i.e. 2% for sy1 and 8% for sy2 .
while this calculation is affected by small - number statistics and the uncertain observational evidence for elusive agns in this luminosity bin , we argue below that it should at least represent a lower limit to the actual fraction of elusive agn in the local universe . the @xmath92 range .
the clearest cases for elusive agn can be found in this luminosity range .
four out of the six objects observed in hard x - rays clearly host a heavily obscured agn .
all galaxies in this luminosity range are from the @xcite sample , with the exception of ngc6240 and ngc1614 which are in @xcite .
there are 34 galaxies in the parent sample _ not _ classified as seyfert , 25 of which were observed with vlbi .
we conservatively assume that galaxies not observed or not detected with vlbi do not contain elusive agn .
sixteen of the galaxies observed with vlbi were found to have @xmath93k ( about half of them have @xmath15k ) .
six of the galaxies with @xmath93k core were observed in the hard x - rays ( table [ tab_tot ] ) and 4 were found to host an obscured agn .
thus , the fraction of x - ray detected elusive agn in this luminosity range can be estimated as @xmath94 .
this is the estimated fraction of elusive agn among galaxies _ not _ classified as seyfert .
@xcite estimate that the fraction of optical hii / liner ( i.e. _ non_-seyfert ) galaxies in this luminosity range is 85% .
therefore the estimated fraction of elusive agn in this luminosity range should be 31% @xmath91 0.85 @xmath52 26% of all galaxies .
however , we have to apply another correction to account for the lack of galaxies with luminosity @xmath95 in the smith et al . sample . by interpolating with
the lower luminosity bin discussed above , we finally estimate that the total fraction of elusive agn in the luminosity range @xmath88 is @xmath5220% .
we again illustrate this result in fig .
[ frac_el ] together with the fraction of seyfert galaxies identified optically by @xcite , i.e. 2% for sy1 and 13% for sy2 .
* ulirgs*. in this paper we do not discuss ultraluminous infrared galaxies ( ulirgs , @xmath96 ) , since they are much sparser and poorly representative of the local population of galaxies .
however , for completeness , we also include in fig .
[ frac_el ] the fraction of elusive agn among ulirgs as derived from x - ray observations in the literature @xcite , the value adopted for this paper .
] along with the fraction of seyferts optically identified by @xcite . combining these data
we obtain : sy1=9% , sy2=21% and elusive agn=14% .
we note that the fractions of elusive agn inferred above are probably lower limits because we have assumed that galaxies not matching the selection criteria of kewley et al . and smith et al .
do _ not _ host elusive agn .
the obvious counter example of ngc4945 demonstrates that this assumption is a conservative one .
moreover , there are some examples of classical seyfert nuclei without a detected radio core , and therefore elusive agn without a radio core may also exist . finally , a large fraction of the elusive agn in table [ tab_tot ] are relatively well - studied objects with better x - ray data ( longer integrations and more detailed analysis ) .
elusive agn may have been missed in other objects which have received less attention .
it is not clear why elusive agns do not produce the typical seyfert - type emission lines in their optical spectra .
dilution by the circumnuclear region or by the host galaxy may be an explanation ; in particular @xcite have shown that this may be the case for some of the x - ray selected agn at high redshift . however , most of the objects in our sample are very nearby and have been studied in great detail , and generally no optical seyfert - like signatures were found ( but see notes to table [ tab_tot ] ) .
a lack of an `` uv bump '' typically observed ( or inferred ) in seyferts is also unlikely : at least some elusive agn show strong emission from hot dust ( @xcite for ugc5101 , @xcite for ngc4945 , @xcite for ngc3690 ) which must be heated by the uv bump associated with the x - ray emission .
the x - ray spectra of elusive agn provide some hints of their nature .
indeed , most elusive agn are compton - thick ( @xmath97 ) .
[ nh ] shows the cumulative n@xmath21 distribution of elusive agn and sy2 ( the latter from * ? ? ?
* ) , under the assumption that elusive agn and sy2 have the same local density .
the comparison of the n@xmath21 distributions indicates that elusive agn are more absorbed than sy2 .
this strongly suggests that the elusive nature of this class of nuclei is associated with heavy obscuration . for sy2 s and elusive agn s , under the assumption that elusive agn s are as numerous as sy2 s . the dark - gray histogram shows the contribution by sy2 s , while the light - gray histogram shows the contribution by elusive agn s .
histograms marked with arrows show the fraction of objects with a lower limit on n@xmath21,width=264 ] one possibility is that the narrow line region ( nlr ) is also heavily obscured along our line of sight .
however , spectroscopic observations in the near - ir and mid - ir , where dust extinction is much reduced , did not detect the expected narrow emission lines in several of these objects ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* maiolino et al . in prep . ) .
moreover , the nlr is generally too extended to be completely obscured .
another possibility is that the nuclear pc - scale absorber is not distributed in the torus - like geometry typically assumed for seyfert 2 nuclei , but covers the nuclear uv source in all directions . in this case
, the uv photons can not escape to produce a classical nlr .
detailed studies of ngc4945 favor this scenario @xcite .
indeed , the clouds located in the cavity produced by the nuclear starburst superwind are very faint and only weakly ionized , implying that they are not exposed to the strong photoionizing uv continuum which is expected to be associated with the x - ray emission .
as suggested by done et al .
( 2003 ) a disk / torus of compton - thick material may extend ( though with lower @xmath85 ) to high latitudes and totally cover the agn .
a similar scenario was also proposed by @xcite , who predict that such embedded agn should be characterized by a deep silicate absorption at 9.7@xmath98 m .
notably , our sample includes some of the objects with the deepest silicate absorption , e.g. ngc4945 @xcite and ngc4418 @xcite .
by combining new and past hard x - ray observations we have defined a small sample of optically elusive agn , i.e. nuclei with optical spectra that do not show any evidence for the presence of a seyfert nucleus , but with x - ray properties indicative of a heavily obscured agn ( with seyfert - type luminosity ) . after accounting for selection effects
, we estimate that the fraction of galaxies with elusive agn may be relatively high , i.e. comparable to or higher then classical , optically classified seyfert nuclei . as a consequence
, the local total density of agn may be a factor of @xmath462 higher than estimated from optical spectroscopic surveys .
our results also imply that the ratio between obscured and unobscured agn is higher than estimated previously , i.e. about 7:1 in the luminosity range @xmath99 and about 16:1 in the luminosity range @xmath100 .
nearly all elusive agn are heavily absorbed in the x - rays ( compton thick ) , suggesting that their elusive nature is associated with heavy obscuration .
the geometry of the obscuring medium may be different than in sy2 galaxies .
in particular , the extension of the compton thick medium may completely embed the nuclear source , preventing uv photons from escaping and producing the narrow line region .
the x - ray spectral properties of elusive agn may have important implications for the x - ray background . indeed , while weak at e@xmath810 kev ( because of the strong absorption ) , the x - ray spectral energy distribution of those elusive agn observed above 10 kev peaks at about 30 kev , due to an absorbing column density in the range @xmath101 , and therefore may provide an important contribution to the peak of the x - ray background at 30 kev . most of these elusive agn would not be detected below 10 kev in chandra and xmm surveys , except in the x - ray brightest cases .
such sources may be the so - called x - ray bright optically normal galaxies ( * ? ? ?
* xbongs , ) . yet ,
an assessment of the contribution of elusive agn to the 30 kev background bump requires a better knowledge of their luminosity distribution .
an xmm program aimed at studying a sample of candidate local elusive agn , recently approved for aot3 , will help to tackle these issues .
we are grateful to d. alexander , m. veron and p. veron for useful comments .
this work was partially supported by the italian space agency ( asi ) under grants i / r/073/02 and i / r/057/02 and by the italian ministry for university and research ( miur ) under grant cofin-00 - 2 - 35 . | a fraction of active galactic nuclei do not show the classical seyfert - type signatures in their optical spectra , i.e. they are optically `` elusive '' .
x - ray observations are an optimal tool to identify this class of objects .
we combine new chandra observations with archival x - ray data in order to obtain a first estimate of the fraction of elusive agn in local galaxies and to constrain their nature .
our results suggest that elusive agn have a local density comparable to or even higher than optically classified seyfert nuclei .
most elusive agn are heavily absorbed in the x - rays , with gas column densities exceeding @xmath0 , suggesting that their peculiar nature is associated with obscuration .
it is likely that in elusive agn , the nuclear uv source is completely embedded and the ionizing photons can not escape , which prevents the formation of a classical narrow line region .
elusive agn may contribute significantly to the 30 kev bump of the x - ray background .
[ firstpage ] galaxies : active galaxies : nuclei galaxies : seyfert x - rays : galaxies |
to understand how bits of information from external or locally computed signals can be specifically distributed through a network or to it s downstream components we first consider a generic stochastic dynamical system that evolves in time @xmath0 according to @xmath1 where @xmath2 denotes the variables of the network nodes , @xmath3 describes the intrinsic dynamics of the network , and @xmath4@xmath5 is a stochastic external input driving instantaneous state variable fluctuations which carry the information to be routed through the network .
we consider a deterministic reference state @xmath6 solving in the absence of signals ( @xmath7 ) . to quantify how bits of information surfing on top of such a dynamical state are routed through the network , we use information theoretic measures that quantify the amount of information shared and transferred between nodes , independent of how this information is encoded or decoded .
more precisely , we measure information sharing between signal @xmath8 and the time @xmath9 lagged signal @xmath10 of nodes @xmath11 and @xmath12 in the network via _ _ the time - delayed mutual information ( dmi ) _ _
@xcite @xmath13 here @xmath14 is the probability distribution of the variable @xmath8 of unit @xmath11 at time @xmath0 and @xmath15 the joint distribution of @xmath8 and the variable @xmath10 lagged by @xmath9 .
_ _ as a second measure we use the delayed _ _ _ _ _ _ transfer entropy ( dte ) _ _ @xcite ( cf .
methods ) that genuinely measures information transfer between pairs of units @xcite .
asymmetries in the dmi and dte curves @xmath16 and _ _ @xmath17 _ _ then indicate the dominant direction in which information is shared or transferred between nodes . to identify the role of the underlying reference dynamical state @xmath6 for network communication a small noise expansion in the signals @xmath18 turns a out to be ideally suited : while the small noise expansion limits the analysis to the vicinity of a specific reference state which is usually regarded as a weakness , in the context of our study , this property is highly advantageous as it directly conditions the calculations on a particular reference state and enables us to extract it s role for the emergent pattern of information routing within the network . for white noise sources
@xmath18 this method yields general expressions for the conditional probabilities @xmath19 that depend on @xmath6 . using this result the expressions for
the delayed mutual information and transfer entropy @xmath20 and @xmath21 become a function of the underlying collective reference dynamical state ( cf .
methods and supplementary section 1 ) . the dependency on this reference state then provides a generic mechanism to change communication in networks by manipulation the underlying collective dynamics . in the following we show how this general principle gives rise to a variety of mechanisms to flexibly change information routing in networks .
we focus on oscillatory phenomena widely observed in networks with a communication function @xcite .
oscillatory synchronization and phase locking @xcite provide a natural way for the temporal coordination between communicating units .
key variables in oscillator systems are the phases @xmath22 at time @xmath0 of the individual units @xmath11 .
in fact , a wide range of oscillating systems display similar phase dynamics @xcite ( cf .
supplementary section 2 ) and phase - based encoding schemes are common , e.g. in the brain @xcite , genetic circuits @xcite and artificial systems @xcite .
we first focus on systems in a stationary state with a stationary distribution for which the expressions for the dmi and dte become independent of the starting time @xmath0 and only depend on the lag @xmath9 and reference state @xmath23 . to assess the dominant direction of the shared information between two nodes we quantify asymmetries in the dmi curve by using the difference @xmath24 between the integrated mutual informations @xmath25 and @xmath26 .
if this is positive , information is shared predominantly from unit @xmath11 to @xmath12 while negative values indicate the opposite direction .
analogously , we compute the differences in dte as @xmath27(cf . methods and supplementary section 3 ) .
the set of pairs @xmath28 or @xmath29 for all @xmath11 , @xmath12 then capture strength and directionality of information routing in the network akin to a functional connectivity analysis in neuroscience @xcite .
we refer to them as information routing patterns ( irps ) . * * @ @ plus 1fil minus @skip * * populations ( wilson - cowan type dynamics ) . for the same network two different collective dynamical states accessed by different initial conditions give rise to two different information sharing patterns ( f h top vs. bottom ) . *
l , * as in a d but for generic oscillators close to a hopf bifurcation ( stuart - landau oscillators ) connected to a larger network . in i and
l connectivity matrices are shown instead of graphs .
two different network - wide information routing patterns arise ( top vs. bottom in j
l ) by changing a small number of connection weights ( purple entries in i and l ) . a range of networks of oscillatory units , with disparate physical interactions , connection topologies and external input signals support multiple irps .
for instance , in a model of a gene - regulatory network with two oscillatory sub - networks ( fig .
[ fig : example_networks]a ) dmi analysis reveals irps with different dominant directions ( fig . [ fig : example_networks]b - d , upper vs. lower sub - panels ) .
the change is triggered by adding an external factor that degrades the transcribed mrna in one of the oscillators and thereby changes its intrinsic frequency ( see methods ) .
more complex changes in irps emerge in larger networks , possibly with modular architecture . in a network of interacting neuronal populations ( fig .
[ fig : example_networks]e ) different initial conditions lead to different underlying collective dynamical states .
switching between them induces complicated but specific changes in the irps ( fig .
[ fig : example_networks]f - h ) .
different irps also emerge by changing a small number of connections in larger networks .
[ fig : example_networks]i - l illustrates this for a generic system of coupled oscillators each close to a hopf bifurcation . in general , several qualitatively different options for modifying network - wide
irps exist , all of which are relevant in natural and artificial systems : ( i ) changing the intrinsic properties of individual units ( fig . [
fig : example_networks]a - d , cf . also fig .
[ fig : control]a - c below ) , ( ii ) modifying the system connectivity ( fig .
[ fig : example_networks]i - l , fig .
[ fig : control]d - f ) and ( iii ) selecting distinct dynamical states of structurally the same system ( fig .
[ fig : example_networks]e - h , see also fig .
[ fig : combinatorics ] below ) .
to reveal how different irps arise and how they depend on the network properties and dynamics , we derive analytic expressions for the dmi and dte between all pairs of oscillators in a network .
we determine the phase of each oscillator @xmath11 in isolation by extending its phase description to the full basin of attraction of the stable limit cycle @xcite . for weak coupling ,
the effective phase evolution becomes @xmath30 where @xmath31 is the intrinsic oscillation frequencies of node @xmath11 and the coupling functions @xmath32 depend on the phase differences only .
the final sum in models external signals as independent gaussian white noise processes @xmath33 and a covariance matrix @xmath34 .
the precise forms of @xmath32 and @xmath34 generally depend on the specific system ( supplementary section 2 ) . as visible from fig .
[ fig : example_networks]e - h , the irp strongly depends on the underlying collective dynamical state .
we therefore decompose the dynamics into a deterministic reference part @xmath35 and a fluctuating component @xmath36 .
we focus on phase - locked configurations for the deterministic dynamics with constant phase offsets @xmath37 .
we estimate the stochastic part @xmath36 via a small noise expansion ( methods , supplementary theorem 1 ) yielding a first - order approximation for the joint probabilities @xmath38 . using together with the periodicity of the phase variables ,
we obtain the delayed mutual information _ @xmath39 _ between phase signals in coupled oscillatory networks ; here @xmath40 is the @xmath41 modified bessel function of the first kind , and @xmath42 is the inverse variance of a von mises distributions ansatz for @xmath43 .
the system s parameter dependencies , including different inputs , local unit dynamics , coupling functions and interaction topologies are contained in @xmath42 . by similar calculations we obtain analytical expressions for @xmath44 ( methods and supplementary theorem 2 ) .
our theoretical predictions well match the numerical estimates ( fig .
[ fig : example_networks]d , h , l , see also fig .
[ fig : mechanisms]c , d below and supplementary figs . 1 , 2 and 7 ) .
for independent input signals ( @xmath45 for @xmath46 ) we typically obtain similar irps determined either by the delayed mutual information or the transfer entropy ( supplementary fig . 2 ) .
* multi - stable dynamics and flexible anisotropic information routing . * _ _ * a , * two identical and symmetrically coupled neuronal circuits of wilson - cowan type ( dark and light green , modular sub - network in fg .
[ fig : example_networks]e ) .
* b , * phase difference @xmath47 between the extracted phases of the two neuronal populations fluctuating around a locked value @xmath48}$ ] of a stable deterministic collective state @xmath49 ( orange ) ; a strong external perturbation ( purple arrow ) induces a switch to stochastic dynamics around an alternate stable state @xmath50 ( brown ) with phase difference @xmath51}$ ] .
* c , * delayed mutual information @xmath52 and * d , * transfer entropy @xmath53 between the phase signals in state @xmath49 ( orange ) and @xmath50 ( brown ) for numerical data ( dots ) and theory ( lines ) .
the change in peak latencies form @xmath54}<0 $ ] to @xmath55}>0 $ ] in the @xmath52 and the asymmetry of the @xmath53 curves show anisotropic information routing .
switching between the two dynamical states reverses the information flow pattern ( graphs , bottom ) .
* e , * phase coupling function @xmath56 ( blue ) and its antisymmetric part @xmath57 ( red ) .
the two zeros of @xmath58 with negative slope indicate the deterministic equilibrium phase differences @xmath48}$ ] and @xmath51}$ ] in states @xmath49 and @xmath50 , receptively .
the directionality in the information routing pattern arises due to the different slopes of @xmath59 ( dashed lines ) at the noiseless phase - locking offsets @xmath48}$ ] and @xmath51}$ ] . ]
to better understand how a collective state gives rise to a specific routing pattern with directed information sharing and transfer , consider a network of two symmetrically coupled identical neural population models ( fig .
[ fig : mechanisms]a ) . due to permutation symmetry , the coupling functions @xmath60 obtained from the phase - reduction of the original wilson - cowan - type equations @xcite ( methods , supplementary section 5 ) , are identical . for biologically plausible parameters
this network in the noiseless - limit has two stable phase - locked reference states ( @xmath49 and @xmath50 ) .
the fixed phase differences @xmath61}$ ] and @xmath62}$ ] are determined by the zeros of the anti - symmetric coupling @xmath63 with negative slope ( fig .
[ fig : mechanisms]e ) . for a given level of ( sufficiently weak ) noise
, the system shows fluctuations around either one of these states ( fig .
[ fig : mechanisms]b ) each giving rise to a different irp .
sufficiently strong external signals can trigger state switching and thereby effectively invert the dominant communication direction visible from the @xmath64 ( fig .
[ fig : mechanisms]c ) and even more pronounced from the @xmath65 ( fig .
[ fig : mechanisms]d ) without changing any structural properties of the network .
the anisotropy in information transfer in the fully symmetric network is due to symmetry broken dynamical states . for independent noise inputs , @xmath66 , that are moreover small , the evolution of @xmath36 , @xmath67 , near the reference state @xmath49 reduces to @xmath68}\left(\phi_{i}^{\left(\mathrm{fluct}\right)}-\phi_{j}^{\left(\mathrm{fluct}\right)}\right)+\varsigma_{i}\xi_{i}\label{eq : firstnoisexpk2}\ ] ] with coupling constants @xmath69}=\gamma'(\delta\phi_{12}^{\left[\alpha\right]})$ ] , @xmath70}}=-\gamma'(2\pi-\delta\phi_{12}^{\left[\alpha\right]})$ ] ( methods ) . as @xmath71}\approx0 $ ] ( fig .
[ fig : mechanisms]e ) , the phase @xmath72 essentially freely fluctuates driven by the noise input @xmath73 .
this causes the system to deviate from the equilibrium phase difference @xmath74}$ ] . at the same time , the strongly negative coupling @xmath69}$ ] dominates over the noise term @xmath75 and unit @xmath76 is driven to restore the phase - difference by reducing @xmath77 .
thus , @xmath78 is effectively enslaved to track @xmath72 and information is routed from unit @xmath79 to unit @xmath76 , reflected in the @xmath64 and @xmath65 curves .
the same mechanism accounts for the reversed anisotropy in communication when the system is near state @xmath50 as the roles of unit @xmath76 and @xmath79 are exchanged . calculating the peak of the @xmath64 curve in this example
also provides a time scale @xmath54}\approx-\log\left(2\right)/g_{1}^{\left[\alpha\right]}$ ] at which maximal information sharing is observed ( methods , eq . ) .
it furthermore becomes clear that the directionality of the information transfer in general need not be related to the order in which the oscillators phase - lock because the phase - advanced oscillator can either effectively pull the lagging one , or , as in this example , the lagging oscillator can push the leading one to restore the equilibrium phase - difference . in summary ,
effective interactions local in state space and controlled by the underlying reference state together with the noise characteristics determine the irps of the network .
symmetry broken dynamical states then induce anisotropic and switchable routing patterns without the need to change the physical network structure .
for networks with modular interaction topology @xcite , our theory relating topology , collective dynamics and irps between individual units can be generalized to predict routing between entire modules . assuming that each sub - network @xmath80 in the noise - less limit has a stable phase - locked reference state , a second phase reduction @xcite generalized to stochastic dynamics characterizes each module by a single meta - oscillator with collective phase @xmath81 and frequency @xmath82 , driven by effective noise sources @xmath83 with covariances @xmath84 .
the collective phase dynamics of a network with @xmath85 modules then satisfies @xmath86 where @xmath87 are the effective inter - community couplings ( supplementary section 4 ) .
the structure of equation is formally identical to equation so that the expressions for inter - node information routing ( @xmath88 , @xmath89 ) can be lifted to expressions on the inter - community level ( @xmath90 , @xmath91 by replacing node- with community - related quantities ( i.e. @xmath31 with @xmath82 or @xmath92 with @xmath93 , etc . ,
supplementary corollary 3 and 4 ) .
importantly , this process can be further iterated to networks of networks , etc .
[ fig : control ] shows examples of information flow patterns resolved at two scales .
the information routing direction on the larger scale reflects the majority and relative strengths of irps on the finer scale .
@ @ plus 1fil minus @skip the collective quantities in the system are intricate functions of the network properties at the lower scales .
intriguingly , the coupling functions @xmath87 not only depend on the non - local interactions @xmath94 between units @xmath95 of module @xmath80 and @xmath96 of cluster @xmath97 but also on purely local properties of the individual clusters .
in particular , the form of @xmath87 is a function of the intrinsic local dynamical states @xmath98 and @xmath99 of both clusters as well as the phase response @xmath100 of sub - network @xmath80 ( see methods and supplementary section 4 ) .
thus irps on the entire network level depend on local community properties .
this establishes several generic mechanisms to globally change information routing in networks via local changes in modular properties , local connectivity , or via switching of local dynamical states . in a network consisting of two sub - networks ( fig .
[ fig : control]a ) the local change of the frequency of a single hopf - oscillator in sub - network @xmath101 induces a non - local inversion of the information routing between cluster @xmath101 and @xmath102 ( fig .
[ fig : control]b - d ) . in fig .
[ fig : control]e - f the direction in which information is routed between two sub - networks * @xmath102 * and @xmath103 of coupled phase oscillators is remotely changed by increasing the strength of a local link in module @xmath101 . the origin in both examples is a non - trivial combination of several factors : the ( small ) manipulations alter the collective cluster frequency @xmath104 , the local dynamical state @xmath105 which in turn change the collective phase response @xmath106 and the effective noise strength @xmath107 of cluster @xmath101 ( supplementary fig . 3 ) .
these changes all contribute to changes in the effective couplings @xmath87 as well as in the inter - cluster phase - locking values @xmath108 . taken together this causes the observed inversions in information routing direction .
interestingly , the transition in information routing has a switch like dependency on the changed parameter ( fig .
[ fig : control]c , d , g ) promoting digital - like changes of communication modes .
* * switching between combinatorially many information routing patterns . * _ _ * a , * modular circuit as in fig . [
fig : example_networks]e . without inter - module coupling , each of the @xmath109 communities @xmath110 exhibits multi - stability between two phase - locked configurations , denoted as states @xmath111 and @xmath112 ( insets ) .
* b , * information routing patterns between the hierarchically reduced sub - networks for different combinations of the local dynamical states @xmath113 $ ] , @xmath114 that give rise to globally phase - locked dynamics . arrows between nodes @xmath80 and @xmath97 indicate strength ( line width ) and sign ( arrow direction ) of the difference in integrated dte curves between the nodes .
the same local dynamical configuration ( @xmath113 $ ] ) can give rise to more than one globally locked collective state marked with dashes , i.e. @xmath113 $ ] , @xmath113'$ ] , @xmath115 .
* c , * the local dynamical state configuration @xmath116 $ ] generates a periodic global dynamical state ( cf . supplementary figure 8) in which the hierarchically reduced information routing pattern ( graphs ) becomes time - dependent ( cf .
also supplementary section 6).,title="fig : " ] * as an alternative to interventions on local properties , also switching between multi - stable local dynamical states @xmath98 can induce global information rerouting . in the example in fig .
[ fig : combinatorics ] , each of the @xmath109 modules @xmath110 exhibits @xmath117 alternative phase - locked states ( labeled @xmath111 and @xmath112 , supplementary section 5.1 ) . for sufficiently weak coupling , this local multi - stability
is preserved in the dynamics of the entire modular network .
consequently each choice of the @xmath118 possible combinations of `` local '' states gives rise to at least one network - wide collective state .
certain combinations of local states can give rise to one or even multiple globally phase - locked states ( e.g. @xmath119 $ ] in fig .
[ fig : combinatorics ] ) .
others support non - phase locked dynamics that gives rise to time - dependent irps ( cf .
[ fig : combinatorics]c and below ) .
thus , varying local dynamical states in a hierarchical network flexibly produces a combinatorial number @xmath120 of different irps in the same physical network .
general reference states , including periodic or transient dynamics , are not stationary and hence the expressions for the dmi and dte become dependent on time @xmath0 .
for example , fig .
[ fig : combinatorics]c shows irps that undergo cyclic changes due to an underlying periodic reference state ( cf . also supplementary figure 8a - c ) . in systems with a global fixed point
systematic displacements to different starting positions in state space give rise to different stochastic transients with different and time - dependent irps ( supplementary figure 8d ) .
similarly , switching dynamics along heteroclinic orbits constitute another way of generating specific progressions of reference dynamics .
thus information surfing on top of non - stationary reference dynamical configurations naturally yield temporally structured sequences of irps , resolvable also by other measures of instantaneous information flow , e.g. @xcite .
the above results establish a theoretical basis for the emergence of information routing capabilities in complex networks when signals are communicated on top of collective reference states .
we show how information sharing ( dmi ) and transfer ( dte ) emerge through the joint action of local unit features , global interaction topology and choice of the collective dynamical state .
we find that information routing patterns self - organize according to general principles ( cf .
[ fig : mechanisms ] , [ fig : control ] , [ fig : combinatorics ] ) and can thus be systematically manipulated . employing formal identity of our approach at every scale in oscillatory modular networks ( eq . vs. ) we identify local paradigms that are capable of regulating information routing at the non - local level across the whole network ( figs .
[ fig : control ] , [ fig : combinatorics ] ) .
we derived theoretical results based on information sharing and transfer obtained via delayed mutual information and transfer entropy curves . using these abstract measures our results are independent of any particular implementation of a communication protocol and
thus generically demonstrate how collective dynamics can have a functional role in information routing .
for example , in the network in fig .
[ fig : mechanisms ] externally injected streams of information are automatically encoded in fluctuations of the rotation frequency of the individual oscillators .
the injected signals are then transmitted through the network and decodable from the fluctuating phase velocity of a target unit precisely along those pathways predicted by the current state - dependent irp ( supplementary section 7 ) .
our theory is based on a small noise approximation that conditions the analysis onto a specific underlying dynamical state . in this way we extracted the precise role of such a reference state for the network
s information routing abilities . for larger signal amplitudes or in highly recurrent networks in which higher - order interactions can play an important role the expansion can be carried out systematically to higher orders using diagrammatic approaches @xcite or numerically to accounting for better accuracy and non - gaussian correlations ( cf .
also supplementary section 3.4 ) . in systems with multi - stable states two signal types need to be discriminated : those that encode the information to be routed and those that indicate a switch in the reference dynamics and consequently the irps .
if the second type of stimuli is amplified appropriately a switch between multi - stable states can be induced that moves the network into the appropriate irp state for the signals that follow .
for example , in the network of fig .
[ fig : mechanisms ] a switch from state @xmath49 to @xmath50 can be induced by a strong positive pulse to oscillator @xmath79 ( and vice versa ) .
if such pulses are part of the input a switch to the appropriate irp state will automatically be triggered and the network auto - regulates its irp function .
more generally a separate part of the network that effectively filters out relevant signatures indicating the need for a different irp could provide such pulses .
moreover , using the fact that local interventions are capable to switch irps in the network the outcomes of local computations can be used to trigger changes in the global information routing and thereby enable context - dependent processing in a self - organized way . when information surfs on top of dynamical reference states the control of irps is shifted towards controlling collective network dynamics making methods from control theory of dynamical systems available to the control of information routing .
for example , changing the interaction function in coupled oscillators systems @xcite or providing control signals to a subset of nodes @xcite are capable of manipulating the network dynamics .
moreover , switch like changes ( cf .
[ fig : control ] ) can be triggered by crossing bifurcation points and the control of information routing patterns then gets linked to bifurcation theory of network dynamical systems .
while the mathematical part of our analysis focused on phase signals , including additional amplitude degrees of freedom into the theoretical framework can help to explore neural or cell signaling codes that simultaneously use activity- and phase - based representations to convey information @xcite .
moreover , separating irp generation , e.g. via phase configurations , from actual information transfer , for instance in amplitude degrees of freedom , might be useful for the design of systems with a flexible communication function . the predicted phenomena , including non - local changes of information routing by local interventions , could be directly experimentally verified using methods available to date , such as electrochemical arrays @xcite or synthetic gene regulatory networks @xcite ( supplementary section 5.3 ) .
in addition our results are applicable to the inverse problem : unknown network characteristics may be inferred by fitting theoretical expected dmi and dte patterns to experimentally observed data . for example , inferring state - dependent coupling strengths could further the analysis of neuronal dynamics during context - dependent processing @xcite .
modifying inputs , initial conditions or system - intrinsic properties may well be viable in many biological and artificial systems whose function requires particular information routing . for instance , on long time scales , evolutionary pressure may select a particular information routing pattern by biasing a particular collective state in gene regulatory and cell signaling networks @xcite ; on intermediate time scales , local changes in neuronal responses due to adaptation or varying synaptic coupling strength during learning processes * * @xcite can impact information routing paths in entire neuronal circuits ; on fast time scales , defined control inputs to biological networks or engineered communication systems that switch the underlying collective state , can dynamically modulate information routing patterns without any physical change to the network .
* methods : * _ transfer entropy . _ the delayed transfer entropy ( dte ) _ _ @xcite from a time - series @xmath121 to a time - series @xmath122 is defined as _
_ @xmath123 with joint probability @xmath124 .
this expression is not invariant under permutation of @xmath11 and @xmath12 , implying the directionality of te . for a more direct comparison with dmi in figure 2
, we define @xmath125 by @xmath17 for @xmath126 and by @xmath127 for @xmath128 . _ _ dynamic information routing via dynamical states . _ _ for a dynamical system the reference deterministic solution @xmath129 starting at @xmath130 is given by the deterministic flow @xmath131 .
the small noise approximation for white noise @xmath18 then yields @xmath132 where @xmath133 denotes the normal distribution with mean @xmath134 and covariance matrix @xmath135 , @xmath136 and @xmath137 . from this and the initial distribution @xmath138
the delayed mutual information and transfer entropy @xmath139 and @xmath140 are obtained via and .
the result depends on time @xmath0 , lag@xmath9 and the reference state @xmath141 .
_ _ oscillator networks . _ _ in fig . 1a
, we consider a network of two coupled biochemical goodwin oscillators @xcite .
oscillations in the expression levels of the molecular products arise due to a nonlinear repressive feedback loop in successive transcription , translation and catalytic reactions .
the oscillators are coupled via mutual repression of the translation process @xcite .
in addition , in one oscillator changes in concentration of an external enzyme regulate the speed of degradation of mrnas , thus affecting the translation reaction , and , ultimately , the oscillation frequency . in fig .
[ fig : example_networks]e , [ fig : mechanisms ] , [ fig : combinatorics ] we consider networks of wilson - cowan type neural masses ( population signals ) @xcite .
each neural mass intrinsically oscillates due to antagonistic interactions between local excitatory and inhibitory populations .
different neural masses interact , within and between communities , via excitatory synapses . in the generic networks in fig .
1i and fig .
3a each unit is modeled by the normal form of a hopf - bifurcation in the oscillatory regime together with linear coupling .
finally , the modular networks analyzed in figures 3a and 3b are directly cast as phase - reduced models with freely chosen coupling functions .
see the supplementary information for additional details , model equations and parameters and phase estimation .
_ _ _ _ analytic derivation of the @xmath64 and @xmath65 curves . _ _ in the small noise expansion @xcite , both @xmath64 and @xmath65 curves have an analytic approximation : for stochastic fluctuations around some phase - locked collective state with constant reference phase offsets @xmath142 the phases evolve as @xmath143 in the deterministic limit , where @xmath144 is the collective network frequency and the @xmath32 are the coupling functions from eq . .
in presence of noise , the phase dynamics have stochastic components @xmath145 . in first order approximation , independent noise inputs @xmath146 yield coupled ornstein - uhlenbeck processes @xmath147 with linearized , state - dependent couplings given by the laplacian matrix entries @xmath148 and @xmath149 . the analytic solution to the stochastic equations provides an estimate of the probability distributions , @xmath150 , @xmath151 and @xmath152 . via this results in a prediction for @xmath16 , eq .
, as a function of the matrix elements @xmath42 specifying the inverse variance of a von mises distribution ansatz for @xmath151 .
similarly via an expression for @xmath17 is obtained . for the dependency of @xmath42 , and @xmath17 on network parameters and further details ,
see the derivation of the theorems 1 and 2 in the supplementary information .
_ _ time scale for information sharing .
_ _ for a network of two oscillators as in fig . with linearized coupling strengths @xmath69}$ ] and @xmath71}$ ] and @xmath69}<g_{2}^{\left[\alpha\right]}$ ] , maximizing @xmath153 ( see supplementary information for full analytic expressions of @xmath64 and @xmath65 in two oscillator networks ) yields _ collective phase reduction .
_ suppose that each node @xmath155 belongs to a specific network module @xmath80 out of @xmath156 non - overlapping modules of a network .
then equation can be simplified to under the assumption that in the absence of noise every community @xmath80 has a stable internally phase - locked state @xmath157 , where @xmath158 are constant phase offsets of individual nodes @xmath95 .
every community can then be regarded as a single meta - oscillator with a collective phase @xmath159 and a collective frequency @xmath160 .
the vector components of the collective phase response @xmath100 , the effective couplings @xmath161 and the noise parameters @xmath162 and @xmath83 are obtained through collective phase reduction and depend on the respective quantities ( @xmath163 ) on the single - unit scale ( see supplementary section 4 for a full derivation ) .
+ * acknowledgements : * we thank t. geisel for valuable discussions . partially supported by the federal ministry for education and research ( bmbf ) under grants no . 01gq1005b
[ ck , db , mt ] and 03sf0472e [ mt ] , by the nvidia corp . , santa clara , usa [ mt ] , a grant by the max planck society [ mt ] , by the fp7 marie curie career development fellowship ief 330792 ( dynvib ) [ db ] and an independent postdoctoral fellowship by the rockefeller university , new york , usa [ ck ] . +
* author contributions : * all authors designed research . c.k . derived the theoretical results , developed analysis tools and carried out the numerical experiments .
all authors analyzed and interpreted the results and wrote the manuscript .
+ * additional information : * the authors declare no competing financial interests .
supplementary information accompanies this paper .
correspondence and requests should be addressed to ck ( [email protected] ) . + | flexible information routing fundamentally underlies the function of many biological and artificial networks . yet ,
how such systems may specifically communicate and dynamically route information is not well understood . here
we identify a generic mechanism to route information on top of collective dynamical reference states in complex networks .
switching between collective dynamics induces flexible reorganization of information sharing and routing patterns , as quantified by delayed mutual information and transfer entropy measures between activities of a network s units .
we demonstrate the power of this generic mechanism specifically for oscillatory dynamics and analyze how individual unit properties , the network topology and external inputs coact to systematically organize information routing . for multi - scale , modular architectures
, we resolve routing patterns at all levels .
interestingly , local interventions within one sub - network may remotely determine non - local network - wide communication .
these results help understanding and designing information routing patterns across systems where collective dynamics co - occurs with a communication function .
attuned function of many biological or technological networks relies on the precise yet dynamic communication between their subsystems .
for instance , the behavior of cells depends on the coordinated information transfer within gene regulatory networks @xcite and flexible integration of information is conveyed by the activity of several neural populations during brain function @xcite .
identifying general mechanisms for the routing of information across complex networks thus constitutes a key theoretical challenge with applications across fields , from systems biology to the engineering of smart distributed technology @xcite .
complex systems with a communication function often show characteristic dynamics , such as oscillatory or synchronous collective dynamics with a stochastic component @xcite .
information is carried in the presence of these dynamics within and between neural circuits @xcite , living cells @xcite , ecologic or social groups @xcite as well as technical communication systems , such as ad hoc sensor networks @xcite . while such dynamics could simply reflect the properties of the interacting unit s , emergent collective dynamical states in biological networks
can actually contribute to the system s function .
for example , it has been hypothesized that the widely observed oscillatory phenomena in biological networks enable emergent and flexible information routing @xcite .
here we derive a theory that shows how information if conveyed by fluctuations around collective dynamical reference states ( e.g. a stable oscillatory pattern ) can be flexibly routed across complex network topologies .
quantifying information sharing and transfer by time - delayed mutual information @xcite and transfer entropy @xcite curves between time - series of the network s units , we demonstrate how switching between multi - stable states enables the rerouting of information without any physical changes to the network . in fully symmetric networks ,
anisotropic information transfer can arise via symmetry breaking of the reference dynamics . for networks of coupled oscillators
our approach gives analytic predictions how the physical coupling structure , the oscillators properties and the dynamical state of the network co - act to produce a specific communication pattern .
resorting to a collective - phase description @xcite , our theory further resolves communication patterns at all levels of multi - scale , modular topologies @xcite , as ubiquitous , e.g. , in the brain connectome and bio - chemical regulatory networks @xcite .
we thereby uncover how local interventions within one module may remotely modify information sharing and transfer between other distant sub - networks .
a combinatorial number of information routing patterns in networks emerge due to switching between multi - stable dynamical states that are localized on individual sub - sets of network nodes .
these results offer a generic mechanism for self - organized and flexible information routing in complex networked systems . for oscillatory dynamics the links made between multi - scale connectivity , collective network dynamics and
flexible information routing has potential applications to the reconstruction and design of gene regulatory circuits @xcite , wireless communication networks @xcite or to the analysis of cognitive functions @xcite , among others . |
the discovery of galaxies with strong [ ] @xmath145007 emission , extreme star formation rates ( sfrs ) , low masses , and low reddening by @xcite was key to finding objects thought to be responsible for the reionization of the universe @xcite .
known as ` green peas ' for their strong emission in the rest - frame optical and compact sizes , similar galaxies have since been found at much higher redshifts .
these green peas seem in many aspects to be more extreme versions of blue compact dwarfs @xcite , having low masses and strong nebular emission lines , albeit with distinctly higher specific star formation rates ( ssfrs ; e.g. , * ? ? ?
these objects exhibit bright optical nebular emission lines such as h@xmath12 and [ ] , with rest - frame equivalent widths of several hundred to over one thousand angstroms , indicating bursty star forming activity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* hereafter s16 ) .
such galaxies are increasingly common at higher redshifts , appear to have enhanced [ ] @xmath145007 relative to both h@xmath2 and [ oii]@xmath143727 , and are often lyman-@xmath12 emitters ( laes ) ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . while the strong emission lines of these objects can be detected with spectroscopy at @xmath15 ( *
* ; * ? ? ?
* s16 ) , the rest - frame optical stellar continuum is quite faint ( @xmath16 ) . indeed for such low - mass objects , this continuum has until recently been too faint for spectroscopy of large samples at these redshifts ( e.g. , * ? ? ?
however , we can use the deep multi - wavelength photometry available in zfourge to construct composite seds @xcite and analyze the faint stellar continuum of a sample of emission - line objects at @xmath0 .
we will refer to the most intense of these compact , [ ] -emitting galaxies , as extreme emission line galaxies ( eelgs ) , and galaxies with less intense , but still significant [ ] emission as strong emission line galaxies ( selgs ) for the duration of the letter .
we assume a @xmath17cdm cosmology of @xmath18 , @xmath19 , and @xmath20 km s@xmath21 mpc@xmath21 .
for such a universe , 1 " = 7.320 kpc at @xmath22 .
we use a chabrier initial mass function @xcite and the ab magnitude system @xcite .
we use data from the medium band near - infrared fourstar galaxy evolution survey ( zfourge ; * ? ? ?
this survey combines imaging from a large number of previous surveys to create the deepest @xmath1-band detection image ( see @xcite for details ) of the chandra deep field south ( cdfs * ? ? ? * ) ) , with a @xmath23 limiting depth of 26.5 ab mag .
imaging in the cosmos @xcite and uds @xcite fields are also quite deep , at 25.5 and 25.7 magnitudes , respectively .
in addition , due to its medium band filters bracketing well - known features , such as the balmer break , [ ] emission , and h@xmath12 emission , zfourge provides precise photometric redshifts .
extensive testing against spectroscopic redshifts yielded @xmath24 @xcite .
the photometric redshift errors for objects without spectroscopic redshifts were also estimated from the photometric redshift differences of close pairs @xcite . for blue star - forming galaxies ,
this method yields @xmath25 at @xmath26 , which agrees with the photo-@xmath27 determinations from eazy @xcite .
forrest , et al .
, in prep , build a set of composite seds from non - agn zfourge galaxies in @xmath0 , with @xmath28 .
agn are removed using the catalogs from @xcite , which account for infrared , x - ray , and radio selected sources .
the composite sed method uses de - redshifted photometry to group galaxies together by sed shape , then scales the photometry of similar galaxies , termed analogs , to trace the intrinsic sed with better spectral resolution than is possible with the photometry of a single galaxy .
medians of these de - redshifted photometric points in wavelength bins are then taken to be the composite sed .
each point has a custom composite sed filter constructed from a weighted combination of the underlying photometry , which is used in fitting programs such as fast @xcite .
this method was introduced by @xcite and details are also provided in @xcite .
the bluest composite seds show significant emission from [ ] + h@xmath2 , corresponding to a sample of eelgs ( 60 galaxies ) , and the other to a population of selgs ( 218 galaxies ) .
the emission lines fall into the @xmath1-band in the redshift range @xmath3 , and this excess flux in the @xmath1-band increases our confidence in the photometric redshifts , which are in strong agreement with the handful of spectroscopic redshifts available for this population .
spectroscopic redshifts for a subset of galaxies in our sample are found by cross - matching with several works including eso / goods spectroscopy from _
fors2 _ @xcite and _ vimos _
@xcite , and the zfire survey @xcite .
these samples total 37 selgs and 7 eelgs with either grism or spectroscopic redshifts .
there are 17 emission line galaxies with _ hst / wfc3/g141 _ grism redshifts from _ 3d - hst _ , with median @xmath29 , and 27 with rest - frame ultraviolet spectroscopic redshifts , with median @xmath30 .
@xmath145007 + h@xmath2 , in the @xmath1-band at @xmath5 .
their de - redshifted and scaled analog photometry is shown in gray , while the median points which make up the composite seds are in purple .
the best - fit model from fast is green ( emission lines are scaled down for clarity ; see section [ fitef ] ) , and the characteristic width of the composite filters is shown at the bottom .
the strong emission in [ ] + h@xmath2 is apparent and the eelg composite also shows ly@xmath12 emission.,scaledwidth=55.0% ] we make use of fast to fit the stellar continuum of the composite seds using the @xcite ( bc03 ) models to obtain equivalent widths for optical emission lines in section [ ew ] .
accounting for emission lines as strong as those observed in the eelgs is of critical importance during this process .
failure to do so will result in fast overestimating the stellar continuum , leading to errors in parameters such as stellar mass , sfr , and age ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* hereafter s15 ; spitler , et al . , in prep ) .
we test three methods for fitting both the composite seds and the individual galaxies using the bc03 models .
declining-@xmath31 models for all three methods have @xmath32 , ages @xmath33 , and dust attenuations @xmath34a@xmath35/@xmath36 .
first , we fit the bc03 models to the set of photometry ( method a ) .
the strength of the emission feature artificially raises the level of the stellar continuum , and thus the mass , for this method .
secondly , we use the same models , but mask out points affected by [ ] + h@xmath2 emission ( method b ) .
the third method uses a set of high resolution ( @xmath37 ) bc03 models with emission lines added fit to the entire composite sed ( method c ) .
these emission line models , as detailed in section 3 of s15 , include 119 sets of relative emission - line strengths with nebular emission taken into account .
while the nebular continuum is not considered in this method , effects from this flux are minimal in the rest - frame optical , and so can be safely ignored ( s15 ) .
we also run sets of models with several metallicity values .
method c yields the best fit to the composite sed , with @xmath38 for the eelg composite , compared to @xmath39 using method a , which is used in the zfourge catalogs .
the best fits also have a sub - solar metallicity , @xmath40 , in agreement with studies suggesting low metallicities for these galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* s15 ; s16 ) . the resulting mean values and
one sigma spreads for the individual elgs are shown in table [ tab : table1 ] .
using method c described above , we find that the eelg population in general is younger than the selg population by 0.3 dex , has 0.5 dex less mass , and has more intense star formation ( ssfr of 55 gyr@xmath21 to 15 gyr@xmath21 ) .
this is consistent with a picture of the selg population being on the same evolutionary path , but slightly more evolved .
we find mean physical sizes of 1.34 kpc and 1.63 kpc for the eelg and selg populations by cross - matching with the @xmath41 sizes from candels ( * ? ? ?
* @xmath42 at @xmath5 ) .
however , we note that at @xmath5 the angular resolution of @xmath43 kpc , and so a number of these galaxies are unresolved .
nonetheless , the small sizes are consistent with measurements in both the local and distant universe .
@xcite find a nuv petrosian radius of @xmath44 kpc at @xmath45 ( @xmath46 ) .
the @xmath47 sample from @xcite has sizes ranging from @xmath48 , also based on data from @xcite ( @xmath49@xmath50 ) .
these data suggest that eelg populations across cosmic time have the same physical sizes at rest - frame uv wavelengths .
additionally , these sizes are consistent with the size - mass relation for @xmath5 derived in @xcite , which when extrapolated to @xmath51 predicts a size of 1.2 kpc .
one of the objects in our eelg sample is the subject of s16 , which analyzes a rest - frame optical spectrum from mosfire .
s16 finds galaxy properties in agreement with those derived in this work , namely a low mass ( @xmath52 ) , high ssfr ( 23 gyr@xmath21 ) , young age ( 160 myr ) , and low dust attenuation ( @xmath53 ) .
critically , they also find a metallicity of 12+(o / h ) = 8.00 , which confirms the low metallicities characteristic of galaxies undergoing early star formation @xcite .
[ cols="<,^,^,^,^,^,^ " , ] while the bc03+emission line models produce good fits to the composite seds ( see figure [ fig : comp ] ) , they still underpredict the strength of the [ ] + h@xmath2 emission feature . to estimate the equivalent width ,
we remove the emission features from the model and add back in emission lines to match the observed composite sed feature .
we take the continuum to be the level of the best - fit spectrum from fast using method c where there are no emission lines , and use a simple linear interpolation to derive the continuum in areas with such features .
since these models have resolution of 1 in the optical , all absorption features ( i.e. , h@xmath2 absorption ) are retained . after obtaining the continuum fit to the composite
sed , we add to it an emission line model with the following flux ratios : @xmath54\lambda5007 } = 3 f_{[oiii]\lambda4959 } = 7 f_{h\beta}$ ] ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the resultant spectrum is convolved with the custom composite sed filters to create synthetic photometry .
the amplitude of the emission line model is varied to minimize the least squares of the composite sed and the synthetic photometry .
we do not have the resolution to separate the effects of the lines but ratios of [ ] @xmath145007 to h@xmath2 from 1 to 20 were tested with negligible change to the resulting equivalent width .
errors on the equivalent width were calculated by fitting a range of line strengths to form a grid of @xmath55 values and one sigma errors were calculated using the relation @xmath56 . for the eelg composite sed ,
we obtain a rest - frame equivalent width @xmath57 + h\beta } = 803\pm228$] , corresponding to an observed - frame equivalent width @xmath57 + h\beta } = 3565$] at the median redshift @xmath58 ( see figure [ fig : ew ] ) . the rest - frame equivalent width for the eelg population
is consistent with those of eelgs in many studies which range from @xmath59 ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the selg population has a rest - frame equivalent width @xmath57+h\beta } = 230\pm90$] .
spectroscopic follow - up of these objects would be useful for not only confirming the accuracy of this method , but also analyzing the spread of equivalent width values and [ ] /h@xmath2 line ratios within the population . to quantify uncertainties in the equivalent widths that are due to uncertainties in the photometric redshifts , we redshift a high equivalent width template and calculate synthetic @xmath1-band photometric values .
we then mimic our composite sed method by de - redshifting the template using the ` correct ' redshift with noise added according to the @xmath60 values from spectroscopy ( see section [ phot ] ) . repeating this process 100 times , we measure the equivalent width to be @xmath61 , which matches very well with our measurement of @xmath6 (table 1 , method c ) .
we also note that both composite seds show emission in the ultraviolet ( uv ) , consistent with many of the analogs having ly@xmath12 emission .
several studies have suggested that laes are similar to the eelgs ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; our sample is consistent with that picture .
[ ] eelgs are less common locally than at higher redshifts ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the redshift range of @xmath0
we detect 60 eelgs with @xmath28 , with the faintest at @xmath62 magnitudes , and a characteristic equivalent width of @xmath63 . separating the eelgs by field we find @xmath64 mpc@xmath65 , @xmath66 mpc@xmath65 , and @xmath67 mpc@xmath65 for cdfs , cosmos , and uds respectively .
selgs have densities of @xmath68 mpc@xmath65 , @xmath69 mpc@xmath65 , and @xmath70 mpc@xmath65 in the same order , and are 3.7 times more common than the eelgs in total . in the redshift range @xmath0 , these eelgs and selgs
dominate the population of the bluest galaxies with @xmath71 .
we select all zfourge galaxies with @xmath72 , @xmath73 and at least 5 photometric detections in the restframe wavelength range @xmath74 , and fit a power law , @xmath75 to obtain the uv slope , @xmath2 @xcite .
of the 58 such galaxies with @xmath71 , 45 ( 78% ) fall into our eelg or selg groups , indicating that the majority of the bluest galaxies have strong emission features .
the eelgs and selgs show an unexpected peak at @xmath5 in cdfs .
the full redshift distribution of galaxies in the zfourge catalog , both photometric and spectroscopic , also confirms the presence of a galaxy overdensity at @xmath5 in cdfs ( see figure [ fig : hist ] , as well as figure 23 of @xcite ) . comparing the cdfs redshift distribution to the combined distribution from cosmos and uds shows a strong difference , greater than @xmath77 as found using a k - s test .
such an overdensity has been suggested by data from _ 3d - hst _
* figure 24 ) , who use eazy as well .
the same is not found in cdf - north , which has similar filter coverage and @xmath1-band depth , and selects based on @xmath78 @xmath79 imaging ( their tables 6 and 7 ) .
this removes the possibility of the overdensity being due to a bias in eazy . in @xmath80 .
, scaledwidth=50.0% ] to map possible overdensities projected on the sky , we use the @xmath81 nearest - neighbor measure ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . for each field
, we create a grid of 1000 points on a side , and calculate the distance to the @xmath81 closest galaxy at each point , which is converted into a density .
each point in all three fields is shaded by the number of standard deviations above the mean density in the respective field ( see figure [ fig : density ] ) .
additional density calculations were made using the @xmath82 , and @xmath83 nearest neighbors - the results are similar to those from the @xmath81 nearest neighbor measure . in the top row of figure
[ fig : density ] , we show all three fields considering galaxies in the range @xmath84 , with a magnitude cut of @xmath85 mag , where all three fields are complete @xcite . the 5 sigma peak in the uds contains 10 galaxies while the 5 sigma peaks in cosmos contain @xmath86 galaxies .
in contrast , the largest overdensity in cdfs contains 53 galaxies .
slightly increasing or decreasing the redshift ranges does not change our results .
we focus on the most significant overdensity at z 3.5 which is in cdfs but note that we do not exclude an overdensity in cosmos . at z=3.5 , the galaxy overdensity in cdfs at ( ra , dec )
= ( 53.08 , -27.85 ) is approximately 1.8 mpc in size .
spectroscopy by @xcite in cdfs confirm five galaxies at z=3.474 ( figure [ fig : density ] , green circles ) .
@xcite also report a massive quiescent galaxy at z=3.56 ( figure [ fig : density ] , blue triangle ) .
previous studies find that blue galaxies such as laes and lbgs trace large scale structure ( e.g. , * ? ? ?
* ; * ? ? ?
our results confirm that galaxies with strong [ ] + h@xmath2 emission also exist in such overdensities .
eelgs have the added advantage that they can be spectroscopically confirmed more easily via their strong emission .
a possible bias in our result is that the nebular emission at @xmath5 is boosting the flux in the @xmath1-band such that we are preferentially detecting lower mass galaxies at this redshift .
however , the redshift and projected distributions of galaxies without these strong emission features show evidence of an overdensity as well .
we utilized multi - wavelength photometry from the zfourge survey to build composite seds based on similar galaxy redshift and sed shape , revealing a population of galaxies with very blue colors and excess emission in the @xmath1-band at @xmath5 .
parameters derived from fast characterize these galaxies as having very young ages ( @xmath87 ) , low masses ( @xmath88 ) , and low dust content ( a@xmath35 @xmath89 ) .
in addition , they are small in size ( @xmath90 kpc ) , and their remarkable emission from [ ] + h@xmath2 , @xmath91 for the strongest emitters , is consistent with the properties of extreme star - forming galaxies that may have reionized the universe .
we observe that these eelgs appear on the order of @xmath92 mpc@xmath65 at @xmath5 , and make up the vast majority ( 78% ) of the galaxies with @xmath71 at this epoch .
finally , we explored the distribution of these galaxies on the sky , and find an overdensity in the overall population of cdfs at @xmath5 .
this giant structure is @xmath93 mpc in projected size and is a candidate progenitor of a galaxy cluster environment .
it also includes a massive quiescent galaxy from @xcite and rest - frame uv spectroscopically confirmed lbgs @xcite in a very narrow redshift range .
further spectroscopic follow - up of these galaxies is critical for understanding the earliest star - forming galaxies , and will also lead to interesting science cases for the next generation of telescopes , including jwst .
such telescopes would be capable of detecting h@xmath12 emission from these galaxies , which currently falls between the @xmath94-band and the _ spitzer _ 3.6@xmath95 band , as well as looking for large scale structure based on optical emission lines .
we wish to thank the mitchell family , particularly the late george p. mitchell , for their continuing support of astronomy .
we also thank the carnegie observatories and the las campanas observatory for their assistance in making the zfourge survey possible .
b. forrest and k. tran acknowledge the support of the national science foundation under grant # 1410728 .
ggk acknowledges the support of the australian research council through the award of a future fellowship ( ft140100933 ) .
finally , we thank the anonymous referee for a series of comments which greatly improved the manuscript . | using deep multi - wavelength photometry of galaxies from zfourge , we group galaxies at @xmath0 by the shape of their spectral energy distributions ( seds ) .
we identify a population of galaxies with excess emission in the @xmath1-band , which corresponds to [ ] + h@xmath2 emission at @xmath3 .
this population includes 78% of the bluest galaxies with uv slopes steeper than @xmath4 .
we de - redshift and scale this photometry to build two composite seds , enabling us to measure equivalent widths of these extreme [ ] + h@xmath2 emission line galaxies ( eelgs ) at @xmath5 .
we identify 60 galaxies that comprise a composite sed with [ ] + h@xmath2 rest - frame equivalent width of @xmath6 and another 218 galaxies in a composite sed with equivalent width of @xmath7 .
these eelgs are analogous to the ` green peas ' found in the sdss , and are thought to be undergoing their first burst of star formation due to their blue colors ( @xmath8 ) , young ages ( @xmath9 ) , and low dust attenuation values .
their strong nebular emission lines and compact sizes ( typically @xmath10 kpc ) are consistent with the properties of the star - forming galaxies possibly responsible for reionizing the universe at @xmath11
. many of the eelgs also exhibit lyman-@xmath12 emission .
additionally , we find that many of these sources are clustered in an overdensity in the chandra deep field south , with five spectroscopically confirmed members at @xmath13 .
the spatial distribution and photometric redshifts of the zfourge population further confirm the overdensity highlighted by the eelgs . |
the two key characteristics of wireless communications that most greatly impact system design and performance are 1 ) the randomly - varying channel conditions and 2 ) limited energy resources . in wireless systems ,
the power of the received signal fluctuates randomly over time due to mobility , changing environment , and multipath fading .
these random changes in the received signal strength lead to variations in the instantaneous data rates that can be supported by the channel @xcite .
in addition , mobile wireless systems can only be equipped with limited energy resources , and hence energy efficient operation is a crucial requirement in most cases . to measure and compare the energy efficiencies of different systems and transmission schemes , one can choose as a metric the energy required to reliably send one bit of information .
information - theoretic studies show that energy - per - bit requirement is generally minimized , and hence the energy efficiency is maximized , if the system operates at low signal - to - noise ratio ( @xmath0 ) levels and hence in the low - power or wideband regimes .
recently , verd in @xcite has determined the minimum bit energy required for reliable communication over a general class of channels , and studied of the spectral efficiency
bit energy tradeoff in the wideband regime while also providing novel tools that are useful for analysis at low @xmath0s . in many wireless communication systems , in addition to energy - efficient operation , satisfying certain quality of service ( qos ) requirements is of paramount importance in providing acceptable performance and quality .
for instance , in voice over ip ( voip ) , interactive - video ( e.g , .
videoconferencing ) , and streaming - video applications in wireless systems , latency is a key qos metric and should not exceed certain levels @xcite . on the other hand , wireless channels , as described above ,
are characterized by random changes in the channel , and such volatile conditions present significant challenges in providing qos guarantees . in most cases , statistical , rather than deterministic ,
qos assurances can be given .
in summary , it is vital for an important class of wireless systems to operate efficiently while also satisfying qos requirements ( e.g. , latency , buffer violation probability )
. information theory provides the ultimate performance limits and identifies the most efficient use of resources .
however , information - theoretic studies and shannon capacity formulation generally do not address delay and quality of service ( qos ) constraints @xcite .
recently , wu and negi in @xcite defined the effective capacity as the maximum constant arrival rate that a given time - varying service process can support while providing statistical qos guarantees .
effective capacity formulation uses the large deviations theory and incorporates the statistical queueing constraints by capturing the rate of decay of the buffer occupancy probability for large queue lengths .
the analysis and application of effective capacity in various settings has attracted much interest recently ( see e.g. , @xcite@xcite and references therein ) . in this paper , we study the energy efficiency in the presence of queueing constraints and channel uncertainty .
we assume that the channel is not known by the transmitter and receiver prior to transmission , and is estimated imperfectly by the receiver through training . in our model
, we incorporate statistical queueing constraints by employing the effective capacity formulation which provides the maximum throughput under limitations on buffer violation probabilities for large buffer sizes .
since the transmitter is assumed to not know the channel , fixed - rate transmission is considered .
we consider a point - to - point wireless link in which there is one source and one destination .
it is assumed that the source generates data sequences which are divided into frames of duration @xmath1 .
these data frames are initially stored in the buffer before they are transmitted over the wireless channel .
the discrete - time channel input - output relation in the @xmath2 symbol duration is given by @xmath3 = h[i ] x[i ] + n[i ] \quad i = 1,2,\ldots.\end{gathered}\ ] ] where @xmath4 $ ] and @xmath5 $ ] denote the complex - valued channel input and output , respectively .
we assume that the bandwidth available in the system is @xmath6 and the channel input is subject to the following average energy constraint : @xmath7|^2\}\le { \bar{p}}/ b$ ] for all @xmath8 .
since the bandwidth is @xmath6 , symbol rate is assumed to be @xmath6 complex symbols per second , indicating that the average power of the system is constrained by @xmath9 . above in ( [ eq : model ] ) , @xmath10 $ ] is a zero - mean , circularly symmetric , complex gaussian random variable with variance @xmath11|^2\ } = n_0 $ ] .
the additive gaussian noise samples @xmath12\}$ ] are assumed to form an independent and identically distributed ( i.i.d . ) sequence .
finally , @xmath13 $ ] , which denotes the channel fading coefficient , is assumed to be a zero - mean gaussian random variable with variance @xmath14 .
we further assume that the fading coefficients stay constant during the frame duration of @xmath1 seconds and have independent realizations for each frame .
hence , we basically consider a block - fading channel model .
we also note that neither the transmitter not the receiver has channel side information prior to transmission .
the system operates in two phases : training phase and data transmission phase . in the training phase ,
known pilot symbols are transmitted to enable the receiver to estimate the channel conditions , albeit imperfectly .
we assume that minimum mean - square - error ( mmse ) estimation is employed at the receiver to estimate the channel coefficient @xmath13 $ ] .
since the mmse estimate depends only on the training energy and not on the training duration @xcite , it can be easily seen that transmission of a single pilot at every @xmath1 seconds is optimal .
note that in every frame duration of @xmath1 seconds , we have @xmath15 symbols and the overall available energy is @xmath16 .
we now assume that each frame consists of a pilot symbol and @xmath17 data symbols .
the energies of the pilot and data symbols are @xmath18 respectively , where @xmath19 is the fraction of total energy allocated to training .
note that the data symbol energy @xmath20 is obtained by uniformly allocating the remaining energy among the data symbols . in the training phase
, the receiver obtains the mmse estimate @xmath21 which is a circularly symmetric , complex , gaussian random variable with mean zero and variance @xmath22 , i.e. , @xmath23@xcite .
now , the channel fading coefficient @xmath24 can be expressed as @xmath25 where @xmath26 is the estimate error and @xmath27 .
consequently , in the data transmission phase , the channel input - output relation becomes @xmath28 = \hat{h}[i ] x[i ] + \tilde{h}[i ] x[i ] + n[i ] \quad i = 1,2,\ldots.\end{gathered}\ ] ] since finding the capacity of the channel in ( [ eq : impmodel ] ) is a difficult task - dependent , discrete distribution with a finite number of mass points . in such cases ,
no closed - form expression for the capacity exists , and capacity values need to be obtained through numerical computations . ] , a capacity lower bound is generally obtained by considering the estimate error @xmath26 as another source of gaussian noise and treating @xmath29 x[i ] + n[i]$ ] as gaussian distributed noise uncorrelated from the input .
now , the new noise variance is @xmath30 x[i ] + n[i]|^2\ } = \sigma_{\tilde{h}}^2 { \mathcal{e}}_s + n_0 $ ] where @xmath31 is the variance of the estimate error . under these assumptions , a lower bound on the instantaneous capacity
is given by @xcite , @xcite @xmath32 where effective @xmath0 is @xmath33 and @xmath34 is the variance of estimate @xmath21 .
note that the expression in ( [ eq : traincap2 ] ) is obtained by defining @xmath35 where @xmath36 is a standard complex gaussian random variable with zero mean and unit variance , i.e. , @xmath37 .
since gaussian is the worst uncorrelated noise @xcite , the above - mentioned assumptions lead to a pessimistic model and the rate expression in ( [ eq : traincap2 ] ) is a lower bound to the capacity of the true channel ( [ eq : impmodel ] ) . on the other hand
, @xmath38 is a good measure of the rates achieved in communication systems that operate as if the channel estimate were perfect ( i.e. , in systems where gaussian codebooks designed for known channels are used , and scaled nearest neighbor decoding is employed at the receiver ) @xcite .
henceforth , we base our analysis on @xmath38 to understand the impact of the imperfect channel estimate .
since the transmitter is unaware of the channel conditions , it is assumed that information is transmitted at a fixed rate of @xmath39 bits / s . when @xmath40 , the channel is considered to be in the on state and reliable communication
is achieved at this rate .
if , on the other hand , @xmath41 , we assume that outage occurs . in this case , channel is in the off state and reliable communication at the rate of @xmath39 bits / s can not be attained . hence
, effective data rate is zero and information has to be resent .
[ fig:00 ] depicts the two - state transmission model together with the transition probabilities . under the block
fading assumption , it can be easily seen that the transition probabilities are given by @xmath42 where @xmath43 and @xmath44 is an exponential random variable with mean @xmath45 , and hence , @xmath46 .
in @xcite , wu and negi defined the effective capacity as the maximum constant arrival rate that a given service process can support in order to guarantee a statistical qos requirement specified by the qos exponent @xmath47 .
if we define @xmath48 as the stationary queue length , then @xmath47 is the decay rate of the tail distribution of the queue length @xmath48 : @xmath49 therefore , for large @xmath50 , we have the following approximation for the buffer violation probability : @xmath51 .
hence , while larger @xmath47 corresponds to more strict qos constraints , smaller @xmath47 implies looser qos guarantees .
similarly , if @xmath52 denotes the steady - state delay experienced in the buffer , then @xmath53 for large @xmath54 , where @xmath55 is determined by the arrival and service processes @xcite .
the effective capacity is given by @xmath56}\}}\end{gathered}\ ] ] where @xmath57 = \sum_{i=1}^{t}r[i]$ ] is the time - accumulated service process and @xmath58 , i=1,2,\ldots\}$ ] denote the discrete - time stationary and ergodic stochastic service process .
note that in the model we consider , @xmath59 = rt \text { or } 0 $ ] depending on the channel state being on or off . in @xcite
, it is shown that for such an on - off model , we have @xmath60 note that @xmath61 in our model .
then , for a given qos delay constraint @xmath47 , the effective capacity normalized by the frame duration @xmath1 and bandwidth @xmath6 , or equivalently spectral efficiency in bits / s / hz , becomes @xmath62 @xmath63 note that @xmath64 is obtained by optimizing both the fixed transmission rate @xmath39 and the fraction of power allocated to training , @xmath19 . in the optimization result ( [ eq : trainopti2 ] ) , @xmath65 and @xmath66 are the optimal values of @xmath39 and @xmath67 , respectively .
it can easily be seen that @xmath68 hence , as the qos requirements relax , the maximum constant arrival rate approaches the average transmission rate . on the other hand , for @xmath69 , @xmath70 in order to avoid violations of buffer constraints . in this paper , we focus on the energy efficiency of wireless transmissions under the aforementioned statistical queueing constraints .
since energy efficient operation generally requires operation at low-@xmath0 levels , our analysis throughout the paper is carried out in the low-@xmath0 regime . in this regime ,
the tradeoff between the normalized effective capacity ( i.e , spectral efficiency ) @xmath64 and bit energy @xmath71 is a key tradeoff in understanding the energy efficiency , and is characterized by the bit energy at zero spectral efficiency and wideband slope provided , respectively , by @xmath72 where @xmath73 and @xmath74 are the first and second derivatives with respect to @xmath0 , respectively , of the function @xmath75 at zero @xmath0 @xcite . @xmath76 and @xmath77 provide a linear approximation of the spectral efficiency curve at low spectral efficiencies .
in this section , we investigate the optimization problem in ( [ eq : trainopti ] ) .
in particular , we identify the optimal fraction of power that needs to be allocated to training while satisfying statistical buffer constraints .
[ theo : optrho ] at a given @xmath0 level , the optimal fraction of power @xmath78 that solves ( [ eq : trainopti ] ) does not depend on the qos exponent @xmath47 and the transmission rate @xmath39 , and is given by @xmath79 where @xmath80 and @xmath81 .
_ proof : _ from ( [ eq : trainopti ] ) and the definition of @xmath67 in ( [ eq : trainthresh ] ) , we can easily see that for fixed @xmath39 , the only term in ( [ eq : trainopti ] ) that depends on @xmath19 is @xmath67 .
moreover , @xmath67 has this dependency through @xmath82 .
therefore , @xmath78 that maximizes the objective function in ( [ eq : trainopti ] ) can be found by minimizing @xmath67 , or equivalently maximizing @xmath82 . substituting the definitions in ( [ eq : trainpower ] ) and the expressions for @xmath83 and @xmath84 into
( [ eq : trainsnr ] ) , we have @xmath85 where @xmath86 .
evaluating the derivative of @xmath82 with respect to @xmath19 and making it equal to zero leads to the expression in ( [ eq : optrho ] ) . clearly , @xmath78 is independent of @xmath47 and @xmath39 .
above , we have implicitly assumed that the maximization is performed with respect to first @xmath19 and then @xmath39 .
however , the result will not alter if the order of the maximization is changed .
note that the objective function in ( [ eq : trainopti ] ) @xmath87 is a monotonically increasing function of @xmath82 for all @xmath39 .
it can be easily verified that maximization does not affect the monotonicity of @xmath88 , and hence @xmath89 is still a monotonically increasing function of @xmath82 .
therefore , in the outer maximization with respect to @xmath19 , the choice of @xmath19 that maximizes @xmath82 will also maximize @xmath90 , and the optimal value of @xmath19 is again given by ( [ eq : optrho ] ) . @xmath91
in this section , we investigate the spectral efficiency bit energy tradeoff as the average power @xmath9 diminishes .
we assume that the bandwidth allocated to the channel is fixed . with the optimal value of @xmath19 given in theorem [ theo : optrho ]
, we can now express the normalized effective capacity as @xmath92 where @xmath93 and @xmath94 note that @xmath95 vanishes with decreasing @xmath9 .
we obtain the following result on the bit energy requirement in the low - power regime as @xmath9 diminishes .
[ theo : imperfect ] in the low - power regime , the bit energy increases without bound as the average power @xmath9 and hence @xmath0 vanishes , i.e. , @xmath96 this result shows us that operation at very low power levels is extremely energy inefficient and should be avoided regardless of the value of @xmath47 . note that the power allocated for training , @xmath97 , decreases with decreasing @xmath9 .
hence , our ability to estimate the channel is hindered in the low - power regime while , as mentioned before , the system operates as if the channel estimate were perfect .
this discrepancy leads to the inefficiency seen as @xmath9 approaches zero .
[ fig:6 ] plots the spectral efficiency vs. bit energy for @xmath98 when @xmath99 hz . as predicted by the result of theorem [ theo : imperfect ] , the bit energy increases without bound in all cases as the spectral efficiency @xmath100 .
consequently , the minimum bit energy is achieved at a nonzero spectral efficiency below which one should avoid operating as it only increases the energy requirements .
another observation is that the minimum bit energy increases as @xmath47 increases and hence as the statistical queueing constraints become more stringent . at higher spectral efficiencies ,
we again note the increased energy requirements with increasing @xmath47 . in fig .
[ fig : ebsnr ] , we plot @xmath101 as a function of @xmath0 for different bandwidth levels assuming @xmath102 .
similarly , we observe that the minimum bit energy is attained at a nonzero @xmath0 value below which @xmath101 requirements start increasing .
furthermore , we see that as the bandwidth increases , the minimum bit energy tends to decrease and is achieved at a lower @xmath0 level . in the rayleigh channel with @xmath103 .
@xmath99 . ]
vs. @xmath0 in the rayleigh channel .
@xmath103 , @xmath47=0.01 . ] in this section , we consider the wideband regime in which the bandwidth is large .
we assume that the average power @xmath9 is kept constant .
note that as the bandwidth @xmath6 increases , @xmath104 approaches zero and we again operate in the low-@xmath0 regime .
we denote @xmath105 .
note that as @xmath106 , we have @xmath107 . with this notation
, we can express the normalized effective capacity as @xmath108 the following result provides the expressions for the bit energy at zero spectral efficiency ( i.e. , as @xmath109 ) and the wideband slope , and characterize the spectral efficiency - bit energy tradeoff in the wideband regime .
[ theo : wideband ] in the wideband regime , the minimum bit energy and wideband slope are given by @xmath110 respectively , where @xmath111 , @xmath112 , and @xmath113 .
@xmath114 is defined as @xmath115 and @xmath114 satisfies @xmath116 we note that the minimum bit energy in the wideband regime is achieved as @xmath109 and hence as @xmath117 .
this is in stark contrast to the results in the low - power regime in which the bit energy requirements grow without bound as @xmath9 and @xmath0 vanishes . in the model we have considered ,
the difference in the performance can be attributed to the fact that increase in the bandwidth , while lowering symbol power @xmath118 , does not necessarily decrease the training power @xmath119 .
[ fig : trainwb ] plots the spectral efficiency bit energy curve in the rayleigh channel for different @xmath47 values . in the figure
, we assume that @xmath120 .
as predicted , the minimum bit energies are obtained as @xmath0 and hence the spectral efficiency approach zero .
@xmath121 are computed to be equal to @xmath122 db for @xmath123 , respectively .
moreover , the wideband slopes are @xmath124 for the same set of @xmath47 values .
as can also be seen in the result of theorem [ theo : wideband ] , the minimum bit energy and wideband slope in general depend on @xmath47 . in fig .
[ fig : trainwb ] , we note that the bit energy requirements ( including the minimum bit energy ) increase with increasing @xmath47 , illustrating the energy costs of stringent queueing constraints . in this paper , we have considered fixed - rate / fixed - power transmissions over imperfectly - known channels . in fig .
[ fig : compair ] , we compare the performance of this system with those in which the channel is perfectly - known and fixed- or variable - rate transmission is employed .
the latter models have been studied in @xcite and @xcite .
this figure demonstrates the energy costs of not knowing the channel and sending the information at fixed - rate .
a. ephremides and b. hajek , information theory and communication networks : an unconsummated union , " _ ieee trans .
inform . theory _
44 , pp .
2416 - 2434 , oct .
d. wu and r. negi `` effective capacity : a wireless link model for support of quality of service , '' _ ieee trans .
wireless commun .
_ , vol.2,no . 4 , pp.630 - 643 .
july 2003 l. liu , p. parag , j. tang , w .- y . chen and j .- f .
chamberland , `` resource allocation and quality of service evaluation for wireless communication systems using fluid models , '' _ ieee trans .
inform . theory _
1767 - 1777 , may 2007 j. tang and x. zhang , `` cross - layer - model based adaptive resource allocation for statistical qos guarantees in mobile wireless networks , '' _ ieee trans .
wireless commun .
_ , vol . 7 , pp.2318 - 2328 , june 2008 .
d. qiao , m.c .
gursoy and s. velipasalar , `` energy efficiency of fixed - rate wireless transmissions under qos constraints , '' to appear at _ ieee international conference on communications ( icc ) _ , jun
chang , _ performance guarantees in communication networks _
, new york : springer , 1995 | energy efficiency of fixed - rate transmissions is studied in the presence of queueing constraints and channel uncertainty .
it is assumed that neither the transmitter nor the receiver has channel side information prior to transmission .
the channel coefficients are estimated at the receiver via minimum mean - square - error ( mmse ) estimation with the aid of training symbols .
it is further assumed that the system operates under statistical queueing constraints in the form of limitations on buffer violation probabilities .
the optimal fraction of of power allocated to training is identified .
spectral efficiency
bit energy tradeoff is analyzed in the low - power and wideband regimes by employing the effective capacity formulation .
in particular , it is shown that the bit energy increases without bound in the low - power regime as the average power vanishes . on the other hand
, it is proven that the bit energy diminishes to its minimum value in the wideband regime as the available bandwidth increases . for this case ,
expressions for the minimum bit energy and wideband slope are derived .
overall , energy costs of channel uncertainty and queueing constraints are identified . |
quantum mechanics ( qm ) and general relativity ( gr ) have modified our understanding of the physical world in depth .
but they have left us with a general picture of the physical world which is unclear , incomplete , and fragmented . combining what we have
learn about our world from the two theories and finding a new synthesis is a major challenge , perhaps _ the _ major challenge , in today s fundamental physics .
the two theories have a opened a major scientific revolution , but this revolution is not completed .
most of the physics of this century has been a sequel of triumphant explorations of the new worlds opened by qm and gr .
qm lead to nuclear physics , solid state physics , and particle physics .
gr to relativistic astrophysics , cosmology and is today leading us towards gravitational astronomy .
the urgency of applying the two theories to larger and larger domains , the momentous developments , and the dominant pragmatic attitude of the middle of this century , have obscured the fact that a consistent picture of the physical world , more or less stable for three centuries , has been lost with the advent of qm and gr .
this pragmatic attitude can not be satisfactory , or productive , in the long run .
the basic cartesian - newtonian notions such as matter , space , time , causality , have been modified in depth .
the new notions do not stay together . at the basis of our understanding of the world
reigns a surprising confusion . from qm
and gr we know that we live in a spacetime with quantum properties , that is , a _
quantum spacetime_. but what is a quantum spacetime ? in the last decade , the attention of the theoretical physicists has been increasingly focusing on this major problem .
whatever the outcome of the enterprise , we are witnessing a large scale intellectual effort for accomplishing a major aim : completing the xxth scientific revolution , and finding a new synthesis . in this effort ,
physics is once more facing _
conceptual _ problems : _ what is matter ? what is causality ? what is the role of the observer in physics ?
what is time ?
what is the meaning of `` being somewhere '' ?
what is the meaning of `` now '' ?
what is the meaning of `` moving '' ? is motion to be defined with respect to objects or with respect to space ? _ these foundational questions , or sophisticated versions of these questions , were central in the thinking and in the results of einstein , heisenberg , bohr , dirac and their colleagues .
but these are also precisely the same questions that descartes , galileo , huygens , newton and their contemporaries debated with passion the questions that lead them to create modern science . for the physicists of the middle of this century ,
these questions were irrelevant : one does not need to worry about first principles in order to apply the schrdinger equation to the helium atom , or to understand how a neutron star stays together .
but today , if we want to find a novel picture of the world , if we want to understand what is quantum spacetime , we have to return , once again , to those foundational issues . we have to find a new answer to these questions different from newton s answer which took into account what we have learned about the world with qm and gr .
of course , we have little , if any , direct empirical access to the regimes in which we expect genuine quantum gravitational phenomena to appear .
anything could happen at those fantastically small distance scales , far removed from our experience . nevertheless , we do have information about quantum gravity , and we do have indications on how to search it .
in fact , we are precisely in one of the very typical situations in which good fundamental theoretical physics has been working at its best in the past : we have learned two new extremely general `` facts '' about our world , qm and gr , and we have `` just '' to figure out what they imply , when taken together .
the most striking advances in theoretical physics happened in situations analogous to this one . here
, i present some reflections on these issues . what have we learned about the world from qm and ,
especially , gr ?
what do we know about space , time and matter ?
what can we expect from a quantum theory of spacetime ?
to which extent does taking qm and gr into account force us to modify the notion of time ?
what can we already say about quantum spacetime ?
i present also a few reflections on issues raised by the relation between philosophy of science and research in quantum gravity .
i am not a philosopher , and i can touch philosophical issues only at the risk of being naive .
i nevertheless take this risk here , encouraged by craig callender and nick huggett extremely stimulating idea of this volume .
i present some methodological considerations how shall we search ?
how can the present successful theories can lead us towards a theory that does not yet exist? as well as some general consideration .
in particular , i discuss the relation between physical theories that supersed each others and the attitude we may have with respect to the truth - content of a physical theory , with respect to the reality of the theoretical objects the theory postulates in particular , and to its factual statements on the world in general .
i am convinced of the reciprocal usefulness of a dialog between physics and philosophy ( rovelli 1997a ) .
this dialog has played a major role during the other periods in which science faced foundational problems . in my opinion , most physicists underestimate the effect of their own epistemological prejudices on their research . and many philosophers underestimate the influence positive or negative they have on fundamental reserach .
on the one hand , a more acute philosphical awarness would greatly help the physicists engaged in fundamental research : newton , heisenberg and einstein could nt have done what they have done if they were nt nurtured by ( good or bad ) philosophy . on the other hand ,
i wish contemporary philosophers concerned with science would be more interested in the ardent lava of the foundational problems science is facing today .
it is here , i believe , that stimulating and vital issues lie .
what is the task of a quantum theory of gravity , and how should we search for such a theory ?
the task of the search is clear and well defined .
it is determined by recalling the three major steps that lead to the present problematic situation .
the first step is in the works of faraday , maxwell and einstein .
faraday and maxwell have introduced a new fundamental notion in physics , the field .
faraday s book includes a fascinating chapter with the discussion of whether the field ( in faraday s terminology , the `` lines of force '' ) is `` real '' .
as far as i understand this subtle chapter ( understanding faraday is tricky : it took the genius of maxwell ) , in modern terms what faraday is asking is whether there are independent degrees of freedom in the electric and magnetic fields .
a degree of freedom is a quantity that i need to specify ( more precisely : whose value and whose time derivative i need to specify ) in order to be able to predict univocally the future evolution of the system .
thus faraday is asking : if we have a system of interacting charges , and we know their positions and velocities , is this knowledge sufficient to predict the future motions of the charges ? or rather , in order to predict the future , we have to specify the instantaneous configuration of the field ( the fields degrees of freedom ) , as well ?
the answer is in maxwell equations : the field has independent degrees of freedom .
we can not predict the future evolution of the system from its present state unless we know the instantaneous field configuration . learning to use these degrees of freedom lead to radio , tv and cellular phone .
to which physical entity do the degrees of freedom of the electromagnetic field refer ?
this was one of the most debated issues in physics towards the end of last century .
the electromagnetic waves have aspects in common with water waves , or with sound waves , which describe vibrations of some material medium .
the natural interpretation of the electromagnetic field was that it too describes the vibrations of some material medium for which the name `` ether '' was chosen .
a strong argument supports this idea : the wave equations for water or sound waves fail to be galilean invariant .
they do so because they describe propagation over a medium ( water , air ) whose state of motion breaks galilean invariance and defines a preferred reference frame .
maxwell equations break galilean invariance as well and it was thus natural to hypothesize a material medium determining the preferred reference frame . but
a convincing dynamical theory of the ether compatible with the various experiments ( for instance on the constancy of the speed of light ) could not be found .
rather , physics took a different course .
einstein _ believed _ maxwell theory as a fundamental theory and _ believed _ the galilean insight that velocity is relative and inertial system are equivalent .
merging the two , he found special relativity .
a main result of special relativity is that the field can not be regarded as describing vibrations of underlying matter .
the idea of the ether is abandoned , and _ the field has to be taken seriously as elementary constituent of reality_. this is a major change from the ontology of cartesian - newtonian physics . in the best description we can give of the physical world
, there is a new actor : the field .
the electromagnetic field can be described by the maxwell potential @xmath0 .
the entity described by @xmath1 ( more precisely , by a gauge - equivalent class of @xmath1 s ) is one of the elementary constituents of the physical world , according to the best conceptual scheme physics has find , so far , for grasping our world .
the second step ( out of chronological order ) is the replacement of the mechanics of newton , lagrange and hamilton with quantum mechanics ( qm ) . as did classical mechanics ,
qm provides a very general framework . by formulating a specific dynamical theory within this framework
, one has a number of important physical consequences , substantially different from what is implied by the newtonian scheme .
evolution is probabilistically determined only ; some physical quantities can take certain discrete values only ( are `` quantized '' ) ; if a system can be in a state @xmath2 , where a physical quantity @xmath3 has value @xmath4 , as well as in state @xmath5 , where @xmath3 has value @xmath6 , then the system can also be in states ( denoted @xmath7 ) where @xmath3 , has value @xmath4 with probability @xmath8 , or , alternatively , @xmath6 with probability @xmath9 ( superposition principle ) ; conjugate variables can not be assumed to have value at the same time ( uncertainty principle ) ; and what we can say about the properties that the system will have the - day - after - tomorrow is not determined just by what we can say about the system today , but also on what we will be able to say about the system tomorrow .
( bohr would had simply said that observations affect the system .
formulations such as bohm s or consistent histories force us to use intricate wording for naming the same physical fact . )
the formalism of qm exists in a number of more or less equivalent versions : hilbert spaces and self - adjoint observables , feynman s sum over histories , algebraic formulation , and others . often , we are able to translate from one formulation to another . however , often we can not do easily in one formulation , what we can do in another .
qm is not the theory of micro - objects .
it is our best form of mechanics .
if quantum mechanics failed for macro - objects , we would have detected the boundary of its domain of validity in mesoscopic physics .
we havent .
the classical regime raises some problems ( why effects of macroscopic superposition are difficult to detect ? ) .
solving these problems requires good understanding of physical decoherence and perhaps more .
but there is no reason to doubt that qm represents a deeper , not a shallower level of understanding of nature than classical mechanics .
trying to resolve the difficulties in our grasping of our quantum world by resorting to old classical intuition is just lack of courage .
we have learned that the world has quantum properties .
this discovery will stay with us , like the discovery that velocity is only relational or like the discovery that the earth is not the center of the universe .
the empirical success of qm is immense .
its physical obscurity is undeniable .
physicists do not yet agree on what qm precisely says about the world ( the difficulty , of course , refers to physical meaning of notions such as `` measurement '' , `` history '' , `` hidden variable '' , ) .
it is a bit like the lorentz transformations before einstein : correct , but what do they mean ? in my opinion , what qm means is that the contingent ( variable ) properties of any physical system , or the state of the system , are relational notion which only make sense when referred to a second physical system .
i have argued for this thesis in ( rovelli 1996 , rovelli 1998 ) .
however , i will not enter in this discussion here , because the issue of the interpretation of qm has no direct connection with quantum gravity . quantum gravity and the interpretation of qm
are two major but ( virtually ) completely unrelated problems .
qm was first developed for systems with a finite number of degrees of freedom .
as discussed in the previous section , faraday , maxwell and einstein had introduced the field , which has an infinite number of degrees of freedom .
dirac put the two ideas together .
he _ believed _ quantum mechanics and he _ believed _
maxwell s field theory much beyond their established domain of validity ( respectively : the dynamics of finite dimensional systems , and the classical regime ) and constructed quantum field theory ( qft ) , in its first two incarnations , the quantum theory of the electromagnetic field and the relativistic quantum theory of the electron . in this exercise ,
dirac derived the existence of the photon just from maxwell theory and the basics of qm .
furthermore , by just _ believing _ special relativity and _ believing _ quantum theory , namely assuming their validity far beyond their empirically explored domain of validity , he predicted the existence of antimatter .
the two embryonal qft s of dirac were combined in the fifties by feynman and his colleagues , giving rise to quantum electrodynamics , the first nontrivial interacting qft .
a remarkable picture of the world was born : quantum fields over minkowski space .
equivalently , la feynman : the world as a quantum superposition of histories of real and virtual interacting particles .
qft had ups and downs , then triumphed with the standard model : a consistent qft for all interactions ( except gravity ) , which , in principle , can be used to predict anything we can measure ( except gravitational phenomena ) , and which , in the last fifteen years has received nothing but empirical verifications .
descartes , in _
le monde _ , gave a fully relational definition of localization ( space ) and motion ( on the relational / substantivalist issue , see earman and norton 1987 , barbour 1989 , earman 1989 , rovelli 1991a , belot 1998 ) . according to descartes
, there is no `` empty space '' .
there are only objects , and it makes sense to say that an object a is contiguous to an object b. the `` location '' of an object a is the set of the objects to which a is contiguous .
`` motion '' is change in location .
that is , when we say that a moves we mean that a goes from the contiguity of an object b to the contiguity of an object c. a consequence of this relationalism is that there is no meaning in saying `` a moves '' , except if we specify with respect to which other objects ( b , c , ) it is moving . thus , there is no `` absolute '' motion .
this is the same definition of space , location , and motion , that we find in aristotle .
relationalism , namely the idea that motion can be defined only in relation to other objects , should not be confused with galilean relativity .
galilean relativity is the statement that `` rectilinear uniform motion '' is a priori indistinguishable from stasis .
namely that velocity ( but just velocity ! ) , is relative to other bodies .
relationalism holds that _ any _ motion ( however zigzagging ) is a priori indistinguishable from stasis .
the very formulation of galilean relativity requires a nonrelational definition of motion ( `` rectilinear and uniform '' with respect to what ? ) .
newton took a fully different course .
he devotes much energy to criticise descartes relationalism , and to introduce a different view . according to him , _
space _ exists .
it exists even if there are no bodies in it .
location of an object is the part of space that the object occupies .
motion is change of location .
thus , we can say whether an object moves or not , irrespectively from surrounding objects .
newton argues that the notion of absolute motion is necessary for constructing mechanics .
his famous discussion of the experiment of the rotating bucket in the _ principia _ is one of the arguments to prove that motion is absolute .
this point has often raised confusion because one of the corollaries of newtonian mechanics is that there is no detectable preferred referential frame . therefore
the notion of _ absolute velocity _ is , actually , meaningless , in newtonian mechanics .
the important point , however , is that in newtonian mechanics velocity is relative , but any other feature of motion is not relative : it is absolute .
in particular , acceleration is absolute .
it is acceleration that newton needs to construct his mechanics ; it is acceleration that the bucket experiment is supposed to prove to be absolute , against descartes . in a sense ,
newton overdid a bit , introducing the notion of absolute position and velocity ( perhaps even just for explanatory purposes ? ) .
many people have later criticised newton for his unnecessary use of absolute position . but
this is irrelevant for the present discussion .
the important point here is that newtonian mechanics requires absolute acceleration , against aristotle and against descartes .
precisely the same does special relativistic mechanics . similarly , newton introduce absolute time .
newtonian space and time or , in modern terms , spacetime , are like a _ stage _ over which the action of physics takes place , the various dynamical entities being the actors .
the key feature of this stage , newtonian spacetime , is its metrical structure .
curves have length , surfaces have area , regions of spacetime have volume .
spacetime points are at fixed _ distance _ the one from the other .
revealing , or measuring , this distance , is very simple .
it is sufficient to take a rod and put it between two points .
any two points which are one rod apart are at the same distance .
using modern terminology , physical space is a linear three - dimensional ( 3d ) space , with a preferred metric . on this space there
exist preferred coordinates @xmath10 , in terms of which the metric is just @xmath11 .
time is described by a single variable @xmath12 .
the metric @xmath11 determines lengths , areas and volumes and defines what we mean by straight lines in space .
if a particle deviates with respect to this straight line , it is , according to newton , accelerating .
it is not accelerating with respect to this or that dynamical object : it is accelerating in absolute terms .
special relativity changes this picture only marginally , loosing up the strict distinction between the `` space '' and the `` time '' components of spacetime . in newtonian
spacetime , space is given by fixed 3d planes .
in special relativistic spacetime , which 3d plane you call space depends on your state of motion .
spacetime is now a 4d manifold @xmath13 with a flat lorentzian metric @xmath14 .
again , there are preferred coordinates @xmath15 , in terms of which @xmath16 $ ] .
this tensor , @xmath14 , enters all physical equations , representing the determinant influence of the stage and of its metrical properties on the motion of anything .
absolute acceleration is deviation of the world line of a particle from the straight lines defined by @xmath14 .
the only essential novelty with special relativity is that the `` dynamical objects '' , or `` bodies '' moving over spacetime now include the fields as well .
example : a violent burst of electromagnetic waves coming from a distant supernova has _ traveled across space _ and has reached our instruments . for the rest , the newtonian construct of a fixed background stage over which physics happen is not altered by special relativity .
the profound change comes with general relativity ( gr ) .
the central discovery of gr , can be enunciated in three points .
one of these is conceptually simple , the other two are tremendous .
first , the gravitational force is mediated by a field , very much like the electromagnetic field : the gravitational field .
second , newton s _ spacetime _ , the background stage that newton introduced introduced , against most of the earlier european tradition , _ and the gravitational field , are the same thing_. third , the dynamics of the gravitational field , of the other fields such as the electromagnetic field , and any other dynamical object , is fully relational , in the aristotelian - cartesian sense .
let me illustrate these three points .
first , the gravitational field is represented by a field on spacetime , @xmath17 , just like the electromagnetic field @xmath1 .
they are both very concrete entities : a strong electromagnetic wave can hit you and knock you down ; and so can a strong gravitational wave . the gravitational field has independent degrees of freedom , and is governed by dynamical equations , the einstein equations .
second , the spacetime metric @xmath14 disappears from all equations of physics ( recall it was ubiquitous ) . at its place we
are instructed by gr we must insert the gravitational field @xmath17 .
this is a spectacular step : newton s background spacetime was nothing but the gravitational field !
the stage is promoted to be one of the actors .
thus , in all physical equations one now sees the direct influence of the gravitational field . how can the gravitational field determine the metrical properties of things , which are revealed , say , by rods and clocks
simply , the inter - atomic separation of the rods atoms , and the frequency of the clock s pendulum are determined by explicit couplings of the rod s and clock s variables with the gravitational field @xmath17 , which enters the equations of motion of these variables .
thus , any measurement of length , area or volume is , in reality , a measurement of features of the gravitational field .
but what is really formidable in gr , the truly momentous novelty , is the third point : the einstein equations , as well as _ all other equations of physics _ appropriately modified according to gr instructions , are fully relational in the aristotelian - cartesian sense .
this point is independent from the previous one .
let me give first a conceptual , then a technical account of it .
the point is that the only physically meaningful definition of location that makes physical sense within gr is relational .
gr describes the world as a set of interacting fields and , possibly , other objects .
one of these interacting fields is @xmath17 .
motion can be defined only as positioning and displacements of these dynamical objects relative to each other ( for more details on this , see rovelli 1991a and especially 1997a ) . to describe the motion of a dynamical object , newton had to assume that acceleration is absolute , namely it is not relative to this or that other dynamical object .
rather , it is relative to a background space .
faraday maxwell and einstein extended the notion of `` dynamical object '' : the stuff of the world is fields , not just bodies .
finally , gr tells us that the background space is itself one of these fields .
thus , the circle is closed , and we are back to relationalism : newton s motion with respect to space is indeed motion with respect to a dynamical object : the gravitational field .
all this is coded in the active diffeomorphism invariance ( diff invariance ) of gr . with the equations of motion
riemann[g]=0 ) might require a detailed analysis ( for instance , hamiltonian ) of the theory .
] because active diff invariance is a gauge , the physical content of gr is expressed only by those quantities , derived from the basic dynamical variables , which are fully independent from the points of the manifold . in introducing the background stage ,
newton introduced two structures : a spacetime manifold , and its non - dynamical metric structure .
gr gets rid of the non - dynamical metric , by replacing it with the gravitational filed .
more importantly , it gets rid of the manifold , by means of active diff invariance . in gr ,
the objects of which the world is made do not live over a stage and do not live on spacetime : they live , so to say , over each other s shoulders .
of course , nothing prevents us , if we wish to do so , from singling out the gravitational field as `` the more equal among equals '' , and declaring that location is absolute in gr , because it can be defined with respect to it . but this can be done within any relationalism : we can always single out a set of objects , and declare them as not - moving by definition .
the problem with this attitude is that it fully misses the great einsteinian insight : that newtonian spacetime is just one field among the others .
more seriously , this attitude sends us into a nightmare when we have to deal with the motion of the gravitational field itself ( which certainly `` moves '' : we are spending millions for constructing gravity wave detectors to detect its tiny vibrations ) .
there is no absolute referent of motion in gr : the dynamical fields `` move '' with respect to each other .
notice that the third step was not easy for einstein , and came later than the previous two .
having well understood the first two , but still missing the third , einstein actively searched for non - generally covariant equations of motion for the gravitational field between 1912 and 1915 . with his famous `` hole argument ''
he had convinced himself that generally covariant equations of motion ( and therefore , in this context , active diffeomorphism invariance ) would imply a truly dramatic revolution with respect to the newtonian notions of space and time ( on the hole argument , see earman and norton 1987 , rovelli 1991a , belot 1998 ) . in 1912
he was not able to take this profoundly revolutionary step ( norton 1984 , stachel 1989 ) . in 1915
he took this step , and found what landau calls `` the most beautiful of the physical theories '' . at the light of the three steps illustrated above
, the task of quantum gravity is clear and well defined .
he have learned from gr that spacetime is a dynamical field among the others , obeying dynamical equations , and having independent degrees of freedom .
a gravitational wave is extremely similar to an electromagnetic wave .
we have learned from qm that every dynamical object has quantum properties , which can be captured by appropriately formulating its dynamical theory within the general scheme of qm .
_ therefore _ , spacetime itself must exhibit quantum properties .
its properties , including the metrical properties it defines , must be represented in quantum mechanical terms .
notice that the strength of this `` therefore '' derives from the confidence we have in the two theories , qm and gr . now , there is nothing in the basics of qm which contradicts the physical ideas of gr .
similarly , there is nothing in the basis of gr that contradicts the physical ideas of qm .
therefore , there is no a priori impediment in searching for a quantum theory of the gravitational fields , that is , a quantum theory of spacetime .
the problem is ( with some qualification ) rather well posed : is there a quantum theory ( say , in one formulation , a hilbert space @xmath18 , and a set of self - adjoint operators ) whose classical limit is gr ? on the other hand , all previous applications of qm to _ field _ theory , namely conventional qft s , rely heavily on the existence of the `` stage '' , the fixed , non - dynamical , background metric structure .
the minkowski metric @xmath14is essentially for the construction of a conventional qft ( in enters everywhere ; for instance , in the canonical commutation relations , in the propagator , in the gaussian measure ) .
we certainly can not simply replace @xmath14 with a quantum field , because all equations become nonsense . therefore , to search for a quantum theory of gravity , we have two possible directions .
one possibility is to `` disvalue '' the gr conceptual revolution , reintroduce a background spacetime with a non - dynamical metric @xmath14 , expand the gravitational field @xmath19 as @xmath20 , quantize only the fluctuations , and hope to recover the full of gr somewhere down the road .
this is the road followed for instance by perturbative string theory .
the second direction is to be faithful to what we have learned about the world so far .
namely to the qm and the gr insights .
we must then search a qft that , genuinely , does not require a background space to be defined .
but the last three decades whave been characterized by the great success of conventional qft , which neglects gr and is based on the existence of a background spacetime .
we live in the aftermath of this success .
it is not easy to get out from the mental habits and from the habits to the technical tools of conventional qft .
still , this is necessary if we want to build a qft which fully incorporates active diff invariance , and in which localization is fully relational . in my opinion
, this is the right way to go .
spacetime , or the gravitational field , is a dynamical entity ( gr ) .
all dynamical entities have quantum properties ( qm ) .
therefore spacetime is a quantum object .
it must be described ( picking one formulation of qm , but keeping in mind that others may be equivalent , or more effective ) in terms of states @xmath21 in a hilbert space .
localization is relational .
therefore these states can not represent quantum excitations localized in some space .
they must define space themselves .
they must be quantum excitations `` of '' space , not `` in '' space .
physical quantities in gr , that capture the true degrees of freedom of the theory are invariant under active diff
. therefore the self - adjoint operators that correspond to physical ( predictable ) observables in quantum gravity must be associated to diff invariant quantities .
examples of diff - invariant geometric quantities are physical lengths , areas , volumes , or time intervals , of regions determined by dynamical physical objects .
these must be represented by operators .
indeed , a measurement of length , area or volume is a measurement of features of the gravitational field .
if the gravitational field is a quantum field , then length , area and volume are quantum observables
. if the corresponding operator has discrete spectrum , they will be quantized , namely they can take certain discrete values only . in this sense
we should expect a discrete geometry .
this discreteness of the geometry , implied by the conjunction of gr and qm is very different from the naive idea that the world is made by discrete bits of something .
it is like the discreteness of the quanta of the excitations of an harmonic oscillator .
a generic state of spacetime will be a continuous quantum superposition of states whose geometry has discrete features , not a collection of elementary discrete objects .
a concrete attempt to construct such a theory , is loop quantum gravity .
i refer the reader to rovelli ( 1997b ) for an introduction to the theory , an overview of its structure and results , and full references . here
, i present only a few remarks on the theory .
loop quantum gravity is a rather straightforward application of quantum mechanics to hamiltonian general relativity .
it is a qft in the sense that it is a quantum version of a field theory , or a quantum theory for an infinite number of degrees of freedom , but it is profoundly different from conventional , non - general - relativistic qft theory . in conventional qft , states are quantum excitations of a field over minkowski ( or over a curved ) spacetime . in loop quantum gravity ,
the quantum states turn out to be represented by ( suitable linear combinations of ) spin networks ( rovelli and smolin 1995a , baez 1996 , smolin 1997 ) .
a spin network is an abstract graphs with links labeled by half - integers .
see figure 1 .
intuitively , we can view each node of the graph as an elementary `` quantum chunk of space '' .
the links represent ( transverse ) surfaces separating the quanta of space .
the half - integers associated to the links determine the ( quantized ) area of these surfaces .
the spin network represent relational quantum states : they are not located in a space .
localization must be defined in relation to them .
for instance , if we have , say , a matter quantum excitation , this will be located on the spin network ; while the spin network itself is not located anywhere .
the operators corresponding to area and volume have been constructed in the theory , simply by starting from the classical expression for the area in terms of the metric , then replacing the metric with the gravitational field ( this is the input of gr ) and then replacing the gravitational field with the corresponding quantum field operator ( this is the input of qm ) .
the construction of these operators requires appropriate generally covariant regularization techniques , but no renormalization : no infinities appear .
the spectrum of these operators has been computed and turns out to be discrete ( rovelli and smolin 1995b , ashtekar lewandowski 1997a , 1997b ) .
thus , loop quantum gravity provides a family of precise quantitative predictions : the quantized values of area and volume .
for instance , the ( main sequence ) of the spectrum of the area is @xmath22 where @xmath23 is any finite sequence of half integers .
this formula gives the area of a surface pinched by @xmath24 links of a spin network state .
the half integers @xmath25 are ones associated with the @xmath24 links that pinch the surface .
this illustrates how the links of the spin network states can be viewed as transversal `` quanta of area '' .
the picture of macroscopic physical space that emerges is then that of a tangle of one - dimensional intersecting quantum excitation , called the weave ( ashtekar rovelli and smolin 1992 ) .
continuous space is formed by the weave in the same manner in which the continuous 2d surface of a t - shirt is formed by weaved threads .
the aspect of the gr s relationalism that concerns space was largely anticipated by the earlier european thinking .
much less so ( as far as i am aware ) was the aspect of this relationalism that concerns time .
gr s treatment of time is surprising , difficult to fully appreciate , and hard to digest .
the time of our perceptions is very different from the time that theoretical physics finds in the world as soon as one exits the minuscule range of physical regimes we are accustomed to .
we seem to have a very special difficulty in being open minded about this particular notion .
already special relativity teaches us something about time which many of us have difficulties to accept . according to special relativity
, there is absolute no meaning in saying `` right now on andromeda '' .
there is no physical meaning in the idea of `` the state of the world right now '' , because which set of events we consider as `` now '' is perspectival .
the `` now '' on andromeda for me might correspond to `` a century ago '' on andromeda for you .
thus , there is no single well defined universal time in which the history of the universe `` happens '' .
the modification of the concept of time introduced by gr is much deeper .
let me illustrate this modifications .
consider a simple pendulum described by a variable @xmath26 . in newtonian mechanics ,
the motion of the pendulum is given by the evolution of @xmath26 in time , namely by @xmath27 , which is governed by the equation of motion , say @xmath28 , which has ( the two - parameter family of ) solutions @xmath29 .
the state of the pendulum at time @xmath30 can be characterized by its position and velocity . from these two
, we can compute @xmath2 and @xmath31 and therefore @xmath27 at any @xmath30 . from the physical point of view
, we are really describing a situation in which there are _ two _ physical objects : a pendulum , whose position is @xmath26 , and a clock , indicating @xmath30 .
if we want to take data , we have to repeatedly observe @xmath26 and @xmath30 .
their _ relation _ will be given by the equation above .
the relation can be represented ( for given @xmath2 and @xmath31 ) by a line in the @xmath32 plane . in newtonian terms
, time flows in its absolute way , the clock is just a devise to keep track of it , and the dynamical system is formed by the pendulum alone .
but we can view the same physical situation from a different perspective .
we can say that we have a physical system formed by the clock and the pendulum together and view the dynamical system as expressing the relative motion of one with respect to the other .
this is precisely the perspective of gr : to express the relative motion of the variables , with respect to each other , in a `` democratic '' fashion . to do that
, we can introduce an `` arbitrary parameter time '' @xmath33 as a coordinate on the line in the @xmath32 plane .
( but keep in mind that the physically relevant information is in the line , not in its coordinatization ! ) . then the line is represented by two functions , @xmath34 and @xmath35 , but a reparametrization of @xmath33 in the two functions is a gauge , namely it does not modify the physics described .
indeed , @xmath33 does not correspond to anything observable , and the equations of motion satisfied by @xmath34 and @xmath35 ( easy to write , but i will not write them down here ) will be invariant under arbitrary reparametrizations of @xmath33 . only @xmath33-independent quantities have physical meaning .
this is precisely what happens in gr , where the `` arbitrary parameters '' , analogous to the @xmath33 of the example , are the coordinates @xmath36 .
namely , the spatial coordinate @xmath37 and the temporal coordinate @xmath12 .
these have no physical meaning whatsoever in gr : the connection between the theory and the measurable physical quantities that the theory predict is only via quantities independent from @xmath37 and @xmath12 .
thus , @xmath37 and @xmath12 in gr have a very different physical meaning than their homonymous in non - general - relativistic physics .
the later correspond to readings on rods and clocks .
the formed , correspond to nothing at all . recall that einstein described his great intellectual struggle to find gr as `` understanding the meaning of the coordinates '' . in the example
, the invariance of the equations of motion for @xmath34 and @xmath35 under reparametrization of @xmath33 , implies that if we develop the hamiltonian formalism in @xmath33 we obtain a constrained system with a ( weakly ) vanishing hamiltonian .
this is because the hamiltonian generates evolutions in @xmath33 , evolution in @xmath33 is a gauge , and the generators of gauge transformations are constraints . in canonical gr
we have precisely the same situation : the hamiltonian vanishes , the constraints generate evolution in @xmath12 , which is unobservable it is gauge .
gr does not describe evolution in time : it describes the relative evolution of many variables with respect to each other .
all these variables are democratically equal : there is nt a preferred one that `` is the true time '' .
this is the temporal aspect of gr s relationalism .
a large part of the machinery of theoretical physics relies on the notion of time ( on the different meanings of time in different physical theories , see rovelli 1995 ) .
a theory of quantum gravity should do without .
fortunately , many essential tools that are usually introduced using the notion of time can equally well be defined without mentioning time at all .
this , by the way , shows that time plays a much weaker role in the structure of theoretical physics than what is mostly assumed .
two crucial examples are `` phase space '' and `` state ''
. the phase space is usually introduced in textbooks as the space of the states of the systems `` at a given time '' . in a general relativistic context
, this definition is useless .
however , it is known since lagrange that there is an alternative , equivalent , definition of phase space as the space of the solutions of the equations of motion .
this definition does not require that we know what we mean by time .
thus , in the example above the phase space can be coordinatized by @xmath2 and @xmath31 , which coordinatize the space of the solutions of the equations of motion .
a time independent notion of `` state '' is then provided by a point of this phase space , namely by a particular solution of the equations of motion .
for instance , for an oscillator a `` state '' , in this atemporal sense , is characterized by an amplitude @xmath2 and a phase @xmath31 .
notice that given the ( time - independent ) state ( @xmath2 and @xmath31 ) , we can compute any observable : in particular , the value @xmath38 of @xmath26 at any desired @xmath30 .
notice also that @xmath38 is independent from @xmath33 .
this point often raises confusion : one may think that if we restrict to @xmath33-independent quantities then we can not describe evolution .
this is wrong : the true evolution is the relation between @xmath26 and @xmath30 , which is @xmath33-independent .
this relation is expressed in particular by the value ( let us denote it @xmath38 ) of @xmath26 at a given @xmath30 .
@xmath38 is given , obviously , by @xmath39 this can be seen as a one - parameter ( the parameter is @xmath30 ) family of observables on the gauge invariant phase space coordinatized by @xmath2 and @xmath31 . notice that this is a perfectly @xmath33-independent expression .
in fact , an explicit computation shows that the poisson bracket between @xmath38 and the hamiltonian constraint that generates evolution in @xmath33 vanishes .
this time independent notion of states is well known in its quantum mechanical version : it is the heisenberg state ( as opposed to schrdinger state ) .
similarly , the operator corresponding to the observable @xmath38 is the heisenberg operator that gives the value of @xmath26 at @xmath30 .
the heisenberg and schrdinger pictures are equivalent if there is a normal time evolution in the theory . in the absence of a normal notion of time evolution ,
the heisenberg picture remains viable , the shrdinger picture becomes meaningless .
in quantum gravity , only the heisenberg picture makes sense ( rovelli 1991c , 1991d ) . in classical gr ,
a point in the physical phase space , or a state , is a solution of einstein equations , up to active diffeomorphisms .
a state represents a `` history '' of spacetime .
the quantity that can be univocally predicted are the ones that are independents from the coordinates , namely that are invariant under diffeomorphisms .
these quantities have vanishing poisson brackets with all the constraints . given a state ,
the value of each of these quantities is determined . in quantum gravity ,
a quantum state represents a `` history '' of quantum spacetime .
the observables are represented by operators that commute with _ all _ the quantum constraints .
if we know the quantum state of spacetime , we can then compute the expectation value of any diffeomorphism invariant quantity , by taking the mean value of the corresponding operator .
the observable quantities in quantum gravity are precisely the same as in classical gr .
some of these quantities may express the value of certain variables `` when and where '' certain other quantities have certain given values .
they are the analog of the reparametrization invariant observable @xmath38 in the example above .
these quantities describe evolution in a way which is fully invariant under the parameter time , unphysical gauge evolution ( rovelli 1991d , 1991e ) .
the corresponding quantum operators are heisenberg operators .
there is no schrdinger picture , because there is no unitary time evolution .
there is no need to expect or to search for unitary time evolution in quantum gravity , because there is no time in which we should have unitary evolution .
a prejudice hard to die wants that unitary evolution is required for the consistency of the probabilistic interpretation .
this idea is wrong .
what i have described is the general form that one may expect a quantum theory of gr to have .
i have used the hilbert space version of qm ; but this structure can be translated in other formulations of qm .
of course , physics works then with dirty hands : gauge dependent quantities , approximations , expansions , unphysical structures , and so on . a fully satisfactory construction of the above does not yet exist . a concrete attempt to construct the physical states and the physical observables in loop quantum gravity is given by the spin foam models approach , which is the formulation one obtains by starting from loop quantum gravity and constructing a feynman sum over histories ( reisenberger rovelli 1997 , baez 1998 , barret and crane 1998 ) . see ( baez 1999 ) in this volume for more details on ideas underlying these developments .
in quantum gravity , i see no reason to expect a fundamental notion of time to play any role .
but the _ nostalgia for time _ is hard to resist . for technical as well as for emotional reasons .
many approaches to quantum gravity go out of their way to reinsert in the theory what gr is teaching us we should abandon : a preferred time .
the time `` along which '' things happen is a notion which makes sense only for describing a limited regime of reality .
this notion is meaningless already in the ( gauge invariant ) general relativistic classical dynamics of the gravitational field . at the fundamental level
, we should , simply , forget time .
i close this section by briefly mentioning two more speculative ideas .
one regards the emergence of time , the second the connection between the relationalism in gr and the relationalism in qm .
\(i ) in the previous section , i have argued that we should search for a quantum theory of gravity in which there is no independent time variable `` along '' which dynamics `` happens '' .
a problem left open by this position is to understand the emergence of time in our world , with its features , which are familiar to us .
an idea discussed in ( rovelli 1993a 1993b , connes and rovelli 1994 ) is that the notion of time is nt dynamical but rather thermodynamical
. we can never give a complete account of the state of a system in a field theory ( we can not access the infinite amount of data needed to completely characterize a state ) .
therefore we have at best a statistical description of the state . given a statistical state of a generally covariant system , a notion of a flow ( more precisely a one - parameter group of automorphisms of the algebra of the observables ) follows immediately . in the quantum context
, this corresponds to the tomita flow of the state .
the relation between this flow and the state is the relation between the time flow generated by the hamiltonian and a gibbs state : the two essentially determine each other . in the absence of a preferred time , however , any statistical state selects its own notion of statistical time .
this statistical time has a striking number of properties that allow us to identify it with the time of non - general relativistic physics .
in particular , a schrdinger equation with respect to this statistical time holds , in an appropriate sense .
in addition , the time flows generated by different states are equivalent up to inner automorphisms of the observable algebra and therefore define a common `` outer '' flow : a one paramater group of outer automorphisms .
this determines a state independent notion of time flow , which shows that a general covariant qft has an intrinsic `` dynamics '' , even in the absence of a hamiltonian and of a time variable .
the suggestion is therefore that the temporal aspects of our world have statistical and thermodynamical origin , rather than dynamical .
`` time '' is ignorance : a reflex of our incomplete knoweldge of the state of the world .
\(ii ) what is qm really telling us about our world ? in ( rovelli 1996 , 1998 ) , i have argued that what qm is telling us is that the contingent properties of any system or : the state of any system must be seen as relative to a second physical system , the `` observing system ''
. that is , quantum state and values that an observables take are relational notions , in the same sense in which velocity is relational in classical mechanics ( it is a relation between two systems , not a properties of a single system ) .
i find the consonance between this relationalism in qm and the relationalism in gr quite striking .
it is tempting to speculate that they are related .
any quantum interaction ( or quantum measurement ) involving a system @xmath2 and a system @xmath5 requires @xmath2 and @xmath5 to be spatiotemporally contiguous .
viceversa , spatiotemporal contiguity , which is the grounding of the notions of space and time ( derived and dynamical , not primary , in gr ) can only be verified quantum mechanically ( just because any interaction is quantum mechanical in nature ) .
thus , the net of the quantum mechanical elementary interactions and the spacetime fabric are actually the same thing .
can we build a consistent picture in which we take this fact into account ? to do that , we must identify two notions : the notion of a spatiotemporal ( or spatial ? ) region , and the notion of quantum system . for intriguing ideas in this direction , see ( crane 1991 ) and , in this volume , ( baez 1999 ) .
part of the recent reflection about science has emphasized the `` non cumulative '' aspect in the development of scientific knowledge .
according to this view , the evolution of scientific theories is marked by large or small breaking points , in which , to put it very crudely , the empirical facts are just reorganized within new theories .
these would be to some extent `` incommensurable '' with respect to their antecedent .
these ideas have influenced physicists .
the reader has remarked that the discussion of quantum gravity i have given above assumes a different reading of the evolution of scientific knowledge .
i have based the above discussion on quantum gravity on the idea that the central physical ideas of qm and gr represent our best guide for accessing the extreme and unexplored territories of the quantum - gravitational regime . in my opinion
, the emphasis on the incommensurability between theories has probably clarified an important aspect of science , but risks to obscure something of the internal logic according to which , historically , physics finds knowledge .
there is a subtle , but definite , cumulative aspect in the progress of physics , which goes far beyond the growth of validity and precision of the empirical content of the theories . in moving from a theory to the theory that supersedes it
, we do not save just the verified empirical content of the old theory , but more .
this `` more '' is a central concern for good physics .
it is the source , i think , of the spectacular and undeniable predicting power of theoretical physics .
let me illustrate the point i am trying to make with a historical case .
there was a problem between maxwell equations and galilei transformations .
there were two obvious way out .
to disvalue maxwell theory , degrading it to a phenomenological theory of some yet - to - be - discovered ether s dynamics . or to disvalue galilean invariance , accepting the idea that inertial systems are not equivalent in electromagnetic phenomena .
both ways were pursued at the end of the century .
both are sound applications of the idea that a scientific revolution may very well change in depth what old theories teach us about the world . which of the two ways did einstein take ?
none of them . for einstein ,
maxwell theory was a source of great awe .
einstein rhapsodizes about his admiration for maxwell theory . for him , maxwell had opened a new window over the world .
given the astonishing success of maxwell theory , empirical ( electromagnetic waves ) , technological ( radio ) as well as conceptual ( understanding what is light ) , einstein admiration is comprehensible .
but einstein had a tremendous respect for galileo s insight as well .
young einstein was amazed by a book with huygens derivation of collision theory virtually out of galilean invariance alone .
einstein understood that galileo s great intuition that the notion of velocity is only relative _ could not be wrong_. i am convinced that in this faith of einstein in the core of the great galilean discovery there is very much to learn , for the philosophers of science , as well as for the contemporary theoretical physicists .
so , einstein _ believed the two theories , maxwell and galileo_. he assumed that they would hold far beyond the regime in which they had been tested .
he assumed that galileo had grasped something about the physical world , which was , simply , _
correct_. and so had maxwell .
of course , details had to be adjusted .
the core of galileo s insight was that all inertial systems are equivalent and that velocity is relative , not the details of the galilean transformations .
einstein knew the lorentz transformations ( found , of course , by lorentz , not by einstein ) , and was able to see that they do not contradict galileo s insight .
if there was contradiction in putting the two together , the problem was ours : we were surreptitiously sneaking some incorrect assumption into our deductions .
he found the incorrect assumption , which , of course , was that simultaneity could be well defined .
it was einstein s faith in the _ essential physical correctness _ of the old theories that guided him to his spectacular discovery .
there are innumerable similar examples in the history of physics , that equally well could illustrate this point .
einstein found gr `` out of pure thought '' , having newton theory on the one hand and special relativity the understanding that any interaction is mediated by a field on the other ; dirac found quantum field theory from maxwell equations and quantum mechanics ; newton combined galileo s insight that acceleration governs dynamics with kepler s insight that the source of the force that governs the motion of the planets is the sun the list could be long . in all these cases , confidence in the insight that came with some theory , or `` taking a theory seriously '' ,
lead to major advances that largely extended the original theory itself . of course , far from me suggesting that there is anything simple , or automatic , in figuring out where the true insights are and in finding the way of making them work together .
but what i am saying is that figuring out where the true insights are and finding the way of making them work together is the work of fundamental physics .
this work is grounded on the _ confidence _ in the old theories , not on random search of new ones .
one of the central concerns of modern philosophy of science is to face the apparent paradox that scientific theories change , but are nevertheless credible .
modern philosophy of science is to some extent an after - shock reaction to the fall of newtonian mechanics .
a tormented recognition that an extremely successful scientific theory can nevertheless be untrue .
but it is a narrow - minded notion of truth the one which is questioned by the event of a successful physical theory being superseded by a more successful one .
a physical theory , in my view , is a conceptual structure that we use in order to organize , read and understand the world , and make prediction about it .
a successful physical theory is a theory that does so effectively and consistently . at the light of our experience
, there is no reason not to expect that a more effective conceptual structure might always exist .
therefore an effective theory may always show its limits and be replaced by a better one . on the other hand , however , a novel conceptualization can not but rely on what the previous one has already achieved .
when we move to a new city , we are at first confused about its geography
. then we find a few reference points , and we make a rough mental map of the city in terms of these points .
perhaps we see that there is part of the city on the hills and part on the plane .
as time goes on , the map gets better .
but there are moments , in which we suddenly realize that we had it wrong .
perhaps there were indeed two areas with hills , and we were previously confusing the two .
or we had mistaken a big red building for the city hall , when it was only a residential construction .
so we adjourn the mental map .
sometime later , we have learned names and features of neighbors and streets ; and the hills , as references , fade away .
the neighbors structure of knowledge is more effective that the hill / plane one the structure changes , but the knowledge increases . and the big red building ,
now we know it , is not the city hall , and we know it forever .
there are discoveries that are forever .
that the earth is not the center of the universe , that simultaneity is relative .
that we do not get rain by dancing . these are steps humanity takes , and does not take back . some of these discoveries amount simply to cleaning our thinking from wrong , encrusted , or provisional credences . but also discovering classical mechanics , or discovering electromagnetism , or quantum mechanics , are discoveries forever . not because the details of these theories can not change , but because we have discovered that a large portion of the world admits to be understood in certain terms , and this is a _ fact _ that we will have to keep facing forever .
one of the thesis of this essay , is that general relativity is the expression of one of these insights , which will stay with us `` forever '' .
the insight is that the physical world does not have a stage , that localization and motion are relational only , that diff - invariance ( or something physically analogous ) is required for any fundamental description of our world .
how can a theory be effective even outside the domain for which it was found ?
how could maxwell predict radio waves , dirac predict antimatter and gr predict black holes ?
how can theoretical thinking be so magically powerful ?
of course , we may think that these successes are chance , and historically deformed perspective .
there are hundred of theories proposed , most of them die , the ones that survive are the ones remembered .
there is alway somebody who wins the lottery , but this is not a sign that humans can magically predict the outcome of the lottery .
my opinion is that such an interpretation of the development of science is unjust , and , worse , misleading .
it may explain something , but there is more in science .
there are tens of thousand of persons playing the lottery , there were only two relativistic theories of gravity , in 1916 , when einstein predicted that the light would be defected by the sun precisely by an angle of 1.75 .
familiarity with the history of physics , i feel confident to claim , rules out the lottery picture .
i think that the answer is simpler .
somebody predicts that the sun will rise tomorrow , and the sun rises .
it is not a matter of chance ( there are nt hundreds of people making random predictions on each sort of strange objects appearing at the horizon ) .
the prediction that tomorrow the sun will rise , is sound .
however , it is not granted either .
a neutron star could rush in , close to the speed of light , and sweep the sun away .
more philosophically , who grants me the right of induction ?
why should i be confident that the sun would rise , just because it has been rising so many time in the past ?
i do not know the answer to _ this _ question .
but what i know is that the predictive power of a theory beyond its own domain is _ precisely of the same sort .
_ simply , we learn something about nature ( whatever this mean ) . and what we learn is effective in guiding us to predict nature s behavior .
thus , the spectacular predictive power of theoretical physics is nothing less and nothing more than common induction . and it is as comprehensible ( or as incomprehensible ) as my ability to predict that the sun will rise tomorrow .
simply , nature around us happens to be full of regularities _ that we understand _ , whether or not we understand why regularities exist at all .
these regularities give us strong confidence -although not certainty- that the sun will rise tomorrow , as well as in the fact that the basic facts about the world found with qm and gr will be confirmed , not violated , in the quantum gravitational regimes that we have not empirically probed .
this view is not dominant nowadays in theoretical physics .
other attitudes dominate .
the `` pragmatic '' scientist ignores conceptual questions and physical insights , and only cares about developing a theory .
this is an attitude , that has been successful in the sixties in getting to the standard model .
the `` pessimistic '' scientist has little faith in the possibilities of theoretical physics , because he worries that all possibilities are open , and anything might happen between here and the planck length .
the `` wild '' scientist observes that great scientists had the courage of breaking with old and respected ideas and assumptions , and explore new and strange hypothesis . from this observation
, the `` wild '' scientist concludes that to do great science one has to explore strange hypotheses , and _ violate respected ideas_. the wildest the hypothesis , the best .
i think wilderness in physics is sterile .
the greatest revolutionaries in science were extremely , almost obsessively , conservative .
so was certainly the greatest revolutionary , copernicus , and so was planck .
copernicus was pushed to the great jump from his pedantic labor on the minute technicalities of the ptolemaic system ( fixing the equant ) .
kepler was forced to abandon the circles by his extremely technical work on the details of mars orbit .
he was using ellipses as approximations to the epicycle - deferent system , when he begun to realize that the approximation was fitting the data better than the ( supposedly ) exact curve . and extremely conservative were also einstein and dirac .
their vertiginous steps ahead were not pulled out of the blue sky .
they did not come from violating respected ideas , but , on the contrary , from respect towards physical insights . in physics
, novelty has always emerged from new data and from a humble , devoted interrogation of the old theories . from turning these theories around and around , immerging into them , making them clash , merge ,
talk , until , through them , the missing gear could be seen . in my opinion ,
precious research energies are today lost in these attitudes .
i worry that a philosophy of science that downplays the component of factual knowledge in physical theories might have part of the responsibility .
if a physical theory is a conceptual structure that we use to organize , read and understand the world , then scientific thinking is not much different from common sense thinking .
in fact , it is only a better instance of the same activity : thinking about the world .
science is the enterprise of continuously exploring the possible ways of thinking about the world , and constantly selecting the ones that work best .
if so , there can not be any qualitative difference between the theoretical notions introduced in science and the terms in our everyday language .
a fundamental intuition of classical empiricism is that nothing grants us the `` reality '' of the referents of the notions we use to organize our perceptions .
some modern philosophy of science has emphasized the application of this intuition to the concepts introduced by science .
thus , we are warned to doubt the `` reality '' of the theoretical objects ( electrons , fields , black holes ) .
i find these warning incomprehensible .
not because they are ill founded , but because they are not applied consistently .
the fathers of empiricism consistently applied this intuition to _ any _ physical object . who grants me the reality of a chair ?
why should a chair be more than a theoretical concept organizing certain regularities in my perceptions ?
i will not venture here in disputing nor in agreeing with this doctrine .
what i find incomprehensible is the position of those who grant the solid status of reality to a chair , but not to an electron .
the arguments against the reality of the electron apply to the chair as well .
the arguments in favor of the reality of the chair apply to the electron as well . a chair , as well as
an electron , is a concept that we use to organize , read and understand the world .
they are equally real .
they are equally volatile and uncertain .
perhaps , this curious schizophrenic attitude of being antirealist with electrons and iron realist with chairs is the result of a complex historical evolution .
first there was the rebellion against `` metaphysics '' , and , with it , the granting of confidence to science alone .
from this point of view , metaphysical questioning on the reality of chairs is sterile
true knowledge is in science .
thus , it is to scientific knowledge that we apply empiricist rigor .
but understanding science in empiricists terms required making sense of the raw empirical data on which science is based . with time , the idea of raw empirical data showed more and more its limits .
the common sense view of the world was reconsidered as a player in our picture of knowledge .
this common sense view should give us a language and a ground from which to start the old anti - metaphysical prejudice still preventing us , however , from applying empiricist rigor to this common sense view of the world as well .
but if one is not interested in questioning the reality of chairs , for the very same reason why should one be interested in questioning the `` reality of the electrons '' ?
again , i think this point is important for science itself .
the factual content of a theory is our best tool .
the faith in this factual content does not prevent us from being ready to question the theory itself , if sufficiently compelled to do so by novel empirical evidence or by putting the theory in relation to other things _ we know _ about the world .
scientific antirealism , in my opinion , is not only a short sighted application of a deep classical empiricist insight ; it is also a negative influence over the development of science .
h. stein ( 1999 ) has recently beautifully illustrated a case in which a great scientist , poincar , was blocked from getting to a major discovery ( special relativity ) by a philosophy that restrained him from `` taking seriously '' his own findings .
science teaches us that our naive view of the world is imprecise , inappropriate , biased .
it constructs better views of the world .
electrons , if anything at all , are `` more real '' that chairs , not `` less real '' , in the sense that they ground a more powerful way of conceptualizing the world . on the other hand ,
the process of scientific discovery , and the experience of this century in particular , has made us painfully aware of the provisional character of _ any _ form of knowledge .
our mental and mathematical pictures of the world are only mental and mathematical pictures .
this is true for abstract scientific theories as well as from the image we have of our dining room . nevertheless , the pictures are powerful and effective and we ca nt do any better than that .
so , is there anything we can say with confidence about the `` real world '' ?
a large part of the recent reflection on science has taught us that row data do not exist , and that any information about the world is already deeply filtered and interpreted by the theory .
further than that , we could even think , as in the dream of berkeley , that there is no `` reality '' outside there .
the european reflection ( and part of the american as well ) has emphasized the fact that truth is always internal to the theory , that we can never exit language , we can never exit the circle of discourse within which we are speaking
. it might very well be so .
but , if the only notion of truth is internal to the theory , then _ this internal truth _ is what we mean by truth .
we can not exit from our own conceptual scheme .
we can not put ourself outside our discourse .
outside our theory . there may be no notion of truth outside our own discourse .
but it is precisely `` from within the language '' that we can assert the reality of the world .
and we certainly do so .
indeed , it is more than that : it is structural to our language to be a language _ about _ the world , and to our thinking to be a thinking _ of _ the world .
therefore , precisely because there is no notion of truth except the one in our own discourse , precisely for this reason , there is no sense in denying the reality of the world .
the world is real , solid , and understandable by science .
the best we can say about the physical world , and about what is there in the world , is what good physics says about it . at the same time , our perceiving , understanding , and conceptualizing the world is in continuous evolution , and science is the form of this evolution . at every stage ,
the best we can say about the reality of the world is precisely what we are saying . the fact we will understand it better later on does not make our present understanding less valuable , or less credible .
a map is not false because there is a better map , even if the better one looks quite different .
searching for a fixed point on which to rest our restlessness , is , in my opinion , naive , useless and counterproductive for the development of science .
it is only by believing our insights and , at the same time , questioning our mental habits , that we can go ahead .
this process of cautious faith and self - confident doubt is the core of scientific thinking . exploring the possible ways of thinking of the world , being ready to subvert ,
if required , our ancient prejudices , is among the greatest and the most beautiful of the human adventures . quantum gravity , in my view , in its effort to conceptualize quantum spacetime , and to modify in depth the notion of time , is a step of this adventure .
* ashtekar a , rovelli c , smolin l ( 1992 ) . weaving a classical metric with quantum threads , _ physical review letters _ * 69 * , 237 .
* ashtekar a , lewandowski j ( 1997a ) . quantum theory of gravity i : area operators , _ class and quantum grav _ * 14 * , a55a81 .
* ashtekar a , lewandowski j ( 1997b ) . quantum theory of geometry ii : volume operators , gr - qc/9711031 . * baez j ( 1997 ) .
spin networks in nonperturbative quantum gravity , in _ the interface of knots and physics _ , ed l kauffman ( american mathematical society , providence ) . * baez j ( 1998 ) .
spin foam models , _ class quantum grav _ , * 15 * , 18271858 . * baez j ( 1999 ) . in `` physics meets philosophy at the planck scale '' ,
c callender n hugget eds , cambridge university press , to appear .
* barbour j ( 1989 ) . _
absolute or relative motion ? _ ( cambridge university press , cambridge ) . * barret j , crane l ( 1998 ) .
relativistic spin networks and quantum gravity , _ journ math phys _ * 39 * , 32963302 . * belot g ( 1998 ) . why general relativity does need an interpretation , _ philosophy of science _ , * 63 * , s80s88 . * connes a and rovelli c ( 1994 ) .
von neumann algebra automorphisms and time versus thermodynamics relation in general covariant quantum theories , _ classical and quantum gravity _ * 11 * , 2899 .
* crane l ( 1991 ) .
2d physics and 3d topology , _ comm math phys _ * 135 * , 615640 . * descartes r ( 1983 ) : _ principia philosophiae _ , translated by vr miller and rp miller ( reidel , dordrecht [ 1644 ] ) .
* earman j ( 1989 ) . _
world enough and spacetime : absolute versus relational theories of spacetime _
( mit press , cambridge ) .
* earman j , norton j ( 1987 ) .
what price spacetime substantivalism ?
the hole story , _ british journal for the philosophy of science _ , * 38 * , 515525 . * isham c ( 1999 ) . in ``
physics meets philosophy at the planck scale '' , c callender n hugget eds , cambridge university press , to appear .
* newton i ( 1962 ) : _ de gravitatione et aequipondio fluidorum _ , translation in ar hall and mb hall eds _ unpublished papers of isaac newton _ ( cambridge university press , cambridge ) . *
norton j d ( 1984 ) .
how einstein found his field equations : 1912 - 1915 , _ historical studies in the physical sciences _ , * 14 * , 253315 . reprinted in _ einstein and the history of general relativity
: einstein studies _ , d howard and j stachel eds . , vol.i , 101 - 159 ( birkhuser , boston ) . *
penrose r ( 1995 ) . _
the emperor s new mind _ ( oxford university press ) * reisenberger m , rovelli c ( 1997 ) . sum over surfaces form of loop quantum gravity , _ physical review _ , * d56 * , 34903508 , gr - qc/9612035 .
* rovelli c ( 1991a ) .
what is observable in classical and quantum gravity ?
, _ classical and quantum gravity _ , * 8 * , 297 .
* rovelli c ( 1991b ) .
quantum reference systems , _ classical and quantum gravity _ , * 8 * , 317 .
* rovelli c ( 1991c ) .
quantum mechanics without time : a model , _ physical review _ , * d42 * , 2638 .
* rovelli c ( 1991d ) .
time in quantum gravity : an hypothesis , _ physical review _
, * d43 * , 442 .
* rovelli c ( 1991e ) .
quantum evolving constants . _ physical review _ * d44 * , 1339 .
* rovelli c ( 1993a ) .
statistical mechanics of gravity and thermodynamical origin of time , _ classical and quantum gravity _ , * 10 * , 1549 .
* rovelli c ( 1993b ) . the statistical state of the universe , _ classical and quantum gravity _ , * 10 * , 1567 .
* rovelli c ( 1993c ) . a generally covariant quantum field theory and a prediction on quantum measurements of geometry , _ nuclear physics _ ,
* b405 * , 797 .
* rovelli c ( 1995 ) .
analysis of the different meaning of the concept of time in different physical theories , _ il nuovo cimento _ , * 110b * , 81 .
* rovelli c ( 1996 ) .
relational quantum mechanics , _ international journal of theoretical physics _ , * 35 * , 1637 .
* rovelli c ( 1997a ) .
half way through the woods , in _ the cosmos of science _ , j earman and jd norton editors , ( university of pittsburgh press and universitts verlag konstanz ) . * rovelli , c. ( 1997b ) loop quantum gravity , living reviews in relativity ( refereed electronic journal ) , http://www.livingreviews.org/ articles / volume1/1998 - 1rovelli ; gr - qc/9709008 * rovelli c ( 1998 )
. incerto tempore , incertisque loci : can we compute the exact time at which the quantum measurement happens ? , _ foundations of physics _
, * 28 * , 10311043 , quant - ph/9802020 .
* rovelli c ( 1999 ) .
strings , loops and the others : a critical survey on the present approaches to quantum gravity , in _ gravitation and relativity : at the turn of the millennium _ , n dadhich j narlikar eds ( poona university press ) , gr - qc/9803024 . *
rovelli c and smolin l ( 1988 ) . knot theory and quantum gravity , _ physical review letters _ , * 61 * , 1155 .
* rovelli c and smolin l ( 1990 ) . loop space representation for quantum general relativity , _ nuclear physics _ , * b331 * , 80 .
* rovelli c , smolin l ( 1995a ) .
spin networks and quantum gravity , _ physical review _ , * d 53 * , 5743 . *
rovelli c and smolin l ( 1995b ) .
discreteness of area and volume in quantum gravity , _ nuclear physics _ * b442 * , 593 .
erratum : _ nuclear physics _ * b 456 * , 734 . *
smolin l ( 1997 ) .
the future of spin networks , gr - qc/9702030 . * stachel j ( 1989 ) .
einstein search for general covariance 1912 - 1915 , in _ einstein studies _ , d howard and j stachel eds , vol 1 , 63 - 100 ( birkhuser , boston ) . *
stein h ( 1999 ) .
physics and philosophy meet : the strange case of poincar . unpublished .
* zeilinger a , in _ gravitation and relativity : at the turn of the millennium _ , n dadhich j narlikar eds ( poona university press ) . | i discuss nature and origin of the problem of quantum gravity .
i examine the knowledge that may guide us in addressing this problem , and the reliability of such knowledge . in particular , i discuss the subtle modification of the notions of space and time engendered by general relativity , and how these might merge into quantum theory .
i also present some reflections on methodological questions , and on some general issues in philosophy of science which are are raised by , or a relevant for , the research on quantum gravity . _ to appear on `` physics meets philosophy at the planck scale '' _ _ c callender n hugget eds , cambridge university press _ |
the star rs ophiuchi ( rs oph ) is located in the direction nearby to the galactic center ( @xmath7 ) at a distance @xmath8 @xcite .
the light of rs oph is significantly attenuated by interstellar extinction .
@xcite determined from hi 21 cm measurements the interstellar absorbing column , @xmath9@xmath10 , and @xcite derived the color excess @xmath11mag from ultraviolet _ iue _ spectra . according to the relationship between @xmath12 and @xmath13 ( e.g. * ? ? ?
* ) , both these parameters are consistent .
rs oph is a recurrent symbiotic nova comprising a late - type , k7iii , giant @xcite and a wd with a mass close to the chandrasekhar limit ( e.g. * ? ? ? * and references therein ) , in a 456-day orbit @xcite .
its nova - like outbursts are characterized with brightening by about 7mag and a recurrence period of about 20 years .
historically , 6 eruptions have been recorded unambiguously . the first one in 1898 and the last one on 2006 february 12.83 @xcite .
the recurrence period of approximately 20 years and a bright peak magnitude , @xmath14 , made rs oph a good target for multifrequency observational campaigns ( e.g. * ? ? ? * and references therein ) . at the beginning of the last two outbursts of 1985 and 2006
, observations revealed a non - spherical shape of the nova ejecta , ( i ) on the radio maps ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , ( ii ) by interferometric technique in the near and mid - ir ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , ( iii ) by the _
hst _ imaging in the optical @xcite , ( iv ) by the _
chandra _ x - ray observatory detected as an extended x - ray emission elongated in the line with the extended infrared emission @xcite and ( v ) by the asymmetric line profiles measured on day 13.9 , interpreted by @xcite as a result of the shock collimation due to equatorial circumstellar density enhancement .
particularly , highly collimated jet - like ejection was indicated by the optical spectroscopy during the first 30 days after the eruption @xcite .
the 2006 outburst was intensively monitored with the _ swift _ and the _ rossi x - ray timing explorer_. prior to day 26 , the behaviour of the x - ray flux was described by the evolution of shock systems established as the high - velocity ejecta impacted the red giant wind @xcite .
the evolving shock in the 2006 outburst of rs oph was analyzed in detail by @xcite using the high - resolution x - ray grating observations with _
chandra _ and _ xmm - newton_. the x - ray observations demonstrated that the supersoft x - ray phase of nova rs oph started rapidly from day 26 ( see also * ? ? ?
the detailed spectral evolution of the 2006 outburst in the optical was reported by @xcite .
he suggested that the profiles of the prominent emission lines resulted from a rapidly expanding ( @xmath15kms@xmath16 ) , low density part and a slowly moving ( @xmath17kms@xmath16 ) high density part in the ejecta , which are probably related to the expansion along the polar and the equatorial regions , respectively . in this contribution
i aim to determine physical parameters of the nova rs oph from its explosion to the first emergence of the supersoft x - rays ( day 26 ) by using the method of multiwavelength modelling the sed . based on the results from spectral fits ,
i also explore the ionization structure of the nova during the first 4 days of its outburst .
section 2 introduces multi - band observations at selected days , and the results of their sed - fitting analysis are given in sect .
discussion and summary are found in sects . 4 and 5 , respectively .
to map the evolution of the fast nova rs oph by the multiwavelength approach , simultaneous observations covering a large wavelength range , taken at different days after the explosion , are required .
rapid changes in the sed following the blast require frequent sets of observations . according to these conditions , it was possible to select ( near-)simultaneous observations at / around day 1 , 3.8 , 7.3 , 14.6 , 19.5 and 26 after the optical maximum of the rs oph outburst .
selection of the appropriate data benefited from a strong similarity of the recent 1985 and 2006 outbursts in the supersoft x - ray , optical and near - ir light curve profiles ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
therefore , observations taken at the same time after the optical maximum of these outbursts were assumed to be simultaneous .
the uv to near - ir observations were dereddened with @xmath13 = 0.73 and the resulting parameters were scaled to the distance of 1.6kpc ( see sect . 1 ) .
the log of the used observations is given in table 1 .
.log of observations [ cols="^,^,^,^",options="header " , ] @xmath18 fixed value , @xmath19 = @xmath20 ( sect .
3.4 , appendix a ) .
the spectral range between @xmath21 and @xmath22 was dominated by a flat harder component .
this part of the x - ray spectrum was compared with a function @xmath23 where the flux of photons @xmath24 at the energy @xmath25 ( kev ) is in photons / s/@xmath26/kev , @xmath27 is a scaling factor and @xmath28 , @xmath29 are fitting parameters
. figure [ fig : seds ] shows a model for @xmath30photons / s/@xmath26/kev , @xmath31 , @xmath32kev and @xmath33@xmath10 . the absorbed flux of this harder ( 6.519.3 ) component was @xmath34@xmath35 , which corresponds to the unabsorbed value of @xmath36@xmath35 .
i note that the total ( 6.531.0 ) absorbed flux of @xmath37@xmath35 is comparable with that obtained by @xcite .
the origin of this component is probably connected with the optically thin , a few millions of kelvins hot , collisionally - ionized plasma @xcite , heated by shock waves in the highly - supersonic mass - outflow .
multiwavelength modelling of the sed revealed the simultaneous presence of a warm stellar ( @xmath38 ) and nebular ( @xmath39 ) components of radiation in the spectrum of rs oph during the first four days after its explosion .
the stellar component is characterized with @xmath40k and @xmath41@xmath42 , which corresponds to production of @xmath43 photons per second that are capable of ionizing hydrogen .
however , the nebular component is characterized with @xmath3@xmath4 , which corresponds to @xmath44 recombinations per second @xmath45 , and thus can not be generated by the observed stellar source of radiation ( @xmath46 ; see eq .
( [ eq : lph ] ) ) .
this implies the presence of a hot ionizing source in the system , which is not seen directly by the outer observer .
such properties of radiative components were revealed in the spectrum of all symbiotic stars with a high orbital inclination during their outbursts .
the corresponding spectrum was called a two - temperature - type ( see sect . 5.3.4 of * ? ? ?
this puts constraints for the ionization and geometrical structure of the circumstellar matter surrounding the burning wd .
according to @xcite the source of the warm stellar component can be ascribed to the wd s pseudophotosphere that is represented by the outer flared rim of an optically thick disk - like formation around the hot star .
the circumstellar material located within the remainder of the sphere is ionized , and thus converts a fraction of the hot star ( here the burning wd ) radiation into the strong nebular emission ( see fig . [
fig : sketch ] ) .
ever since the very beginning of the nova explosion ( day 1.16 ) , very broad emission lines of hydrogen were observed in the optical spectrum ( e.g. * ? ? ?
@xcite fitted the broad wings of h@xmath28 from day 1.38 by the wind model of @xcite , in which a fraction of the wind from the central star is blocked by a disk around the wd s equator , and thus causes its bipolarity .
the very high level of the continuum around the h@xmath28 line of @xmath47@xmath48 ( see fig . [
fig : seds ] , left top panel ) yields its luminosity , @xmath49@xmath50 and the mass loss rate of a few @xmath51@xmath6 @xcite .
consequently , the h@xmath28 luminosity was generated in a volume with the emission measure _ _
em__@xmath52@xmath4 for @xmath53erg@xmath54s@xmath16 , which is consistent with that of the nebular continuum .
as both the nebular continuum and the hydrogen emission lines originate in the same region , the h zone , producing the strong nebular component of radiation in the sed , can be ascribed to the fast ionized wind from the outbursting wd .
in addition , the biconical structure and the nature of the ionized zone was also supported by the presence of satellite emission components located at @xmath55kms@xmath16 with respect to the main cores of hydrogen lines , measured from the very beginning of the outburst and being recognizable to day @xmath56 @xcite .
these emission features can be interpreted as a result of highly collimated bipolar mass outflow .
this was confirmed by the radio 43ghz image from day 55 that showed directly highly collimated outflows produced by free - free emission from the thermal plasma @xcite .
the presence of bipolar jets requires the presence of the inner disk , throughout which sufficient amount of mass must be accreted to balance the outflow via the jets ( e.g. * ? ? ? * ) .
thus the presence of a large disk - like formation encompassing the burning wd could also be responsible for keeping the super - eddington luminosity of the burning wd for a long time .
\(i ) the bolometric luminosity of the burning wd can not be determined unambiguously during this period of the nova evolution .
only its lower limit can be estimated from the components of radiation determining the observed two - temperature - type of the spectrum .
the disk - like shell around the wd converts a fraction of its radiation into the warm stellar component ( here approximated by blackbody ) .
its luminosity at day @xmath57 was @xmath58@xmath42 ( table 2 ) .
another fraction of the burning wd s luminosity is reprocessed by ionization / recombination events into the nebular component .
its luminosity can be expressed as ( see eq .
( 8) in paper i ) @xmath59 where the nebula is assumed to radiate under conditions of case b. the _ em_(day @xmath60@xmath4 corresponds to @xmath61@xmath42 .
this also represents a lower limit , because a fraction of ionizing photons is not detectable ( they are not converted into the nebular photons , i.e. @xmath62 , see appendix a ) and , in addition , the nebula can be partially optically thick ( i.e. the measured _ em _ is smaller than that originally created by ionizations ) .
thus the bolometric luminosity of the burning wd is @xmath63 at day @xmath57 .
note that such adding of luminosities from different sources to the bolometric luminosity is possible only in the case of the two - temperature - type of the spectrum ( i.e. the warm pseudophotosphere is far below the capability of giving rise to the observed nebular component of radiation ) .
\(ii ) according to the bipolar ionization structure of the burning wd , the vertical extension of the warm pseudophotosphere is limited by the orbital inclination to get hidden the ionizing source in direction of the observer .
its luminosity can be written as @xmath64 where the surface @xmath65 of the pseudophotosphere is represented by a belt on a sphere with the radius @xmath66 , located around its equator and extended to the latitude @xmath67 ( see fig .
[ fig : sketch ] ) .
@xmath68 is its effective radius ( see eq . ( 6 ) in paper i ) .
then the radial extension of such the pseudophotosphere can be expressed as @xmath69 for @xmath70@xmath1 ( table 2 ) and the limiting case @xmath71 , where @xmath72 , we get @xmath73@xmath1 . other possibilities , @xmath74 yield @xmath75@xmath1 . at day 3.8 after the nova explosion
, @xcite used the kin to find the structure of the circumstellar material around the star .
they determined the angular size of the mid - infrared continuum emitting material to 4.06.2mas , which corresponds to linear size of 6.49.9(d/1.6kpc)au .
as the n - band region is dominated by the nebular emission ( see fig . [
fig : seds ] ) , which is due to the fast ionized wind from the wd ( see sect .
4.1.1 ) , its maximum extension ( 13au for the velocity of 3000kms@xmath16 and 3.8 days ) is consistent with that detected by the kin .
the measurements obtained by the kin in the inner spatial regime showed evidence of enhanced neutral atomic hydrogen emission and atomic metals located near the wd .
this could reflect the h zone in the ionization structure inferred from the model sed ( fig . [
fig : sketch ] ) .
@xcite measured the near - ir size of rs oph using the iota and keck interferometer at the same day .
they found a significantly asymmetric emission with the characteristic size of fwhm@xmath76mas .
observations for the following 60 days did not indicate any change , suggesting thus that the ir emission is due to the @xmath77 and @xmath78 transitions in thermal plasma rather than from a hot gas in the expanding shock as considered in current popular models .
also these results the size and the nature of the ir emission are consistent with the model sed here . using the amber / vlt interferometer , @xcite measured the extension of the milliarcsecond - scale emission in the k band continuum , in the br@xmath79 and he2.06@xmath80 m lines at day 5.5 .
observations showed that the emission from all investigated spectral regions is highly flattened with the ratio @xmath81 , having the same position angle and a typical gaussian extension of @xmath82mas for the continuum .
thus , it is possible to conclude that the biconical ionization structure of the circumstellar matter around the wd in rs oph , as inferred from multiwavelength modelling the sed , is supported by direct interferometric measurements using iota , keck and amber / vlti performed during a few days after the eruption .
the model sed for day 7.3 showed that the shift of the sed maximum from the optical to the far - uv ( or even beyond ) happened during a relatively short time of @xmath83 days ( see fig .
[ fig : seds ] ) .
a sudden decrease of the optical depth causing the shrinkage of the inflated wd from @xmath84@xmath1 to less than 2.1@xmath1 ( table
2 ) would require removal of a significant amount of dense , optically thick , material in the line of sight .
the corresponding time of such the change should be proportional to the dynamical time scale of the inflated wd , @xmath85 where @xmath86 is the density of the matter giving rise to the gravitational force .
so , for the wd mass of @xmath87@xmath88 , its effective radius 160@xmath1 would collapse during @xmath89 days , which is by a factor of @xmath90 larger than the time scale of the observed change in the effective radius . therefore , a gradual shrinking of the wd pseudophotosphere due to the expansion of the outflowing material can not be the cause of its sudden disappearance between day 3.8 and 7.3 .
the sudden change in the sed can be a result of a change in the position of the ionization boundary with respect to the line of sight . if the opening angle of the h zone @xmath91 ( see fig . [
fig : sketch ] ) shrinks so that @xmath92 , the burning wd will rise up the horizon represented by the h / h boundary .
a shrinkage / enlargement of the @xmath93 angle can be a result of a decreasing mass loss rate from the wd .
recently , @xcite found that the ionization structure of the hot components in symbiotic binaries during outbursts is similar to that sketched in the figure [ fig : sketch ] here ( see their figs . 1 and 6 ) .
in their model , the opening angle of the h zone , @xmath94 , is related to the mass loss rate from the wd as @xmath95 .
this implies that a relatively small decrease in the @xmath96 can yield a large increase in the @xmath94 angle ( see fig . 1 of * ? ? ?
during the first days , when the @xmath96 was at a maximum , @xmath97 ( i.e. the hot wd was below the h / h boundary ) , because we observed the warm pseudophotosphere .
a small decrease in @xmath96 could lead to opening of the h zone so that @xmath98 , i.e. the burning wd raised up the h / h horizon and thus we could observe directly its hot @xmath99@xmath1 photosphere at day 7.3 .
this could happen in a short time , because of a relatively low value of @xmath100 .
a transient increase of the _ em _ at day 7.3 supports this view .
multiwavelength modelling the sed revealed that the luminosity of the wd was super - eddington during the whole period of its nuclear burning ( table 2 here and table 2 of the following paper ) .
prior to the emergence of supersoft x - rays , only the lower limit of the wd s luminosity could be determined . during the presence of the two - temperature - type of the spectrum ( i.e. for the first 4 days )
, it was given by the integral of directly measured fluxes ( or their model ) scaled with the distance .
the model sed at day @xmath57 and 3.8 corresponded to the luminosity of 6 and 2@xmath101 , respectively , for @xmath102@xmath88 .
however , its true value was higher , because both the warm shell and the nebula converted only a fraction of the wd radiation into direction of the observer ( sect .
4.1.2 , eq .
( [ eq : lbol ] ) , fig .
[ fig : sketch ] ) . from day 7.3 , measuring only the short rayleigh - jeans tail of the hot wd radiation , only the lower limit of its temperature and luminosity could be determined ( see appendix a ) . using the emission measure from the model
sed and the @xmath38 flux from the far - uv spectrum , allowed to determine the minimum temperature of the wd
photosphere , @xmath103k ( eq . ( [ eq : tmin2 ] ) , table 2 ) and the minimum of its luminosity , @xmath104 ( eq . ( [ eq : lth ] ) ) , at which the burning wd is just capable of producing the observed _ em _ during day 7.3 19.5 .
note that the luminosity of the nebular component itself corresponds to 1.24@xmath101 at day 7.3 ( table 2 ) . the emergence of the first supersoft x - rays from the wd at day 26 allowed to determine its luminosity unambiguously to @xmath105@xmath101 ( sect .
such a highly super - eddington luminosity contradicts theoretical predictions , treating the nova event as a simple nuclear bomb explosion .
according to @xcite the luminosity can exceed the eddington limit only at the very beginning of the nova eruption with a rapid decline below it on a time - scale of hours .
however , the opposite cases were reported by some authors already during 70 s and 80 s decades of the last century .
for example , @xcite showed that the energy output of the nova fh ser was well above the eddington limit for about two months after its optical maximum .
the super - eddington state was theoretically investigated by @xcite , who considerred a reduction of the effective opacity due to the rise of a ` porous layer ' above the convective zone of the burning wd ( see fig . 1 of * ? ? ?
these conditions enlarges the eddington luminosity well above its classical value , calculated for the thomson scattering opacity .
the super - eddington state also accelerates a thick continuum driven wind at high rates . the mass loss through the wind at a high rate of a few @xmath106@xmath6 during the hydrogen burning period of rs oph @xcite is consistent with that suggested for the super - eddington state ( eq .
( 30 ) of * ? ? ?
* for @xmath107 = 20 - 50kms@xmath16 ) .
in addition , the presence of the bipolar jets for the first 30 days @xcite , which suggests accretion throughout the disk at a high rate , also supports the super - eddington state .
for example , accretion at @xmath108@xmath6 can release its binding energy @xmath109@xmath42 , and thus help to sustain the high luminosity of the accretor for a long time .
finally , it is important to note that the high energy output released during the outburst of rs oph , as given by the global model sed of this paper , is more than a factor of 10 larger than that so far derived .
this requires a very high accretion rate , which is far above that can be obtained from a spherical wind mass loss of the giant ( e.g. * ? ? ?
considering that the primary source of the material powering the outburst is the wind from the giant , then a different , more efficient mechanism of the mass transfer onto the wd must be in the effect .
this could be realized by the so - called wind roche - lobe overflow ( wrlof ) , which can occur when the wind acceleration zone is stretched out to the roche lobe of the giant ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in this case
the wind of the giant is focused towards the orbital plane and in particular towards the wd .
the accretion rate predicted in the wrlof regime can be 100 times larger than the rate expected from standard bondi - hoyle accretion ( see * ? ? ?
more detailed discussion on the accretion mechanism is presented in the following paper , _
rs ophiuchi : the supersoft x - ray phase and beyond _ , where other critical parameters are determined from the model seds during the sss phase and following quiescence .
in this paper i applied the method of multiwavelength modelling the sed ( see paper i ) to the recurrent symbiotic nova rs oph from its explosion to the sss phase ( to day 26 ) . the model seds revealed the presence of a strong stellar as well as nebular component of radiation in the spectrum .
the contribution from the cool giant was negligible with respect to these components ( sect . 3 , fig .
[ fig : seds ] , table 2 ) .
the main results can be summarized as follows . 1
. the radiation from the giant corresponds to its effective temperature @xmath110k , the radius , @xmath111@xmath1 and the luminosity , @xmath112@xmath50 .
the giant underfills deeply its roche lobe for d = 1.6kpc ( sect .
2 . during the first 4 days after the explosion , the model sed identified a biconical ionization and geometrical structure of the nova .
the @xmath0k warm pseudophotosphere was represented by a belt on a sphere with the radius of 160200@xmath1 , located around the wd s equator and extended to the latitude @xmath113 .
the remaining space around the wd s poles was ionized , producing a strong nebular emission ( @xmath3@xmath4 ) via the fast ionized wind from the wd ( see fig .
[ fig : sketch ] , sect .
the luminosity of the burning wd was super - eddington for the whole investigated period , although it was possible to determine only its lower limit prior to the emergence of the supersoft x - ray radiation ( sect .
4.3 , table 2 ) .
+ during the first 4 days , the luminosity @xmath114 for @xmath102@xmath88 . here
the lower limit is given by the fit of the model sed to the total measured spectrum , because both the neutral and the ionized region convert only a fraction of the wd s radiation into direction of the observer ( sect .
+ from day 7.3 , the two - temperature - type of the spectrum suddenly disappeared , because the h zone opened so that we could observe directly the hot wd ( sect .
4.2 ) . in this case
lower limits of the wd s temperature and luminosity were determined from the condition that the corresponding flux of ionizing photons was just capable of producing the observed _ em _ ( sect . 4.3 , eq . (
[ eq : lph ] ) ) .
here the model seds suggested @xmath115 .
+ at day 26 , the first detection of the supersoft x - ray fluxes allowed to determine the wd s luminosity unambiguously to @xmath105@xmath101 ( sect .
3.5 , fig .
[ fig : seds ] ) .
the long - lasting super - eddington luminosity and the mass loss at a very high rate are consistent with theoretical considerations treating the super - eddington state as a result of a reduction the effective opacity of the wd atmosphere .
however , the appearance of highly super - eddington luminosity just at the beginning of the sss phase ( day 26 ) is enigmatic , unless this could also be connected with a relevant reduction of the effective opacity .
the presence of collimated jets during the burning phase suggests an accretion throughout a disk during this period , which can help to balance the high energy output for a long time .
this means that the disk was not destroyed by the eruption .
5 . the high energy output released during the outburst of rs oph requires a very high accretion rate , which could be realized in the wrlof regime . further analysis and discussion on the accretion mechanism
is found in the following paper , _
rs ophiuchi : the supersoft x - ray phase and beyond_.
_ iue _ spectra presented in this paper were obtained from the mast .
the author thanks manfred hanke for extracting the _ xmm - newton _ observation from the archive and richard barry for valuable discussion about the keck interferometer nuller observing regimes .
this research has been in part supported by the project no .
sla/103115 of the alexander von humboldt foundation and by the slovak academy of sciences under a grant vega no .
2/0002/13 .
abate , c. , pols , o. r. , izzard , r. g. , mohamed , s. s. , de mink , s. e. : 2013 , a&a , 552 , a26 anupama , g. c. : 2008 in asp conf .
401 , rs ophiuchi ( 2006 ) and the recurent nova phenomenon , eds .
a. evans , m. f. bode , t. j. obrien and m. j. darnley , ( san francisco : asp ) , 251 banerjee , d. p. k. , das , r. k. , ashok , n. m. : 2009 , mnras , 399 , 357 barry , r. k. , mukai , k. , sokoloski , j. l. , danchi , w. c. , hachisu , i. , evans , a. , gehrz , r. , mikolajewska , j. : 2008a , in evans a. , bode m. f. , obrien t. j. , darnley m. j. , eds , asp conf . ser .
401 , rs ophiuchi ( 2006 ) and the recurrent nova phenomenon .
asp , san francisco , p. 52
barry , r. k. , w. c. danchi , w. a. , traub , et al . : 2008b , apj , 677 , 1253 bode , m. f. : 1987 , in bode m. f. , ed , rs ophiuchi ( 1985 ) and the recurrent nova phenomenon .
vnu science press , utrecht , p. 241
bode , m. f. , obrien , t. j. , osborne , j. p. , page , k. l. , senziani , f. , skinner , g. k. , starfield , s. , ness , j .- u . : 2006 , apj , 652 , 629 bode , m. f. , harman , d. j. , obrien , t. j. , et al . : 2007 , apj , 665 , l63 brandi , e. , quiroga , c. , mikolajewska , j. , ferrer , o. e. , garca , l. g. : 2009 , a&a , 497 , 815 carikov , z. , & skopal , a. : 2012 , a&a , 548 , a21 chesneau , o. , nardetto , n. , millour , f. , et al . : 2007 , a&a , 464 , 119 diplas , a. , savage , b. d. : 1994 , apj , 427 , 274 drake , j. j. , laming , j. m. , ness , j .- u .
, et al . : 2009 , apj , 691 , 418 dumm , t. , schild , h. : 1998 , new astron .
, 3 , 137 evans , a. , callus , c. m. , albinson , j. s. , whitelock , w. a. , glass , i. s. , carter , b. , roberts , g. : 1988 , mnras , 234 , 755 evans , a. , woodward , c. e. , helton , l. a. , gehrz , r. d. , lynch , d. k. , rudy , r. j. , russell , r. w. , kerr , t. : 2007 , apj , 663 , l29 fekel , f. c. , joyce , r. r. , hinkle , k. h. , skrutskie , m. : 2000 , aj , 119 , 1375 friedjung , m. : 1987 , a&a , 179 , 164 hauschildt , p. h. , allard , f. , ferguson , j. , baron , e. , alexander , d. r. : 1999 , apj , 525 , 871 hjellming , r. m. , van gorkom , j. h. , seaquist , r. e. , taylor , a. r. , padin , s. , davis , r. j. , bode , m. f. : 1986 , apj , 305 , l71 iijima , t. : 2009 , a&a , 505 , 287 lane , b. f. , sokoloski , j. l. , barry , r. k. , et al . : 2007 , apj , 658 , 520 livio , m. , pringle , j. e. , king , a. r. : 2003 , apj , 593 , 184 luna , g. luna , g. j. m. , montez , r. , sokoloski , j. l. , mukai , k. , kastner , j. h. : 2009 , apj , 707 , 1168 mohamed , s. , podsiadlowski , p. : 2007
, in 15th european workshop on white dwarfs , eds .
r. napiwotzki , & m. r. burleigh , asp conf .
372 , 397 monnier , j. d. , barry , r. k. , traub , w. a. , et al . : 2006 , apj , 647 , l127 mrset , u. , schmid , h. m. : 1999 , a&as , 137 , 473 narumi , h. , hirosawa , k. , kanai , k. , et al . : 2006 , iau circ . , 8671 nelson , t. , orio , m. , cassinelli , j. p. , leibowitz , e. , mucciarelli , p. : 2008
, apj , 673 , 1067 ness , j .- u . , starrfield , s. , page , k. l. , osborne , j. p. , beardmore , a. p. , drake , j. j. : 2007a
, pthps , 169 , 187 ness , j .- u . , drake , j. j. , starrfield , s. , et al . : 2009 , aj , 137 , 3414 obrien , t. , bode , m. f. , porcas , r. w. , muxlow , t. w. b. , eyres , s. p. s. , beswick , r. j. , garrington , s. t. , davis , r. j. , evans , a. : 2006 , nature , 442 , 279 osborne , j. p. , page , k. l. , beardmore , a. p. , et al . : 2011 , apj , 727:124 ( 10pp ) pavlenko , ya .
v. , evans , a. , kerr , t. , et al . : 2008 , a&a , 485 , 541 rosino , l. : 1987 , in bode m. f. , ed , rs ophiuchi ( 1985 ) and the recurrent nova phenomenon .
vnu science press , utrecht , p. 1
rupen , m. p. , mioduszewski , a. j. , sokoloski , j. l. : 2008 , apj , 688 , 559 rushton , m. t. , kaminsky , b. , lynch , d. k. , et al . : 2010 , mnras , 401 , 99 schaefer , b. e. : 2009 , apj , 697 , 721 shaviv , n. j. : 2001 , mnras , 326 , 126 shaviv , n. j. , dotan , c. : 2010 , mem .
s.a.it . , 81 350 skopal , a. : 2001 , a&a , 366 , 157 skopal , a. : 2005 , a&a , 440 , 995 skopal , a. : 2006 : a&a , 457 , 1003 skopal , a. : 2007 , new astron . , 12 , 597 skopal , a. : 2014 , new astron . , ( in press , paper i ) skopal , a. , vako , m. , komk , r. , chochol , d. : 2008a , in evans , a. , bode , m. f. , obrien , t. j. , darnley , m. j. , eds , asp conf
401 , rs ophiuchi ( 2006 ) and the recurrent nova phenomenon .
asp , san francisco , p. 101
skopal , a. , pribulla , t. , buil , ch . ,
vittone , a. , errico , l. : 2008b , in evans , a. , bode , m. f. , obrien , t. j. , darnley , m. j. , eds , asp conf .
401 , rs ophiuchi ( 2006 ) and the recurrent nova phenomenon .
asp , san francisco , p. 227 snijders , m. a. j. : 1987 , in in bode m. f. , ed , rs ophiuchi ( 1985 ) and the recurrent nova phenomenon . vnu science press , utrecht , p. 51
sokoloski , j. l. , luna , g. j. m. , mukai , k. , kenyon , s. j. : 2006 , nature , 442 , 276 sokoloski , j. l. , rupen , m. p. , mioduszewski , a. j. : 2008 , apj , 685 , l137 sostero , g. , guido , e. : 2006 , iau circ .
8673 starrfield , s. , iliadis , ch . ,
hix , w. r. : 2008 , in m. f. bode , & a. evans , eds .
, classical novae . cup , cambridge , p. 77 taylor , a. r. , davis , r. j. , porcas , r. w. , bode , m. f. : 1989 , mnras , 237 , 81 van belle , g. t. , lane , b. f. , thompson , r. r. , boden , a. f. , colavita , m. m. , dumont , p. j. , mobley , d. w. , palmer , d. , et al . : 1999 , aj , 117 , 521 west , j. d. : 2006 , iau circ .
8683 wilms , j. , allen , a. , mccray , r. : 2000 , apj , 542 , 914
here we determine the lower limit of the temperature and luminosity of the ionizing source , at which the corresponding flux of ionizing photons gives rise to the observed @xmath118 .
the method is based on the fact that the nebular component of radiation represents a fraction of the hot stellar radiation reprocessed via the ionization / recombination events .
the result of this process depends on the number of ionizing photons produced by the hot star and the number of particles on their path that are subject to ionization . in the limiting case , when _ all _
the ionizing photons are converted into the diffuse radiation , we can approximate the equilibrium condition between the flux of ionizing photons @xmath119 and the rate of recombinations in the nebula as @xmath120 where @xmath121 ( @xmath54s@xmath16 ) is the recombination coefficient to all but the ground state of hydrogen ( i.e. for case @xmath122 ) .
this equation is valid for the hydrogen plasma , heated by photoionizations , and characterized with a constant @xmath123 and particle concentration .
thus , the @xmath118 constrains a minimum luminosity of @xmath119 photons , which is a function of @xmath116 and @xmath117 of the hot ionizing source . using the expression for the @xmath119 function ( see eq . ( 11 ) of * ? ? ?
* ) , the luminosity of the ionizing source , which is capable of producing the observed _ em _ , can be written as @xmath124 where the function @xmath125 examples of the @xmath126 function parametrized with different values of _ em _ are plotted in fig .
[ fig : app ] .
its minimum corresponds to the temperature , at which the source produces maximum flux of ionizing photons ( for hydrogen it is 73000k ) . for
higher / lower @xmath116 it increases , because of decreasing production of the ionizing photons for a given @xmath117 . for @xmath127k we can observe only the jeans tail of the hot source radiation in the ultraviolet .
this implies that a small part of the far - uv fluxes increasing to higher energies can be compared with any blackbody radiation from this range of temperatures .
if we have available an appropriate spectrum , a @xmath128-tepmerature can be adopted . here , according to eq .
( [ eq : lph ] ) , we determine a minimum temperature @xmath20 of the ionizing source , at which the blackbody radiation , scaled to the far - uv fluxes , is just capable of producing the observed _
em_. according to @xcite we can get @xmath20 by solving equation @xmath129 where @xmath130 and @xmath123 are fitting parameters ( see sect .
2.4 of paper i ) and @xmath131 scales the blackbody flux emitted by the ionizing source to its observed flux @xmath38 .
then , eq . ( [ eq : tmin1 ] ) can be expressed as @xmath132 where @xmath38 is selected at a wavelength @xmath133 , at which the ionizing source dominates the spectrum ( usually in the far - uv ) .
its solution for the input quantities of @xmath130 , @xmath38 , @xmath133 and @xmath121 provides @xmath134 . according to eq .
( [ eq : lth ] ) , this temperature determines the minimum luminosity @xmath135 , at which the ionizing source is just capable of producing the observed _ em_. in the real case , a fraction of the ionizing photons can escape the star without being converted into the nebular radiation . also the measured _ em _ represents a lower limit of what was originally created by ionizations , because we observe only the optically thin part of the nebula . under these circumstances
the luminosity of the @xmath119 photons , originally emitted by the source , is larger than that given by the equilibrium condition ( [ eq : lph ] ) , i.e. @xmath136__em _ _ , which according to eq .
( [ eq : lth ] ) implies @xmath137 even for @xmath138 . considering that @xmath116 can be @xmath139 and
the nebula during outbursts is ionization bounded , it is probable that @xmath140 . | rs oph is a recurrent symbiotic nova that undergoes nova - like outbursts on a time scale of 20 years .
its two last eruptions ( 1985 and 2006 ) were subject of intensive multiwavelengths observational campaign from the x - rays to the radio .
this contribution aims to determine physical parameters and the ionization structure of the nova from its explosion to the first emergence of the supersoft x - rays ( day 26 ) by using the method of multiwavelength modelling the sed . from the very beginning of the eruption , the model sed revealed the presence of both a strong stellar and nebular component of radiation in the spectrum . during the first 4 days , the nova evinced a biconical ionization structure .
the @xmath0k warm and 160200@xmath1 extended pseudophotosphere encompassed the white dwarf ( wd ) around its equator to the latitude @xmath2 .
the remaining space around the wd s poles was ionized , producing a strong nebular continuum with the emission measure @xmath3@xmath4 via the fast wind from the wd .
the luminosity of the burning wd was highly super - eddington for the whole investigated period .
the wind mass loss at rates of @xmath5@xmath6 and the presence of jets suggest an accretion throughout a disk at a high rate , which can help to sustain the super - eddington luminosity of the accretor for a long time .
+ * physical parameters of the nova rs oph were determined from its x - ray / ir spectrum . *
the model revealed the presence of a strong stellar and nebular component . *
the luminosity was super - eddington during the whole burning phase . *
the high luminosity was sustained by a super - critical accretion .
stars : fundamental parameters individual : rs oph binaries : symbiotic novae , cataclysmic variables x - rays : binaries |
the proximity effect between a superconductor and a normal metal has been thoroughly investigated using various techniques : ginzburg - landau theory @xcite , quasi - classical green function methods @xcite , gorkov equation methods @xcite , and tight - binding bdg methods @xcite . from an experimental point of view ,
one of the better suited experiments is the measurement of the josephson critical current in weak links @xcite . in a recent experiment
@xcite an unusually large proximity effect is reported , and the authors argue that it can not be explained by the conventional proximity effect . the system used in the experiment is a c - axis oriented one . the c - axis
josephson critical current is measured through a thin film system made of doped lco with @xmath2k , sandwiched between optimally doped lsco with @xmath3k .
the thin film is considered to be in the clean limit and because of the epitaxial growth of the films the transmission at the interfaces is close to unity and interface roughness is on the order of the lattice constant . in a particular setup ,
the lco thin film used had a thickness of @xmath4 .
fitting the critical current around @xmath5 the authors extract a coherence length in the lco film which is two orders of magnitude larger than expected . because of this discrepancy and the observation of non - zero critical current for @xmath6k
the authors reported this effect to be a giant proximity effect " .
although the josephson junction has been thoroughly investigated in the past for both s - wave @xcite and d - wave symmetries @xcite , we feel that the calculation of the cooper pair leaking distance in the case of clean limit and superconducting weak links needs further investigation .
we are interested to observe if the leaking distance will be influenced by the finite critical temperature of the weak superconductor .
we propose the use of the numerical solutions of the bdg equations in a tight binding formulation in order to obtain a direct calculation of the coherence length and of the josephson critical current . in the clean limit
the bdg equations are particularly easy to solve because impurity averaging is not required .
this method is complementary to the quasi - classical methods used in the dirty limit , namely the usadel equations @xcite . for coherent transport in the c - axis direction ,
the properties of the josephson current for d - wave superconductors will be similar to the properties of the current for s - wave superconductors .
for planar interfaces , with @xmath7 the direction perpendicular to the interfaces , the d - wave order parameter will have no @xmath8 dependence , @xmath9 and therefore will have properties similar to a superconductor with s - wave symmetry .
when fourier transforming the @xmath10 and @xmath11 directions , and considering an effective 1d problem in the @xmath7 direction , the d - wave order parameter will be due to an effective on - site interaction within each ab - plane .
we will calculate the josephson current in the c - axis direction for a 3d d - wave superconductor .
we will also show calculations of the josephson critical current and the cooper pair leaking distance for a 2d s - wave superconductor and for the 100 interface of a 2d d - wave superconductor .
the giant proximity effect is observed in underdoped cuprates , for temperatures @xmath12 for which the middle layer is considered to be in the pseudo - gap state .
previous theoretical investigations of the giant proximity effect @xcite considered the n layer to be comprised of pockets of superconductivity . in a recent theoretical study @xcite ,
interstitial oxygen dopants are considered to modify locally the pairing interactions .
the disordered dopants are enhancing the pairing interactions , thus increasing the size of the local gap .
this was observed in recent stm experiments @xcite in bscco , which showed that the regions of enhanced superconductivity are correlated with the positions of the interstitial oxygen atoms . because of the proximity effect , the superconducting pockets will be coupled and current will flow through percolating paths .
the presence of these pockets will effectively shorten the length of the weak link and the strong external superconductors will be coupled for values of the effective length comparable with the leaking distance .
the modification of the leaking distance due to the finite value of @xmath5 will have an important influence on the effective length .
considering equally spaced areas of strong superconductivity with critical temperature @xmath0 embedded in the weak superconductor with @xmath5 we will calculate the critical josephson current and find its dependence on the length of the weak link and on the volume of the embedded superconducting pockets .
if one considers disordered regions of strong superconductivity in the @xmath13 planes of a high - t@xmath14 superconductor , then the distance between two pockets from different @xmath15 planes will be normally distributed .
the equally spaced pockets scenario should be the one that gives maximal josephson current and will give insight about the influence of these pockets on the current .
this paper is organized as follows . in the next section
we will present our method .
while the bdg procedure on a lattice is now well known , we nonetheless include some details , as some standard " approximations are included for clarity
. the treatment of infinite surfaces will be outlined . in the third section
we apply the bdg equations to a tri - layer system and show results of the calculation of the order parameter , the leaking distance and the dc josephson current .
both the cases of s - wave and d - wave symmetries of the superconducting order parameters are considered .
we find that the proximity effect can be considerably enhanced at temperatures close to ( but above ) the critical temperature of the weak superconductor .
the presence of randomly distributed pockets of superconductivity in n enhances dramatically the josephson critical current and leads to a `` giant proximity effect '' .
in order to describe the superconducting state we use the tight binding extended hubbard hamiltonian : @xmath16 where @xmath17 is the nearest neighbor hopping amplitude which describes the kinetic energy , @xmath18 is the chemical potential used to fix the filling of the system , @xmath19 is the on - site interaction , @xmath20 is the nearest neighbor interaction and @xmath21 is the density operator at site @xmath22 corresponding to spin @xmath23 .
the properties of this hamiltonian have been studied previously @xcite ; it should be viewed as an effective hamiltonian with which one can describe s - wave and d - wave symmetries of the superconducting order parameter . for an s - wave superconductor we choose an attractive on - site interaction @xmath24 and no nearest - neighbor interaction @xmath25 , while for a d - wave superconductor we set the nearest neighbor interaction to be attractive @xmath26 and the on - site interaction to vanish or be repulsive .
the interaction parameters @xmath19 and @xmath20 are dependent on position , breaking translational invariance .
this will allow us to describe interfaces between different types of materials . using the hartree - fock mean field decomposition
this hamiltonian can be transformed into a one - particle mean - field hamiltonian : @xmath27 for planar junctions , infinite surfaces can be considered and therefore translational invariance in the direction parallel to the surface is recovered . by doing a fourier transform of the hamiltonian in the direction parallel to the surface we only have to solve one - dimensional inhomogeneous problems . for any point in k
- space the problem becomes a one - dimensional inhomogeneous problem . in the case of a 2d superconductor with an infinite surface along the @xmath11 direction
the hamiltonian becomes : @xmath28 c_{i\sigma}^{k_y\dagger } c_{j\sigma}^{k_y } \nonumber\\ & + & \sum_{k_y}\sum_{i } [ \delta_i+2\delta_{i}^{||}cos(k_ya)]c_{i{\uparrow}}^{k_y\dagger } c_{i{\downarrow}}^{k_y\dagger } + h.c .
\nonumber\\ & + & \sum_{k_y}\sum_{<ij > } \delta_{ij}^{\perp } [ c_{i{\uparrow}}^{k_y\dagger } c_{j{\downarrow}}^{k_y\dagger } + c_{i{\downarrow}}^{k_y\dagger } c_{j{\uparrow}}^{k_y\dagger } ] + h.c.,\end{aligned}\ ] ] where @xmath22 and @xmath29 are now in the direction perpendicular to the surface and @xmath30 is the lattice constant .
@xmath31 and @xmath32 are the hopping amplitude and pair potential in the direction perpendicular to the surface and @xmath33 and @xmath34 are the hopping amplitude and the pair potential in the direction parallel to the surface .
the mean - field order parameters are to be calculated self - consistently : @xmath35 where @xmath36 is the s - wave order parameter , @xmath37 is the d - wave order parameter of a link in the direction parallel to the surface and @xmath38 is the d - wave order parameter of a link in the direction parallel to the surface . in the 3d c - axis geometry ,
the surface is considered to be in the @xmath39 plane .
after fourier transforming in these directions , the hamiltonian becomes : @xmath40\ } c_{i\sigma}^{k_x k_y \dagger } c_{j\sigma}^{k_x k_y } \nonumber\\ & + & \sum_{k_x k_y } \sum_{i } \{\delta_i+2\delta_{i}^{||}[cos(k_xa)-cos(k_ya)]\}c_{i{\uparrow}}^{k_x k_y \dagger } c_{i{\downarrow}}^{k_x k_y \dagger } + h.c .. \end{aligned}\ ] ] the self - consistency in the order parameters is now given by the following equations : @xmath41 \,\,\,\,\,\,\,\,\,\end{aligned}\ ] ] where @xmath22 is now taken to be the site index in the @xmath7 direction .
@xmath42 is the on - site s - wave order parameter while @xmath34 is the d - wave order parameter which has components only in the @xmath10 and @xmath11 directions .
we follow the standard procedure @xcite of introducing a canonical transformation of the electron operators : @xmath43 this transformation will diagonalize the hamiltonian and one obtains the bdg equations for each pair of momentum vectors @xmath44 , @xmath45 : @xmath46 these equations describe the quasi - particle states in inhomogeneous superconductors .
the bdg equations are equivalent to an eigenvalue problem with parameters that require self - consistent calculation .
we start with an initial guess for the order parameter profile and we diagonalize the resulting hamiltonian . in our infinite - surface setup , we need to diagonalize a one - dimensional hamiltonian for every point in momentum space . using the self - consistency equations ( 4)-(6 ) we recalculate the order parameter profile .
the solution is obtained when the difference in the order parameters between two steps is smaller than a desired accuracy .
the self - consistent calculation of the order parameter ensures that the order parameter in the normal metal " region has knowledge about the pair potential in this layer . if the initial guess is a step function , i.e. the order parameter in the middle region is zero , after one iteration the pair amplitude will become non - zero because of the proximity effect . in the case
@xmath47 , the order parameter will remain zero : @xmath48 , while for the case @xmath49 the new order parameter has a finite value throughout the layer .
if we were to fix the order parameter in the superconducting regions ( we do not ) , the @xmath50 solution would need only one iteration to converge .
the bdg formalism allows us to calculate the dc current in the absence of applied voltages . in the tight - binding formulation
the current operator is : @xmath51 the expectation value of the current will be non - zero only if the order parameters in the two superconducting layers have different phases . for the mean - field hamiltonian with s - wave order parameters
one gets for the continuity equation : @xmath52 > \nonumber\\ & = & \sum_\sigma t_{ij}(<c_{i\sigma}^\dagger c_{j\sigma } > - < c_{j\sigma}^\dagger c_{i\sigma } > ) \nonumber \\ & + & < c_{i{\downarrow}}c_{i{\uparrow}}>\delta_i^\star - < c_{i{\uparrow}}^\dagger c_{i{\downarrow}}^\dagger > \delta_i\end{aligned}\ ] ] if the order parameter is calculated self - consistently @xmath53 , then we recover the continuity equation , @xmath54 .
otherwise , if the order parameters are not calculated self - consistently but set to a desired value ( the case of hard boundary ) then the last two terms in eq .
( 14 ) can be seen as current source terms . in our calculation
the coherence factors , @xmath55 and @xmath56 , are complex numbers and they will give the magnitude and the phase of the order parameters .
the magnitude of the order parameters is calculated self - consistently and after each iteration the phase of the external layers is set to a desired value .
for the calculation of the dc josephson current we restrict the phase of the two superconductors to a desired phase difference , while for the weak link , the phase is calculated self - consistently .
for the @xmath58 tri - layers the interactions are only on - site attractive interactions . the value of the parameter @xmath19 will set the magnitude of the order parameter throughout the sample .
we consider the following setup ( fig .
1 ) , @xmath59 for @xmath60 and @xmath61 , while @xmath62 for @xmath63 . in this particular case
@xmath20 is vanishing , because we ignore the d - wave symmetry .
the value of @xmath64 is chosen so that @xmath65 , allowing us to describe the @xmath66 material with a lower critical temperature @xmath67 . for temperatures greater than @xmath67 the @xmath13 region can not sustain superconductivity by itself .
the order parameter will leak from the stronger superconductors , and the characteristic length is called the leaking " distance . 0 fig.1 shows the order parameter profile for @xmath68 , @xmath69 and @xmath70 . note that @xmath71 for the weak superconductor , while @xmath72 for the strong one .
similar to the @xmath73 case the order parameter has an exponential dependence on distance away from the interface , @xmath74 .
this is true only for temperatures much larger than @xmath67 and for distances from the interface greater than the coherence length of the stronger superconductors .
the coefficient of the exponential decay is given by the leaking distance , @xmath75 .
in the normal metal case ( @xmath76 k ) the clean limit leaking distance is inverse proportional to the temperature : @xmath77 for the case of a weak superconductor , the relevant temperature scale is @xmath78 . plotting the order parameter versus distance from the interface on a semi - log scale ( fig .
2 ) for different temperatures , we can extract the leaking distance .
as expected from the @xmath76 k case , the leaking distance is decreasing with increasing temperature . -90
-90 if we plot the order parameter as a function of temperature for different lengths ( l ) of the weak link ( fig .
3 ) , we can observe two main effects .
first , at @xmath79 k , the proximity effect will modify the order parameter at @xmath80 . for lengths smaller than 10 lattice constants
this effect is important . because the @xmath66 layer is superconducting at @xmath79k , the main length scale in this layer is the superconducting coherence length @xmath81 .
the second effect is observed at temperatures close to @xmath67 .
if the @xmath66 layer was not connected to the superconducting layers , then , according to the mean - field behavior , the order parameter would vanish at @xmath67 . for temperatures higher than @xmath67
the @xmath66 layer can not sustain superconductivity by itself .
it is only in the presence of the @xmath82 layers , that the order parameter at @xmath80 has non - zero values .
note that _ the length @xmath83 for which we obtain non - zero values of the order parameter above @xmath5 is much larger than the value of the length beyond which effects are unobservable at @xmath79 k_. in fig .
4 , we compare the order parameters at @xmath80 for two cases : @xmath50 and @xmath69 .
we observe that the order parameter for the case @xmath69 is larger and that close to @xmath67 the discrepancy is enhanced .
this is a clear indication that the cooper pair leaking distance is larger in the case of a non - zero @xmath67 . -90
-90 in order to investigate further the dependence of the leaking distance on the magnitude of the superconducting correlations in the @xmath66 layer , in fig . 5
, we summarize the extracted leaking distance obtained for different parameters .
the @xmath76k line(dashed ) is inversely proportional to the temperature , as expected . for @xmath84
k the leaking distance is diverging at @xmath5 ; this , of course leads to a giant proximity effect near these temperatures , as the figure visually demonstrates .
another feature of the calculation is that for any given temperature , a higher @xmath5 will result in a larger leaking distance . for repulsive on - site interactions , @xmath85 ,
the leaking distance is even smaller than in the normal metal case .
this is a clear demonstration of the fact that interactions in the @xmath66 layer will influence the way cooper pairs leak from the superconducting side
. such feedback " will not be captured in calculations that are not self - consistent .
the proximity effect can be observed either by growing superconducting thin films on top of normal metals and measuring the critical temperature of the system , or by forming a josephson junction and measuring the josephson current through the weak link .
if the two superconducting sides are not coupled then there is no josephson current .
as we bring the superconducting sides closer to one another , the proximity effect will influence the value of the order parameter in the @xmath66 layer .
a non - zero value of the order parameter throughout the whole system will result in a non - zero value of the josephson current .
the bdg equations are well suited for calculating the dc josephson current . in order
to have current between the two superconducting sides , the order parameters in the two sides have to have different phases .
-90 in our calculation we fix the phases of the order parameter on the @xmath82 layers , and our self - consistent calculation will give the magnitude and the phase of the order parameter in the @xmath66 side .
the results of such a calculation are shown in fig .
6a and 6b .
the phase of the order parameter in @xmath66 will vary continuously from @xmath86 to @xmath87 and the dc josephson current will be constant throughout the layer .
an interesting case is the one where @xmath88 , for which there is a phase - slip point at @xmath80 .
right at the phase - slip the order parameter vanishes . in order to extract only the proximity effect from the current calculation
, we need to find the phase difference for which the current is maximal . for a point contact josephson junction
the current has the following behavior @xcite : @xmath89 while for a long junction it deviates from the sinusoidal behavior @xcite .
-90 we calculate the dc josephson current for different lengths of the weak link and for different temperatures for @xmath90 and @xmath68 .
the results are summarized in fig .
when the two superconducting layers ( s ) are close together the proximity effect modifies the magnitude of the order parameter at l/2 in the n@xmath57 layer .
a large value of the order parameter will give a large value for the current .
as l increases the order parameter decreases exponentially .
this results in a decay of the current as a function of l. the main result is that the behavior of the josephson current as a function of l and t reflects the existence of a leaking distance larger than the one expected from a normal metal . for @xmath90 and @xmath91 ,
the normal metal gives a leaking distance of @xmath92 while the self - consistently calculated one gives a value of @xmath93 .
this is seen in fig .
7 , where for @xmath94 the current is non - zero for temperatures close to but greater than @xmath5 , and it has a linear dependence on temperature near @xmath5 . as shown in previous attempts to explain the
giant proximity effect " , the presence of pockets of superconductivity in the n@xmath57 layer will greatly enhance the current through the system , even for long weak links . coupled with the enhancement of the leaking distance around @xmath5
the presence of the superconducting pockets will effectively decrease the length of the weak link .
we consider equally spaced superconducting areas with on - site interactions of strength @xmath95 embedded in the weak link with interaction strength @xmath96 . in fig .
8 we show the josephson current for different lengths of the weak link and with superconducting pockets occupying a volume percentage @xmath97 of the weak link .
the size of the considered pockets is one lattice site .
the effect on the josephson current is drastic the current has non - zero values well above @xmath5 .
we also notice that for this volume of embedded superconductivity , the current has a weak dependence on the length of the junction . -90
-90 the strength of the coupling between the exterior superconductors will be given by the volume of these pockets .
this is seen in fig .
9 , where we plot the josephson current for a weak link of length @xmath98 as a function temperature for different percent volumes of strong superconducting pockets embedded in the weak superconductor .
as expected , for @xmath99 the junction is too long to couple the strong superconductors and the current vanishes above but very close to @xmath5 . increasing @xmath100 , the junction will effectively shorten and
the two exterior superconductors will be coupled well above @xmath5 .
the d - wave symmetry of the order parameter is attained if we consider nearest - neighbor interactions @xmath101 and vanishing or repulsive on - site interactions .
the d - wave order parameter is calculated in the following way : @xmath102 where @xmath103 describes superconducting correlations in the @xmath10 direction . in a similar manner ,
as detailed in the sn@xmath57s case , we set up @xmath104 so that @xmath105 for @xmath60 and @xmath61 , while @xmath106 for @xmath63 .
this will allow us to describe a weak link with a non - zero critical temperature . -90
-90 -90 for the 100 interface ( between the @xmath13 planes of a high-@xmath0 superconductor ) , the dependence of the order parameter is very similar to the s - wave case .
10 shows the semi - log plot of the order parameter as a function of distance from the interface for different temperatures .
again , we can observe the exponential decay and define the leaking distance @xmath75 .
the dependence of the order parameter on temperature for different lengths of the weak link is shown in fig .
11 and the two manifestations of the proximity effect are seen .
first at @xmath79k the order parameter is modified if @xmath80 is of the order of the superconducting coherence length in the n@xmath57 layer .
secondly , above @xmath67 the order parameter decays with increasing temperature but has non - zero value even if @xmath80 is greater than the conventional leaking distance defined by the @xmath73k case .
the self - consistently calculated leaking distance is shown in fig .
similar to the s - wave case it diverges at @xmath67 and , for the same temperature , larger interactions in the weak superconductor will increase the leaking distance .
-90 the coherent transport in the c - axis direction will be described by the hopping amplitude in the @xmath7 direction , @xmath107 .
13 shows the josephson critical current as a function of temperature for different lengths of the weak superconducting layer .
the general behavior is similar to the one of the s - wave junction .
the d - wave order parameter has no @xmath8 dependence , and thus in the @xmath7 direction it is an effective on - site order parameter . for short weak links the current does not vanish abruptly above @xmath5 but rather has a smooth dependence on temperature
this dependence on temperature above @xmath5 shows that the length of the weak link is comparable to the leaking distance in this layer .
the increase of the leaking distance due to the finite @xmath5 can not explain by itself the observed giant proximity effect " .
it is only the conjunction with the presence of disordered pockets of superconductivity in the weak link that makes this effect possible .
the calculation of the josephson current in the presence of the disordered pockets from the previous section stands also for the c - axis geometry .
the extra dimension will only affect the necessary volume of superconductivity needed to observe a giant proximity effect " .
in summary , using a tight binding formulation of the extended hubbard hamiltonian , we solve the bdg equations for a system composed of three layers : two superconducting layers ( either s - wave or d - wave ) , and a weaker superconductor sandwiched in between .
we examined the proximity effect induced by the exterior superconducting layers in the normal metal " interior layer .
we observed that , in agreement with previous calculations , the order parameter has an exponential decay behavior , the characteristic decay length being the leaking distance . in both s - wave and d - wave cases
the leaking distance is only dependent on the properties of the n@xmath57 layer . for @xmath73k ( normal metal ) for both s - wave and d - wave symmetries the leaking distance is inversely proportional to the temperature . if @xmath108k , the leaking distance diverges at @xmath67 and at the same temperature larger attractive interactions in the middle layer will increase the leaking distance . essentially , the bdg formalism provides a means for the normal layer to feel pairing fluctuations above its critical temperature , @xmath5 .
these are not spontaneous , in that they arise from an ` applied ' pairing field produced by the outer layers .
this accounts for the much higher leaking distance for a weak superconductor .
we also calculated the dc josephson current , and extracted the maximum value .
we observed that the current has a non - zero value for lengths of the weak link much larger than the conventional " leaking distance , and for temperatures well above @xmath67 .
although the divergence of the leaking distance at the critical temperature of the n@xmath57 layer enhances the josephson current for temperatures above @xmath5 , it is not enough to explain the experimental measurement of the giant proximity effect " @xcite . as prompted by previous attempts to explain the
giant proximity effect " @xcite , we considered areas of superconductivity with critical temperature @xmath109 , which are embedded in the n@xmath57 layer .
further enhancement of the josephson current is observed .
depending on the volume of the superconducting pockets , non - zero values of the josephson current are obtained even for temperatures @xmath110 .
these results form the basis for a qualitative understanding of the giant proximity effect observed by bozovic et al .
@xcite we thank wonkee kim and fatih doan for useful discussions .
this work was supported in part by the natural sciences and engineering research council of canada ( nserc ) , by icore ( alberta ) , and by the canadian institute for advanced research ( ciar ) .
i. bozovic et al .
, , 157002 ( 2004 ) .
v. l. ginzburg and l. d. landau , zh .
. fiz . * 20 * , 1064 ( 1950 ) .
jian hua chen , , 3952 ( 1990 ) ; , 3957 ( 1990 ) . v. g. kogan , * 26 * , 88 ( 1982 ) .
g. de gennes , rev .
mod . phys . * 36 * , 225 ( 1964 ) .
n. r. werthamer , phys . rev . * 132 * , 2440 ( 1963 ) . w. l. mcmillan , phys . rev . * 175 * , 559 ( 1968 ) ; phys . rev . * 175 * , 537 ( 1968 ) .
yao , c.s ting , w.k .
chu , * 46 * , 14059 ( 1992 ) .
j. e. hirsch , physica c * 194 * , 119 ( 1992 ) .
x . zhu and c. s. ting , * 61 * , 1456 ( 2000 ) .
k. halterman and o. t. valls , * 65 * , 014509 ( 2002 ) . k. k. likharev , rev . mod
* 51 * , 101 ( 1979 ) .
golubov et al . , rev .
phys . * 76 * , 411 ( 2004 ) . y. tanaka , s. kashiwaya , , 3451 ( 1995 ) .
y. tanaka , s. kashiwaya , , 892 ( 1997 ) .
y. tanaka , s. kashiwaya , , r11957 ( 1997 ) .
y. tanaka , yu .
v. nazarov , a. a. golubov , s. kashiwaya , , 144519 ( 2004 ) . y. tanaka , yu .
v. nazarov , s. kashiwaya , , 167003 - 1 ( 2003 ) . k. a. delin , a. w. kleinsasser , supercond .
* 9 * ( 1996 ) 227 .
k. d. usadel , , 507 ( 1970 ) . for a more recent application of the usadel formalism ,
see , for example , j.c .
cuevas , j. hammer , j. kopu , j.k .
viljas , and m. eschrig , cond - mat/0507247 .
v. kresin , yu .
ovchinnikov , s. wolf , appl .
. lett . * 83 * 722 ( 2003 ) .
g. alvarez , m. mayr , a. moreo , e. dagotto , , 014514 ( 2005 ) .
zhu , cond - mat/0508646 .
k. mcelroy _
et al . _ ,
science * 309 * , 1048 ( 2005 ) .
r. micnas , j. ranninger and s. robaszkiewicz , rev .
phys . * 62 * , 113 ( 1990 ) .
de gennes , _ superconductivity of metals and alloys _
( benjamin , new york , 1966 ) . v. ambegaokar and a. baratoff , , 486 ( 1963 ) .
a. martn - rodero , f. j. garca - vidal , and a. l. yeyati , 554 ( 1994 ) . | recent measurements of the josephson critical current through lsco / lco / lsco thin films showed an unusually large proximity effect . using the bogoliubov - de gennes ( bdg ) equations for a tight binding hamiltonian we describe the proximity effect in weak links between a superconductor with critical temperature @xmath0 and one with critical temperature @xmath0 , where @xmath1. the weak link ( n ) is therefore a superconductor above its own critical temperature and the superconducting regions are considered to have either s - wave or d - wave symmetry .
we note that the proximity effect is enhanced due to the presence of superconducting correlations in the weak link .
the dc josephson current is calculated , and we obtain a non - zero value for temperatures greater than @xmath0 for sizes of the weak links that can be almost an order of magnitude greater than the conventional coherence length . considering pockets of superconductivity in the n layer ,
we show that this can lead to an even larger effect on the josephson critical current by effectively shortening the weak link . |
recently several designs for silicon - based quantum computer architectures have been proposed.@xcite in this work we concentrate our efforts on the kane model,@xcite which exploits a qubit array of nuclear spins of @xmath0p dopants embedded within a silicon crystal matrix .
the model is based on the use of @xmath0p nuclear spins as qubits , with the donor electrons functioning to mediate control of single qubit operations via the hyperfine interaction , and interaction between individual qubits via the exchange interaction , and permit read - out of nuclear spin states .
performing logical operations on either electron - spin or nuclear - spin solid - state qubits requires precise control over single and two - qubit unitary operations , which corresponds to precise control over the electron - electron exchange interaction and the electron - nucleus hyperfine interaction in the kane quantum computer .
here we calculate the exchange interaction as a function of the two donors relative positions in the lattice and strain .
we use multi - valley effective mass theory to calculate the single donor electron wave functions , these single donor orbitals combine to form our two - electron basis .
this theory incorporates the si crystal lattice effects by including the si crystal bloch functions into our single donor electron basis . instead of using the heitler - london ( h - l ) approximation ,
which has been used extensively in the literature so far for impurities in si,@xcite we describe our two - donor system using a rigorous molecular orbital method which employs our multi - valley single donor orbitals to form our two - electron basis , to calculate the exchange coupling more accurately .
an important feature necessary for quantum computing is to have well - characterized qubits , and for the two qubit case this is the ground state singlet and triplet two - electron states .
it is meaningful to study the degree of proximity these targeted ground state orbitals are to the rest of the unwanted excited state hilbert space.@xcite this energy separation gives us an estimate for the conditions under which adiabaticity can be attained .
we have pursued this goal using a molecular orbital method , which enables us to calculate a large number of two - electron energy levels ( 144 ) , in the energy spectrum for our two donor system .
we used a molecular orbital method which includes the single donor ground state and first five excited states at each donor to form the two - electron basis .
this yields a basis of 78 singlet states and 66 triplet states .
this method not only gives us the exchange coupling , which is the difference between the ground singlet and triplet two - electron states , but also the spectrum of energy levels for the two donor system .
for comparison we also calculated the h - l exchange coupling using just the symmetrized and anti - symmetrized products of the single donor ground states , and the hund - mulliken ( h - m ) exchange coupling which in addition to the h - l states , includes the two doubly occupied single donor ground states at each donor , in our two - electron basis .
we calculated the exchange coupling and energy spectrum for our two - electron system as a function of donor position and strain .
in addition , we also calculated the probability that both electrons will be on the same donor .
this is also an important parameter for quantum gate operations , as these doubly occupied states can become a potential source of error .
several authors have similarly studied the exchange coupling , double - occupancy errors and adiabaticity of spin qubits in a number of different solid - state quantum computing architectures.@xcite much attention has been devoted to modeling the hyperfine and exchange interactions in these devices.@xcite inter valley interference between degenerate conduction band minima in si has been shown to lead to oscillations in the exchange coupling as a function of the donor pair positioning in the lattice.@xcite this poses serious problems for the fabrication of these devices , and leads to an extreme sensitivity of the exchange energy on the relative orientation of the p atoms .
koiller _ _
et al.__@xcite demonstrated that the introduction of external strain on the si lattice partially lifts the valley degeneracy in the bulk si .
they showed that the inter valley effects could be reduced in some cases depending on the relative orientations of the donor pairs , whilst in other cases the donor exchange coupling remains oscillatory .
the molecular orbital method we employ not only improves the calculation of the two donor electron wave functions , but we also use a more flexible basis than previous studies@xcite to calculate the single donor electron wave functions , which are used to construct the two - electron basis .
we advance beyond the simple h - l model for the two - electron wave function in the kane device that has been previously considered.@xcite in the molecular orbital method we use the single donor wave functions to form our basis states , and solve the 6-d schr@xmath1dinger equation for the two electrons through a direct matrix diagonalisation.@xcite in the simplest case , the h - l approximation , the donor pair wave function is modeled as the symmetrized and anti - symmetrized products of the two single donor ground state wave functions ( `` @xmath2 '' states ) at each p nucleus , to form our singlet and triplet states respectively . in the h - m approximation , in addition to the two h - l states , the h - m basis incorporates the two `` ionised '' or `` polarised '' doubly occupied ground states , at each donor . for the molecular orbital calculation we extended these bases to also include the first five excited states for each donor in our basis , in addition to the single donor ground states .
this was chosen so that our basis included the six symmetry ground states for the single p donor , ( @xmath3 and @xmath4 states ) .
we performed calculations for the exchange coupling to see the effect that this larger basis has on lowering the energy of both the singlet and triplet ground states , and to improve upon and test the validity of using h - l theory to model the two - donor system over a range of device parameters .
we get a basis for our two - electron system which consists of the spatially symmetric singlet states , and anti - symmetric triplet states . because the spin part of the singlet and triplet states are orthogonal
, we can consider the singlet and triplet bases independently .
using six single donor orbitals on both qubits , we can form 78 singlet states and 66 triplet states .
the molecular orbital method has advantages over some quantum chemical methods as it includes the correlation between the two electrons , by virtue of including many two - electron orbitals to minimise the energy of the system .
[ htp ] ( 60,25 ) ( 10,0)(0,0)*@xmath5 * ( 50,0)(0,0)*@xmath6 * ( 10.5,5)(1,0)10 ( 10.5,5)(0,1 ) 10 ( 7.5 , 13)(0,0 ) @xmath7 ( 19,2)(0,0)@xmath8 ( 10.5,5)(1,0 ) 40 ( 42,2)(0,0)@xmath9 ( 10.5,5)(1,1 ) 13.33 ( 50.5,5)(-2,1 ) 13.33 ( 37.17,11.665)(-2,1 ) 13.33 ( 50.5,5)(-1,1 ) 13.33 ( 10.5,5)(2,1 ) 13.33 ( 23.83,11.665)(2,1 ) 13.33 ( 23.83,18.33)(1,0 ) 13.34 ( 37.17,18.33 ) ( 23.83,18.33 ) ( 10.5,5 ) ( 50.5,5 ) ( 21,20)(0,0)@xmath10 ( 40,20)(0,0)@xmath11 ( 18,17)(0,0)@xmath12 ( 24,9)(0,0)@xmath13 ( 34,10)(0,0)@xmath14 ( 44,17)(0,0)@xmath15 ( 30,20.5)(0,0)@xmath16 we show the geometry of our two - electron problem in fig . [
fig : fig1 ] for two p donors , @xmath5 and @xmath6 , embedded in the si lattice , ( the origin is at @xmath5 ) . for the singlet ( symmetric spatial orbitals )
basis we form the following two - electron wave functions , from our basis of single donor orbitals , @xmath17 and @xmath18 , ( @xmath19th state at @xmath5 , and @xmath20th state at @xmath6 respectively ) : @xmath21 , \\ & & \mbox{for } \quad n=0 \mbox { to } 5 \quad \mbox{and } \quad m = n\mbox { to } 5 . \\ \psi^s_{22 - 42 } & = & \frac{1}{\sqrt{2(1+\delta_{nm } ) } } \left [ \psi_{q_2}^{e_n}(\bm{r_1-r } ) \psi_{q_2}^{e_m}(\bm{r_2-r } ) + \psi_{q_2}^{e_n}(\bm{r_2-r } ) \psi_{q_2}^{e_m}(\bm{r_1-r } ) \right ] , \\ & & \mbox{for } \quad n=0\mbox { to } 5 \quad \mbox{and } \quad m = n\mbox { to } 5 .
\\ \psi^s_{43 - 78 } & = & \frac{1}{\sqrt{2(1+|s_{nm}|^2 ) } } \left [ \psi_{q_1}^{e_n}(\bm{r_1 } ) \psi_{q_2}^{e_m}(\bm{r_2-r } ) + \psi_{q_1}^{e_n}(\bm{r_2 } ) \psi_{q_2}^{e_m}(\bm{r_1-r } ) \right ] , \\ & & \mbox{for } \quad n= 0\mbox { to } 5 \quad \mbox{and } \quad m = 0\mbox { to } 5 , \end{aligned}\ ] ] @xmath22 here we see that the two - electron singlet donor wave functions @xmath23 and @xmath24 , are the doubly occupied singlet states located at @xmath5 and @xmath6 respectively .
the two - electron states , @xmath25 , are the `` heitler - london like '' singlet states formed from the single donor ground state and excited state wave functioins . similarly for the triplet ( anti - symmetric spatial orbitals )
basis we obtain the following two - electron wave functions : @xmath26 , \\ & & \mbox{for } \quad n=0\mbox { to } 5 \quad \mbox{and } \quad m = n+1\mbox { to } 5 , \quad ( m\neq n ) .
\\ \psi^t_{16 - 30 } & = & \frac{1}{\sqrt{2 } } \left [ \psi_{q_2}^{e_n}(\bm{r_1-r } ) \psi_{q_2}^{e_m}(\bm{r_2-r } ) - \psi_{q_2}^{e_n}(\bm{r_2-r } ) \psi_{q_2}^{e_m}(\bm{r_1-r } ) \right ] , \\ & & \mbox{for } \quad n=0\mbox { to } 5 \quad \mbox{and } \quad m = n+1\mbox { to } 5 , \quad ( m\neq n ) .
\\ \psi^t_{31 - 66 } & = & \frac{1}{\sqrt{2(1-|s_{nm}|^2 ) } } \left [ \psi_{q_1}^{e_n}(\bm{r_1 } ) \psi_{q_2}^{e_m}(\bm{r_2-r } ) - \psi_{q_1}^{e_n}(\bm{r_2 } ) \psi_{q_2}^{e_m}(\bm{r_1-r } ) \right ] , \\ & & \mbox{for } \quad n=
0\mbox { to } 5 \quad \mbox{and } \quad m = 0\mbox { to } 5.\end{aligned}\ ] ] it is clear that the singlet and triplet bases contain the original h - l states , @xmath27 in the singlet basis , and @xmath28 in the triplet basis . for the h - m calculation we include the two additional `` ionised '' or doubly occupied ground states , @xmath29 and @xmath30 in our singlet basis .
in the extended molecular orbital basis , we consider all 78 singlet states and 66 triplet states in our two donor electron hamiltonians for the singlet and triplet bases respectively .
to obtain our two - electron states we first need to evaluate the single donor wave functions at each donor to use in the two - electron basis .
we did this in the case of no strain and with uniaxially strained si . we calculated the single donor orbitals using multi - valley effective mass theory .
we use a basis for the multi - valley single donor wave function which includes the full bloch structure at each conduction band minimum , in our basis functions .
the kohn - luttinger@xcite form of the wave function for a donor ( where we define the p nucleus to be at the origin on a substituted si atom site ) is given by : @xmath31 where we choose @xmath32,@xcite as we take the origin to be at an substitutional p donor site , here @xmath33 nm is the length of the unit cell .
the term @xmath34 arises because si has two atoms per lattice point in the unit cell , where the si atoms are chosen to be displaced a distance of @xmath35 away from each lattice point in the unit cell.@xcite for substitutional donors @xmath36 .
@xmath37 is the donor envelope function at the conduction band minimum , @xmath38 , and @xmath39 is the silicon crystal bloch function at the conduction band minimum , @xmath38 , where @xmath40 etc . and @xmath41 . the multi - valley effective mass equation for a p donor in si under strain is:@xcite @xmath42 & & \nonumber \\ \times f^{(\mu)}(\mathbf{r } ) = 0 , & & \label{eq : s1a } \end{aligned}\ ] ] where : @xmath43 here @xmath44 are the anisotropic kinetic energy terms , due to the anisotropy of the conduction band minima in si .
the impurity potential , @xmath45 , is the potential term due to the effective @xmath46 charge of the p nucleus in the si lattice .
here we model the impurity potential as a screened coulombic potential , @xmath47 .
we are using atomic units , where the unit of length @xmath48 = \hbar^2 \epsilon / m_\perp e^{'2}= 31.667\dot{\mbox{a } } $ ] and unit of energy @xmath49 = m_{\perp } e^{'4 } / 2 \hbar^{2 } \epsilon^2 = 19.9436$]mev , where @xmath50a.u . and
@xmath51.@xcite @xmath52 is the potential due to uniaxial strain along the @xmath7-direction which we will define later .
we expanded the donor electron envelope wave function , @xmath53 , in a basis of the single - valley zero field envelope wave functions , @xmath54 , at each of the six conduction band minimum , @xmath55 . here
@xmath54 are the eigenfunctions of the single - valley zero field hamiltonian , @xmath56 .
we have discussed previously@xcite how we obtained the single - valley zero - field wave functions , @xmath54 , by expanding these single - valley wave functions in a basis of deformed hydrogenic orbitals . in eq .
( [ eq : s1 ] ) we expanded the donor electron wave function , @xmath57 , in a basis of the donor electron envelope functions , @xmath53 , at each minimum .
in addition we can also expand @xmath53 in our basis of single - valley donor electron wave functions , @xmath54 : @xmath58 and @xmath59 are the expansion co - efficients for our basis functions @xmath60 .
we see that the single - valley envelope functions , @xmath54 , are in turn , also a sum of basis functions : the deformed hydrogenic orbitals , @xmath61 , given already in a previous paper.@xcite the co - efficients @xmath62 are determined already since @xmath63 are the eigenfunctions of the single - valley hamiltonians , @xmath64 , ie . @xmath65 .
note that including the expansion co - efficients @xmath59 in eq .
( [ eq : s2 ] ) is a generalisation of the calculations of wellard _ _
et al.__@xcite where we have removed the restriction that the donor wave function be composed of equal contributions from the six conduction band minima . clearly this restriction breaks down when an external strain is applied as this will break the degeneracy of the six conduction band minima .
so now eq . ( [ eq : s1a ] ) becomes : @xmath66 f_j^{(\mu)}(\mathbf{r } ) & = & 0 .
\nonumber \\
\label{eq : s3}\end{aligned}\ ] ] we now multiply eq .
( [ eq : s3 ] ) by @xmath67 and integrate over @xmath68 .
the orthonormality of this basis is enforced by the @xmath69 terms which appear in the matrix elements , and due to their rapidly oscillating nature average to zero unless @xmath70.@xcite in the standard effective mass treatment , the inter valley mixing terms which couple the envelope functions at different conduction band minima in the above approximation are neglected , and six independent equations are obtained .
for the higher donor excited states this is a valid approximation , as their energies agree quite well with calculations using only single valley effective mass theory.@xcite however , we need to consider the inter valley coupling for the donor ground state , which has the effect of lifting the six - fold degeneracy of the @xmath71 states predicted by the one - valley effective mass equations . in order to obtain the correct symmetry states for the donor ground state , of a singlet ( @xmath2 ) , a triplet ( @xmath72 ) and doublet ( @xmath4 ) , we add empirically determined parameters to our hamiltonian , as was done by koiller _
_ et al.__@xcite hence the multi valley effective mass eq .
( [ eq : s3 ] ) becomes : @xmath73 f_j^{(\nu ) } ( \mathbf{r } ) \nonumber \\ & = & e \delta_{\mu\nu } \delta_{ij } c_i^{(\nu)}. \label{eq : s4}\end{aligned}\ ] ] where @xmath74 here we also scale the single valley ground state energy , @xmath75mev , to reproduce accurately the experimental splitting for the p donor ground state .
we only scaled the single valley ground state energy because the single valley calculation reproduces the higher excited state energies reasonably accurately.@xcite when we consider the effect of strain on the single donor orbitals , we consider its effect only on the lowest six energy states , because later we construct our two - electron basis from these lowest six single donor states . here
we follow the treatment of koiller _ _ et al.__@xcite and introduce the relative energy shifts due to uniaxial strain along the @xmath7-direction , in terms of a dimensionless valley strain parameter , @xmath76 . in their paper
they discuss the physical relevance and tuning of this parameter . for our purposes
we consider four cases of the strain parameter , corresponding to @xmath77 and @xmath78 .
negative values of @xmath76 correspond to tensile strain , which favours the @xmath7-envelopes energetically , and @xmath79 represents the realistic situation of si grown over relaxed si@xmath80ge@xmath81.@xcite to evaluate the @xmath52 terms we first need to define @xmath82 to correspond to the @xmath83 valleys respectively . now the @xmath52 terms in eq .
( [ eq : s4 ] ) become : @xmath84 f_1^{(\mu ) } ( \mathbf{r } ) } \nonumber \\ & = & \delta_{1j } \left ( \sum_{\mu=1}^2 c_1^{(\mu ) } ( 2 \chi \delta_c ) + \sum_{\mu=3}^6 c_1^{(\mu ) } ( -\chi \delta_c)\right ) , \label{eq : hs1}\end{aligned}\ ] ] where @xmath85mev is used to be consistent with koiller _ _ et al.__@xcite [ t ! ] contour plot of the ground state electron density in the @xmath86-plane for @xmath2 state without any strain applied
. here the p nucleus is located at the origin.,width=240 ] [ t ! ] contour plot of the ground state electron density in the @xmath86-plane with a strain parameter @xmath79 . here
the p nucleus is located at the origin.,width=240 ] [ t ! ] contour plot of the ground state electron density in the @xmath87-plane with a strain parameter @xmath79 . here
the p nucleus is located at the origin.,width=240 ] we plot the ground state electron density without any external strain , @xmath88 , and with strain applied , @xmath89 , in fig .
[ fig : s1 ] to [ fig : s3 ] using eq .
( [ eq : s2 ] ) . here
the bloch functions are obtained using the empirical pseudo - potential technique @xcite and the p donor envelope functions are obtained from the multi - valley effective mass equations .
figure [ fig : s1 ] is a contour plot of the ground state electron density in the @xmath86-plane for the symmetric @xmath2 state corresponding to zero strain , where the contribution from all six valleys are equivalent .
however with a strain applied in the @xmath7-direction , we see in fig .
[ fig : s2 ] , where we plotted the electron density in the @xmath86-plane , the effective bohr radius in the @xmath7-direction is reduced .
this is because with an external strain applied , the six - valley degeneracy of the symmetric @xmath2 ground state is broken .
this can be seen from eq .
( [ eq : hs1 ] ) , and the lowest energy state is the one in which the effective bohr radius in the direction parallel to the strain is reduced , ie . the @xmath90 states .
in contrast in fig .
[ fig : s3 ] , where we plotted the ground state density in the @xmath87-plane , we see that the strain ( applied in the @xmath7-direction ) is equivalent in these two directions , and the effective bohr radii has increased .
table [ table2 ] reports the energy splitting between the ground state and first excited state for a single donor electron for different magnitudes of strain applied .
the energy levels become closer together when a strain is applied , and the ground `` @xmath2 '' state is no longer degenerate in the six valleys , and we find that the @xmath91 valleys become more favored .
this leads to a smaller effective bohr radius in the @xmath7-direction , and larger effective bohr radii in the @xmath92-directions , which was demonstrated already in fig .
[ fig : s2 ] and [ fig : s3 ] .
.[table2 ] energy splitting between the ground state and the first excited state for a single donor electron .
[ cols="^,^",options="header " , ] + energy & two - electron & one - electron + ( mev ) & basis states & states involved + -109.18685 & @xmath27 & @xmath2 + & + -97.77240 & @xmath93/@xmath94/@xmath95/@xmath96 & @xmath2/@xmath72(@xmath97 ) + & and @xmath2/@xmath72(@xmath7 ) + -97.76809 & @xmath93/@xmath94/@xmath95/@xmath96 & @xmath2/@xmath72(@xmath97 ) + & and @xmath2/@xmath72(@xmath7 ) + -96.38458 & @xmath98/@xmath99/@xmath100/@xmath101 & @xmath2/@xmath4 + & + -96.32310 & @xmath102/@xmath103/@xmath100/@xmath101/@xmath99 & @xmath2 , @xmath2/@xmath4 + & and @xmath2/@xmath72(@xmath8 ) + -95.94663 & + & @xmath2 , @xmath2/@xmath4 + & and @xmath2/@xmath72(@xmath8 ) + + energy & two - electron & one - electron + ( mev ) & basis states & states involved + -106.90440 & @xmath28 & @xmath2 + -97.77305 & @xmath104/@xmath105 & @xmath2/@xmath72(@xmath7 ) + & + -97.76847 & @xmath106/@xmath107 & @xmath2/@xmath72(@xmath97 ) + & + -96.50967 & @xmath108/@xmath109/@xmath110/@xmath111 & @xmath2/@xmath4 + & + -96.19400 & @xmath112/@xmath113 & @xmath2/@xmath72(@xmath8 ) + & + -95.76107 & @xmath112/@xmath27/@xmath110 & @xmath2/@xmath72(@xmath8 ) + & and @xmath2/@xmath4 + + energy & two - electron & one - electron + ( mev ) & basis states & states involved + -99.98734 & @xmath27 & @xmath2 + -88.14626 & @xmath95/@xmath96 & @xmath2/@xmath72(@xmath97 ) + -88.14312 & @xmath93/@xmath94 & @xmath2/@xmath72(@xmath7 ) + -88.12780 & @xmath102/@xmath103 & @xmath2/@xmath72(@xmath8 ) + -88.12567 & @xmath102/@xmath103 & + -88.11610 & @xmath93/@xmath95/@xmath94/@xmath96 & @xmath2/@xmath114 + -88.11399 & @xmath93/@xmath95/@xmath94/@xmath96 & and @xmath2/@xmath72 ( @xmath97 ) + -86.83179 & @xmath99/@xmath100/@xmath101 & @xmath2/@xmath4 + -86.82667 & @xmath98/@xmath99/@xmath100/@xmath101 & + -86.82358 & @xmath98/@xmath99/@xmath100 & + -86.80250 & @xmath98/@xmath99/@xmath100/@xmath101 & + + energy & two - electron & one - electron + ( mev ) & basis states & states involved + -99.96806 & @xmath28 & @xmath2 + -88.14459 & @xmath106/@xmath107 & @xmath2/@xmath72(@xmath97 ) + -88.14316 & @xmath104/@xmath105 & @xmath2/@xmath72(@xmath7 ) + -88.12656 & @xmath112/@xmath113 & @xmath2 /@xmath72(@xmath8 ) + -88.12504 & @xmath112/@xmath113 & + -88.11693 & @xmath106/@xmath107 & @xmath2/@xmath72 ( @xmath97 ) + -88.11485 & @xmath104/@xmath105 & @xmath2/@xmath72(@xmath7 ) + -86.83227 & @xmath108/@xmath115/@xmath110/@xmath111 & @xmath2/@xmath4 + -86.81847 & @xmath108/@xmath115/@xmath110/@xmath111 & + -86.79819 & @xmath108/@xmath115/@xmath110/@xmath111 & + -86.79310 & @xmath108/@xmath115/@xmath110/@xmath111 & + energy & two - electron & one - electron + ( mev ) & basis states & states involved + -255.87009 & @xmath27 & @xmath116 + -252.49159 & @xmath93/@xmath94 & @xmath117 + -252.45132 & @xmath93/@xmath94 & @xmath118 + -249.11343 & @xmath119 & @xmath120 + -225.97320 & @xmath121/@xmath30 & doubly occupied + -225.82787 & @xmath121/@xmath30 & @xmath122 + tables [ tableref1b ] and [ tableref1 ] list the lowest set of singlet and triplet two - electron eigenvalues and eigenvectors for @xmath88 and @xmath123 and @xmath124 nm respectively , and the corresponding single donor basis states contributing .
similarly , table [ tableref1a ] shows the singlet two - electron states for @xmath89 .
these tables clearly show the difference in the eigenvector basis components with and without strain applied . for @xmath125 nm , the ground singlet state is the heitler - london state @xmath27 , which is composed of the single donor ground states at @xmath5 and @xmath6 .
but we see in table [ tableref1a ] for @xmath89 that this single donor ground state is no longer the six - valley degenerate @xmath2 symmetry state , as the strain has broken the degeneracy of the six valleys .
the lowest energy states are when the effective bohr radius in the direction parallel to the strain is decreased , ie .
the single donor @xmath126 basis states . for @xmath88
the first two singlet and triplet excited states are nearly degenerate and these two - electron eigenstates involve significant contributions from the single donor @xmath2 and @xmath72 symmetry states . here
we see in table [ tableref1b ] and [ tableref1 ] that for @xmath88 these two states involve the @xmath72 states in the @xmath127 and @xmath128 valleys , and are lower in energy than the next states involving the @xmath72 states in the @xmath129 valleys .
this is because the overall two - electron / two - nuclei coupling leads to a more stable configuration when the donor electron densities are centered toward each other along the inter donor axis ( @xmath8-axis ) , ie .
the @xmath130 and @xmath114 states .
in contrast , table [ tableref1a ] shows that for @xmath89 that all the higher excited states involve significant contributions from the @xmath126 basis states .
this is because with such a large strain applied , the lowest single - donor eigenstates are when the effective bohr radius in the direction parallel to the strain is decreased , ie .
the single donor @xmath126 basis states .
tables [ tableref1b ] and [ tableref1 ] compare the degeneracy lifting of the @xmath131 and @xmath4 states for small and large inter donor separations .
this may be interesting for or relevant to the proposed optical raman experiments in the unstrained case.@xcite we find that lowest few energy excited states are formed when @xmath2 mixes with either @xmath72 or @xmath4 symmetry states . for @xmath123 nm , the two - electron @xmath132 state is even lower in energy than the @xmath133 state .
as we noted earlier this is because the @xmath134 state has a smaller effective bohr radius along the inter donor axis .
the two - electron donor pair is more stable when the electron wave functions are more centered toward each other along the inter donor axis .
for example , at the energies corresponding to the states containing the @xmath4 one - electron symmetry states , the two - electron wave - functions are very mixed and we are no longer able to assign the energy levels to pure basis states
. for the smaller donor separation with @xmath135 , and with a strain applied , the two - electron wave functions contain significant contributions from the doubly occupied orbitals .
this study shows that these doubly occupied states are important basis functions to include .
we have improved upon a previous study@xcite where only h - l type orbitals were considered . because we use an extended basis which includes both doubly occupied and h - l type two - electron orbitals , we find that the two - electron wave functions are often mixed states .
the two - electron energies given in table [ tableref1a ] and plotted in fig .
[ fig : fign2 ] to [ fig : fign4 ] are not scaled so that the conduction band bottom is at zero - energy .
we evaluated relative energy shifts using the valley strain parameter , @xmath76 , for the single - donor strain hamiltonian matrix and neglected any shift proportional to identity in it , to be consistent with the calculations of koiller et al.@xcite therefore , the calculated energies do not refer to the zero - energy to be at the bottom of the conduction band . but the energy eigenvalues shown in table [ tableref1a ] and fig .
[ fig : fign2 ] to [ fig : fign4 ] give the correct relative splitting among the eigenstates .
even just considering the four h - m states in our two - electron basis , gives an exchange energy which is very close to the extended basis calculation .
for the range of device parameters we consider , the h - m calculation is the most convenient method to use since it is relatively inexpensive and very accurate .
for this reason we use this method exclusively in the next section to obtain accurate results expediently and rapidly . [ t ! ] comparison of the hund - mulliken exchange coupling for different values of strain parameter , @xmath76 , for @xmath136 in the @xmath137 $ ] or @xmath8 direction .
we also plot the strain dependence of the double occupancy probability in ( b).,title="fig:",width=192 ] comparison of the hund - mulliken exchange coupling for different values of strain parameter , @xmath76 , for @xmath136 in the @xmath137 $ ] or @xmath8 direction .
we also plot the strain dependence of the double occupancy probability in ( b).,title="fig:",width=192 ] in figure [ fig : figh1](a ) we show the h - m calculation for the exchange energy for a range of inter donor separations along the @xmath137 $ ] direction to compare with the previous section .
this plot demonstrates the oscillations in the exchange energy due to the inter valley interference between the degenerate conduction band minima .
this oscillatory nature of the exchange coupling has already been reported by wellard _ _ et al.__@xcite and koiller _ _ et al.__@xcite using a h - l calculation .
we have improved upon and checked the h - l approximation , by extending our basis to include the h - m states , and found that the h - m basis offers some improvement for the close inter donor separations , and also allows us to calculate the ground state double occupancy probability of both electrons on the same donor .
we observe that for the zero strain case the oscillations in the exchange are the most conspicuous , as the inter valley interference is highest , because there are equal contributions from all six valleys in the ground singlet and triplet states .
however , when a strain is applied in the @xmath7-direction , we find that the @xmath128 valleys are favored energetically ( see fig .
[ fig : s2 ] ) , which implies a larger effective bohr radius along the inter donor axis ( @xmath8 ) , and hence the exchange coupling is improved substantially for this particular orientation of the p donor atoms along the @xmath137 $ ] axis.@xcite in addition to calculating the variation of the exchange coupling with inter donor separation in fig .
[ fig : figh1 ] , we also calculated the probability of the ground state double occupation of the h - m singlet states ( @xmath29 and @xmath30 ) in ( b).@xcite the exchange coupling with only the h - m states , matches the exchange coupling with the full spectrum of states very accurately .
thus , based on this we expect that the most appreciable contribution of the doubly occupied states will be from the h - m doubly occupied singlet states in our basis , which is why we only report the double occupancy probability of these states .
this double occupation probability is also an important parameter for quantum gate operations using electron spins as qubits . for non - zero exchange coupling the spin degrees of freedom
are also correlated with the orbital motion , and there is some probability that both electrons will be on the same donor.@xcite these doubly occupied states are not part of the targeted computational space , and as such , are a potential source of leakage error in solid - state quantum computers .
these plots show that as the strain parameter decreases , ( and the effective bohr radii in the @xmath97 and @xmath8-directions increases ) , this double occupancy probability increases as one may expect .
schliemann _ _ et al.__@xcite showed that gate operations on coupled quantum dot pairs which temporarily increase the exchange splitting , in order to swap electronic spins , inevitably lead to a finite double occupancy probability for both dots .
however they showed that this double occupancy amplitude does not lead to significant errors in quantum computing , provided that after the gate action is completed , the double occupancy probability is vanishingly small .
but if the double occupancy probability occurs to any sizable extent before or as a result of the gating action , then any quantum computer based on this hardware is likely to fail .
thus for the fabrication of these devices we want to minimise the double occupation probability for all states at zero voltage . the exchange coupling increases consequently with a uniaxial strain applied in a direction perpendicular to the inter donor axis , which achieves faster gating times .
however , the double occupancy probability also increases correspondingly , which increases the error requirement subsequently . in the last section we saw that both these features also lead to a greater mixing of the ground state with the higher excited states , which causes the energy levels to become closer together .
this energy splitting between the targeted h - l orbitals and the rest of the excited two - electron states informs us if during the gating action , the coupled donor system is well isolated and the higher excited states can be safely neglected .
this can also give us an estimate for gating times , so that the gating operation remains adiabatic .
[ t ! ] plot of the three singlet and one triplet energy levels calculated with the h - m basis , for @xmath88 in ( a ) and @xmath79 in ( b).,title="fig:",width=192 ] plot of the three singlet and one triplet energy levels calculated with the h - m basis , for @xmath88 in ( a ) and @xmath79 in ( b).,title="fig:",width=192 ] in fig .
[ fig : figh2 ] we plot the four two - electron energy levels predicted using our h - m basis , in order to compare with our more rigorous evaluation of the higher excited states in fig .
[ fig : fign1 ] and [ fig : fign4 ] , parts ( b ) and ( c ) .
the insets in these plots magnify the splitting between the energy levels .
the bottom inset in both plots corresponds to the ground state singlet and triplet energy levels , and compares favorably with the results shown for this splitting using the molecular orbital basis , in fig .
[ fig : fign1](b ) and [ fig : fign4](b ) for @xmath138 nm .
we observe that although the h - m basis provides an adequate description of the ground singlet and triplet ``
h - l '' states and exchange coupling , it is unable to predict the higher energy levels accurately .
this is because as we reported earlier in table [ tableref1 ] , using our full molecular orbital calculations , the first excited singlet state does not include significant contributions from the doubly occupied h - m singlet basis states .
this leads to an erroneously large energy splitting between the targeted ground `` h - l '' orbitals and the higher excited states .
furthermore we also investigated the variation of the exchange coupling for @xmath5 and @xmath6 displaced at small distances away from a targeted inter donor separation along the @xmath137 $ ] or @xmath8-axis . in these calculations we fixed the magnitude of the inter donor separation to be 14 nm , and varied the two angular variables , @xmath139 and @xmath140 defined in fig .
[ fig : figh4 ] . [ htp ] ( 60,25 ) ( 17,5)(0,0)*@xmath5 * ( 55,5)(0,0)*@xmath6 * ( 20.5,5)(1,0)10 ( 20.5,5)(-1,-1)7 ( 20.5,5)(0,1 ) 10 ( 12 , 0)(0,0 ) @xmath97 ( 17.5 , 13)(0,0 ) @xmath7 ( 29,2)(0,0)@xmath8 ( 20.5,5)(1,0 ) 30 ( 42,2)(0,0)@xmath9 ( 18.5,3)(22.5,2)(25.5,5 ) ( 20.5,8.5)(24.5,7)(25.5,5 ) ( 22.5 , 1)(0,0 ) @xmath139 ( 24.5,9)(0,0 ) @xmath140 ( 20.5,5 ) ( 50.5,5 ) [ h ! ] plot of the h - m exchange coupling for @xmath141 nm and @xmath88 and -20 . in ( a ) we calculate @xmath142 in the @xmath87-plane , where @xmath143 .
in ( b ) we calculate @xmath142 in the @xmath86-plane , where @xmath144 . in ( c )
we calculate @xmath142 in the @xmath145$]-plane , where @xmath146 . here
the marked points correspond to actual data points evaluated using the h - m method , the lines drawn are a guide for the reader only . , title="fig:",width=192 ] plot of the h - m exchange coupling for @xmath141 nm and @xmath88 and -20 . in ( a )
we calculate @xmath142 in the @xmath87-plane , where @xmath143 . in ( b ) we calculate @xmath142 in the @xmath86-plane , where @xmath144 . in ( c )
we calculate @xmath142 in the @xmath145$]-plane , where @xmath146 . here
the marked points correspond to actual data points evaluated using the h - m method , the lines drawn are a guide for the reader only . , title="fig:",width=192 ] plot of the h - m exchange coupling for @xmath141 nm and @xmath88 and -20 . in ( a )
we calculate @xmath142 in the @xmath87-plane , where @xmath143 . in ( b ) we calculate @xmath142 in the @xmath86-plane , where @xmath144 . in ( c )
we calculate @xmath142 in the @xmath145$]-plane , where @xmath146 . here
the marked points correspond to actual data points evaluated using the h - m method , the lines drawn are a guide for the reader only . , title="fig:",width=192 ] the results presented earlier in this section show the exchange coupling for varying magnitudes of @xmath147 only along the @xmath8 axis .
when the inter donor separation contains non - zero @xmath148 , @xmath149 and @xmath150 terms , we see a marked difference from the relative smoothness of the exchange coupling curves with only a non - zero @xmath149 term .
we show the exchange coupling for @xmath151 nm and @xmath88 and @xmath78 , for varying @xmath139 and @xmath140 in fig .
[ fig : figh3 ] .
the inter valley interference causing the wild oscillations in @xmath142 is highest when the inter donor separation contains all non - zero @xmath148 , @xmath149 and @xmath150 terms ( i.e. in ( c ) where both @xmath139 and @xmath140 are varied ) .
similar results have already been reported@xcite using h - l theory , and we have confirmed these results using our more extensive h - m basis . for @xmath89
we see that the exchange coupling changes dramatically for non - zero @xmath150 parts in fig .
[ fig : figh3 ] ( i.e. when @xmath140 varies ) .
this is because for @xmath89 the inter valley interference terms come only from the @xmath128 valleys , as the single donor ground state orbitals favour the @xmath91 valleys.@xcite if we study the exchange coupling closely we can identify where the peaks and troughs occur in the exchange coupling as a function of @xmath139 or @xmath140 .
for example , if we examine fig .
[ fig : figh3](a ) we can identify where the peaks and troughs occur for @xmath152 . in calculating the two - electron hamiltonian matrix and overlap matrix elements
, we find that several of the integrals involving @xmath153 in the integrand , have a common factor of @xmath154 , in the sum over @xmath38 and @xmath155 . in fig .
[ fig : figh3](a ) , we are varying @xmath147 in the @xmath87-plane , and consider @xmath156 , where @xmath157 and @xmath158 , and @xmath159 .
we recognise that for the range of @xmath139 we consider , @xmath149 does not vary as much as @xmath148 , and @xmath148 ranges over both positive and negative values .
thus we found that @xmath148 is the most significant factor in determining the peaks and troughs for this plot . in the simplest case
when @xmath160 and @xmath161 , and @xmath162 , there is a peak in the exchange coupling .
we find that this is due to the fact that the real part of @xmath163 is the maximum value of 1 , for 18 out of the 36 possible combinations of @xmath38 and @xmath155 .
similarly we observe peaks in the exchange coupling when @xmath164 , where m is any integer , and @xmath41 .
if @xmath165 , we find the real part of @xmath166 equals 1 .
again we find that the real part of @xmath163 is 1 for 18 out of the 36 possible combinations of @xmath38 and @xmath155 .
it is difficult to determine the magnitude of the peaks in the exchange coupling , because this magnitude also depends on terms involving @xmath167 which is a complicated function of @xmath149 and @xmath139 .
however in general we observe local maxima at values of @xmath139 for which @xmath168 , where m is any integer . conversely we find troughs in the exchange coupling occurring at values of @xmath139 for which @xmath169 , where m is an odd integer . here
we find that the real part of @xmath154 is 1 for 10 out of the 36 possible combinations of @xmath38 and @xmath155 , and @xmath170 for 8 of the remaining combinations .
thus at these values of @xmath139 we observe local minima in the exchange coupling .
because we are only able to evaluate the exchange coupling at finite grid points , the peaks and troughs were best matched to the data points available , thus enabling us to identify these trends .
for @xmath79 in fig .
[ fig : figh3](a ) we find that the exchange coupling is relatively constant .
this is because for @xmath79 the dominant contributions in the ground state come from the two @xmath171 valleys , and only very small contributions from the other four valleys .
thus when @xmath172 , the real part of @xmath173 is 1 for all the combinations of @xmath171 , and since the ground state has its largest components only in the @xmath171 valleys , the exchange coupling is maximised and almost constant , for this orientation .
the small fluctuations in the exchange coupling are most likely due to the fact there may be small contributions from the other four conduction band minima , which oscillate as a function of @xmath139 as we saw earlier . in ( b ) and ( c ) of this figure
we observe that when @xmath147 contains non - zero @xmath150 part , the exchange coupling may oscillate even more wildly for @xmath79 , as a result of the inter valley interference from only the two conduction band minima , @xmath171 .
it is imperative to know precisely , the single electron wave function and two - electron states to determine accurately the parameter regime necessary for a nuclear spin or electron spin quantum computer . in this paper
we provide detailed electronic structure calculations using a molecular orbital method , for a pair of p donor electrons in relaxed and uniaxially strained si .
we have determined the excitation spectrum of two electrons on p donors in relaxed and strained si , and studied its dependence on donor positioning in the si lattice .
in particular , we concentrated on the targeted ground state singlet and triplet `` h - l '' orbitals , and examined the isolation of these ground states , from the rest of the excited hilbert space .
furthermore we calculated the exchange coupling and double occupancy probability as a function of strain and donor position .
both the exchange splitting and double occupancy probability have a similar dependence on inter donor distance and lattice positioning of the p atoms .
thus , a compromise is needed to maintain a very small double occupancy probability in zero field , while realizing a sizable exchange coupling during a gating action .
in unison with previous theoretical studies of the exchange coupling of p donors in relaxed and strained si,@xcite we found that the exchange coupling , ( and thus double occupancy probability ) , is extremely sensitive to the relative orientation of the two p donors in the si lattice .
however , the energy level spectrum appears not to be affected by the relative orientation of the p donors , as these energies are on a much larger scale than the exchange coupling .
the oscillations in the exchange coupling due to inter valley interference , have serious implications for any quantum computer architecture that relies on the exchange interaction to couple qubits .
this sensitivity can be reduced in the presence of strain , for displacements of the donors within the plane perpendicular to the direction of the uniaxial strain .
we have also identified the values of @xmath147 that lead to the peaks and troughs in the exchange coupling .
it would be useful to investigate the effect of an applied voltage on these oscillations , to examine the exchange coupling during and after a gate operation , to determine what device parameters are required for fault - tolerant quantum computation .
it is also important to study if the gating action can be performed adiabatically , i.e. if during the evolution of the two - electron system there remains a finite gap between the ground and excited states.@xcite wellard _ _ et al.__@xcite have extended these calculations to include the voltage dependence of the exchange coupling within the heitler - london framework .
we hope to develop their results further by using our extended basis to calculate the voltage dependence of not only the exchange coupling , but also the energy spectrum of the two - electron system , and double occupancy probability of the ground states .
the energy spectrum informs us if during the gating action , the coupled donor system is well isolated , and the higher excited states can be safely neglected .
in addition , the double occupancy probability gives us an estimate of the error rate .
however , including the electric field is a very computationally intensive task , as it requires evaluating the basis functions on our grid , over a much greater range of device parameters , which is by far the most time consuming part for our calculations .
the results presented here using effective mass theory provide a solid foundation for future device modelling . ongoing work on this project
is focusing on extending the multi valley calculations to investigate the effect of gate voltage on the oscillations in the exchange coupling .
one would expect that as the gate voltage is turned on , it will become more favorable for the donor wave function to distort toward the gate . as a result
, the donor wave function no longer has equal contributions from all six valleys , and the oscillations may smooth out as the inter valley interference effects decrease .
we will need to implement the full molecular orbital approach to obtain the higher excited two - electron states accurately . thus using this method we gain insight not only into the exchange coupling and double occupancy probability , but also the conditions required to perform adiabatic gate operations . to make these calculations more tractable
, we envisage that the hund - mulliken basis should suffice in calculating the exchange coupling over a greater range of device parameters , as we have shown here that this is the best compromise between accuracy and speed .
l.m.k . and
would like to thank g.j .
milburn , c.j .
wellard and l.c.l .
hollenberg for valuable discussions relating to this work .
l.m.k . would like to thank the centre for quantum computer technology and the centre for computational molecular science at the university of queensland , where most of the computational work was performed .
we acknowledge supercomputing time from the following facilities : the australian partnership for advanced computing national facility , university of queensland computational molecular science cluster facility , and the university of queensland s high performance computers .
l.m.k . and h.s.g .
acknowledge support from the national science council , taiwan ( roc ) , under grant numbers ncs94 - 2811-m-002 - 050 and nsc94 - 2112-m-002 - 028 .
kane , nature * 393 * , 133 ( 1998 ) .
r. vrijen , e. yablonovitch , k. wang , h.w .
jiang , a. balandin , v. roychowdhury , t. mor , and d. divincenzo , phys .
a * 62 * , 012306 ( 2000 ) .
skinner , m.e .
davenport , and b.e .
kane , phys .
90 * , 087901 ( 2003 ) .
ladd , j.r .
goldman , f. yamaguchi , y. yamamoto , e. abe , and k.m .
itoh , phys .
lett . * 89 * , 017901 ( 2002 ) .
r. de sousa , j.d .
delgado , and s. das sarma , phys .
. a * 70 * , 052304 ( 2004 ) .
m. friesen , p. rugheimer , d.e .
savage , m.g .
lagally , d.w .
van der weide , r. joynt , and m.a .
eriksson , phys .
b * 67 * , 121301(r ) ( 2003 ) . l.c.l .
hollenberg , a.s .
dzurak , c. wellard , a.r .
hamilton , d.j .
reilly , g.j .
milburn , and r.g .
clark , phys .
b * 69 * , 113301 ( 2004 ) . l.m
. kettle , h .- s .
goan , sean c. smith , c. j. wellard , and l.c.l .
hollenberg , j. phys . :
matter * 16 * , 1011 ( 2004 ) .
wellard , l.c.l .
hollenberg , f. parisoli , l. kettle , h .- s .
goan , j.a.l .
mcintosh , and d.n .
jamieson , phys .
b * 68 * , 195209 ( 2003 ) . c.j .
wellard , l.c.l .
hollenberg , l.m .
kettle , and h .- s .
goan , j. phys . :
matter * 16 * , 5697 ( 2004 ) .
b. koiller , x. hu , and s. das sarma , phys . rev .
b * 66 * , 115201 ( 2002 ) .
b. koiller , x. hu , and s. das sarma .
lett . * 88 * , 027903 ( 2002 ) .
x. hu and s. das sarma , phys .
a * 61 * , 062301 ( 2000 ) .
x. hu and s. das sarma , phys .
a * 66 * , 012312 ( 2002 ) .
j. schliemann , d. loss , and a.h .
macdonald , phys .
b * 63 * , 085311 ( 2001 ) .
s.d . barrett and c.h.w .
barnes , phys . rev .
b * 66 * , 125318 ( 2002 ) . l.m .
kettle , h .- s .
goan , sean c. smith , c. j. wellard , l.c.l .
hollenberg , and c.i .
pakes , phys .
b * 68 * , 075317 ( 2003 ) .
larionov , l.e .
fedichkin , a.a .
kokin and k.a .
valiev , nanotechnology * 11 * , 392 ( 2000 ) .
wellard , l.c.l .
hollenberg , c.i .
pakes , nanotechnology * 13 * , 570 ( 2002 ) .
smit , s. rogge , j. caro , and t.m .
klapwijk , phys .
b * 68 * , 193302 ( 2003 ) .
a. fang , y.c .
chang , and j. tucker , phys .
b * 66 * , 155331 ( 2002 ) . w. kohn and j.m .
luttinger , phys . rev . * 97 * , 1721 ( 1955 ) .
yu and m. cardona , _ fundamentals of semiconductors . _ ( springer - verlag , berlin , 1996 ) .
m.l . cohen and j.r .
chelikowsky , _ electronic structure and optical properties of semiconductors . _
( springer - verlag , berlin , 1988 ) .
pantelides , rev .
phys . * 50 * , 797 ( 1978 ) .
faulkner , phys . rev .
* 184 * , 713 ( 1969 ) .
m. cohen and t. bergstresser , phys . rev . *
141 * , 789 ( 1966 ) .
slater , _ quantum theory of molecules and solids .
( mcgraw - hill book company , new york , 1963 ) b. koiller , x. hu , h.d . drew , and s. das sarma .
lett . * 90 * , 067401 ( 2003 ) . | we theoretically study the hilbert space structure of two neighbouring p donor electrons in silicon - based quantum computer architectures . to use electron spins as qubits , a crucial condition is the isolation of the electron spins from their environment , including the electronic orbital degrees of freedom .
we provide detailed electronic structure calculations of both the single donor electron wave function and the two - electron pair wave function .
we adopted a molecular orbital method for the two - electron problem , forming a basis with the calculated single donor electron orbitals .
our two - electron basis contains many singlet and triplet orbital excited states , in addition to the two simple ground state singlet and triplet orbitals usually used in the heitler - london approximation to describe the two - electron donor pair wave function .
we determined the excitation spectrum of the two - donor system , and study its dependence on strain , lattice position and inter donor separation .
this allows us to determine how isolated the ground state singlet and triplet orbitals are from the rest of the excited state hilbert space .
in addition to calculating the energy spectrum , we are also able to evaluate the exchange coupling between the two donor electrons , and the double occupancy probability that both electrons will reside on the same p donor .
these two quantities are very important for logical operations in solid - state quantum computing devices , as a large exchange coupling achieves faster gating times , whilst the magnitude of the double occupancy probability can affect the error rate . |
the basic heliospheric structure created by the interaction between the fully ionized solar wind and the partially ionized local interstellar medium ( lism ) can to first order be modeled by considering plasma interactions alone , and therefore many models both new and old have been presented that ignore the neutral lism particles entirely ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , the charge exchange mechanism , whereby an interstellar neutral loses its electron to a proton , allows the neutrals to take part in the solar - wind / lism interaction in many important ways ( see review by * ? ? ?
for example , many interstellar neutrals that penetrate the heliopause are ionized by charge exchange with solar wind protons and are then picked up by the outflowing solar wind .
these `` pick - up ions '' dominate the internal energy of the solar wind beyond about 10 au @xcite , and the acceleration of some of these ions at the termination shock is believed to be the origin of anomalous cosmic rays @xcite .
heliospheric models that treat the neutral gas and plasma in a self - consistent manner predict somewhat different properties than plasma - only models , such as shorter distances to the termination shock , heliopause , and bow shock @xcite .
these models also suggest that neutral hydrogen in the heliosphere should be very hot , with temperatures of order 20,00040,000 k. the importance of this prediction is that this high temperature gas should produce h i lyman-@xmath0 absorption broad enough to be separable from the interstellar absorption observed toward nearby stars , meaning hydrogen in the outer heliosphere can potentially be directly detectable .
this absorption has in fact been observed using observations by the _ hubble space telescope _ ( hst ) .
@xcite detected excess lyman-@xmath0 absorption in hst spectra of the nearby stars @xmath0 cen a and b. they and @xcite both demonstrated that the properties of this excess absorption are consistent with a heliospheric origin .
in upwind directions , such as that toward @xmath0 cen , the heliospheric h i column density is dominated by compressed , heated , and decelerated material just outside the heliopause , which constitutes the so - called `` hydrogen wall . ''
this material accounts for most of the non - lism absorption observed toward @xmath0 cen .
an additional detection of heliospheric h i absorption only @xmath2 from the upwind direction was provided by hst observations of 36 oph @xcite . for downwind lines of sight
the h i density is much lower than in the hydrogen wall , but the sightline through the heated heliospheric h i is longer , potentially allowing heliospheric lyman-@xmath0 absorption to be observed downwind as well as upwind @xcite .
@xcite report a detection of heliospheric absorption along a downwind line of sight toward the star sirius .
in addition to detections of hot h i surrounding the sun , there have also been many reported detections of analogous `` astrospheric '' material surrounding other sun - like stars @xcite . @xcite
have illustrated how these observations might potentially be used to infer properties of the winds of these stars and their interstellar environments .
models that self - consistently treat the neutrals and plasma are essential to help interpret these observations , but such models are a complex theoretical and computational problem .
the fundamental difficulty is that neutrals in the heliosphere are far from equilibrium , and their velocity distribution at a given location can be highly non - maxwellian .
one approach is to treat the plasma as one fluid and the neutral hydrogen as three separate fluids , one for each distinct environment in which charge exchange occurs @xcite .
@xcite used such four - fluid models to reproduce the lyman-@xmath0 absorption observed toward @xmath0 cen .
however , the four - fluid treatment of the problem remains an approximation and may not be entirely successful in reproducing the actual h i velocity distribution @xcite , or the expected lyman-@xmath0 absorption for a given line of sight .
@xcite adopted a particle - mesh method for solving the neutral boltzmann equation , which should yield accurate particle distribution functions , and @xcite further developed this code ( see also * ? ? ? * ) . in this paper , we use heliospheric models created by this boltzmann code to predict heliospheric lyman-@xmath0 absorption profiles that can be compared with the hst observations .
models are created assuming different mach numbers for the inflowing lism to see which best matches the observed amount of absorption . for the @xmath0 cen line of sight
, a similar analysis has been carried out by @xcite but using four - fluid rather than boltzmann models , and assuming different lism parameters .
we compare the absorption predicted by the models with lines of sight observed by hst , including @xmath0 cen and the aforementioned 36 oph and sirius lines of sight , which all show evidence for heliospheric absorption @xcite .
by now , at least twenty or so useful lines of sight through the lism have been observed by hst ( for a partial list , see * ? ? ? * ) .
evidence for heliospheric absorption has only been found for three of them , but some of the other lines of sight with low lism column densities might still be useful for providing upper limits for heliospheric absorption .
thus , in addition to the three lines of sight that suggest heliospheric absorption ( 36 oph a , @xmath0 cen , and sirius ) we also work with three other lines of sight ( 31 com , @xmath3 cas , and @xmath1 eri ) that sample different directions through the heliosphere . in figure 1 , we display the lyman-@xmath0 lines of these six stars .
the @xmath4 values shown in the figure are the angles from the upwind direction , which range from the nearly upwind line of sight toward 36 oph a ( @xmath5 ) to the nearly downwind line of sight toward @xmath1 eri ( @xmath6 ) .
all of the spectra except that of 36 oph a were taken using the goddard high resolution spectrograph ( ghrs ) instrument .
of these spectra , all but that of sirius utilized the high resolution echelle - a grating .
the sirius spectrum was taken using the lower resolution g140 m grating , explaining the larger binsize of those data .
the 36 oph a data were obtained by the space telescope imaging spectrograph ( stis ) , which replaced ghrs in 1997 .
the resolution of that spectrum , which was observed with stis s e140h grating , is comparable to that of the echelle - a ghrs spectra .
the spectra in figure 1 are plotted on a heliocentric velocity scale .
all show broad , saturated h i absorption near line center , and narrower deuterium ( d i ) absorption about 80 km s@xmath7 blueward of the h i absorption .
also plotted are the assumed intrinsic stellar lyman-@xmath0 lines ( solid lines ) and the best estimates for the interstellar absorption ( dotted lines ) , which fit the data only for 31 com and @xmath3 cas .
these lism absorption estimates are based on previously published work , which we now discuss briefly . as mentioned in 1
, the @xmath0 cen data provided the first evidence for heliospheric absorption . both members of the @xmath0 cen binary system were observed .
the existence of two independent data sets for the same line of sight allowed for a very nice consistency check , and @xcite demonstrated that the results of their analysis were the same for both the @xmath0 cen a and @xmath0 cen b data .
it is the @xmath0 cen b spectrum that we display in figure 1 , since it has somewhat higher signal - to - noise ( s / n ) than the @xmath0 cen a data . using the d i line to define the central velocity and temperature of the interstellar material , @xcite found they could not fit the @xmath0 cen h i profile with lism absorption alone .
i absorption existed on both sides of the lyman-@xmath0 line ( see fig . 1 ) , especially on the red side .
heliospheric absorption is believed to be responsible for the excess absorption on the red side , and stellar astrospheric absorption may be responsible for the less prominent excess absorption on the blue side @xcite .
the lism h i column density toward @xmath0 cen could not be constrained well at all , so two models of the absorption were presented : one with a deuterium - to - hydrogen ( d / h ) ratio similar to previously measured values for the local interstellar cloud , @xmath8 @xcite , and one with a much smaller d / h value ( @xmath9 ) . in figure 1 , we show the model with the generally accepted d / h value .
the analysis of the 36 oph data proved to be very similar to that of @xmath0 cen , with the apparent existence of excess lyman-@xmath0 absorption on both the red and blue sides of the line ( see fig . 1 ) .
@xcite argue that only contributions of both the heliospheric and astrospheric absorption can account for all of the excess absorption .
the sirius data were first presented by @xcite .
the sirius lyman-@xmath0 spectrum has lower resolution and lower s / n than the other spectra in figure 1 .
furthermore , unlike the other lines of sight , there are two interstellar components instead of just one .
nevertheless , @xcite found that if they constrained the h i absorption by assuming the lism temperature and local d / h ratio reported by @xcite toward capella ( t=7000 k and @xmath10 , respectively ) , they could not account for excess h i absorption on both the blue and red sides of the line ( see fig . 1 ) .
they interpreted the larger excess on the blue side to be due to an optically thick stellar wind , and the red side excess to be due to absorption from an evaporative interface between the local cloud and surrounding hot ism .
@xcite presented a different interpretation of these data .
they proposed that the red excess is due to heliospheric absorption and the blue excess is due to analogous astrospheric absorption .
they supported this interpretation using a self - consistent kinetic / gasdynamic model , which suggests that a combination of heliospheric and astrospheric material can produce the required amount of absorption .
the estimated lism absorption shown in figure 1 is from @xcite . the 31 com , @xmath3 cas , and @xmath1 eri data shown in figure 1 were first analyzed by @xcite .
the intrinsic stellar lyman-@xmath0 lines we assume in the figure are somewhat different than those derived by @xcite , but the lism parameters used to compute the interstellar absorption profiles are nearly identical . the 31 com and @xmath3 cas spectra can be fitted beautifully by single interstellar absorption components , but the @xmath1 eri spectrum is more complex .
the h i absorption is significantly blueshifted relative to the d i absorption , indicating a substantial amount of excess h i absorption on the blue side of the line ( see fig . 1 ) .
@xcite interpreted this excess absorption to be astrospheric in origin .
we construct many models of the heliosphere using the boltzmann technique described by @xcite , @xcite , and @xcite , and we also present a four - fluid model for comparison ( see table 1 ) .
lyman-@xmath0 absorption is computed for each model for comparison with the hst data .
the models assume standard values for the fully ionized solar wind at 1 au : @xmath11 @xmath12 , @xmath13 km s@xmath7 , and @xmath14 k. for the inflowing partially ionized lism , we assume a velocity of 26 km s@xmath7 and a neutral hydrogen density of @xmath15 @xmath12 , which are consistent with observations @xcite . other input parameters
are varied , and the parameters assumed for the various models are listed in table 1 .
observations suggest a likely value of @xmath16 k for the interstellar temperature @xcite and a value of @xmath17 @xmath12 for the interstellar proton density @xcite , so these are the numbers assumed for most of our models . in models 8 and 9 , we also experiment with smaller values that are still plausible assumptions for the real lism ( see table 1 ) . however , the parameter we vary the most is a parameter , @xmath0 , relating the proton pressure to the total lism plasma pressure , @xmath18 . for the known velocity of the inflowing lism ,
the mach number m is related to @xmath0 by @xmath19 , with @xmath20 measured in k. for a pure hydrogen plasma , @xmath21 , based on the assumption that the electrons and the protons of the plasma have equal number densities and temperatures .
the existence of significant cosmic ray and magnetic field pressure in the lism could potentially increase @xmath0 well above 2 .
neither of these additional sources of pressure is well constrained by observation , so we experiment with a wide range of @xmath0 s ( see table 1 ) . in figure 2 , we display the temperature and density distributions for protons and neutral hydrogen in the heliosphere for model 1 .
panel 2a shows the locations of the heliospheric boundaries , and streamlines in panel 2b indicate the paths of protons in the model .
the lism plasma is heated at the bow shock ( bs ) and diverted around the heliosphere ( see fig .
the tangential discontinuity of the heliopause ( hp ) separates the lism plasma from the solar wind . in model 1 , the distances from the sun to the hp and to the bs in the upwind direction are 100 au and 260 au , respectively . inside the heliosphere ,
the cool , fast solar wind goes through a termination shock ( ts ) , which is at 62 au in model 1 .
the solar wind is heated to temperatures in excess of @xmath22 k and diverted tailward .
the neutral temperatures and densities are computed by taking moments of the particle distributions computed by the boltzmann code ( see * ? ? ?
* ) . in the region between hp and bs ,
the overdensity in neutral hydrogen referred to as the hydrogen wall is clearly visible in figure 2d .
downwind of this hydrogen wall , neutral hydrogen is depleted but hot ( @xmath23 k , see fig .
note that if @xmath24 then @xmath25 ( assuming @xmath26 k ) , in which case there will be no bow shock ( models 47 ) . a more in - depth description of models 110 and their properties will be provided in a future paper ( mller , zank , & wood 2000 , in preparation ) .
the lism protons are shock heated at the bow shock , and if there is no bow shock they are still heated by adiabatic compression at the heliopause .
the interstellar neutrals are then heated by charge exchange with these protons , creating the high temperatures seen in figure 2c .
charge exchange with outflowing solar wind protons also occurs , resulting in a flux of hot neutrals back through the heliopause and bow shock .
further charge exchanges can deposit additional energy both in the hydrogen wall and in front of the bow shock , and such processes can have measurable effects on the heliospheric structure @xcite . in general , higher heliospheric hydrogen temperatures and densities are observed for models with lower values of @xmath0 , which is not surprising since a lower @xmath0 corresponds with a higher mach number for the inflowing lism and therefore more heating . because higher temperatures and column densities will produce broader lyman-@xmath0 absorption profiles , models with higher mach numbers will also generally produce more absorption .
the boltzmann models provide us with velocity distributions throughout the heliosphere that we can use to compute the absorption profiles . in figure 3
, radial velocity distributions are displayed for neutral hydrogen particles in the upwind and downwind directions for model 1 . in order to create these distributions ,
particles are summed within @xmath27 au boxes centered on the locations indicated in the figure .
poissonian error bars ( @xmath28 ) are displayed for each velocity bin .
the bin size of the histogram is 6 km s@xmath7 .
gaussians have been fitted to the distributions ( dashed lines ) , and the poor fits illustrate the non - maxwellian character of the distributions . the complexity of the distributions is due in part to the fact that the neutral particles are created by charge exchange with protons throughout the heliosphere with very different temperatures and flow velocities .
the main peak consists mostly of neutral particles created by charge exchange with heated lism protons in the hydrogen wall . following the nomenclature used in @xcite ,
these are called `` component 1 '' neutrals .
the distributions also contain some particles created by charge exchange with the very hot , decelerated solar wind protons in between the termination shock and heliopause , which for the upwind case in figure 3 are seen after flowing back across the heliopause .
these particles , the `` component 2 '' neutrals , are particularly prevalent in the far wings of the distributions . the component 2 neutrals
are far more abundant in downwind directions , explaining why the downwind distribution is broader and has more extended wings than the upwind distribution . finally , there are a few particles in the distributions created by charge exchange with solar wind protons near the sun , thereby creating a small population of neutrals with speeds of about @xmath29 km s@xmath7 ( see fig . 3 ) , which are referred to as `` component 3 '' neutrals . at the upwind location shown in figure 3 ,
the peak of the distribution lies at about @xmath30 km s@xmath7 , which represents a significant deceleration from the @xmath31 km s@xmath7 speed of the undisturbed lism . at the downwind location shown in the figure
, the average speed is about @xmath32 km s@xmath7 , slightly faster than the @xmath33 km s@xmath7 lism flow speed . in figure 4
, we show the h i lyman-@xmath0 absorption predicted by model 1 for an upwind and downwind line of sight .
in computing such absorption profiles , we first divide the radial path through the heliosphere into many segments .
we assign an h i density to each segment based on the number of particles found within a @xmath34 au box containing the segment . to get viable statistics for the distribution of radial velocities for the segment
, we must use a larger area of influence to determine the distribution histogram .
we use a circle with a radius of 30 au containing the segment for this purpose .
velocities can be linearly mapped onto wavelengths using the familiar relation @xmath35 where @xmath36 is the rest wavelength of lyman-@xmath0 .
thus , the velocity distribution defines a line profile function , @xmath37 , which is normalized so that @xmath38 the opacity profile in frequency space is ( in cgs units ) @xmath39 where @xmath40 is the oscillator absorption strength and @xmath41 is the column density ( i.e. , the density times the pathlength of the line segment in question ) . for lyman-@xmath0 ,
@xmath42 @xcite . since @xmath43 , the opacity profile in wavelength space is @xmath44 we compute these opacity profiles for each segment of the line of sight and then add them all up .
the absorption profile is then simply @xmath45 , where @xmath46 is the assumed background lyman-@xmath0 flux . in figure 4
, we simply assume @xmath47 for all wavelengths .
in the procedure described above , we have assumed that the line profile is regulated solely by doppler motions .
this is a very good approximation here since the velocity distributions are much broader than the core of the natural line broadening profile , and column densities are too low for the lorentzian wings of the natural profile to become evident .
we have also ignored the fine structure of lyman-@xmath0 , but this has only a very minor effect since the two fine structure components are separated by only 1.3 km s@xmath7 .
the h i properties displayed in figures 2c - d were computed by taking moments of the velocity distribution functions . in figure 4 , we show absorption profiles computed directly from these moments ( dashed lines ) .
these profiles , however , are significantly different from those computed directly from the distributions , indicating that flow velocities and temperatures computed from moments do not define the characteristics of the distribution well enough to be used to accurately determine absorption profiles .
accurate absorption profiles can be computed only from the distributions themselves .
in figures 5 and 6 we show the heliospheric absorption predicted for the lines of sight observed by hst for models 17 .
the heliospheric absorption is combined with the lism absorption ( dotted lines ) and then convolved with the instrumental line spread function before being plotted . for the sake of clarity , we display four of the models in figure 5 and four in figure 6 , with model 4 being shown in both .
we have zoomed in on the red side of the absorption profiles since that is where most of the heliospheric absorption is located , astrospheric absorption being the more likely explanation for any excess absorption on the blue side ( see 2 ) .
the models certainly have no trouble producing observable absorption .
in fact , the models tend to predict too much absorption , especially downwind
. this downwind discrepancy may be even more dramatic than the figures suggest , because our models only extend 1000 au in that direction , which may be far enough for lines of sight with @xmath48 but is probably not far enough to incorporate all the heliospheric h i further downwind .
thus , we will underestimate the amount of heliospheric absorption toward sirius and @xmath1 eri .
the absorption exhibits some interesting nonlinear behavior , which is particularly pronounced in the downwind direction .
as expected , the amount of absorption decreases with increasing @xmath0 between @xmath21 and @xmath49 , but the absorption _ increases _ between @xmath49 and @xmath50 before decreasing once again between @xmath50 and @xmath51 ( see , e.g. , the @xmath1 eri panels of figs .
similar behavior is observed upwind , although it is not as dramatic .
the transition between decreasing absorption and increasing absorption at @xmath49 is presumably associated with the boundary between a supersonic and a subsonic lism wind , which is at @xmath52 . unfortunately
, none of the models does a very good job in reproducing the observations . some of the apparent disagreement is illusory , however , because it can be fixed by tweaking either the assumed stellar lyman-@xmath0 profile or the assumed lism absorption , or both . in figure 7
, we show how this is done for two examples : the @xmath53 model for the 36 oph line of sight and the @xmath54 model for the 31 com line of sight . in both cases ,
we allow the assumed lism absorption parameters to be altered within the uncertainties derived in the initial analyses of these data , which are determined partially from the analyses of other lism lines @xcite .
we also alter the shapes of the assumed stellar profiles to try to make the models fit the data ( solid lines : original , dashed lines : altered profiles ) . however , there are limits to the shapes a reasonable profile can have .
this is especially true for 31 com .
the 31 com lyman-@xmath0 emission line is very broad , meaning we can not allow the slope of the line to become very large within the absorption region .
this limits how effective profile changes are in helping us to force the heliospheric absorption model to fit the data .
the narrower 36 oph emission line allows us more leeway in altering the assumed profile .
thus , the changes from the original profiles to the new profiles in figure 7 are much larger for 36 oph than for 31 com .
another problem for 31 com is that there is no evidence for excess absorption on the blue side of the line that would suggest the presence of astrospheric absorption , meaning that we can not completely ignore the quality of fit on that side of the line , in contrast with the 36 oph example .
the quality of the 31 com fit in figure 7 is noticeably worse than that of the simpler lism - only fit in figure 1 .
nevertheless , we conclude that the 31 com fit in figure 7 is just good enough to claim that the @xmath54 model ( model 6 ) is barely consistent with the 31 com observations at @xmath55 .
based on figure 7 , we also claim that the @xmath53 model ( model 1 ) is consistent with the 36 oph observations at @xmath5 .
we have performed an analysis like that in figure 7 for all combinations of models and data , and in figure 8 we show the best fits that can be obtained by such alterations of stellar profile and lism absorption for models 17 .
many of the models that did not appear to agree well with the data in figures 56 do fit the data well in figure 8 , but many still do not . in general ,
discrepancies near the base of the absorption ( e.g. , at @xmath56 km s@xmath7 for @xmath0 cen ) are much harder to remedy than discrepancies farther in the wing of the absorption ( e.g. , at @xmath57 km s@xmath7 for @xmath0 cen ) .
large discrepancies near the base can generally only be fixed by introducing unreasonable fine structure into the assumed stellar lyman-@xmath0 profile .
based on the results displayed in figure 8 , we provide in table 1 our evaluation of which observed lines of sight are inconsistent with which models .
none of the models discussed to this point ( models 17 ) are consistent with every line of sight .
the low @xmath0 models do better upwind and the high @xmath0 models do better downwind . in general
, the problem with the models that do not fit the data is that they predict too much absorption , the exception being the 36 oph line of sight for which most of the models predict too little absorption .
model 7 is the only model that does not predict way too much absorption along the downwind line of sight to @xmath1 eri . for models 8 and 9 in table 1
, we repeated the @xmath53 model but with a lower lism proton density and temperature , respectively , in order to see what effect this has on the predicted absorption ( see 3 ) .
the n(h@xmath58 ) and t values assumed for these models are not the most likely values based on our current understanding of the lism , but are still within the realm of possibility ( see table 1 ) .
figure 9 compares the absorption predicted by all the @xmath53 models .
the figure shows that lowering the proton density decreases the amount of absorption downwind but has little effect elsewhere .
decreasing the temperature decreases the amount of absorption in all directions .
the changes in absorption indicated in figure 9 are fairly modest , however , and certainly not enough to change the results shown in figure 8 , which suggest that @xmath53 models are consistent only with the 36 oph line of sight .
if lower proton densities and/or temperatures were assumed for the models with other assumed @xmath0 s , it might allow some of them to become consistent with some of the observed lines of sight .
the @xmath54 model could be made consistent with the @xmath1 eri data , for example ( see fig .
8) , but it would worsen the disagreement with the 36 oph data .
thus , we conclude that small changes in n(h@xmath58 ) and t will not solve the overall disagreement between the current boltzmann models and the data .
the models predict too much absorption downwind and sidewind relative to the amount of absorption observed upwind .
given the non - linear response of heliospheric models to different boundary conditions , we can not completely exclude better agreement between data and models in another area of parameter space , perhaps one where the solar wind parameters are different than those that we assume .
moreover , barring unaccounted systematic errors in the assumed stellar profiles and the ism absorption component , the heliospheric models employed in this study do not yet include all possibilities of interaction in the heliosphere .
all of the models neglect the magnetic fields of both the solar wind and interstellar medium , the latter being very poorly constrained observationally .
the boltzmann kinetic models neglect neutral - neutral collisions , as well as proton - neutral collisions that are not accompanied by charge exchange , and neutral depletion due to photoionization close to the sun .
ironically , the one model listed in table 1 that is consistent with all lines of sight is the four - fluid model ( model 10 ) , which uses a less sophisticated treatment of the heliospheric h i than the boltzmann models .
figure 10 shows the comparison of predicted and observed absorption for model 10 , both before and after the assumed stellar profile and ism parameters are tweaked to maximize the quality of fit ( as was done in fig . 8 for models 17 ) .
the agreement is good enough to claim that the model is consistent with all six lines of sight .
note that the absorption predicted by the four - fluid model 10 is vastly different from that predicted by the boltzmann model 1 ( see fig .
10 and @xmath53 model in fig . 5 ) , despite the fact that their input parameters are identical ( see table 1 ) .
this illustrates just how sensitive the predicted absorption is to how the h i velocity distributions are treated in the models .
there are several ways to interpret the results illustrated in figure 10 .
it should be understood that the four - fluid model and the boltzmann / kinetic models represent somewhat different physics for the neutral h distributions .
the underlying assumption inherent in the four - fluid model is that some scattering of neutral h is present to ensure that the distributions of the individual components , _ if not the full distribution _ , is approximately maxwellian .
this process is absent in the kinetic models ( see zank et al.1996 and williams et al .
1997 for discussions concerning the four - fluid model assumptions ) .
figure 10 indicates that a good fit to the lyman-@xmath0 data along the six sightlines requires that the heliospheric neutral h distribution be very like that produced by the four - fluid model .
that boltzmann model 1 and the four - fluid model 10 produce quite different absorption profiles suggests the possibility that additional physical processes with respect to h may need to be incorporated in the boltzmann codes .
unfortunately , we have not yet explored parameter space in sufficient detail to rule out the possibility that a boltzmann model does in fact exist that results in acceptable agreement with the lyman-@xmath0 data , assuming reasonable input parameters .
in addition , other factors not included , explicitly or implicitly , in either the boltzmann or four - fluid models ( such as magnetic fields ) may still be required to fully address these issues .
we have compared h i lyman-@xmath0 absorption profiles computed using fully self - consistent kinetic / hydrodynamic models of the heliosphere with profiles observed in uv spectra obtained by hst .
we make this comparison for six different lines of sight through the heliosphere toward six nearby stars .
our results are summarized as follows : 1 .
: : most of our models use a boltzmann particle code for the neutrals , allowing us to estimate neutral velocity distributions throughout the heliosphere .
these velocity distributions are far from maxwellian , having extended wings that lead to broad lyman-@xmath0 absorption profiles
. 2 . : : average temperatures and flow velocities can be computed by taking moments of the velocity distributions , and lyman-@xmath0 absorption profiles can be computed from these moments .
however , we find that such profiles do not generally agree well with more accurate profiles computed directly from the distributions .
thus , the moments of the distributions apparently do not provide enough information about their shapes to yield accurate absorption profiles .
3 . : : the amount of absorption predicted by the models generally increases as the assumed mach number of the interstellar wind is increased ( i.e. , the @xmath0 parameter is decreased ) . however , there are some interesting non - linear effects near the transition between subsonic and supersonic models ( i.e. , near @xmath59 ) .
4 . : : the boltzmann models tend to predict too much absorption in sidewind and downwind directions , and too little absorption upwind .
this problem is exacerbated at higher mach numbers .
varying the assumed interstellar temperature and proton density does not seem to help .
5 . : : in contrast to the boltzmann models , a four - fluid model that uses a less sophisticated , multi - fluid treatment of the neutrals is consistent with the data .
the absorption predicted by this model is very different from that predicted by a boltzmann model with the same input parameters , emphasizing how crucial the treatment of the neutrals is for accurately predicting lyman-@xmath0 absorption profiles .
: : possible reasons for the lack of success for the boltzmann models in reproducing the data include the neglect of collisions not involving charge exchange in those models , or perhaps an insufficient exploration of parameter space .
there may also be other physical processes that have not been considered in either the boltzmann or four - fluid models which could be important and affect our results , such as processes involving magnetic fields .
baranov , v. b. , izmodenov , v. v. , & malama , y. g. 1998 , j. geophys .
res . , 103 , 9575 baranov , v. b. , & malama , y. g. 1993 , j. geophys .
res . , 98 , 15157 baranov , v. b. , & malama , y. g. 1995 , j. geophys .
res . , 100 , 14755 bertin , p. , lamers , h. j. g. l. m. , vidal - madjar , a. , ferlet , r. , & lallement , r. 1995b , a&a , 302 , 899 bertin , p. , vidal - madjar , a. , lallement , r. , ferlet , r. , & lemoine , m. 1995a , a&a , 302 , 889 burlaga , l. f. , ness , n. f. , belcher , j. w. , szabo , a. , isenberg , p. a. , & lee , m. 1994 , j. geophys .
res . , 99 , 21511 dring , a. r. , linsky , j. , murthy , j. , henry , r. c. , moos , w. , vidal - madjar , a. , audouze , j. , & landsman , w. 1997 , apj , 488 , 760 fisk , l. a. , kozlovsky , b. , & ramaty , r. 1974 , apj , 190 , l35 gayley , k. g. , zank , g. p. , pauls , h. l. , frisch , p. c. , & welty , d. e. 1997 , apj , 487 , 259 gloeckler , g. , et al .
1993 , science , 261 , 70 gloeckler , g. , et al .
1994 , j. geophys .
, 99 , 17637 holzer , t. e. 1989 , ara&a , 27 , 199 izmodenov , v. v. , lallement , r. , & malama , y. g. 1999 , a&a , 342 , l13 klecker , b. 1995 , space sci .
, 72 , 419 lallement , r. , ferlet , r. , lagrange , a. m. , lemoine , m. , & vidal - madjar , a. 1995 , a&a , 304 , 461 linsky , j. l. , brown , a. , gayley , k. , diplas , a. , savage , b. d. , ayres , t. r. , landsman , w. , shore , s. n. , & heap , s. r. 1993 , apj , 402 , 694 linsky , j. l. , redfield , s. , wood , b. e. , & piskunov , n. 2000 , apj , 528 , 756 linsky , j. l. , & wood , b. e. 1996 , apj , 463 , 254 linsky , j. l. 1998 , space sci .
, 84 , 285 lipatov , a. s. , zank , g. p. , & pauls , h. l. 1998 , j. geophys .
res . , 103 , 20631 morton , d. c. 1991 , apjs , 77 , 119 mller , h. -r . , & zank , g. p. 1999
, in solar wind 9 , ed
. s. r. habbal , et al .
( new york : aip ) , 819 .
mller , h. -r . ,
zank , g. p. , & lipatov , a. s. 2000 , j. geophys .
, submitted pauls , h. l. , zank , g. p. , & williams , l. l. 1995 , j. geophys .
, 100 , 21595 qumerais , e. , bertaux , j .- l . ,
sandel , b. r. , & lallement , r. 1994 , a&a , 290 , 941 steinolfson , r. s. , pizzo , v. j. , & holzer , t. 1994 , geophys .
lett . , 21 , 245 wang , c. , & belcher , j. w. 1998 , j. geophys .
res . , 103 , 247 williams , l. l. , hall , d. t. , pauls , h. l. , & zank , g. p. 1997 ,
apj , 476 , 366 witte , m. , banaszkiewicz , m. , & rosenbauer , h. 1996 , space sci .
, 78 , 289 wood , b. e. , alexander , w. r. , & linsky , j. l. 1996 , apj , 470 , 1157 wood , b. e. , & linsky , j. l. 1998 , apj , 474 , l39 wood , b. e. , & linsky , j. l. 1998 , apj , 492 , 788 wood , b. e. , linsky , j. l. , & zank , g. p. 2000 , apj , to appear july 1 zank , g. p. 1999
, space sci .
, 89 , 413 zank , g. p. , pauls , h. l. , williams , l. l. , & hall , d. t. 1996 , j. geophys . res . , 101 , 21639 cccccccccccc 1 & boltzmann & 0.10 & 8000 & 2.0 & 1.7 & y & n & n & n & n & n + 2 & boltzmann & 0.10 & 8000 & 3.5 & 1.3 & y & n & n & n & n & n + 3 & boltzmann & 0.10 & 8000 & 5.0 & 1.1 & n & y & y & n & n & n + 4 & boltzmann & 0.10 & 8000 & 7.6 & 0.9 & n & y & y & n & n & n + 5 & boltzmann & 0.10 & 8000 & 9.6 & 0.8 & n & y & n & n & n & n + 6 & boltzmann & 0.10 & 8000 & 12.5 & 0.7 & n & y & y & y & y & n + 7 & boltzmann & 0.10 & 8000 & 18.0 & 0.6 & n & y & y & y & y & y + 8 & boltzmann & 0.05 & 8000 & 2.0 & 1.7 & y & n & n & n & n & n + 9 & boltzmann & 0.10 & 6000 & 2.0 & 2.0 & y & n & n & n & n & n + 10 & four - fluid & 0.10 & 8000 & 2.0 & 1.7 & y & y & y & y & y & y + | we use self - consistent kinetic / hydrodynamic models of the heliosphere to predict h i lyman-@xmath0 absorption profiles for various lines of sight through the heliosphere .
these results are compared with lyman-@xmath0 absorption lines of six nearby stars observed by the _ hubble space telescope_. the directions of these lines of sight range from nearly upwind ( 36 oph ) to nearly downwind ( @xmath1 eri ) .
only three of the lyman-@xmath0 spectra ( 36 oph , @xmath0 cen , and sirius ) actually show evidence for the presence of heliospheric absorption , which is blended with the ubiquitous interstellar absorption , but the other three spectra still provide useful upper limits for the amount of heliospheric absorption for those lines of sight .
most of our models use a boltzmann particle code for the neutrals , allowing us to estimate neutral velocity distributions throughout the heliosphere , from which we compute model lyman-@xmath0 absorption profiles . in comparing these models with the data ,
we find they predict too much absorption in sidewind and downwind directions , especially when higher mach numbers are assumed for the interstellar wind .
models created assuming different values of the interstellar temperature and proton density fail to improve the agreement . somewhat surprisingly , a model that uses a multi - fluid treatment of the neutrals rather than the boltzmann particle code is more consistent with the data , and we speculate as to why this may be the case . |
at extremely high energies , nucleus - nucleus collisions may be described by parton interactions in the framework of perturbative qcd ( pqcd)-inspired models @xcite . in this framework , hard or semihard scatterings among partons dominate the reaction dynamics .
they can liberate partons from the individual confining nucleons , thus producing large amount of transverse energy in the central region @xcite , and drive the initially produced parton system toward equilibrium @xcite . in principle
, the same kind of hard processes , such as open charm production @xcite , direct photon and dilepton production @xcite , can also be used as direct probes of the early parton dynamics and the evolution of the quark - gluon plasma . unlike strange quarks , charm quarks can not be easily produced during the mixed and hadronic phases of the dense matter since
the charm mass is much larger than the corresponding temperature scale .
the only period when charm quarks can be easily produced is during the early stage of the parton evolution when the effective temperature is still high . at this stage ,
the parton gas is still not fully equilibrated yet so that the temperature is only an effective parameter describing the average momentum scale . by measuring this pre - equilibrium charm production
, one can thus probe the initial parton density in phase space and shed light on the equilibration time @xcite . roughly speaking ,
ultrarelativistic heavy - ion collisions in a partonic picture can be divided into three stages : ( 1 ) during the early stage , hard or semihard parton scatterings , which happen on a time scale of about 0.2 fm/@xmath0 , produce a hot and dilute parton gas .
this parton gas is dominated by gluons and its quark content is far below the chemical equilibrium value .
multiple hard scatterings suffered by a single parton during this short period of time when the beam partons pass through each other are suppressed due to the interference embedded in the glauber formula for multiple scatterings @xcite .
this leads to the observed disapparence of the cronin effect at high energy and at large transverse momentum @xcite .
interference and parton fusion also lead to the depletion of small-@xmath1 partons in the effective parton distributions inside a nucleus @xcite .
this nuclear shadowing of parton distributions reduces the initial parton production @xcite .
( 2 ) after two beams of partons pass through each other , the produced parton gas in the central rapidity region starts its evolution toward ( kinetic ) thermalization mainly through elastic scatterings and expansion .
the kinematic separation of partons through free - streaming gives an estimate of the time scale @xmath2 fm/@xmath0 @xcite , when local isotropy in momentum distributions is reached .
( 3 ) further evolution of the parton gas toward a fully ( chemically ) equilibrated parton plasma is dictated by the parton proliferation through induced radiation and gluon fusion . due to the consumption of energy by the additional parton production ,
the effective temperature of the parton plasma cools down considerably faster than the ideal bjorken s scaling solution .
therefore , the life time of the plasma is reduced before the temperature drops below the qcd phase transition temperature @xcite .
similarly , charm production can also be divided into three different contributions in the history of the evolution of the parton system : ( 1 ) initial production during the overlapping period ; ( 2 ) pre - thermal production from secondary parton scatterings during the thermalization , @xmath3 ; ( 3 ) and thermal production during the parton equilibration , @xmath4 , in the expanding system . in this paper
, we will first review the equilibration of the initially produced parton gas in sec .
ii , incorporating the result of an improved perturbative qcd analysis of landau - pomeranchuk - migdal ( lpm ) effect @xcite .
then we will discuss the three stages of open charm production in a reversed order , starting with the charm production during the final stage of parton equilibration in sec .
iii . for pre - thermal charm production , we will consider the space - momentum correlation in the initial parton phase - space distributions , which will suppress open charm production during this period as compared to previous estimates @xcite . in sec .
iv , we will compare the results to the charm production during the initial hard or semi - hard scatterings and also to the results in geiger s calculation @xcite which is about 40 - 50 times higher than our estimates here .
we will also discuss the change in charm production due to the uncertainties in the initial parton density and effective temperature .
finally we give our conclusions in sec . v.
at collider energies ( @xmath5 gev ) , hard or semihard parton scatterings are believed to be the dominant mechanism for transverse energy production in the central region @xcite .
these hard processes happen on a short time scale and they generally break color coherence inside individual nucleons @xcite . after the fast parton
pass through each other and leave the central region , a partonic gas will be left behind which is not immediately in thermal and chemical equilibrium .
the partons inside such a system will then undergo further interactions and free - streaming . neglecting parton scatterings in this period of time ,
the kinematic separation of partons with different rapidities in a cell establishes local momentum isotropy at the time of the order of @xmath6 fm/@xmath0 @xcite .
if we assume this is the actual kinetic equilibration ( or thermalization ) time for the partonic system , the subsequent chemical equilibration can then be described by a set of rate equations . in this section
we will review parton equilibration following ref .
@xcite with improved estimate of the gluon equilibration rate .
currently there are many models for incorporating hard and semi - hard processes in hadronic and nuclear collisions @xcite .
we will use the results of the hijing monte carlo model @xcite to estimate the initial parton production . in this model , multiple hard or semi - hard parton scatterings with initial and final state radiation
are combined together with lund string phenomenology @xcite for the accompanying soft nonperturbative interactions .
let us first estimate the initial conditions at time , @xmath7 , from the hijing results .
since we are here primarily interested in the chemical equilibration of the parton gas which has already reached local isotropy in momentum space , we shall assume that the parton distributions can be approximated by thermal phase space distributions with non - equilibrium fugacities @xmath8 : @xmath9 where @xmath10 is the four - velocity of the local comoving reference frame .
when the parton fugacities @xmath8 are much less than unity as may happen during the early evolution of the parton system , we can neglect the quantum corrections in eq .
( [ eq : eq1 ] ) and write the momentum distributions in the factorized form , @xmath11 using this form of distributions , one has the parton and energy densities , @xmath12 where @xmath13 , @xmath14 , for a bose distribution , @xmath15 and @xmath16 for a dirac distribution . for a baryon symmetric system , @xmath17 .
since boost invariance has been demonstrated to be a good approximation for the initially produced partons @xcite , we can then estimate the initial parton fugacities , @xmath18 and temperature @xmath19 from @xmath20 where @xmath21 is the average transverse momentum .
the quark fugacity is taken as @xmath22 , corresponding to a ratio 0.14 of the initial quark(antiquark ) number to the total number of partons .
table [ table1 ] shows these relevant quantities at the moment @xmath7 , for au + au collisions at brookhaven national laboratory relativistic heavy ion collider ( rhic ) and cern large hadron collider ( lhc ) energies .
one can observe that the initial parton gas is rather hot reflecting the large average transverse momentum .
however , the parton gas is very dilute as compared to the ideal gas at the same temperature .
the gas is also dominated by gluons with its quark content far below the chemical equilibrium value .
we should emphasize that the initial conditions listed here result from hijing calculation of parton production through semihard scatterings .
soft partons , _
e.g. _ , due to parton production from the color field @xcite , are not included . in general , chemical reactions among partons
can be quite complicated because of the possibility of initial and final - state gluon radiations .
interference effects due to multiple scatterings inside a dense medium , _
i.e. _ , lpm suppression of soft gluon radiation has to be taken into account .
one lesson one has learned from lpm effect @xcite is that the radiation between two successive scatterings is the sum , _ on the amplitude level _ , of both the initial state radiation from the first scattering and the final state radiation from the second one .
since the off - shell parton is space - like in the first amplitude and time - like in the second , the picture of a time - like parton propagating inside a medium in the parton cascade simulations @xcite shall break down . instead , we shall here consider both initial and final state radiations together associated with a single scattering ( to the leading order , a single additional gluon is radiated , such as @xmath23 ) , in which we can include lpm effect by a radiation suppression factor .
the analysis of qcd lpm effect in ref .
@xcite has been done for a fast parton traveling inside a parton plasma .
we will use the results for radiations off thermal partons who average energy is about @xmath24 , since we expect the same physics to happen . in order to permit the approach to chemical equilibrium
, the reverse process , _ i.e. _ , gluon absorption , has to be included as well , which is easily achieved making use of detailed balance .
we consider only the dominant process @xmath23 .
radiative processes involving quarks have substantially smaller cross sections in pqcd , and quarks are considerably less abundant than gluons in the initial phase of the chemical evolution of the parton gas . here
we are interested in understanding the basic mechanisms underlying the formation of a chemically equilibrated quark - gluon plasma , and the essential time - scales .
we hence restrict our considerations to the dominant reaction mechanism for the equilibration of each parton flavor .
these are just four processes @xcite : @xmath25 other scattering processes ensure the maintenance of thermal equilibrium @xmath26 , etc . ) or yield corrections to the dominant reaction rates @xmath27 , etc . ) . restricting to the reactions in eq .
( [ eq : eq4 ] ) and assuming that elastic parton scatterings are sufficiently rapid to maintain local thermal equilibrium , the evolution of the parton densities is governed by the master equations @xcite : @xmath28 where @xmath29 denote the densities with unit fugacities , @xmath30 , @xmath31 and @xmath32 are thermally averaged , velocity weighted cross sections , @xmath33 we have also assumed detailed balance and a baryon symmetric matter , @xmath34 .
if we neglect effects of viscosity due to elastic scattering @xcite , we then have the hydrodynamic equation @xmath35 which determines the evolution of the energy density . for a time scale @xmath36
, we can neglect the transverse expansion and consider the expansion of the parton plasma purely longitudinal , which leads to the bjorken s scaling solution @xcite of the hydrodynamic equation : @xmath37 we further assume the ultrarelativistic equation of motion , @xmath38 with @xmath39 and @xmath40 given by eq .
( [ eq : eq2 ] )
. we can then solve the hydrodynamic equation , @xmath41^{3/4 } t^3\tau = \hbox{const . } \;\ ; , \label{eq : eq10}\ ] ] and rewrite the rate equation in terms of time evolution of the parameters @xmath42 , @xmath43 and @xmath44 , @xmath45 where the density weighted reaction rates @xmath46 and @xmath47 are defined as @xmath48 notice that for a fully equilibrated system ( @xmath49 ) , eq .
( [ eq : eq10 ] ) corresponds to the bjorken solution , @xmath50 . to take into account of the lpm effect in the calculation of the reaction rate @xmath46 for @xmath51
, we simply impose the lpm suppression of the gluon radiation whose effective formation time @xmath52 is much longer than the mean - free - path @xmath53 of multiple scatterings to each @xmath51 process . in the mean time
, the lpm effect also regularizes the infrared divergency associated with qcd radiation .
however , @xmath31 still contains infrared singularities in the gluon propagators . for an equilibrium system
one can in principle apply the resummation technique developed by braaten and pisarski @xcite to regularize the electric part of the propagators , though the magnetic sector still has to be determined by an unknown magnetic screening mass which can only be calculated nonperturbatively @xcite up to now .
since we are dealing with a nonequilibrium system , braaten and pisarski s resummation may not be well defined . as an approximation
, we will use the debye screening mass @xcite , @xmath54 to regularize all singularities in both the scattering cross sections and the radiation amplitude . to further simplify the calculation we approximate the lpm suppression factor in ref .
@xcite by a @xmath55-function , @xmath56 , where @xmath57 is the effective formation time of the gluon radiation in qcd which depends on the second casimir of the beam parton representation in @xmath58 , _
e.g. _ , @xmath59 for a gluon . in the previous calculation of the interaction rate @xcite ,
this color factor was not taken into account .
the modified differential cross section for @xmath51 is then , @xmath60 where the second step - function accounts for energy conservation , and @xmath61 is the average squared center - of - mass energy of two gluons in the thermal gas .
the regularized gluon density distribution induced by a single scattering is @xcite , @xmath62}. \label{eq : dng}\ ] ] similarly , the regularized small angle @xmath63 scattering cross is , @xmath64 the mean - free - path for elastic scatterings is then , @xmath65 which depends very weekly on the gluon fugacity @xmath66 as compared to the independent one used in a previous study @xcite . using , @xmath67 we can complete part of the integrations and have , @xmath68 where @xmath69 is a function of @xmath66 , @xmath70 ^ 2 - 4x\;z\;x_d}}\right .
\nonumber \\ & + & \left .
\frac{1}{s\lambda_f^2}{\cosh^{-1}(\sqrt{x } ) \over \sqrt{[1+x(1+z)y_d]^2 - 4x\;z\;y_d}}\right\},\end{aligned}\ ] ] where @xmath71 and @xmath72 .
we can evaluate the integration numerically and find out the dependence of @xmath73 on the gluon fugacity @xmath66 . in fig .
[ fig1 ] , @xmath73 is plotted versus @xmath66 for a coupling constant @xmath74 . the gluon production rate increases with @xmath66 and then saturates when the system is in equilibrium .
note that in principle one should multiply the phase - space integral by @xmath75 to take into account of the symmetrization of identical particles in @xmath76 as in ref .
however , for the dominant soft radiation we consider here , the radiated soft gluon does not overlap with the two incident gluons in the phase - space .
thus we only multiply the cross section by @xmath77 to obtain the equilibration rate .
the calculation of the quark equilibration rate @xmath47 for @xmath78 is more straightforward .
estimate in ref .
@xcite gives , @xmath79 the dashed line in fig .
[ fig1 ] shows the normalized rate @xmath80 for @xmath81 , taking into account the reduced phase space of strange quarks at moderate temperatures , as a function of the gluon fugacity . with the parton equilibration rates which in turn depend on the parton fugacity
, we can solve the master equations self - consistently and obtain the time evolution of the temperature and the fugacities .
shown in figs .
[ fig2 ] and [ fig3 ] are the time dependence of @xmath24 , @xmath66 , and @xmath82 for initial conditions listed in table [ table1 ] at rhic and lhc energies .
we find that the parton gas cools considerably faster than predicted by bjorken s scaling solution @xmath83 = const . ) shown as dotted lines , because the production of additional partons approaching the chemical equilibrium state consumes an appreciable amount of energy .
the accelerated cooling , in turn , slows down the chemical equilibration process , which is more apparent at rhic than at lhc energies .
therefore , the parton system can hardly reach its equilibrium state before the effective temperature drops below @xmath84 mev in a short period of time of 1 - 2 fm/@xmath0 at rhic energy . at lhc energy ,
however , the parton gas becomes very close to its equilibrium and the plasma may exist in a deconfined phase for as long as 4 - 5 fm/@xmath0 .
another important observation is that quarks never approach to chemical equilibrium at both energies .
this is partially due to the small initial quark fugacity and partially due to the small quark equilibration rate .
we note that the initial conditions used here result from the hijing model calculation in which only initial direct parton scatterings are taken into account . due to the fact that hijing is a pqcd motivated phenomenological model , there are some uncertainties related to the initial parton production , as listed in ref .
we can estimate the effect of the uncertainties in the initial conditions on the parton gas evolution by multiplying the initial energy and parton number densities at rhic energy by a factor of 4 .
this will result in the initial fugacities , @xmath85 and @xmath86 . with these high initial densities
, the parton plasma can evolve into a nearly equilibrated gluon gas as shown in fig .
the deconfined phase will also last longer for about 4 fm/@xmath0 .
though , the system is still dominated by gluons with few quarks and antiquarks as compared to a fully chemical equilibrated system .
if the uncertainties in the initial conditions are caused by the soft parton production from the color mean fields , the initial effective temperature will decrease .
therefore , we can alternatively increase the initial parton density by a factor of 4 and decrease @xmath19 to 0.4 gev at the same time .
this leads to higher initial fugacities , @xmath87 and @xmath88 . as shown in fig .
[ fig4 ] by the curves with stars , this system evolves faster toward equilibrium , however , with shorter life - time in the deconfined phase due to the reduced initial temperature .
we thus can conclude that perturbative parton production and scatterings are very likely to produce a quark - gluon plasma ( or more specifically a gluon plasma ) in ultrarelativistic heavy ion collisions at lhc energy .
however , the fate of the quark - gluon plasma at rhic energy has to be determined by a more careful examination of the uncertainties in the initial conditions .
these uncertainties will surely affect the open charm production during the equilibration as we shall discuss .
with the given evolution of the parton gas , we can now calculate open charm production during the parton equilibration .
similar to light quarks , charm quarks are produced through gluon fusion @xmath89 and quark antiquark annihilation @xmath90 during the evolution of the parton plasma . however , since the number of charm quarks is very small as compared to gluons and light quarks , we can neglect the back reactions , @xmath91 and their effect on the parton evolution . given the phase - space density of the equilibrating partons , @xmath92 ,
the differential production rate is @xcite , @xmath93 , \label{eq : therm1}\end{aligned}\ ] ] where @xmath94=16 , @xmath95 , are the degeneracy factors for gluons and quarks ( antiquarks ) respectively , @xmath96 , @xmath97 are the _ averaged _ matrix elements for @xmath89 and @xmath90 processes , respectively , @xmath98 -\frac{6}{\hat{s}}(2m^2-\hat{t}-\hat{u } ) \nonumber \\ & + & \frac{6}{\hat{s}}\frac{m^2(\hat{t}-\hat{u})^2 } { ( m^2-\hat{t})(m^2-\hat{u } ) } -\frac{2}{3}\frac{m^2(\hat{s}-4m^2)}{(m^2-\hat{t})(m^2-\hat{u})},\\ \frac{|\overline{\cal m}_{q\bar{q}\rightarrow c\bar{c}}|^2}{\pi^2\alpha_s^2 } & = & \frac{64}{9\hat{s}^2 } \left[(m^2-\hat{t})^2+(m^2-\hat{u})^2 + 2m^2\hat{s}\right],\end{aligned}\ ] ] due to the small charm density , we can neglect the pauli blocking of the final charm quarks . for large charm quark mass , @xmath99 , we can approximate the phase - space density @xmath92 by a boltzmann distribution .
we further assume that the distributions are boost invariant , _
i.e. _ , @xmath100 where @xmath101 is the spatial rapidity of a space - time cell at @xmath102 . neglecting the transverse expansion
, the above assumption implies that the space - time cell at @xmath102 have a flow velocity , @xmath103 .
we can now complete the integral over @xmath104 in @xmath105 and obtain , @xmath106 , \label{eq : therm2}\end{aligned}\ ] ] where @xmath107 is the modified bassel function and @xmath108 is the time when the temperature , @xmath24 , drops below 200 mev .
the kinematic variables are chosen such that , @xmath109 the center - of - mass momentum , @xmath110 , is defined as @xmath111 , and @xmath112,\nonumber\\ q_{\perp}^2&=&p_{\perp}^2+p_{\perp 2}^2 + 2p_{\perp}p_{\perp 2}\cos\phi_2,\nonumber
\\ q_{\perp}^2&=&q^2+q_{\perp}^2=m_{t}^2+m_{\perp 2}^2 + 2m_{t}m_{\perp 2}\cosh(y - y_2).\end{aligned}\ ] ] using these variables and the energy - momentum conservation , we have , @xmath113/q_{\perp}.\end{aligned}\ ] ] in the integral over @xmath114 , we shall use the time evolution of the temperature , @xmath42 , and fugacities , @xmath115 , as given in the previous section .
before the parton distributions reach local isotropy in momentum space so that the rate equations can be applied to describe the equilibration of the parton system , scatterings among free - streaming partons can also lead to charm production . since the system during this period consists dominantly of gluons
, we shall only consider gluon fusions . to model the phase - space distribution ,
we take into account the distribution of the initial production points which spread over a region of width , @xmath116 in @xmath117 coordinate .
following ref .
@xcite , we assume free - streaming until @xmath7 and neglect the expansion in the transverse direction .
the correlated phase - space distribution function is given by @xmath118 where @xmath119 is the parametrization of the parton spectrum given by hijing simulations , @xmath120 the phase - space distribution is normalized such that @xmath121 .
the function @xmath122 and the rapidity width @xmath123 are given in table [ table2 ] for central @xmath124 collisions at rhic and lhc energies which also gives the initial conditions as listed in table [ table1 ] .
substituting the phase - space distribution , into eq .
( [ eq : therm1 ] ) , and integrate over space and time , we obtain the charm production distribution in the pre - thermal period , @xmath125 @xmath126 where the kinematic variables are similarly defined as in eq .
( [ eq : therm2 ] ) , and in addition , @xmath127 , \nonumber \\ \sinh y_{k2}&=&[m_{\perp}\sinh y + m_{\perp 2}\sinh y_2-k_{\perp 1}\sinh y_{k1}]/k_{\perp 2}.\end{aligned}\ ] ] note that the correlation between momentum and space - time in the phase - space distribution was not considered in a previous calculation @xcite . as we will show
this correlation is very important and will reduce the pre - thermal charm production as compared to the uncorrelated distributions .
similar effect was recently discussed by lin and gyulassy in ref .
@xcite , where formation time effect is also included which is expected to further suppress pre - thermal charm production .
during the initial interaction period , charm quarks are produced together with minijets through gluon fusion and quark anti - quark annihilation .
like gluon and light quark production , charm production through the initial fusion is very sensitive to the parton distributions inside nuclei .
in addition , the cross section is also very sensitive to the value of charm quark mass , @xmath99 . if higher order corrections are taken into account , the production cross section depends also on the choices of the renormalization and factorization scales .
detailed studies on the next - leading - order calculation@xcite shows , however , that higher order corrections to the total charm production cross section can be accounted for by a constant @xmath128-factor of about 2 .
this is what we will use next . for consistency
we use @xmath129 gev for all calculations . shown in figs .
[ fig5 ] and [ fig6 ] as solid lines are the initial charm production given by hijing calculations at rhic and lhc energies , with mrsd@xmath130 @xcite parton distributions .
the corresponding total integrated cross sections are , @xmath131 ( 5.75 ) mb at rhic ( lhc ) energy , where nuclear shadowing of the gluon distribution function is also taken into account . in hijing calculations ,
high order corrections are included via parton cascade in both initial and final state radiations .
the resultant distributions in @xmath132-pair momentum are very close to the explicit higher order calculations @xcite . plotted in figs .
[ fig5 ] and [ fig6 ] as dot - dashed and dashed lines are the pre - thermal and thermal production . in the calculation ,
a factor of 2 is also multiplied to the lowest order matrix elements of charm production .
both contributions are much smaller than the initial charm production at both energies .
the pre - thermal contributions shown are also much smaller than what was found in ref . @xcite .
this is because momentum and space - time correlation was not taken into account in ref .
@xcite which suppresses the pre - thermal charm production .
similar results are also found in a study by lin and gyulassy @xcite .
as we have already discussed , the initial conditions in tables [ table1 ] and [ table2 ] given by hijing calculations have many uncertainties .
if one increases the initial parton number density at rhic energy by a factor of 4 with the same initial temperature , charm production from both pre - thermal and thermal sources will increase about a factor of 12 as shown in fig .
[ fig7 ] , leading to a total secondary contribution comparable to the initial charm production . in the extreme limit
, a fully equilibrated parton plasma ( @xmath49 ) at the same initial temperature would give an enhancement of charm production about 4 times higher than the initial production , shown as dotted lines in figs .
[ fig5 ] and [ fig6 ] . in this case
, the enhancement not only comes from higher parton densities , but also from the much longer life time of the parton plasma ( cf . figs .
[ fig2 ] and [ fig3 ] ) .
much higher enhancements predicted in ref .
@xcite are due to the overestimate of the intrinsic charm production as pointed by gyulassy and lin in a recent paper @xcite .
though the intrinsic charm production is important in the forward direction at large @xmath133 @xcite , it is strongly suppressed in the mid - rapidity region due to the interference among pqcd amplitudes to the same order @xcite . to test the sensitivity of open charm production to uncertainties in initial fugacities and temperature separately
, we consider an alternative scenario as we have discussed in the parton evolution .
we assume the initial parton densities to be 4 times higher than given in table [ table1 ] at rhic energy but with lower initial temperature , @xmath134 gev .
accordingly , the initial phase - space distribution is also modified to : @xmath135 from the one in table [ table2 ] , which gives 4 times of the initial parton density but smaller average transverse momentum , @xmath136 gev . the reduced average transverse momentum corresponds to lower initial effective temperature .
this system with higher initial fugacities evolves faster toward equilibrium but the life - time of the deconfined phase is shorter due to the reduced temperature as we have discussed .
the corresponding open charm production is shown in fig .
[ fig7 ] by the lines with stars .
we observe that open charm production from both pre - thermal and thermal contribution is reduced due to the reduction in initial temperature and life - time of the parton plasma , even though the initial fugacities are much higher and the evolution toward equilibrium is faster .
thus , open charm production is much more sensitive to the change in the initial temperature than the parton fugacities .
we also note from eqs .
( [ eq : therm2 ] ) and ( [ eq : preth ] ) that the pre - thermal and thermal charm production depends on the thermalization time @xmath7 and the life time of the parton plasma . therefore , by measuring the charm enhancement , we can probe the initial parton phase - space distribution , initial temperature and the thermalization and equilibration time .
in this paper , we have calculated open charm production in an equilibrating parton plasma , taking into account the evolution of the effective temperature and parton fugacities according to the solution of a set of rate equations . in the evaluation of the interaction rate @xmath46 for induced gluon radiation , a color dependent effective formation time
was used which reduces the gluon equilibration rate through lpm suppression of soft gluons . in the calculation of the pre - thermal contribution to open charm production , correlation between momentum and space - time
was also included .
this correlation reduces the pre - thermal charm production as compared to the uncorrelated one used in a previous estimate @xcite .
we found that both the thermal contribution during the parton equilibration and pre - thermal contribution with the current estimate of the initial parton density from hijing monte carlo simulation are much smaller than the initial direct charm production .
however , the final total charm production is very sensitive to the initial condition of the parton evolution . if uncertainties in the initial parton production can increase the initial parton density , _
e.g. _ , by a factor of 4 , the total secondary charm production will become comparable or larger than the initial production , due to both the increased production rate and longer life time of the parton plasma . we also found that open charm production is more sensitive to the initial temperature of the parton system than the initial parton fugacities .
therefore , open charm production is a good probe of the initial parton distribution in phase - space and the thermalizaion and equilibration time of the parton plasma .
p. l. and x .-
n . w. would like to thank m. asakawa and m. gyulassy for helpful discussions .
b. m. and x .-
w. thank t. s. bir , e. van doorn , and m. h. thoma for their early collaboration in the study of parton equilibration .
this work was supported by the director , office of energy research , division of nuclear physics of the office of high energy and nuclear physics of the u.s .
department of energy under contract no .
de - ac03 - 76sf00098 and de - fg05 - 90er40592 .
p. l. and x .-
w. were also supported by the u.s . -
hungary science and technology joint fund j. f. no .
wang and m. gyulassy , phys .
d * 44 * , 3501 ( 1991 ) ; comp .
( 1994 ) , in press .
k. geiger and b. mller , nucl .
b369 * , 600 ( 1992 ) ; k. geiger , phys . rev .
d * 47 * , 133 ( 1993 ) .
h. j. moehring and j. ranft , z. phys .
c * 52 * , 643 ( 1991 ) ; p.
aurenche,_et al .
rev . d*45 * , 92 ( 1992 ) .
aurenche,_et al .
_ , comp . phys .
* 83 * , 107 ( 1994 ) .
blaizot , a.h .
mueller , nucl .
* b289 * , 847 ( 1987 ) . k. kajantie , p. v. landshoff and j. lindfors , phys
lett . * 59 * , 2517 ( 1987 ) ; k. j. eskola , k. kajantie and j. lindfors , nucl . phys . *
b323 * , 37 ( 1989 ) ; k. j. eskola , k. kajantie and j. lindfors , phys . lett .
b * 214 * , 613 ( 1989 ) .
e. shuryak , phys .
lett . * 68 * , 3270 ( 1992 ) . t. s. bir , e. van doorn , b. mller , m. h. thoma , and x .-
wang , phys .
c * 48 * , 1275 ( 1993 ) .
l. xiong and e. shuryak , phys .
c * 49 * , 2207 ( 1994 ) .
b. mller and x .-
wang , phys .
* 68 * , 2437 ( 1992 ) .
k. geiger , phys .
d * 48 * , 4129 ( 1993 ) .
z. lin and m. gyulassy , cu - tp-638 , nucl - th-9409007 .
e. shuryak and l. xiong , phys .
70 * , 2241 ( 1993 ) .
m. t. strickland , phys .
b * 331 * , 245 ( 1994 ) .
wang , in proceedings of the workshop on finite temperature qcd and quark - gluon transport theory , wuhan , china , april 18 - 26 , 1994 , edited by lianshou liu ( world scientific ) .
b. straub _ et al .
* 68 * , 452 ( 1992 ) . s. j. brodsky and h. j. lu , phys .
lett . * 64 * , 1342 ( 1990 ) .
k. j. eskola , j. qiu and x .-
wang , phys .
lett . * 72 * , 36 ( 1994 ) .
wang and m. gyulassy , phys .
* 68 * , 1480 ( 1992 ) .
k. j. eskola and x .-
wang , phys .
d * 49 * , 1284 ( 1994 ) .
m. gyulassy and x .-
wang , nucl . phys .
b * 420 * , 583 ( 1994 ) .
wang , m. gyulassy and m. plmer , lbl-35980 , hep - ph/9408344 .
t. s. bir , b. mller , and x .-
wang , phys .
b283 * , 171 ( 1992 ) .
b. andersson , g. gustafson , g. ingelman and t. sjstrand , phys .
rep . * 97*,31 ( 1983 ) .
k. j. eskola and m. gyulassy , phys .
c * 47 * , 2329 ( 1993 ) .
t. matsui , b. svetitsky , and l. mclerran , phys .
d * 34 * , 783 ( 1986 ) .
p. danielewicz and m. gyulassy , phys .
d * 31 * , 53 ( 1985 ) ; a. hosoya and k. kajantie , nucl .
phys . * b250 * , 666 ( 1985 ) ; s. gavin , nucl . phys .
* a435 * , 826 ( 1985 ) . j. d. bjorken , phys .
d * 27 * , 140 ( 1983 ) .
e. braaten and r. d. pisarski , nucl . phys . *
b337 * , 569 ( 1990 ) .
t. s. bir and b. mller , nucl . phys . *
a561 * , 477 ( 1993 ) .
j. f. gunion and g. bertsch , phys .
d * 25 * , 746 ( 1982 ) .
p. nason , s. dawson , and r.k .
ellis , nucl .
b303 * , 607 ( 1988 ) ; _ ibid . _ * b327 * , 49 ( 1989 ) ; m.l .
mangano , p. nason , and g. ridolfi , nucl .
b373 * , 295 ( 1992 ) .
w. beenakker , h. kuijf , w.l .
van neerven , and j. smith , phys .
* d40 * , 54 ( 1989 ) ; w. beenakker , w.l .
van neerven , r. meng , g.a .
schuler , and j. smith , nucl .
* b351 * , 507 ( 1991 ) .
i. sarcevic and p. valerio , phys .
b*338 * , 426 ( 1994 ) ; azph - th-94 - 20 , 1994 .
r. vogt , _ et al .
_ , in hard processes in hadronic interactions , edited by h. satz and x .-
n . wang , to be published .
a. d. martin , w. j. striling and r. g. roberts , phys . lett . *
306b * , 145 ( 1993 ) .
r. vogt , s. j. brodsky abd p. hoyer , nucl .
bf b383 , 643 ( 1992 ) ; s. j. brodsy , p. hoyer , a. h. mueller and w. k. tang , _ ibid . _
* b 369 * , 519 ( 1992 ) .
j. c. collins , d. e. soper and g. sterman , nucl .
b263 * , 37 ( 1986 ) . | open charm production during the equilibration of a gluon dominated parton plasma is calculated , with both the time - dependent temperature and parton densities given by a set of rate equations .
including pre - thermal production , the total enhancement of open charm production over the initial gluon fusion depends sensitively on the initial parton density and the effective temperature .
the dependence of the pre - thermal charm production on the space - momentum correlation in the initial parton phase - space distribution is also discussed . |
the european strategy forum on research infrastructures ( esfri ) , a strategic initiative to develop the scientific integration of europe , has identified four facilities whose science cases are so outstanding that they can be considered as the main ( ground - based ) priorities of the european astronomy and astro - particles communities .
these are the square kilometre array ( ska ) , the cherenkov telescope array ( cta ) , km3net ( km@xmath0 neutrino telescope ) and the european extremely large telescope ( e - elt ) . to address the common challenges of these infrastructures through synergy , the asterics ( astronomy esfri and research infrastructure cluster ) project was proposed to the european commission ( ec ) and funded .
its major objectives are to support and accelerate the implementation of the esfri telescopes , to enhance their performance beyond the current state - of - the - art , and to see them interoperate as an integrated , multi - wavelength and multi - messenger facility .
an important focal point is the management , processing and scientific exploitation of the huge datasets the esfri facilities will generate .
asterics will seek solutions to these problems outside of the traditional channels by directly engaging and collaborating with industry and specialised small and medium - sized enterprises ( smes ) .
the various esfri pathfinders and precursors and other selected world - class projects ( including space - borne facilities ) will present the perfect proving ground for new methodologies and prototype systems .
in addition , asterics will enable astronomers from across the member states to have broad access to the reduced data products of the esfri telescopes via a seamless interface to the virtual observatory framework .
this is expected to massively increase the scientific impact of the telescopes , and greatly encourage use ( and re - use ) of the data in new and novel ways , typically not foreseen in the original proposals . by demonstrating cross - facility synchronicity , and by harmonising various policy aspects
, asterics will realise a distributed and interoperable approach that ushers in a new multi - messenger era for astronomy . through an active dissemination programme , including direct engagement with all relevant stakeholders , and via the development of citizen scientist mass participation experiments , asterics has the ambition to be a flagship for the scientific , industrial and societal impact esfri projects can deliver .
the work packages have been built around the astronomy related esfri projects with a sharp focus on advancing and contributing to their design , construction and implementation .
futhermore , the work package activities and their deliverables have been selected to be of high impact and broadly relevant to as wide a range of the astronomy related esfri projects as possible , addressing common technical challenges and collective issues such as harmonisation , interoperability , exchange and cross - facility multi - wavelength / multi - messenger integration .
* asterics management support team ( amst ) * + this work package will establish the asterics management support team ( amst ) , and will thus guarantee the smooth execution of all financial , administrative and reporting elements of the project
. it will also permit the amst to exercise central control and oversight of the scientific and technical progress of the project , as measured by secured milestones and the successful receipt of deliverables . a high - level policy forum ( involving the esfri projects and other large astronomy research infrastructures )
will also be established in order to coordinate and agree new models for joint time allocation , observing and data access / sharing , in addition to other more general policy matters of common interest .
the culmination of asterics will be an integrating event to show - case the results of the project and their relevance to the esfri telescopes and all other relevant stakeholders .
* dissemination , engagement and citizen science ( decs ) * + the objective of decs is to promote asterics and the esfri astronomy facilities it aims to serve . in particular
, it aims to open - up the esfri facilities to all relevant stakeholders and the widest possible audience , including the public : i ) production of high quality branding and promotional outreach materials ; more ambitiously , decs also embraces the adoption of the principles of science 2.0 ; ii ) create web - based interfaces that will open - up the astronomy related esfri facilities to the general public via a harmonised suite of citizen science mass participation experiments ( mpes ) and online video material ; iii ) attract young people to science by networking the esfri facilities via citizen science initiatives , and through coordinating open educational resources .
* observatory e - environments linked by common challenges ( obelics ) * + the aim of obelics is to enable interoperability and software re - use for the data generation , integration and analysis of the esfri and pathfinder facilities through the creation of an open innovation environment for establishing open standards and software libraries for multiwavelength / multi - messenger data .
the specific objectives are : i ) train researchers and data scientists in the asterics - related projects to apply state - of - the - art parallel software programming techniques , to adopt big - data software frameworks , to benefit from new processor architectures and e - science infrastructures .
this will create a community of experts that can contribute across facilities and domains ; ii ) maximise software re - use and co - development of technology for the robust and flexible handling of the huge data streams generated by the asterics - related facilities ( this involves the definition of open standards and design patterns , and the development of software libraries in an open innovation environment ) ; iii ) adapt and optimise extremely large database systems to fulfil the requirements of the asterics - related projects ( this requires the development of use cases , prototypes and benchmarks to demonstrate scalability and deployment on distributed non - homogeneous resources ) ; cooperation with the esfri pathfinders , computing centres , e - infrastructure providers and industry will be organised and managed to fulfil this objective ; iv ) study and demonstrate data integration across asterics - related projects using data mining tools and statistical analysis techniques on petascale data sets ( this will require adaptable and evolving workflow management systems , to allow deployment on existing and future e - science infrastructures ) .
all tasks are built upon the state - of - the - art in ict , in cooperation with major european e - infrastructures and are conceived to minimise fragmentation .
communications and links with other communities and e - science service providers are considered in order to contribute to the effectiveness of the proposed objectives . * data access , discovery and interoperability ( dadi ) * + the virtual observatory ( vo ) framework is a key element in successfully clustering the esfri projects . with the esfri facilities and their pathfinders included in the vo
, astronomers will be able to discover , access , use and compare their data , combining it with data from other ground- and space - based observatories as well as theoretical model collections .
the goal of dadi is to make the esfri and pathfinder project data available for discovery and usage in the international vo framework , and accessible with the set of vo - enabled common tools . more specifically : i ) train and support esfri projects staff in the usage and implementation of the vo framework and tools , and make them active participants in the development of the vo framework definition and updates , thus contributing to relevance and sustainability of the framework ; ii ) train and support the wider astronomical community in scientific use of the framework , in particular for pathfinder data , and gather their requirements and feedback ; iii ) adapt the vo framework and tools to the esfri projects needs , in continuous cooperation with the international virtual observatory alliance ( ivoa ) . * connecting locations of esfri observatories and partners in astronomy for timing and real - time alerts ( cleopatra ) * + the partners in asterics share an ambition to use modern communication methods , such as fast broadband connectivity , to improve the scientific capabilities of their research infrastructures .
the research activities aim specifically at synergetic observing modes , and fast and reliable access to large data streams .
these aspects are covered in cleopatra : i ) develop technology for the enabling of long - haul and many - element time and frequency distribution over fibre connections .
this has the potential to increase the efficiency and affordability of all radio astronomy facilities ( ska , lofar , evn ) ; such developments are also highly relevant for astroparticle facilities ( cta , km3net ) and can enable novel realtime multi messenger observations ; ii ) develop methods for relaying alerts , which will signal transient event detections between the facilities and enable joint observing programmes ; the focus is both on interchange formats and on scientific strategies and methods for joint observing ; iii ) further development of existing data streaming software , building on previous e - vlbi projects , and providing tools for robust and efficient data dissemination for all facilities in the user domain , including eso facilities such as alma and the e - elt ; iv ) foster the development of advanced scheduling algorithms , using ai approaches for optimal usage of the esfri facilities , so to achieve a consistent set of enhancements of the facilities based on developments in connectivity and data transport .
* interactions among work packages * + obelics and dadi have a strong and interrelated focus on delivering common solutions , standards and analysis to the management and exploitation of large volume and high velocity data streams .
key goals include interoperability between facilities such as the ska , cta , euclid , ego , km3net etc .
both obelics and dadi also entail significant training opportunities for external stakeholders , in order to ensure that esfri projects staff and users are fully engaged with the asterics
programme.there is a direct connection between these activities , and a specific task in obelics is dedicated to the interface and coordination with dadi .
quite naturally , as all work packages need a dissemination activity to be carried out , and need to be managed , all of them will have a direct connection to decs and amst .
using an inclusive , engaging and open approach in the preparation of the project work plan , the asterics partners have arrived at a concept that is relevant to the implementation of the astronomy related esfri facilities , and that is supported and endorsed by the main players of the multi - disciplinary astronomy and astroparticle physics communities .
asterics is an ambitious project but the goals are clear and attainable .
the ultimate measure of its success is how often the project results and products are incorporated into the esfri observatories . on timescales similar to the duration of the project ( 4 years )
, success can also be measured via implementation on the esfri pathfinder telescopes the latter provide existing platforms on which asterics technologies can be tested and proven .
the incorporation of other related projects ( e.g. euclid , ego ) provides further testing opportunities , and additionally addresses the demand to realise the potential offered by combining astrophysics and cosmology together .
asterics is a project funded by the european commission under the horizon2020 programme ( i d 653477 ) .
all the partner institutions and individual participants are gratefully acknowledged for their work in the project . | the large infrastructure projects for the next decade will allow a new quantum leap in terms of new possible science .
esfri , the european strategy forum on research infrastructures , a strategic initiative to develop the scientific integration of europe , has identified four facilities ( ska , cta , km3net and e - elt ) deserving priority in support .
the asterics project aims to address the cross - cutting synergies and common challenges shared by the various astronomy esfri and other world - class facilities .
the project ( 22 partners across europe ) is funded by the eu horizon 2020 programme with 15 meuro in 4 years .
it brings together for the first time the astronomy , astrophysics and particle astrophysics communities , in addition to other related research infrastructures . |
1 . momentum correlation between the two emitted electrons given by present calculations . | * abstract * in this paper , we have investigated the correlated electron emission of the nonsequence double ionization ( nsdi ) in an intense linearly polarized field .
the theoretical model we employed is the semiclassical rescattering model , the model atom we used is the helium .
we find a significant correlation between magnitude and direction of the momentum of two emission electrons , and give a good explanation for this striking phenomenon by observing the classical collisional trajectories .
we argue that this correlation phenomenon is universal in nsdi process , as revealed by the recent experiment on the argon .
the excessive double ionization observed in helium experiments @xcite draws much attention to the multiple - electron dynamics in the laser - atom interaction . in these experiments the single ionization of he in a linearly polarized field
is accurately predicted by the single active electron ( sae ) approximation @xcite , well described by the ammosov - delone - krainov ( adk ) tunnelling theory @xcite . however , the case of double ionization is more complicated . in the regime of very high intensities ( @xmath0w/@xmath1 ) where strong double
ionization occurs , the double ionization keeps in good agreement with the sequential sae models as that in the lower intensities regime(@xmath2w/@xmath1 ) .
the double ionization deviates seriously from the sequential sae model and shows a great enhancement in a `` knee '' regime @xmath3@xmath4w/@xmath5 $ ] .
this surprising large yields of the double ionization obviously indicates that the sequential ionization is no longer the dominating process in this regime and the electron - electron correlation has to be taken into account .
intense efforts to model the two - electron process of the double ionization in a laser field have reproduced the main feature of the knee structure in the double ionization yield as a function of laser peak intensity and , moreover , yielded quantitative agreement with the experiments in some cases @xcite .
the physical mechanism behind this nonsequential process is , however , still debatable .
both the `` shake - off '' model and the `` recollision '' model are suggested to describe the electron s correlation @xcite . however
, none of the two nonsequence double ionization ( nsdi ) mechanisms can completely explain the experimental observations .
for the `` shake - off '' model , it can not give the reason for the decrease in the double ionization yields as the polarization of the laser field departs from linear @xcite .
in the `` recollision '' model , the returning electrons are known to have a maximum classical kinetic energy of @xmath6 @xmath7 , so one can determine a minimum intensity required for the rescattering electron to have enough energy to excite the inner electron .
but the double ionization yields observed in experiments has no such an intensity threshold .
in fact , the double ionization process is rather complicated and subtle , both of the two nsdi processes and the sequential ionization have contributions to it and may dominate in the different regimes .
in another aspect , becker and faisal proposed a `` correlated energy sharing '' model to describe the nsdi processes and nuclei recoil experiment @xcite .
the model is based on the so - called intense - field many - body s - matrix theory derived by a rearrangement of the usual s - matrix series and include time electron correlation and the rescattering mechanism .
we have employed the semiclassical model to study the double ionization of helium in intense linearly polarized field @xcite .
our calculations reproduced the excessive double ionization and the photoelectron spectra observed experimentally both quantitatively and qualitively , and we argue that the classical collisional trajectories is the main source of the nonsequence double ionization of helium in the `` knee '' regime .
recently , the observation of the correlated electron emission in laser - induced double ionization of argon @xcite provided new insights into the nsdi process .
these authors reported a strong correlation between the direction and the magnitude of the momenta of two electrons emitted from an argon atom : the momenta of the two emission electrons tend to have the same magnitude and sign in the polarization direction . on the theoretical side , by solving the time - dependent schrdinger equation for two electrons in three dimensions , taylor et al @xcite gave that the most of double ionization probability flux tends to emerge to the same side of the ion .
similar conclusions have been drawn from one - dimensional model @xcite . in this paper ,
based on the 3d semiclassical rescattering model developed recently @xcite , we investigate the dynamical behavior of the correlated electrons in the double ionization process by analyzing their classical trajectories .
this investigation , as shown later , is very helpful to understand the physical mechanism behind the momentum correlation .
the model atom we use is the helium , however we argue that our discussions are available to the other multi - electron atoms , like the argon in recent experiment@xcite .
firstly , we briefly present the semiclassical rescattering model adopted in our calculations .
the ionization of the first electron from bound state to the continues state is treated by the tunnelling ionization theory generalized by delone et al . @xcite .
the subsequent evolution of the ionized electron and the bound electron in the combined coulomb potential and the laser fields is described by a classical newtonian equation . to emulate the evolution of the electron , a set of trajectories
is launched with initial conditions taken into from the wave function of the tunnelling electron .
the evolution of the two electrons after the first electron tunnelled are described by the classical equations ( in atomic unit ) : @xmath8 here @xmath9 is the laser field .
the indices @xmath10 and @xmath11 refer to the tunnel ionized and bound electron respectively .
the potentials are @xmath12 the initial condition of the tunnelled electron , under the sae approximation of he@xmath13 , is determined by a equation including the effective potential given in ref .
@xcite and a generalized tunnelling formula developed by delone et al .
. in parabolic coordinates , the schrdinger equation for a hydrogen - like atom in a uniform field @xmath14 is written ( in atomic unit ) , @xmath15 in which @xmath16 @xmath17 is the negative ionization potential of the outer electron .
the above equation has the form of the one - dimensional schrdinger equation with the potential @xmath18 and the energy @xmath19 the evolution of the outer electron is traced by launching a set of trajectories with different initial parameters @xmath20 and @xmath21 , where @xmath22 is the initial velocity perpendicular to the polarization of the electric field .
the initial position of the electron born at time @xmath20 is given by @xmath23 , @xmath24 from the eq .
( [ sch ] ) .
the initial velocity is set to be @xmath25 @xmath26 .
thus , the weight of each trajectory is evaluated by @xcite @xmath27 @xmath28 and where @xmath29 is the tunnelling rate in the quasistatic approximation @xcite .
the initial state of the bounded electron is described by assuming that the electron is in the ground state of he@xmath13 with energy @xmath30 and its initial distribution is microcanonical distribution @xcite . in our calculation , the eqs .
( [ eq1 ] ) are solved in a time interval between @xmath20 and @xmath31 by employing the standard runge - kuta algorithm@xmath32 during the first ten optical cycles the electric field amplitude is constant , and then the field is switched off using a @xmath33 envelope during three cycles , and during the last two optical cycles the electrons is free from the electric field .
the wavelength is @xmath34 @xmath35 , which is so chosen to match the experiment @xcite , and the intensity of laser is @xmath36 which is right at the ` knee ' region of the double ionization of helium . @xmath37 or more initial points are randomly distributed in the parameter plane @xmath38 @xmath39 for the outer electron and in the microcanonical distribution for the inner electron .
the distribution for the ionization electron can be obtained by making statistics on an ensemble of classical trajectories weighed by the ( [ wei1 ] ) .
the results have been tested for numerical convergence by increasing the number of trajectories .
figure 1 shows the momentum correlation between the two emission electrons in the double ionization of the present calculations .
the horizontal axis shows the momentum component of the first electron in the direction of polarization @xmath40 and the vertical axis the same momentum component of the second electron @xmath41 this figure shows a strong correlation between the momenta of the two electrons .
there is a clear maximum for both electrons being emitted with the same momentum component in the direction of polarization axis of about @xmath42 @xmath17 , and emission to opposite half planes is strongly suppressed , i.e. , both two electrons tend to fly to same side of ion in the direction of polarization .
this phenomena has been observed in the ` knee ' region for argon @xcite . on the other hand , from fig . 1
, we see that the maximum momentum of both electron is about @xmath43 @xmath44which is consist with the electron - ion coincidence experiment observation of helium @xcite , in which the maximum energy of emission electron in nsdi process is @xmath45 since the perpendicular component of momentum is small , the maximum momentum component in the polarization direction can be approximate obtained as @xmath46 @xmath47 an useful alternative perspective on fig .
2 is obtained by rotating the distribution by @xmath48 .
then we can get two new distributions . in fig .
2(a ) , we show the distribution of the sum momentum @xmath49 ; in fig .
2(b ) , we show the distribution of the difference momentum @xmath50 @xmath51 . owing to momentum conservation , the sum momentum of emission electrons is equal and opposite to the he@xmath52 recoil ion momentum @xcite , so its distribution , as we know , shows a characteristic double - hump structure with a central minima .
a vast amount of literature has devoted to discuss the double - hump structure of momenta of recoil ions @xcite , here we only pay attention to the correlation between the two emission electrons . comparing the fig.2(a ) with fig.2(b ) , one finds that the sum momentum at the peak is about @xmath53 , almost above the maximum momentum of each electron , and the maximum sum momentum is almost @xmath54 @xmath44 about twice of the maximum momentum of one electron ; furthermore the distribution width of difference moment is much smaller than the one of the sum momentum .
these features indicate the momenta parallel to polarization axis of the two emission electrons likely have the same direction . on the other hand ,
the peak of the distribution of difference momentum is at zero , so the momenta of the two emission electron tend to have the equal magnitude . to study the origin of the correlation of electrons emission , we show two trajectories of the electrons in fig .
3 . fig .
3(a ) shows a typical trajectory of which the difference momentum in polarization direction is small and fig .
3(b ) shows a typical trajectory of which the difference momentum in polarization direction is nearly the maximum . from these trajectories
, it can be concluded that after the second electron ionized the momentum transfer received from the laser field is almost identical for both electron . because the electron can obtain very high energy from the laser
, the momenta of the two thus be accelerated to the same direction . on the other hand , this part of the acceleration only adds to the sum of the momenta of the two electrons , but does not change the difference of the momenta
. therefore the difference of the momenta is only determined by the ionization process of the second electron . as we have discussed before
, the second electron is mostly ionized by the `` recollision '' mechanism : the second electron is ionized by a collision with the tunnelled electron @xcite . because the collision between the two electrons is almost instantaneous so that the energy is conserved approximately when the collision happens .
the total energy of the two electron can be expressed by @xmath55 where @xmath56 includes the potentials and the kinetic energy of the perpendicular parts . to the best of our knowledge
, so far the energy distribution of the returning electron has not been reported , but we believe that it exhibits a peak at zero energy and decreases as the energy increasing , on the other hand , only the returning electron which energy is above the ionization potential of the bound electron @xmath57 can cause the second electron ionized , so the total energy of the two electrons soon after the collision occurred is small and its distribution peaks at zero .
therefore the @xmath58 is more likely zero , i.e. , the momentum components in the polarization direction of the two emission electrons are likely equal to each other .
the returning electrons are known to have a maximum classical kinetic energy of @xmath59 therefore the maximum difference momentum must satisfy @xmath60 so @xmath61 . on the other hand ,
since the total energy of the two electrons soon after the collision occurring is small , the total final energy of the two emission electrons mostly reflect the energy transfer received from the field .
the field acceleration make the sum momentum increase .
the maximum energy for an electron in the double ionization process is about @xmath62 @xcite , from eq .
( [ energy ] ) , assuming the two emission electron have the same energy , we can obtain the maximum sum momentum @xmath63 .
this indicates that the joint acceleration of the electrons in the laser field clearly dominates over the influence of the electron repulsion , and both electrons ionized are driven by the laser to the same side . in conclusion
, we have investigated the momentum correlation between magnitude and direction of the two emission electrons in nonsequence double ionization .
the numerical results on the helium show a significant correlation on the momentum of the two electrons : the emission electrons tend to have the same momentum component in the polarization direction .
the phenomena can be directly comprehended from the classical collisional trajectories .
these discussions suggest that the correlated electrons emission in double ionization process is a semiclassical process .
we also evaluated the width of distribution of the sum and difference momentum .
because the difference momentum is only determined by the ionization process , so it is important to verify the dominating process in the `` knee '' regime .
based on the rescattering model , we argue that the maximum difference momentum of the two emission electrons is @xmath64 the predictions coincide with the argon experiment @xcite .
we hope our discussions will stimulate the experimental works on the helium in the direction . we are indebt to dr . jing chen for reading our paper and proposing many suggestive opinions
this work was supported by the 973 project of china .
d.n .
fittinghoff , p.r .
bolton , b. chang , and k.c .
kulander , phys .
rev .
lett . * 69 * , 2642 ( 1992 ) .
b. walker , b. sheehy , l.f .
dimauro , p. agostini , k.j .
schafer , and k.c .
kulander , phys .
rev .
lett .
* 73 * , 1227 ( 1994 ) b. sheehy , r. lafon , m. widmer , b. walker , l.f .
dimauro , p.a .
agostini , and k.c .
kulander , phys .
rev .
a * 58 * , 3942 ( 1998 ) m.v .
ammosov , n.b .
delone , and v.p .
krainov , sov .
phys .
jetp * 64 * , 1191 ( 1986 ) a. becker and f.h.m .
faisal , j. phys .
b * 32 * , l335 ( 1996);becker and f.h.m .
faisal , j. phys .
b * 29 * , l197 ( 1996 ) li - bin fu , jie liu , jing chen and shi - gang chen , phys .
rev .
a * 63 * , 043416(2001 ) p.b .
corkum , phys .
rev .
lett . * 71 * , 1994 ( 1993 ) k.c .
kulander , j. cooper , and k.j .
schafer , phys .
rev .
a * 51 * , 561 ( 1995 ) b. walker , e. mevel , b. yang , p. breger , j.p .
chambaret , a. antonetti , l.f .
dimauro , and p.a .
agostini , phys .
rev .
a * 48 * r894 ( 1993 ) d.n .
fittinghoff , p.r .
bolton , b. chang , and k.c .
kulander , phys .
rev .
a * 49 * , 2174 ( 1994 ) k. kondo , a. sagisaka , t. tamida , y. nabekawa , and s. watanabe , phys .
rev .
a * 48 * , r2531 ( 1993 ) a. becker and f.h.m .
faisal , phys .
rev .
a * 59 * , r1742 ( 1999 ) a. becker and f.h.m .
faisal , phys .
rev .
lett . * 84 * , 3546 ( 2000 ) j. chen , j. liu , l.b .
fu and w.m .
zheng , phys .
rev .
a * 63 * , 011404(r ) ( 2001 ) th .
weber , h. giessen , m. weckenbrock , g. urbasch , a. staudte , l. spielberger , o. jagutzki , v. mergel , m. vollmer and r. drner , nature 405 , 658 ( 2000 ) k.t .
taylor , j.s .
parker , d. dundas , e. smyth , and s. vitirito , laser phys .
* 9 * , 98 ( 1999 ) m. lein , e.k.u .
gross and v. engel , j. phys .
b * 33 * , 433 ( 2000 ) ; m. drr , optics express * 6 * , 111 ( 2000 ) n. b. delone , and v. p. krainov , j opt .
soc . am .
b * 8 * , 1207 ( 1991 ) l.d .
landau , e. m. lifishitz , _ quantum mechanics .
_ ( rergamon , oxford , 1977 ) a.m. perelomov , v.s . popov and v.m .
terenev , zh . eksp .
teor .
fiz .
50 , 1393 ( 1966 ) ; m.v .
ammosov , n.b .
delone , and v.p .
krainov , ibid 91 , 2008 ( 1986 ) j.s .
cohen , phys .
rev .
a * 26 * , 3008 ( 1982 ) th .
weber et al . , phys .
rev .
lett . * 84 * , 443 ( 2000 ) r. lafon , j.l .
chaloupka , b. sheehy , p.m. paul , p.agostini , k.c .
kulander , and l.f .
dimauro , phys .
rev .
lett . * 86 * , 2762 ( 2001 ) |
despite the numerous cosmological and astrophysical indications of the presence of non - baryonic dark matter ( dm ) , the particle nature of this substance remains unknown .
if the dm consists of weakly interacting massive particles ( wimps ) , an important tool for inferring their properties could be `` indirect detection '' , i.e. astrophysical observations of the annihilation ( or decay ) products of dm in our galaxy or beyond . for wimps with masses at or around the electroweak scale , @xmath0(0.1 - 1 ) tev
, the annihilation products are typically found at gev - tev energies , the domain of high - energy astrophysics . of the different annihilation products , gamma rays and neutrinos
have the important advantage of retaining directional information while not suffering energy losses .
the very small cross sections of neutrinos , however , make their flux from the region of the galactic center very difficult to detect . on the contrary
, the gamma - ray spectrum from dm annihilation or decay may be detectable with sufficient statistics , energy resolution , and over an extended angular distribution , to provide a very distinctive set of information related to both the particle identity of the wimp and its astrophysical distribution .
a major challenge is in separating the dm signal from any astrophysical backgrounds , whose energy spectrum and angular distribution are not well known .
a major change in the prospects for dm detection has occurred in the last few years , following the discovery of a bright astrophysical source of tev gamma rays from the galactic center . as a result of this discovery , we now know that dm emission from the galactic center will not be detectable in a ( quasi ) background - free regime , and unless one turns the attention to other targets the peculiar spectral shape and angular distribution of the signal must be used to extract it from this and other backgrounds .
this topic is the main subject of this review .
the remainder of this paper is structured as follows : in sec .
[ instr ] we briefly describe the capabilities of current high energy gamma - ray telescopes . in sec .
[ signal ] we review the features of the dm annihilation signal , both in its energy ( sec . [ espectra ] ) and angular dependence ( sec .
[ angshape ] ) . in sec .
[ gc ] we review the prospects for detecting or constraining dm annihilations in the galactic center region , while in sec .
[ innerhalo ] we briefly discuss some interesting aspects of the diffuse signal from the inner halo . in sec .
[ conclusions ] we report our conclusions . appendix a briefly illustrates the qualitative changes to the considerations we develop in the main text when decaying ( as opposed to annihilating ) dm is considered .
appendix b describes a subtlety regarding the angular shape of the signal arising for a velocity - dependent annihilation cross section .
the ability of gamma - ray experiments to identify dm annihilation radiation from the galactic center region relies on the effective area , angular and energy resolution of the existing telescopes , as well as the rejection of other ( mostly hadronic ) cosmic ray events contaminating the gamma - ray sample . since the mean free path of gamma rays
is much shorter than the atmospheric slant depth , direct observations in the gev region and above can only be done from space which is the strategy pursued by the lat detector on the fermi gamma - ray space telescope ( formerly glast ) @xcite or indirectly by ground - based imaging atmospheric cerenkov telescopes ( iacts ) such as hess @xcite , magic @xcite , veritas @xcite and cangaroo - iii @xcite . in the latter category ,
the direction and energy of the primary particle hitting the atmosphere is reconstructed from the cherenkov emission of the secondary charged particles generated in the atmospheric shower .
these differences lead the two classes of experiments to adopt different strategies in the search for dm .
fermi - lat is very effective in rejecting hadronic events , and continuously monitors a large fraction of the sky , but has an effective area of only @xmath1m@xmath2 , far smaller than that of ground - based telescopes , @xmath3m@xmath2 .
on the other hand , iacts study small angular fields and have a lower rejection capability , but much greater overall exposure . as a consequence , diffuse gamma - ray signals are better probed by fermi - lat .
any unidentified sources detected by fermi - lat which lack a low - energy counterpart could be potentially attributed to dm substructure .
iacts would be very effective in providing detailed follow - up observations of such sources .
in addition , the accessible energy range is very different between these two classes of experiments : @xmath4100 mev to 300 gev for fermi - lat , and above @xmath4100 gev for acts .
this difference makes fermi - lat most sensitive to dm particles lighter than a few hundreds gev , while iacts are better suited for tev - scale or heavier wimps . on the other hand ,
the fermi - lat has poorer angular resolution than iacts , so it is less accurate in the localization of point - like sources . for both instrument classes ,
the search for indirect dm signatures is among the top physics priorities .
the reach of fermi - lat and current and future iacts has been recently assessed in ref .
@xcite and ref .
@xcite , respectively .
the differential flux of gamma rays ( photons per unit area , time , energy and steradian ) produced in dm annihilations is described by @xmath5\int_{\rm{los } } \rho^2(\ell,\omega)\ , \d \ell , \label{flux1}\ ] ] where @xmath6 is the wimp annihilation cross section multiplied by the relative velocity of the two wimps ( averaged over the wimp velocity distribution ) , @xmath7 is the mass of the wimp , @xmath8 is the position - dependent dm density , and the integral is performed over the line - of - sight ( los ) in the direction of the sky , @xmath9 . the gamma - ray spectrum generated per wimp annihilation is @xmath10 , it has units of energy@xmath11 and its integral over energy is equal to 1 . if the dm is not its own antiparticle as assumed here , eq .
( [ flux1 ] ) should be multiplied further by a factor 1/2 ( if @xmath12 and @xmath13 are equally abundant ) .
the factor in square brackets in eq .
( [ flux1 ] ) depends only on particle physics : in particular , cross section , mass , and the spectrum of gamma rays produced through dm annihilations depends on the nature of the wimp .
the integral over the line - of - sight determines instead the angular dependence of the signal and is controlled by the astrophysical distribution of dm .
we shall discuss each of these two terms in the following , assuming that this factorization holds ( see [ entanglingeandang ] for a discussion of possible violations of this hypothesis ) .
in convenient units , eq .
( [ flux1 ] ) can be recast as : @xmath14 the dimensionless function @xmath15 depends only on the dm distribution in the halo and is defined by convention as @xcite @xmath16 for typical halo models ( see sec .
3.2 ) this is a function strongly peaked towards the galactic center ( for an illustration , see for example fig .
a1 ) . the benchmark value for the cross - section , @xmath171 picobarn ( or in cgs units @xmath18 ) ,
is motivated by the fact that a wimp annihilating with such a cross section during the freeze - out epoch will be generated as a thermal relic with a density similar to the measured dm abundance ( for a review , see ref .
wimps constituting the _ cold _ dm annihilate in the non - relativistic limit
. if annihilations take place largely through @xmath19-wave processes , then the annihilation cross section of wimps in the galactic halo ( i.e. in the low velocity limit ) will also be approximately equal to this value , which justifies the benchmark value used in eq .
( [ flux2 ] ) . yet , it is important to stress that much lower signals are possible ( e.g. if @xmath20wave annihilation dominates the freeze - out process ) , as well as significantly enhanced ones in models where the dm is non - thermally produced in the early universe ( see e.g. ref .
@xcite ) . some further considerations and refs . can be found in [ entanglingeandang ] .
it is also worth noting that for s - wave annihilating thermal relics the indirect detection signal has two advantages compared to direct detection via nuclear recoils in underground detectors : i ) it is proportional to a relatively large annihilation cross section ; ii ) it is less dependent from the particle physics details .
for example , even under the ( possibly unrealistic ) assumption that dm annihilates mostly into quarks of the first generation , the natural expectation value for dm - nucleon elastic cross section is at the level of @xmath21pb@xmath22cm@xmath2 . for dm annihilating predominantly into heavier particles ,
a further suppression is expected .
[ spectra ] the gamma - ray spectrum from dm annihilation originates from several different contributions . typically , the most abundant source of photons is the hadronization and/or decays of unstable particles .
for example , neutralinos in the minimal supersymmetric standard model ( mssm ) , dominantly annihilate to final states consisting of heavy fermions @xmath24 , @xmath25 , @xmath26 ( i.e. bottom quarks , top quarks and tau leptons , respectively ) or bosons @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , where @xmath35 are the gauge bosons mediating the weak interactions and @xmath36 , @xmath37 , @xmath38 and @xmath39 are the higgs bosons of the mssm @xcite . with the exception of the @xmath40 channel , each of these annihilation modes result in a very similar spectrum of gamma rays , dominated eventually by the decay of mesons , especially @xmath41 , generated in the cascade . in fig . 1 we show the predicted gamma - ray spectrum per annihilation , for several possible wimp annihilation modes .
the harder spectrum here shown from annihilation into the @xmath42 channel is not typical of susy models , although it might be a distinctive feature of dm annihilating dominantly to charged leptons pairs ; this arises , for example , in kaluza - klein dm models ( for a review of particle dm models , see e.g. @xcite ) .
although these secondary photons provide the dominant emission , other important channels can exist at the one - loop level .
a particularly striking signature would be the mono - energetic photons resulting from final states such as @xmath43 , @xmath44 or @xmath45 ( see ref .
@xcite for a discussion of these processes within the context of supersymmetry ) .
unfortunately , such processes are expected to produce far fewer events than continuum emission and in typical models can not easily be detected ( for a counterexample , see ref .
bremsstrahlung ( with an additional photon appearing in the final state ) is automatically present when annihilations produce charged final states , and can dominate the high energy region of the spectrum when those charged annihilation products are much lighter than the wimps .
this is particularly important when the tree - level processes to a pair of light fermions are disfavored by `` selections rules '' , but no suppression is present for three body final states @xcite . for more details ,
see the review by l. bergstrom in this issue @xcite .
note that gamma - ray emission from dm annihilating into standard model particles is unavoidable , even in the most extreme case when only neutrino final states are allowed at tree level , due to @xmath46 and @xmath47-strahlung @xcite . the gamma - ray spectrum and angular distribution predicted by eq .
( [ flux1 ] ) is rather general ( see , however , appendix b ) , and could be applied to the case of a smoothly distributed galactic halo , or alternative targets such as dwarf spheriodal galaxies , microhalos / clumps , density spikes around intermediate mass black holes , or the integrated extragalactic diffuse background ( described in other contributions to this special issue ) . while some of these targets have interesting observational prospects , the intensity of these signals is strongly dependent on often unknown cosmological and astrophysical properties , such as the quantity of small - scale structures in dm halos or the population of intermediate mass black holes in the milky way . in this article , we will focus on the smooth galactic halo . naively , by inputing the dm density profile inferred from kinematical observations of the milky way into eq .
( [ flux1 ] ) , one could obtain an approximate lower bound on the gamma - ray flux from dm annihilation . in principle , this would enable us to translate the observations from ground or space - based gamma - ray telescopes into constraints on the particle physics properties of the wimp ( mass , annihilation cross section and dominant modes ) . unfortunately ,
even the average , smooth distribution of the dm particles in our galaxy is not well known , especially in the volume within the solar circle .
this is due to the fact that the baryonic material dominates the gravitational potential in the inner galaxy , and lacking a detailed knowledge of its distribution , a reconstruction of the dm distribution `` by subtraction '' is unfeasible .
it is not surprising , then , that very different profiles have been claimed to fit the observations ( for example , compare the results of ref .
@xcite and ref .
@xcite ) . on general grounds , a class of spherically symmetric , smooth halo distributions can be used to approximately fit both the observed rotation curves of galaxies and the results of numerical simulations of dm halos . a non - trivial angular dependence of the gamma - ray signal results from the off - center position of the sun within the halo ( see e.g. ref .
@xcite and references therein for a discussion of sub - leading effects determining the angular distribution of the signal ) .
the function @xmath48 introduced in eq .
( [ flux2 ] ) and eq .
( [ jpsi ] ) then depends only on the angle @xmath49 between the observed direction of the sky and the galactic center or , in terms of galactic latitude @xmath50 and longitude @xmath51 , only on @xmath52 .
the radial variable , @xmath53 , can be expressed in terms of the relevant quantities @xmath54 as @xmath55 where @xmath56kpc @xcite is the distance of the solar system from the galactic center .
typically , one considers a distribution of the form @xmath57^{(\beta-\gamma)/\alpha}},\label{prof}\ ] ] where @xmath58 is a normalization constant and @xmath59 is a characteristic radius below which the profile scales as @xmath60 .
a very well known , universal profile of this class fit to dm - only ( _ i.e. _ neglecting baryons ) n - body simulations has been proposed by navarro , frenk and white ( nfw ) @xcite , corresponding to the choice @xmath61 .
steeper or softer profiles have also been extensively discussed , such as that proposed by moore et al .
@xcite and kravtsov et al .
@xcite , respectively .
while simulations ( and data as well ) typically agree on the shape of profile in the outskirts of the halos , a disagreement clearly exists concerning the inner slope , @xmath62 .
more recent simulations @xcite suggest that halo density profiles are better represented by a function with a continuously - varying slope , as the one proposed by einasto @xcite @xmath63}\,,\label{einasto}\ ] ] with clear hints for non - universality in the form of halo - to - halo variations in the quantity @xmath64 . the quantity @xmath65kpc denotes the radius at which the logarithmic slope of the profile , @xmath66 , assumes the value @xmath67 ; the other free parameters are @xmath64 and the overall normalization , here chosen as the density @xmath68 . at @xmath69 kpc
, these newly proposed fitting formulae provide only marginal improvement with respect to the more traditional ones . at smaller
radii , however , these recent results lead us to expect the inner slope of dm halos to be shallower that that predicted by moore _ et al . _ , and probably shallower than nfw , as well . [ tableii ]
.parameters describing some common halo profiles of the form described by eq .
( [ prof ] ) , where @xmath70 is the dm density at the solar distance from the gc .
see text for more details . [
cols="^,^,^,^,^,^ " , ] there are a number of astrophysical processes that may potentially modify the dm distribution , none of which are taken into account in the above - mentioned profiles .
it is very difficult to reliably account for these effects in simulations , and only a few results are available ( see e.g. refs . @xcite ) .
qualitatively , since baryons can cool and contract , one expects them to steepen the gravitational potential in the central regions galaxies and , as a result , enhance the dm density @xcite . on the other hand , other feedback or frictional effects
have been proposed that could reduce the dm density in the inner halos and bring the prediction closer to observations ( for a recent review , see ref .
@xcite ) . for some galaxies , it has actually been argued that flat - cored profiles fit the observations better than cusped profiles .
even greater uncertainties exist concerning the dm distribution at the very center of the milky way , in the region immediately surrounding the central supermassive black hole .
adiabatic accretion may lead to the formation of a spike in the dm distribution , resulting in a very high dm annihilation rate in the innermost parsecs of the galaxy @xcite .
mergers as well as scattering on the dense stellar cusp around the central black hole may potentially destroy density enhancements , however . in general , these effects only affect regions too close to the galactic center to be resolved angularly by present detectors , leaving only the energy spectrum to be used for separating the dm signal from the background .
yet , such a spike might lead to a measurable gamma - ray flux from the innermost angular bin , even in presence of relatively large astrophysical backgrounds . for more details , we direct the reader to the references cited in ref .
@xcite or ref .
@xcite . given these considerably uncertainties ,
we have chosen to use three illustrative choices for the dm halo profiles : the moore _
_ , nfw , and kravtsov profiles ( with the parameters given in table i ) . in each case , as in ref .
@xcite , the normalization has been chosen so that the mass contained within the solar circle provides the appropriate dm contribution to the local rotational curves . in all of the cases we have discussed , even for the most conservative cored profile , the dm signal peaks at the galactic center . in presence of an isotropic ( or even vanishing ) astrophysical background , the galactic center region thus becomes the natural location to look for a dm signal .
unfortunately , there is also a higher background density as we look toward the inner region of our galaxy .
still , as a first step , one can consider the detection prospects for this region , a task we address in sec .
[ gc ] , before turning to more general arguments regarding the optimal regions of the sky for dm searches in sec .
[ innerhalo ] .
the galactic center has long been considered to be among the most promising targets for the detection of dm annihilation , particularly if the halo profile of the milky way is cusped in its inner volume @xcite .
this has been complicated , however , by the recent discovery of astrophysical gamma - ray sources from the galactic center . following an earlier claim by the whipple iact @xcite , very high - energy gamma rays from the galactic center
have been detected by hess @xcite , magic @xcite , and cangaroo - ii @xcite .
this source is consistent with point - like emission and is located at @xmath71 ( stat ) , @xmath72 ( stat ) with a systematic pointing error of 28@xmath73 @xcite , coincident with the position of sgr a@xmath74 , the black hole constituting the dynamical center of the milky way .
the spectrum of this source is well described by a power - law with a spectral index of @xmath75 over the range of approximately 160 gev to 20 tev .
although speculations were initially made that this source could be the product of annihilations of very heavy ( @xmath410 to 50 tev ) dm particles @xcite , this interpretation is disfavored by the power - law form of the observed spectrum and the wide energy range over which it extends ( see fig . [
fig : edis ] ) .
spectral energy density of the galactic center source as measured by hess in 2004 ( full points ) and 2003 @xcite ( open points ) .
upper limits are 95% cl .
the shaded area shows the power - law fit @xmath76 , with @xmath77(stat.)@xmath78(syst . ) .
the dashed line illustrates typical spectra of phenomenological mssm dm annihilation for best fit neutralino masses of 14 tev .
the dotted line shows the distribution predicted for kaluza - klein dm with a mass of 5 tev .
the solid line gives the spectrum of a 10 tev dm particle annihilating into @xmath79 ( 30% ) and @xmath80 ( 70% ) . from ref .
] the source of these gamma rays is , instead , likely an astrophysical accelerator associated with our galaxy s central supermassive black hole @xcite . in recent analyses ,
this source has been treated as a background for dm searches @xcite .
given the presence of this background , the prospects for detecting dm annihilation products from the galactic center appear considerably less promising than they had a few years ago . the main challenge involved in dm searches with fermi will be to distinguish the signal from this and other backgrounds .
this is made particularly difficult by our ignorance regarding the nature of these backgrounds .
the galactic center is indeed a complex region of the sky at all wavelengths , the gamma - ray window being no exception @xcite .
an attempt to use angular and spectral information to separate dm annihilation products from these backgrounds was performed in ref .
@xcite , studying a @xmath81 region arond the galactic center ( see also the analysis of the fermi dm team in ref .
this study considered two known point - source backgrounds : a yet unidentified source detected by egret approximately 0.2@xmath82 away from the dynamical center of our galaxy @xcite , and the iact source discussed above .
in addtion , a diffuse spectrum with a free power - law index was also included .
more in detail , we describe the spectrum of the source revealed by iacts at the galactic center as a power - law given by : @xmath83 the flux from the egret source slightly off - set is instead modeled as : @xmath84 where the exact value of the cutoff energy ( here 30 gev ) is somewhat arbitrary , but reflects the fact that this source has not been observed yet by iacts at @xmath85gev .
finally , we allow for a diffuse / unresolved flux with spectrum @xmath86 where @xmath64 is allowed to vary between 1.5 and 3.0 .
we adopt an overall normalization , @xmath38 , such that the integrated flux of the diffuse background between 1 gev and 300 gev in a @xmath81 field of view around the galactic center is equal to @xmath87 . we do not , however , assume that this normalization is known in our analysis , leaving open the possibility that some of the diffuse gamma rays observed are the product of dark matter annihilations . a multi - parameter @xmath88 analysis of the simulated sky against models including a contribution from dm annihilation radiation yields the projected exclusion limits at the 95% confidence level shown in fig
. [ limitnfw ] ( for ten years of collection time by fermi - lat ) .
the simulated sky used to produce this figure contains the two resolved sources and the diffuse background but no dark matter , and to derive the exclusion limits we compare the signal thus obtained to a model which includes both the backgrounds and a signal from dark matter .
the left and right panels refer to the cases of an nfw profile and moore _
_ profile ( replaced by a flat core within @xmath89 pc of the galactic center to avoid a divergency ) , respectively .
the wimp is always assumed to annihilate dominantly to @xmath24 .
this assumption is not particularly restrictive , since for many annihilation modes the spectrum would look very similar ( see fig
apart for some cases , as possibly kaluza klein dm , the gamma - ray spectrum alone does not help much in discriminating among several dm candidates , see e.g. @xcite for details . to give a feeling for the numbers involved , the integral of the function @xmath48 over the considered region leads to a factor @xmath90 for the nfw model ; in this model , for the benchmark 100 gev wimp with @xmath91 @xmath92/s one would expect less than 3000 events from dm above 1 gev , to be compared with more than 2.3@xmath93 background photons .
the background is almost evenly split between the egret point source and the diffuse flux ; the iact source contributes less than 7000 events , but it is located at the gc and , differently from the egret source , is not cut - off at high energy , so it is important to include it especially for high wimp masses .
the solid line in each frame represents the limit found if the diffuse background is assumed to be distributed isotropically , while the dashed line represents the conservative limit obtained if the diffuse background has the same angular distribution as the dm signal ( _ i.e. _ the case in which angular information is not useful in disentangling the signal from the diffuse background ) . for values of @xmath6 below the corresponding lines
, a pure background model is expected to be consistent with the data .
the fact that the limits are significantly stronger in the uniform background case is the manifestation of the improved sensitivity which can be achieved by an analysis including both energy and angular information . for comparison , in fig .
[ limitnfw ] it is also shown the region already excluded by egret @xcite ( above the dotted line ) and the mass and cross section of neutralino models found in a random scan over supersymmetric parameters , as calculated using darksusy @xcite .
as expected , many of the models cluster around @xmath91 @xmath92/s , the value required of a thermal relic annihilating via an @xmath19-wave amplitude .
each point shown represents a model which respects all direct collider constraint and generates a thermal dm abundance consistent with the observed dm density . in the scan
the susy parameters varied were @xmath94 , @xmath95 and @xmath96 up to 2 tev , @xmath97 and @xmath98 up to 1 tev and @xmath99 up to 60 .
also , the gaugino masses were assumed to evolve to a single unified scale , such that @xmath100 , @xmath101 .
should gamma rays be identified as having been produced in dm annihilations , such observations could then be used to measure the characteristics of the dm particle , including its mass , annihilation cross section and spatial distribution .
such determinations are an important step toward identifying the particle nature of dm .
a calculation similar to the one leading to the results of fig .
[ limitnfw ] can be performed , this time including in the template a contribution from a fiducial model of dm annihilation , and asking the accuracy by which a reconstruction of its input properties is possible . in fig .
[ ellipsenfw ] , the ability of fermi - lat to determine the wimp mass and annihilation cross section for wimps distributed with an nfw halo profile is illustrated , for the case of an annihilation cross section of @xmath102 and a mass of 100 gev . in each frame
the projected 2 and 3@xmath103 constraints on the input parameters are reported , assuming an isotropic diffuse background ( in addition to background point sources ) . in the left frame ,
we treat the shape of the halo profile ( nfw ) as if it is known in advance .
of course , this is not a realistic assumption , and a less accurate determination of the wimp mass must be expected in a more realistic treatment . in the right panel of fig .
[ ellipsenfw ] we report the results obtained by marginalizing over the inner slope of the halo profile , @xmath62 . in the absence of an assumption on the inner halo slope , the constraint on the dm mass worsens by a factor @xmath1042 .
if the spectrum and angular distribution of gamma rays from dm annihilations in the galactic center region are sufficiently well measured , it will also be possible to measure the underlying dm distribution .
[ ellipseslope ] shows the results from the right frame of fig .
[ ellipsenfw ] in the @xmath105 plane , marginalized over the annihilation cross section . in the above benchmark model
, the inner slope of the halo profile can be determined at approximately the @xmath106 level .
a few caveats are in order : we are possibly oversimplifying the spectral shape of the background , since we are extrapolating the known point - like source properties from lower and higher energies .
also , the bounds are slightly optimistic , in the sense that we are considering the longest plausible lifetime of fermi ( ten years ) and that further effects degrading the angular resolution or accounting for dead - time may loosen the constraints by up to a factor@xmath107(2 ) .
finally , we have included only the hess and egret sources ( in addition to the diffuse background ) in our analysis , but it is likely that other astrophysical point sources with different spectra will be discovered by fermi - lat .
the above estimates should be quite realistic as long as a limited number of discrete sources will `` contaminate '' the inner galactic center angular bins , so that they could be removed effectively by the fermi team . on the other hand ,
it is worth taking a lesson from the `` sudden '' discovery of an astrophysical source just at the galactic center : in the worst case , the galactic center region might so crowded with ( yet unknown ) sources that the separation of background and signal might be degraded with respect to expectations based on present knowledge . still , it is worth mentioning that the bulk of the statistical significance of the dark matter annihilation signal typically does not come from the inner @xmath108 around the galactic center ( where the iact source dominates ) , rather from the surrounding angular region .
it is then interesting to see what are the perspectives to detect a dm signal from a more extended region , which we address in the following .
the emission of radiation per unit solid angle from dm annihilation is expected to be maximized at the galactic center . yet
, geometric factors and the presence of point - like and diffuse backgrounds make the choice of the optimal window size a non - trivial problem @xcite .
unless the dm halo is very cuspy towards the galactic center ( say , cuspier than nfw ) , the optimal strategy for dm searches is never to focus on the inner sub - degree around the galactic center .
rather , a window size up to @xmath109 or more is preferred .
the optimal shape of the window depends on the angular distribution of the signal and backgrounds , but also on the details of the analysis ( like energy cuts , astrophysical foreground removal , etc . ) .
it could be optimized from morphological studies of the low - energy emission measured by fermi , but a circular annulus around the galactic center or a `` rectangular '' window in galactic coordinates with an inner rectangular mask generally work fairly well . here
we illustrate this issue within a simplified model : based on egret data , in the diffuse background one can identify an isotropic , extragalactic component @xmath110 and a galactic component scaling as @xmath111 , which is however dominant at gev energies and peaks towards the galactic plane ( see @xcite for details ) .
the energy spectrum of the dm signal thus mostly enters the game in determining which one of the two spectra ( of different angular shape ) dominates the background : the harder the dm spectrum , the closer the background is to an isotropic emission , the smaller the optimal angular window .
still , even in the quite extreme case of @xmath112gev monochromatic lines , the optimal angular window is very extended .
this is illustrated in fig .
[ fig2 ] , where we plot the relative signal - to - noise ( @xmath113 ) as a function of the maximum galactic latitude , @xmath114 , for a galactic region @xmath115 , @xmath116 .
the solid lines refer to an hypothetical gamma - line emission for the benchmark `` model i '' in the inert higgs doublet scenario ( see ref .
@xcite for details ) for three different halo profiles .
the dot - dashed line refers to the continuum spectrum of a 100 gev neutralino annihilating into @xmath117 with an nfw distribution , where indeed we see that the optimal window is even more extended than for the line signal : this is due to the fact that the background at the typical energies of the continuum dm photons , @xmath118 , is mostly dominated by the galactic unresolved background , which peaks towards the galactic plane / center . to give a feeling of the quantitative difference with respect to the gc case treated in the previous section ,
let us note that for our usual benchmark model ( @xmath119gev , @xmath120pb with @xmath80 final state ) , a `` naive '' count statistics above 1 gev ( including all backgrounds ) would lead to a @xmath121 for the @xmath81 window considered in sec .
4 , while for the region considered above and @xmath122 one would find almost 40000 events from dm vs. about seven millions background events , resulting in a @xmath123 . of course , this analysis does not take into account angular and energy cuts that ( as shown for the gc case above ) do improve the diagnostic power , as well as other possible systematics .
yet , the importance of this diffuse signal can not be underestimated ( and it is actually confirmed by other analyses , see sec .
4.2 in ref .
[ 6 ] ) , especially in the case no signal is revealed from the gc , which may be due to a inner halo profile more cored than nfw .
the extended signal has indeed a milder dependence on the profile . also , from fig .
[ fig2 ] it is clear that at least for the simple @xmath113 estimator
it is only important to adopt the right angular cut within a factor of two or so in order not to degrade the sensitivity by more than@xmath124 .
but `` blindly '' focusing on a too narrow window ( degree scale ) around the gc might be overly penalizing ; especially for relatively cored profiles , the loss in sensitivity may reach a factor of two or more . besides improving the prospects for detection , other advantages in focusing on an extended region around the galactic center include : * enabling an empirical determination of the dark matter profile slope outside of the region dominated by the gravitational potential of the central supermassive black hole . * obtaining independent evidence that an `` excess '' signal with respect to backgrounds is the product of dm annihilation , by comparing the emission from many angular regions .
in contrast , if the spectrum is found to vary with location , it is most likely the product of astrophysical backgrounds . * constraining the quantity of dark matter substructure in the halo , by observing the angular distribution of the emission .
the shape of the smooth , unresolved inner profile @xcite could be studied , in addition to an anisotropy / multipole analysis @xcite .
this , in turn , may have important implications for dm cosmology ; see the review by t. bringmann in this issue .
the most challenging task for indirect dark matter searches is not to detect a few events , but to confidently identify those events as the products of dark matter annihilations . in particular
, any signal must be separated from astrophysical backgrounds if it is to be reliably claimed to be a detection of dark matter .
this is certainly true in the case of gamma - ray telescopes hoping to observe dark matter annihilations in the region of the galactic center ( gc ) .
the discovery of a bright tev source of astrophysical origin at the gc by iacts has changed the prospects for such searches considerably .
one strategy in light of this is to focus on different targets with lesser astrophysical background ( such as dwarf galaxies ) .
another is to search for a signal in the `` noisy '' gc region , by taking advantage of the peculiar spectral and angular properties expected from dark matter annihilation products . in this paper , we have reviewed the latter strategy , reporting on recent studies of the gc and inner galaxy .
given the characteristics of this search , the lat instrument on board of the fermi satellite is better suited than existing iacts .
fermi - lat instrument will detect a number of astrophysical sources in the region of the sky around the gc , including the point sources previously identified by hess and egret , and perhaps others .
a diffuse gamma - ray background will also likely be present .
although predictions of fermi s sensitivity are unavoidably limited by our incomplete knowledge of these backgrounds , we have shown that the spectral and angular differences between the signal and backgrounds should be distinctive enough to allow one to separate signal from background over a significant region of the parameter space , at least for a sufficiently cusped dark matter profile ( nfw - like or steeper ) . in the optimistic case where dark matter annihilation products are identified by fermi ,
then it may also be possible to measure or constrain the properties of the dark matter , including its mass , annihilation cross section , and spatial distribution .
it is unlikely that fermi will determine the wimp s mass with high precision , however .
for example , for the case of a 100 gev wimp with an annihilation cross section of @xmath125 @xmath92/s , distributed with an nfw halo profile , the mass could be determined to lie within approximately 50 - 300 gev . in the same benchmark model , the inner slope of the dark matter halo profile
could be determined to @xmath106 precision .
the combination of several indirect detection channels will be crucial to both confirm such a detection , and to best constrain the wimp s properties .
on the other hand , it is not excluded that fermi will lead to a radical revision of the present gamma - ray picture of the gc , revealing a more complicated zoo of astrophysical accelerators than envisaged in the present estimates . in the case where either the dm signal from the gc is too low or the the background is too large / complex ,
a dm discovery in gamma rays is still possible by looking at the emission from an extended region in the inner halo with fermi , or from other dark matter substructures with both fermi - lat and iacts
. in particular , the morphology and the spectral properties of the unresolved galactic background at @xmath126gev will be useful to optimize the angular and energy - cut templates for searches of the dm emission from an annulus of several tens of degrees around the gc .
ps would like to thank the galileo galilei institute for theoretical physics for the hospitality and the infn for partial support during the completion of this work .
dh is supported in part by the fermi research alliance , llc under contract no .
de - ac02 - 07ch11359 with the us department of energy and by nasa grant nnx08ah34 g .
finally , we thank sergio palomares ruiz for noticing a typo in the v2 of the manuscript .
in the case of decaying dark matter , eq . ( [ flux1 ] ) is modified to @xmath127 where @xmath128 is the decay width ( inverse lifetime ) and the spectrum now refers to the photons generated in the decay process . unlike with the cross section in the case of annihilating dark matter , one does not have any strong theoretical motivation for considering any particular lifetime for an unstable dm particle . in any case
, arguments have been put forward justifying the typical range of the lifetimes needed for significant signatures in astrophysics with @xmath4tev mass particles and gut - scale physics mediating the process ( in analogy with the expected proton decay in guts ) , see e.g. @xcite . from the phenomenological point of view
, there are a couple of points worth mentioning regarding decaying dm candidates : * the dm distribution and the role of substructures in particular is of little importance in determining the level of the signal . *
the angular distribution of the gamma - ray signal is very distinctive , and much flatter than the corresponding annihilation signal , as illustrated for a nfw profile in fig .
( [ fig : angular ] ) . should gamma rays be detected from dm , a comparison between the emission in the inner galaxy and the emission at high latitude would immediately reveal the nature of the particle physics process ( annihilation or decay ) responsible for the emission @xcite .
notice that this information is very difficult to extract with other cosmic ray probes .
although the factorization between the particle physics term and astrophysical term in eq . ( [ flux1 ] ) is a useful approximation and valid in most practical cases , there are exceptions .
more correctly , one should write @xmath129 \rho^2(\ell,\omega)\ , \d \ell , \label{flux3}\ ] ] where @xmath130 is the relative velocity between the two particles ( with @xmath131 ) and @xmath132 its distribution function ( not necessarily isotropic ) , whose integral over @xmath133 is normalized to unity .
the factorization assumed in eq .
( [ flux1 ] ) holds only if the integral in square brackets which is nothing but @xmath6is independent of position . a sufficient condition for this
is that @xmath134 is velocity - independent . in general , the integral depends on the kinematical structure of the halo via the position - dependent velocity dispersion , anisotropy , etc .
then , both the astrophysical distribution of the dm and the particle physics contribute in determining the angular shape of the signal .
one case in which the factorization is not valid can be found when the wimp annihilations mainly through a @xmath135-wave process , such that @xmath136 @xcite .
this is a largely academic case , however , since whenever @xmath135-wave annihilation is dominant the gamma - ray signal is expected to be suppressed .
more interesting is the case in which non - perturbative processes lead to large `` sommerfeld enhancements '' to the annihilation cross section at low velocities @xcite .
this effect can be thought of as the distortion of the wave - function due to a relatively long - range attraction between the wimps . in this case , a further steepening of the signal towards the galactic center is expected , which in turn should ease the detection of gamma rays from the inner galaxy @xcite .
00 j. buckley _ et al .
_ , arxiv:0812.0795 [ astro - ph ] . l. bergstrom , p. ullio and j. h. buckley , astropart . phys .
* 9 * , 137 ( 1998 ) [ arxiv : astro - ph/9712318 ] .
g. bertone , d. hooper and j. silk , phys .
rept . * 405 * , 279 ( 2005 ) [ arxiv : hep - ph/0404175 ] .
t. moroi and l. randall , nucl .
b * 570 * , 455 ( 2000 ) [ arxiv : hep - ph/9906527 ] .
l. bergstrom and p. ullio , nucl .
b * 504 * , 27 ( 1997 ) [ hep - ph/9706232 ] ;
p. ullio and l. bergstrom , phys . rev .
d * 57 * , 1962 ( 1998 ) [ hep - ph/9707333 ] .
m. gustafsson , e. lundstrom , l. bergstrom and j. edsjo , phys .
lett . * 99 * , 041301 ( 2007 ) [ arxiv : astro - ph/0703512 ] .
j. j. binney and n. w. evans , mon . not .
* 327 * , l27 ( 2001 ) [ arxiv : astro - ph/0108505 ] .
a. klypin , h. zhao and r. s. somerville , astrophys . j. * 573 * , 597 ( 2002 ) [ arxiv : astro - ph/0110390 ] .
d. hooper and p. d. serpico , jcap * 0706 * , 013 ( 2007 ) [ astro - ph/0702328 ] .
s. gillessen , f. eisenhauer , s. trippe , t. alexander , r. genzel , f. martins and t. ott , arxiv:0810.4674 [ astro - ph ] .
a. v. kravtsov , a. a. klypin , j. s. bullock and j. r. primack , astrophys . j. * 502 * , 48 ( 1998 ) [ astro - ph/9708176 ] .
c. power _ et al .
_ , mon . not .
soc . * 338 * , 14 ( 2003 ) [ astro - ph/0201544 ] .
j. f. navarro _ et al .
_ , mon . not .
astron . soc . * 349 * , 1039 ( 2004 ) [ astro - ph/0311231 ] .
d. reed _ et al .
_ , mon . not .
* 357 * , 82 ( 2005 ) [ astro - ph/0312544 ] .
d. merritt , j. f. navarro , a. ludlow and a. jenkins , astrophys .
j. * 624 * , l85 ( 2005 ) [ astro - ph/0502515 ] .
j. f. navarro _ et al .
_ , arxiv:0810.1522 [ astro - ph ] .
j. einasto , trudy inst .
alma - ata , 51 , 87 ( 1965 ) o. y. gnedin , a. v. kravtsov , a. a. klypin and d. nagai , astrophys .
j. * 616 * , 16 ( 2004 ) [ astro - ph/0406247 ] ; m. gustafsson , m. fairbairn and j. sommer - larsen , phys .
d * 74 * , 123522 ( 2006 ) [ arxiv : astro - ph/0608634 ] .
j. i. read , l. mayer , a. m. brooks , f. governato and g. lake , arxiv:0902.0009 [ astro-ph.ga ] .
f. prada , a. klypin , j. flix , m. martinez and e. simonneau , phys .
lett . * 93 * , 241301 ( 2004 ) [ arxiv : astro - ph/0401512 ] . g. bertone and d. merritt , mod .
a * 20 * , 1021 ( 2005 ) [ astro - ph/0504422 ] ; g. bertone and d. merritt , phys . rev .
d * 72 * , 103502 ( 2005 ) [ astro - ph/0501555 ] .
j. a. sellwood , arxiv:0807.1973 [ astro - ph ] .
p. gondolo and j. silk , phys
. rev .
* 83 * , 1719 ( 1999 ) [ astro - ph/9906391 ] ; p. ullio , h. zhao and m. kamionkowski , phys
. rev .
d * 64 * , 043504 ( 2001 ) [ astro - ph/0101481 ] ; g. bertone , g. sigl and j. silk , mon . not .
soc . * 337 * , 98 ( 2002 ) [ astro - ph/0203488 ] .
h. yuksel , s. horiuchi , j. f. beacom and s. ando , phys .
d * 76 * , 123506 ( 2007 ) [ arxiv:0707.0196 [ astro - ph ] ] .
l. bergstrom , j. edsjo and p. ullio , phys . rev
. lett .
* 87 * , 251301 ( 2001 ) [ astro - ph/0105048 ] ; v. berezinsky , a. bottino and g. mignola , phys .
b * 325 * , 136 ( 1994 ) [ hep - ph/9402215 ] ; a. cesarini , f. fucito , a. lionetto , a. morselli and p. ullio , astropart .
phys . * 21 * , 267 ( 2004 ) [ astro - ph/0305075 ] ; p. ullio , l. bergstrom , j. edsjo and c. g. lacey , phys .
d * 66 * , 123502 ( 2002 ) [ astro - ph/0207125 ] .
k. kosack _ et al . _
[ the veritas collaboration ] , astrophys .
j. * 608 * , l97 ( 2004 ) [ astro - ph/0403422 ] .
f. aharonian _ et al .
_ [ the hess collaboration ] , astron .
astrophys .
* 425 * , l13 ( 2004 ) [ astro - ph/0408145 ] .
j. albert _ et al .
_ [ magic collaboration ] , astrophys . j. * 638 * , l101 ( 2006 ) [ astro - ph/0512469 ] . c. van eldik , o. bolz , i. braun , g. hermann , j. hinton and w. hofmann [ for the hess collaboration ] , arxiv:0709.3729 [ astro - ph ] .
d. hooper and j. march - russell , phys .
b * 608 * , 17 ( 2005 ) [ hep - ph/0412048 ] ; d. hooper , i. de la calle perez , j. silk , f. ferrer and s. sarkar , jcap * 0409 * , 002 ( 2004 ) [ astro - ph/0404205 ] ; s. profumo , phys .
d * 72 * , 103521 ( 2005 ) [ astro - ph/0508628 ] ; l. bergstrom , t. bringmann , m. eriksson and m. gustafsson , phys .
lett . * 94 * , 131301 ( 2005 ) [ astro - ph/0410359 ] .
f. aharonian and a. neronov , astrophys .
j. * 619 * , 306 ( 2005 ) [ astro - ph/0408303 ] ; astro - ph/0503354 ; aip conf .
proc . * 745 * , 409 ( 2005 ) ; a. atoyan and c. d. dermer , astrophys .
j. * 617 * , l123 ( 2004 ) [ astro - ph/0410243 ] .
g. zaharijas and d. hooper , phys .
d * 73 * , 103501 ( 2006 ) [ arxiv : astro - ph/0603540 ] . c. van eldik , arxiv:0811.0931 [ astro - ph ] .
d. hooper and b. l. dingus , phys .
d * 70 * , 113007 ( 2004 ) [ astro - ph/0210617 ] ; d. hooper and b. dingus , proc . of the 34th cospar scientific assembly ,
houston , texas ( 2002 ) , astro - ph/0212509 .
m. pohl , astrophys .
j. * 626 * , 174 ( 2005 ) .
p. gondolo , j. edsjo , p. ullio , l. bergstrom , m. schelke and e. a. baltz , jcap * 0407 * , 008 ( 2004 ) [ astro - ph/0406204 ] .
f. stoehr , s. d. m. white , v. springel , g. tormen and n. yoshida , mon . not .
soc . * 345 * , 1313 ( 2003 ) [ astro - ph/0307026 ] .
n. w. evans , f. ferrer and s. sarkar , phys .
d * 69 * , 123501 ( 2004 ) [ astro - ph/0311145 ] ; s. k. lee , s. ando and m. kamionkowski , arxiv:0810.1284 [ astro - ph ] . m. fornasa , l. pieri , g. bertone and e. branchini , arxiv:0901.2921 [ astro - ph ] .
a. arvanitaki , s. dimopoulos , s. dubovsky , p. w. graham , r. harnik and s. rajendran , arxiv:0812.2075 [ hep - ph ] .
g. bertone , w. buchmuller , l. covi and a. ibarra , jcap * 0711 * , 003 ( 2007 ) [ arxiv:0709.2299 ] .
m. a. amin and t. wizansky , phys .
d * 77 * , 123510 ( 2008 ) [ arxiv:0710.5517 [ astro - ph ] ] .
j. hisano , s. matsumoto , m. m. nojiri and o. saito , phys .
d * 71 * , 063528 ( 2005 ) [ arxiv : hep - ph/0412403 ] . | in this article , we review the prospects for the fermi satellite ( formerly known as glast ) to detect gamma rays from dark matter annihilations in the central region of the milky way , in particular on the light of the recent astrophysical observations and discoveries of imaging atmospheric cherenkov telescopes .
while the existence of significant backgrounds in this part of the sky limits fermi s discovery potential to some degree , this can be mitigated by exploiting the peculiar energy spectrum and angular distribution of the dark matter annihilation signal relative to those of astrophysical backgrounds .
+ + _ keywords _ : dark matter , gamma rays , galactic center cern - ph - th/2009 - 025 , fermilab - pub-09 - 045-a |
many users have asked for background data sets with which to analyse extended objects , particularly where determination of background is difficult from their own data sets .
we have developed some event lists for each epic camera that could be used _
ad interim_. these files have an extension containing a calibrated event list in the same format as produced by sas .
compared with previous efforts , the new sets ( ca .
december 2001 ) are significantly longer to allow improved signal : noise ( typically the exposure duration is about an order of magnitude longer than the normal go data set ) ; the data are homogeneously collected from high galactic fields with no bright sources ; the data set uses homogenously the thin filter only and exposure , stdgti , badpix and events extensions of the files have all been concatenated and made coherent , so that in principle the observer could use them as a coherent data set for extractions via various @xmath2 or @xmath3 procedures , without making the manual corrections and calculations that were required for the previous incarnation .
many diverse aspects of the xmm - newton radiation environment were discussed extensively at a workshop held in december 2000 .
the reader is invited to look at the links under /docs
/ documents/ pages of the vilspa web site , particularly the document : cal-min-0002-0-0.html .
this has pointers to viewgraphs and minutes from the workshop .
the present note concerns _ quiescent _ background ( i.e. after removal of proton flares via .
gti filtering ) .
we have had to resort to compiling data from blank sky fields .
to do so with realistic s : n ratios for each of 4 instrument modes would impose an unacceptable penalty on guest observer science programme time , so we concentrated on the full frame imaging modes which are generally used for the faint extended objects . to minimise statistical uncertainties ,
the effective exposure duration in our data sets should be an order of magnitude longer than that of the typical guest observer exposure .
again it was an unrealistic proposition to observe with the 3 epic optical blocking filters for @xmath4300ks each , specifically to obtain this required data , therefore we decided to make use of a variety of guaranteed and calibration time observations of `` blank '' fields to compile our data serendipitously .
suitable fields were almost all taken with the filter in thin position , although this imposes additional complications for those users who might need to use a thicker optical blocking filter .
nevertheless as the soft diffuse x - ray component is more spatially and spectrally variable than the harder x - ray emission components this was not the major driver for our analysis
. table 1 summarises the location of the fields selected .
l l l c c l l + ra & dec & date of & duration & n@xmath5 & l ii & b ii + [ + 5pt ] & & observation&(ks ) & ( 10@xmath6@xmath0 ) & & + [ + 5pt ] + 02:18:00&-05:00:00&2000 - 07 - 31&60&2.5&169.7&-59.8 + 02:19:36&-05:00:00&2000 - 08 - 04&60&2.55&170.35&-59.5 + 02:25:20&-05:10:00&2001 - 07 - 03&25&2.7&172.3&-58.6 + 02:28:00&-05:10:00&2001 - 07 - 06&25&2.7&173.5&-58,2 + 10:52:44&+57:28:59&2000 - 04 - 29&70&0.56&149.3&53.1 + 12:36:57&+62:13:30&2001 - 06 - 01&90&1.5&125.9&54.8 + 13:34:37&+37:54:44&2001 - 06 - 23&80&0.83&85.6&75.9 + 22:15:31&-17:44:05&2000 - 11 - 18&55&2.3&39.3&-52.9 + the data sets were processed using the pipeline processing of xmm - newton science analysis subsystem 5.2 , in order to generate calibrated event lists for each epic camera .
proton flare gti filtering was made on time bins of @xmath4100s at count rates @xmath7 45 ( 20 ) events / bin in pn ( mos ) .
next we removed the signature of bright field sources by forming images in the 0.5 - 2kev band and running sas task @xmath8 .
approximately 10 objects per field were identified and an exclusion of 25 arcsec radius applied around the detected centroids .
this exclusion can not be directly associated with a source flux level , but individual analysis of the fields confirmed the estimates based on logn - logs curves ( @xcite ) , that sources of flux brighter than about 1 - 2 @xmath9 10@xmath10 ergs @xmath0 s@xmath1 ( 0.5 - 2 kev ) have had about 80% of their flux removed with this recipe .
finally the screened data were co - added , and the exposure and gti extensions of the fits data files carefully added and concatenated together .
the passage of charged particles through the cameras is associated with generation of fluorescent x - ray emission .
this emission is most clearly seen in the form of emission line energies characteristic of the camera body materials ( aluminium and stainless steel components for example ) .
the construction of the two epic camera types is quite different , leading to substantially different manifestations of these features .
the outer 6 of 7 ccds detect more al k radiation due to their closer proximity to the aluminium camera housing .
si k emission however is concentrated along the edges of some ccds .
in contrast the pn camera has a relatively intense contribution from energies around the cu k line .
what is notable is that this emission is spatially variable ( @xcite )
. the consequences for ignoring the spatial variation of these background features could be dramatic .
xmm - newton is possibly the observatory of choice for spectrally - resolved imaging of large clusters , for example to map radial temperature and element distributions .
however the variable al and si background lines would compromise abundance determinations of cluster emission lines with moderate redshifts , while the variable high energy background would bias temperature measurements at large radii .
these difficulties should be alleviated if the proposed background templates prove to be representative .
a background component that is relatively constant in spectrum and with position , is the remnant events generated by charged particles which are not rejected by on - board or ground processing . compared with
a maximum predicted primary cosmic ray particle rate of 4 @xmath0 s@xmath1 , the measured rates in all the ccd cameras on since launch have lain in the range 2 - 2.5 @xmath0 s@xmath1 .
the internal spectrum in the mos1 camera , below 10kev , after selection for x - ray event characteristics ( the sas attributes _ ( # xmmea_em&&pattern in [ 0:12 ] ) _ ) is 0.026 events @xmath0s@xmath1 . of this 0.021 ( @xmath110.0022 ) @xmath0s@xmath1
is in the flat spectrum component , the remainder in emission line components and noise component increasing to lower energies .
the flat spectrum count rate implies a cosmic ray rejection efficiency of @xmath499% as expected . the equivalent spectrum in the pn cameras after applying the sas attributes _
# xmmea_ep & & pattern in [ 0:4 ] _ is 0.039 @xmath0s@xmath1 is in the flat spectrum component , and 0.034 @xmath0s@xmath1 in the fluorescent lines .
this could imply a reduced rejection efficiency than mos or a higher detection efficiency to local @xmath12 s as a consequence of the much thicker pn detection thickness .
the most obvious features in the lowest energy band are effects of `` bad pixels '' .
a pixel which is consistently bad is flagged for removal on - board by loading a table of positions to be blanked out .
some pixels `` flicker '' on and off ( @xcite)with low recurrence rate ( @xmath131% ) , and the efficiency for finding them post - facto in the sas pipeline is dependent on many factors , so that some such events may occur in the background template and not the observer s data set and vice versa .
in the pn camera there are occasional blocks of bright pixels , typically 4 pixels in height .
their presence varies from observation to observation .
they arise from an artefact of the ccd offset bias level calculation at the start of each observation .
in principle one can use data from outside the filter rim as a measure of the true internal background .
however as noted , this is not strictly true due to the spatial variation in fluorescent emission .
this was easiest to perform in mos , and therefore we modeled the internal background , with multiple gaussian functions to characterise these emission lines , imposed upon a broken power - law to describe the continuum due to unrejected particle backgrounds , and low energy noise . in fig
[ dlumb - wa2b_fig : cxbint ] we show the comparison of this internal background component with the total spectrum including the cxb .
the internal component should be scaled with its particle continuum according to the ratio of in- and out - of - field detector areas , but the emission line intensities must be allowed to vary ( in order to account for the aforementioned spatial variation ) . at energies @xmath135kev the cxb dominates , with a clear signature of low energy emission lines at e@xmath131 kev not present in the internal background , but most likely from a thermal cosmic plasma spectrum . in the case of the mos data files , where we are more secure of a subtraction of the internal background component
, we attempted a simple spectral fit to the diffuse background spectrum .
we adopted an empirical two temperature mekal ( mewe et al @xcite ) plus power - law component to describe the cosmic diffuse x - ray background .
the former represent the contribution from the supposed galactic halo , and the latter the extragalactic unresolved agn population .
temperatures of @xmath40.08kev and @xmath40.21kev together with a power law 1.4 could be recovered .
we made an estimate of the sub - kev emission in some of the different fields noting differences in the measured temperature and flux .
this variability in emission and lack of detailed temperature measurement capability hampers a more accurate determination of galactic background properties .
for example the measured mean deviation from field to field of 2 - 10 kev flux is about 3.5% , consistent with a uniform diffuse extragalactic background . on the other hand the mean deviation of 0.2 - 1
kev flux is about 35% from field to field .
the recovered spectral fitting parameters are consistent with a number of investigations of the diffuse x - ray background , which leads us to believe that the user should be able to use the sets with suitable caveats to analyse more general conditions with different filters and galactic latitude , as described below .
the calculated normalisation of extragalactic background was 11.4 kev @xmath0 s@xmath1 kev@xmath1 sr@xmath1 , which is comparable with determination by sax , but places the fraction of chandra - resolved harder x - ray background towards the lower limits of claimed fractions ( 70% rather 95% ) .
the epic cameras allow a choice of optical blocking filter to prevent contamination by optically bright targets . for
most extended and/or faint extragalactic targets , such contamination is negligible , and the thinnest filter can be employed , as indeed was the case for all the observations compiled in this work .
however should the observer have chosen a thicker filter , the transmission of diffuse background x - rays ( and no doubt any remnant proton flux ) will be reduced .
therefore the observer would have to model the differences in soft component based on the response matrices and limited knowledge of the expected galactic emission .
the observer must take care to extract background from an appropriate location in the field of view .
most simply the extraction region should be defined in detector co - ordinates to match closely the region of the desired science target . in some applications
the user would subtract data in sky co - ordinates . in this case
the template event lists could be recast to mimic the nominal pointing direction of the observer s field ( for example using sas task @xmath14 invoking the attributes @xmath15 @xmath16 @xmath17 ) we noted that despite the standard recipe for filtering proton flares , the background rate as measured in 1 - 10kev bands exhibited some remnant flares .
this implies that at lower energies , the proton flares turn on more slowly , yet before the main flare component .
we have not chosen to further force stringent data cuts in the background template files , so that the general observer has more leeway to make an additional selection of the template files to match his / her own data selection .
we emphasise that in analysing the spectrum of diffuse x - rays in the field , the recovered spectral slopes steepen with more thorough flare mitigation .
careful comparison of the recipes used for gti creation must be made .
nevertheless it is possible that the lowest level proton fluxes are spectrally variable , and no complete subtraction can be made .
although we noted that point sources have been removed or significantly diluted , careful examination of the images derived from these event lists reveals intensity fluctuations . over scales of arcminutes appropriate to extended sources , it is not expected to be a significant problem , and indeed _ representative _ of unresolved background .
however if the user should try to extract spectra from regions comparable with the mirror point spread function scale , then manual inspection might be necessary to guard against a local deficit or excess of counts arising from treatment of point sources in the template files .
as noted previously there are particular defects to be expected in the lowest energy spectral ranges .
furthermore , at the time of writing , the calibration of the epic soft x - ray spectral response awaits completion .
the transmission of filters at energies @xmath13250ev is difficult to measure , the ccds calibration at the ground synchrotron facility was not performed at energies @xmath13150ev and the detailed redistribution of signal from photons of energies @xmath41kev into partially collected events in the softest band was also not determined completely in ground measurements . for the time being the extension of spectral analysis to _
any _ data below 250ev should be treated with caution .
for the highest fidelity determination of background appropriate to the observer s own data the compensation for galactic soft x - ray component , changing cosmic ray rates and different filters must be made .
it is intended to provide tools within the sas environment to achieve this , but most steps should be achievable manually .
* following suitable flare screening , define a background region ( b ) and extract the observed spectrum ( c@xmath18 ) from the observer s data set . from an identical region in the template
file the observed spectrum ( t@xmath18 ) should provide a measure of variability of cr component by checking count rates for e@xmath75kev and/or the fluorescent emission line normalisations . *
to estimate a better internal background spectrum for the observer s data set ( c@xmath19 ) , determine a predicted cosmic background spectrum for the observer s region based on rosat ass maps , hydrogen column etc .. an experimental tool is available at heasarc web site .
create response matrices for the background region ( here is where you can introduce the effect of different filters ) .
fold this cosmic spectrum through the response matrices to obtain a _ predicted _ cosmic component for the background region , ( c@xmath20 ) .
c@xmath19 = c@xmath18 - c@xmath20 * a similar approach with the template files showed that with a weighted average n@xmath5 of 1.7 10@xmath6 leads to a r@xmath21 pspc count rate of 1.3s@xmath1 in 144arcmin@xmath22 and a 0.47 - 1.21 flux of 1.67 10@xmath23 ergs @xmath0 s@xmath1 for a 0.2kev r - s spectrum .
this could likewise be used to make an estimate of the internal background of the template file region in order to better estimate the scaling factor ( k ) for cr component .
t@xmath19 = t@xmath18 - t@xmath20 and k @xmath4c@xmath19/t@xmath19 * repeating the same exercise for the source region in both template and observed data sets could lead to a _ predicted _ background data spectrum comprising the scaled internal component , and the predicted galactic component with the appropriate filter responses .
freyberg , m. et al 2002 in _ new visions of the x - ray universe _ this volume hasinger , g. , altieri , b. , arnaud , m. et al .
, 2001 , a&a 365 , l45 hopkinson , g.r . , 1992 ,
ieee trans nucl sci 39 , 2018 mewe , r. , gronenschild , e.h.b.m . and vd oord , g.h.j . , 1985 , a&as 62 , 197 | we briefly describe the methods used in compiling a set of high galactic latitude background data . the characteristics and limitations of the data which affect their use as a template for analysing
extended objects are described .
we briefly describe a spectral fitting analysis of the data which reveals a normalisation for the extragalactic background of 11.4 kev @xmath0 s@xmath1 sr@xmath1 , which implies the fraction of hard x - ray background presently resolved by various chandra observations is towards the lower limit of their estimates . |
the lack of accepted microscopic theory of superconductivity in high - t@xmath0 cuprates did not prevent successful research in this field since 1986 .
we have in several respects the repetition of the situation ca 1938 , but with a clear advantage of already having bcs theory to provide insight and language for phenomenological treatment . on this level ,
high - t@xmath0 superconductors are successfully described by gorkov equations for normal and anomalous green s functions@xcite , which in matsubara representation are defined in the usual way : @xmath3 the only difference from conventional superconductivity is in the nontrivial symmetry of the pairing potential , @xmath4 as usual , we write equations of motion for each of these functions and use the `` anomalous mean field '' recipe to decouple the four - operator products , @xmath5 .
this brings out modified self - consistency relations for the order parameter , @xmath6 and gorkov equations , which in the spatially uniform , stationary system read : @xmath7 here @xmath8 is the fourier transform of the kinetic energy operator , @xmath9 , @xmath10 being the chemical potential .
it was established , that in high - t@xmath0 cuprates , like ybco , the order parameter is a spin singlet with @xmath11-wave orbital symmetry , @xmath12 ( with axes chosen along crystallographic directions ( 1,0,0 ) , ( 0,1,0 ) in the cuprate layer)@xcite . for the following ,
the most important consequence of this symmetry is the sign change of @xmath13 for certain directions .
this means , first , that there exists an _ intrinsic phase shift _ of @xmath14 between different directions in the crystal ; second , that in certain ( nodal ) directions the order parameter is zero , and therefore the quasiparticle excitation spectrum is not gapped . on the spatial scale exceeding the coherence length , @xmath15 ,
it is more convenient to use the eilenberger equations@xcite , which follow from ( [ eq_gorkov ] ) in the quasiclassical limit .
this is certainly justified in high - t@xmath0 cuprates with their small @xmath15 .
the eilenberger equations are conveniently written in matrix form , @xmath16=0.\end{aligned}\ ] ] here the matrix green s function and order parameter , @xmath17 depend both on position @xmath18 and on ( direction of ) fermi velocity @xmath19 .
( the layered structure of high - t@xmath0 cuprates allows us to reduce the problem to two dimensions to a good accuracy , unless we have to consider , for example , a twist junction@xcite , or tunneling in the @xmath20-direction . ) the components of @xmath21 , obtained from gorkov s functions by integration over energies , satisfy the normalization condition , @xmath22 and the self - consistency relation ( [ eq_self - consistency ] ) becomes @xmath23 here the angle averaging @xmath24 . in a little different language , the same results are obtained with the andreev approximation in the bogoliubov - de gennes equations for the components of the single - bogolon wave function , @xmath25 .
the original bogoliubov - de gennes equations are obtained in the process of diagonalization of pairing bcs hamiltonian @xcite : @xmath26 here the quasimomentum @xmath27 labels the bogolon state , @xmath28 is the excitation energy , and the self - consistency relation reads @xmath29}{2t}.\ ] ] unlike the gorkov equations ( [ eq_gorkov ] ) ( or the initial bogoliubov - de gennes equations ) , the equations ( [ eq_eilenberger],[eq_bdg ] ) are of the first order in gradients , which allows us to introduce quasiclassical trajectories ( characteristics ) along @xmath19 and solve the corresponding equations by integration along these trajectories , with proper boundary conditions .
it is known that a quasiparticle ( electron or hole , described , in terms of bogoliubov - de gennes equations , by a vector @xmath30 ( @xmath31 ) respectively ) impinging on the superconductor from the normal metal can undergo an _
andreev reflection _ , switching the branch of the excitation spectrum , acquiring an additional phase , and almost exactly reversing the direction of its group velocity ( this happens because for an electron and a hole with the same momentum @xmath27 group velocities are opposite ) : @xmath32 where @xmath33 the phase @xmath34 is the phase of the superconducting order parameter ; the ( @xmath35)-shift is exact in the limit when the quasiparticle energy is much less than the superconducting gap ( generally it is some energy - dependent function @xmath36 ) . now consider a slab of normal conductor sandwiched between two superconductors ( _ sns junction _ ) .
if we neglect the spatial dependence of the order parameter in superconductors , we do nt need to solve the self - consistency equations ( [ eq_self - consistency],[eq_self - consistency_bdg ] ) .
therefore the problem reduces to a single - particle one and is most naturally solved in bogoliubov - de gennes language ( eq.([eq_bdg ] ) becomes a schrdinger equation for a two - component wave function ) .
solutions of this equation with the boundary condition ( [ eq_andreev_bc ] ) are standing waves . obviously in the act of andreev reflection a charge of @xmath37 is transferred to the superconductor , therefore every standing wave ( andreev level ) carries supercurrent .
quasiclassically , in order to find andreev levels in a normal layer of thickness @xmath38 , sandwiched between `` left '' and `` right '' superconductors , with phase difference @xmath34 , we write the bohr - sommerfeld quantization condition , @xmath39 here the kinematic phase gain of the quasiparticle along the closed trajectory , @xmath40 , takes into account electron - hole ( or vice versa ) conversion .
the positions of levels , and therefore the supercurrent , depend on the phase difference @xmath34 between the superconducting banks , and we arrive at josephson effect in sns structures @xcite .
actually , the language of andreev levels can be successfully used to describe the josephson effect in general ( for a review see @xcite ) . the crucial experiments(@xcite ;
see also review@xcite ) which confirmed @xmath11-wave pairing symmetry in high - t@xmath0 cuprates were directed at catching the intrinsic phase @xmath14-shift .
the general idea of the experiment follows from the fluxoid quantization condition in a superconductor : if a superconducting contour @xmath41 is penetrated by the magnetic flux @xmath42 , then @xcite @xmath43 ( in cgs units ; @xmath44wb in si ) . in the case of a massive superconducting ring
the contour can be chosen well inside the superconductor , where there is no current and therefore no superconducting phase gradient .
then the magnetic flux is quantized in units of @xmath45 . if there is a josephson junction in the ring , the phase change will concentrate there , yielding @xmath46 here @xmath34 is the phase difference across the junction .
the equilibrium value of @xmath34 is determined by the interplay between josephson and magnetic energy of the system .
the josephson energy is related to the josephson current via @xmath47 and in the simplest case of a tunneling junction , @xmath48 , and @xmath49 for a conventional josephson junction @xmath50 is at a minimum when @xmath51 ; therefore in the absence of an external field the total energy of the ring , @xmath52 has a single global minimum at @xmath51 for any value of @xmath53 ( here @xmath38 is the self - inductance of the ring ) .
not so if the pairing symmetry is @xmath11-wave : if we choose the configuration of the loop in such a way , that the opposite sign lobes contact across the junction ( so called @xmath14-junction ) , the current - phase relation switches to @xmath54 , and @xmath55 therefore if @xmath56 the system has two degenerate minima ; if this ratio is so big that the magnetic energy can be neglected compared to the josephson energy , the equilibrium phase difference is @xmath57 , and we obtain from ( [ eq_fluxoid_quantization_2 ] ) , that @xmath58 thus in equilibrium there is a spontaneous flux @xmath59 in the ring .
its direction ( up or down ) allows us to distinguish the two ground states of the system , where the _ time - reversal symmetry is thus broken . _ such behaviour was indeed observed in an experiment @xcite : a tri - crystal ring of ybco generated a half - flux quantum , which was detected by squid microscopy .
a natural question is whether @xmath14-junctions are the only possibility provided by @xmath11-wave symmetry , and whether only @xmath59 fluxes can be spontaneously generated .
the answer is no : in principle , any equilibrium phase difference can be realized in a @xmath11-wave junction , and time - reversal symmetry breaking can be accompanied by generation of an arbitrary magnetic flux , or none at all .
let us return to an sns junction ( assuming a rectangular normal part , @xmath60 ) .
each quasiclassical trajectory connecting two superconductors acts as a conduit of supercurrent between them , an `` andreev tube '' of diameter @xmath61 , which carries supercurrent , determined by the phase difference between its ends .
if we neglect the normal scattering at ns interfaces , there is no mixing between different trajectories , since andreev reflections simply reverse the velocity , sending the reflected particle along the same path .
it can be shown , that if @xmath62 , every such trajectory passing through the point @xmath18 contributes a partial supercurrent density@xcite @xmath63 in eq.([eq_partial_current_density ] ) , @xmath64(@xmath19 ) is the length of the trajectory , and @xmath65 is the so called normal metal coherence length .
( we also included effects of weak impurity scattering with scattering time @xmath66 . )
the physical meaning of @xmath67 is that in a clean normal metal an electron and andreev - reflected hole ( or vice versa ) with energy @xmath68 maintain phase coherence across distance @xmath67 simply because they travel along the same trajectory .
indeed , the momentum of an electron ( hole ) with energy @xmath69 is @xmath70 at a distance @xmath71 from the point of andreev reflection ( measured along the trajectory ) they would gain phase difference @xmath72 if the phase difference is of order @xmath14 , the coherence is effectively lost , so we get @xmath73 .
the factor of @xmath74 appears in accurate treatment , like in eq.([eq_partial_current_density ] ) . as an aside , in case of very strong scattering in the normal part of the system , when the motion of electrons / holes is diffusive with diffusion coefficient @xmath75 , we can still use the same argument .
now the observable length scale is given by the displacement of quasiparticle , @xmath76 while the phase difference between the electron and the hole is gained along the crooked path of length @xmath77 they take .
so the condition @xmath78 yields @xmath79 such coherence in a normal metal is a purely _ kinematic _ effect , since there is no interaction in the normal ( non - magnetic ) metal , which would either support or suppress superconductivity .
( `` normal metal is neutral with respect to superconducting correlations.''-c.w.j . beenakker . )
nevertheless its effects are quite real , e.g. the very possibility of coherent supercurrent flow ( josephson effect ) in sns junctions .
let us apply the approach of ( [ eq_partial_current_density ] ) to calculation of josephson current in a clean sns contact .
in every point of ns boundary ( e.g. @xmath80 ) , we integrate ( [ eq_partial_current_density ] ) over directions of @xmath19 ( such that @xmath81 ) .
as it should be expected , the result reproduces ishii s sawtooth @xcite : in the limit of zero temperature and no scattering inside the normal layer ( @xmath82 ) , @xmath83 ( periodically extended ) . and
@xmath84 indicate andreev zero- and @xmath14-levels . in equilibrium their contributions to the josephson current ( normal to the ns boundary )
cancel . by switching current directions ( dashed arrows )
we come to the other ground state , with the opposite direction of the spontaneous current , @xmath85 . ]
consider now an snd junction , that is , an sns contact , where one of the superconductors has @xmath11-wave symmetry @xcite .
now , when integrating over the directions of @xmath19 we encounter two kinds of trajectories : `` zero''- and `` @xmath14''-trajectories , which link the conventional superconductor with the lobes of the @xmath11-wave order parameter of opposite sign ( intrinsic phase @xmath14-shift ) .
the josephson current is therefore a sum of two contributions like ( [ eq_ishii ] ) , one of which has the phase argument @xmath86 .
the relative weight of these contributions depends on the orientation of the sd boundary with respect to the crystal axes of the cuprate ( to which the order parameter is nailed ) . in the case of a @xmath87 orientation ,
when the boundary is along the ( 110 ) plane and is therefore normal to the nodal direction of the @xmath11-wave order parameter , the two groups contribute equally , yielding ( @xmath88 ) @xmath89 all odd harmonics cancel , and we obtain a @xmath14-periodic sawtooth : at @xmath90 , @xmath91 @xmath92 ( periodically extended ) .
( this is simply a sum of two identical , @xmath74-periodic sawtooth functions , shifted by @xmath14 , see fig.[fig_sum_sawtooth ] .
) there are two stable equilibrium phase differences across the junction : @xmath93 this means , that the time - reversal symmetry is broken .
( in addition , the frequency of ac josephson effect in the system will be doubled @xcite ) .
-levels to the josephson current in a 0 - 45@xmath94 dd(sd ) junction . b)resulting current - phase dependence .
c ) same in more general case .
d ) spontaneous current in a 0 - 45@xmath94 junction . ]
the snd junction may seem simply a `` superposition '' of two sns junctions , with phase differences shifted by @xmath14 , but the situation is more interesting . in equilibrium ,
spontaneous currents flow in the normal layer , parallel to the boundary@xcite .
this is clear from our `` andreev tube '' treatment .
take , for example , two trajectories , with @xmath95 and @xmath96 , carrying equal partial currents .
their contribution to the josephson current ( normal to the boundary ) is zero , since their currents projections on this direction cancel each other . on the contrary , the projections on the direction , parallel to the boundary , add .
the phase dependence of this current , @xmath97 , can be obtained in the same way as for the josephson current . in the limit of zero temperature and clean normal layer it reduces to a _ difference _ of two sawtooth functions , yielding @xmath98
again periodically extended .
states with @xmath99 and @xmath35 thus carry spontaneous currents in opposite directions . if an snd junction is closed on itself ( annular geometry )
, these currents translate into spontaneous magnetic fluxes , normal to the plane of the system @xcite .
note that all of the above is only possible if higher harmonics of current - phase dependence are not all zero , since all the odd harmonics ( including the standard josephson term , @xmath100 ) cancel .
( the expected `` total depairing '' in the ( 110 ) plane . ) in the case of arbitrary orientation of the nd plane , the weights of zero- and @xmath14-sawtooth functions will be different ( fig.[fig_sum_sawtooth]c ) .
then the current - phase dependence regains @xmath74-periodicity . at zero temperature and in the absence of scattering , the t - breaking is generally still present .
the equilibrium phase difference is no longer @xmath101 and depends on the geometry .
( the exception is when the boundary is along ( 100 ) or ( 010 ) .
then the contribution of only one group survives , and we have either standard , or @xmath14-junction .
of course , it is impossible to tell , whether a single junction is a @xmath14-junction - it is necessary to look at the contour , in which it is included@xcite . ) finite temperature and scattering suppress higher harmonics first , and therefore the transition to non - t - breaking state is possible @xcite .
essentially the same interplay of zero- and @xmath14-levels takes place in dnd junctions .
there we have a somewhat richer picture .
for example , the time - reversal symmetry can be broken without producing spontaneous currents ( fig.[fig_ddcurrents]c ) . )
c ) symmetric ( 22.5@xmath94 - 22.5@xmath94 ) junction .
directions corresponding to zero - levels are shaded . in the case
c ) the contributions of zero- and @xmath14-levels to the current in @xmath102-direction are identically zero ; therefore the spontaneous current is absent , but the time - reversal symmetry is still broken .
d ) orientation of the @xmath11-wave order parameter with respect to the crystalline axes . ] .
the mismatch angles are @xmath103 and @xmath104 ( 1 ) , @xmath105 ( 2 ) , @xmath106 ( 3 ) , @xmath107 ( 4 ) @xcite . ] in a clean grain boundary junction as a function of the mismatch angle : @xmath103 ; @xmath108 ( fig .
[ fig_ddcurrents ] ) .
parameter @xmath109 is the dimensionless temperature .
triangles and circles correspond to self - consistent and non - self - consistent solutions of eilenberger equations , respectively .
the 0 - 45@xmath94 junction is a @xmath14/2-junction at any temperature , since by symmetry all odd components of the josephson current are cancelled .
compare this to the behaviour observed in ybco junctions ( fig.[fig_ilichev01 ] ) . ]
one of the reliable methods of fabrication of josephson devices in high - t@xmath0 cuprates is based on forming grain boundary junctions ( see e.g. @xcite ) .
the order parameter is suppressed within @xmath110 around the boundary , and this region can be considered `` normal '' .
therefore the snd / dnd model applies , but only qualitatively , since the above equations are derived in the limit @xmath62 .
a more accurate approach , based on eilenberger equations ( [ eq_eilenberger ] ) , confirms the qualitative similarity between dnd and dd junctions@xcite . integrating the current - phase dependences of fig.[fig_sum_sawtooth ]
, we obtain the josephson energy of the junction , eq.([eq_josephson_relation ] ) , which has two minima , corresponding to its degenerate ground states .
this bistability plays the crucial role in qubit applications of high - t@xmath0 superconductors .
this prediction was confirmed in grain boundary ybco junctions@xcite ( fig.[fig_ilichev01 ] ) . and
( b ) symmetric ( 22.5@xmath94 - 22.5@xmath94 ) junction . curves ( top to bottom ) correspond to temperatures ( a ) 30 , 20 , 15 , 10 , and 4.2 k ; ( b ) 20 , 15 , 11 , 10 , 5 , and 1.6 k , respectively . ]
so far we did not take into account normal scattering on ns interfaces , and therefore missed an important point .
let us first consider a boundary of a @xmath11-wave superconductor with a vacuum or an insulator .
the order parameter near the surface is suppressed , and a qualitative understanding of the situation can be obtained by `` inserting '' a normal layer of thickness @xmath111 between the insulator and the bulk superconductor , similar to the dnd model@xcite .
energy levels in such a layer can be found from ( [ eq_bohr_sommerfeld ] ) for every quasiparticle trajectory ( assuming specular reflection ) , fig.[fig_ins]a : @xmath112 here @xmath113 or @xmath114 depending on whether the trajectory connects the lobes of the same or opposite sign .
note that @xmath115 and @xmath116 . therefore solution @xmath117 exists if and only if @xmath114 .
this is the zero - energy , or midgap , state ( zes , mgs ) , which obviously can not exist in conventional superconductors , where @xmath118 always equalis zero ( see review @xcite ) .
now consider a dd junction with finite transparency , @xmath119 ( instead of the case of ideal transparency , @xmath120 , which we dealt with earlier ) .
we can use the dnd model , inserting in the middle of the normal layer an infinitely thin barrier with transparency @xmath121 , dependent on the incidence angle of the quasiparticle trajectory@xcite .
bohr - sommerfeld quantization conditions for quasiparticle trajectories shown in fig.[fig_ins]b yield the energy levels of the bound states in the junction . in the limit of low transparency , @xmath122 ,
the critical current is much larger ( @xmath123 , not @xmath124 ) , if the orientations of the @xmath11-wave order parameter allow formation of zero energy states on both sides of the barrier .
even when transparency is not small , the presence of zes may be qualitatively important .
for example , the junction can become a @xmath14-junction at low enough temperature ( see review @xcite ) ; you can see such behaviour in fig.[fig_ilichev01]b , where below 11 k the potential wells are centered around @xmath14 rather than zero .
-wave superconductor ( a ) and model of a dd junction with finite transparency ( b ) ( after @xcite ) . ]
the appearance of spontaneous currents in sd and dd junctions can be considered as due to a time - reversal symmetry breaking order parameter with @xmath125 or @xmath126 symmetry , which is formed in the junction area due to the proximity effect ( here @xmath127 is the equilibrium phase difference across such junction ) . in certain conditions , such combinations could appear near the surface of a @xmath11-wave superconductor , leading to spontaneous currents and magnetic moments .
so far there is no conclusive evidence for such currents ( see @xcite and references therein ) .
qubits are the basic building blocks of future quantum computers . essentially they are two - state quantum systems which can be put in an arbitrary superposition of states ( `` initialized '' ) , coupled to each other , undergo desired quantum evolution and measured ( `` read out '' ) before losing quantum coherence . here
we concentrate on _ superconducting phase qubits_. the simplest example of such a qubit is an rf squid , that is , a loop with a single josephson junction ( like in section [ sec_1_pi_junction ] ) .
we saw , that in certain circumstances the system has two degenerate minima , corresponding to a flux of @xmath128 through the loop .
the hamiltonian of such a system is @xmath129 where @xmath130 is the coulomb energy of charge @xmath131 on the junction ( which has some finite capacitance @xmath132 ) .
the charge operator @xmath133 can be expressed in terms of the phase difference across the junction as @xmath134 ( see e.g. @xcite ) . due to the presence of the coulomb term
, phase is no longer a `` good '' quantum variable , and the eigenstates of the hamiltonian ( [ eq_hamiltonian_phase_qubit ] ) become linear combinations of `` up '' and `` down '' , or `` left '' and `` right '' , states ( with spontaneous flux @xmath128 ) . in other words , the qubit can now tunnel between the wells of the josephson potential , corresponding to certain phases ( fig.[fig_ilichev01 ] ) .
( this description is appropriate as long as the coulomb energy does not exceed the josephson energy , otherwise the natural starting point would be the states with definite _ charge _ on the junction ; we do not consider such systems ( _ charge _ qubits ) here , but they were successfully implemented experimentally@xcite . )
coherent quantum tunneling was indeed observed in an rf squid@xcite .
simultaneously , this effect was obtained in a different system , consisting of a small inductance loop with _ three _ josephson junctions ( the so called persistent current qubit , @xcite ) .
the advantage of the latter design is as follows . as we have seen in section [ sec_1_pi_junction ] , the degenerate states appear in the rf squid only if the self - inductance of the loop is large enough , and they carry comparatively large spontaneous fluxes , @xmath128
. therefore they will couple to the external degrees of freedom and reduce the time @xmath135 during which the system maintain its quantum coherence .
even more important is the fact that the resulting potential barrier is comparatively high , which may prohibit the tunneling we are after . in an experiment @xcite the coherent tunneling
was indeed observed not between the lowest , but between the excited states in the wells . in case of the three - junction loop of negligible self - inductance ,
the fluxoid quantization condition ( [ eq_fluxoid_quantization_2 ] ) leaves two independent josephson phase differences in the circuit : @xmath136 the josephson energy , @xmath137 of the system forms a 2d potential profile ( e.g. as a function of @xmath138 , @xmath139 ) , which depends on the external flux @xmath140 as a parameter .
if @xmath141 , and the @xmath142 s are comparable , this potential has two degenerate minima ; unlike the case of the rf squid , the potential barrier can be small , as are the spontaneous fluxes corresponding to the `` left '' and `` right '' states .
one of the main problems with the designs of the previous section is the necessity to artificially break the -symmetry of the system by putting a half flux quantum through it .
estimates show @xcite that the required relative accuracy is @xmath143 .
the micron - size qubits must be positioned close enough to each other to make possible their coupling ; the dispersion of qubit parameters means that applied fields must be locally calibrated ; this is a formidable task given such sources of field fluctuations as fields generated by persistent currents in qubits themselves , which depend on the state of the qubit ; field creep in the shielding ; captured fluxes and magnetic impurities . moreover ,
the circuitry which produces and tunes the bias fields is an additional source of decoherence in the system .
these problems are avoided if the qubit is _ intrinsically bistable_. the most straightforward way to achieve this is to substitute the external flux by a static _ phase shifter _ , a josephson junction with unconventional superconductors with nonzero equilibrium phase shift @xmath144 .
for example , three - junction ( persistent current , mooij ) qubits would require an extra @xmath14-junction ( @xmath145 ) @xcite .
the only difference compared to the case of an external magnetic field bias is in the prevalent decoherence sources : instead of noise from field - generating circuits we will have to take into account intrinsic decoherence from nodal quasiparticles and interface bound states ( see below ) .
qubit @xcite .
the ring d formed of high - t@xmath0 film contains one @xmath14-junction and therefore supports a spontaneous flux @xmath128@xcite .
the ring s of conventional superconductor screens this flux from the environment . ] another suggestion @xcite is based on the same tricrystal high - t@xmath0 ring geometry as in exp eriments@xcite
( fig[fig_newns ] ) .
the spontaneous flux @xmath146 generated by such a structure labels the qubit states @xmath147 , @xmath148 .
tunnelling between them , necessary for quantum operations , is made possible by applying a magnetic field in the plane of the system . indeed , then the states @xmath147 , @xmath148 are no longer the eigenstates of the hamiltonian .
the tricrystal ring d is surrounded by an s - wave superconductor ring s , aimed at screening the spontaneous flux from the environment ( including other qubits ) . indeed , due to the fluxoid quantization condition ( [ eq_fluxoid_quantization ] ) in the ring s , the total flux through it must be an _ integer _
multiple of @xmath45 , and therefore the states of the rings d and s become entangled ( e.g. @xmath149 ) .
this entanglement puts forward an interesting problem . in order to perform two - qubit operations , as well as initialization and readout ,
it is necessary to make the qubits `` visible '' to each other and the outside world . to do this
, it is suggested in @xcite to locally destroy superconductivity in the screening ring by using a superconducting field effect transistor ( sufet ) ( not shown ) ( i.e. by applying a gate voltage to the part of the screening ring ) .
will this transition collapse the wave function of the qubit , destroying quantum coherence between @xmath150 and @xmath151 ? and
@xmath118 connect the sds structure to a conventional josephson junction , @xmath152 , and a large capacitor , @xmath153 . ] now let us consider the case when the bistable @xmath11-wave system is employed @xmath154 , that is , when its phase is allowed to tunnel between the degenerate values . in a so called `` quiet '' qubit @xcite an sds junction ( effectively two sd junctions in the ( 110 ) direction ) put in a small - inductance squid loop in parallel with a conventional josephson junction and a large capacitor , fig.[fig_ioffe ] .
one of the sd junctions plays the role of a @xmath99-phase shifter .
the other junction s capacitance @xmath132 is small enough to make possible tunneling between the @xmath99 and @xmath35-states due to the coulomb energy @xmath155 .
two consecutive sd junctions are effectively a single junction with equilibrium phases @xmath156 and @xmath14 ( which are chosen as working states of the qubit , @xmath147 and @xmath148 ) . the proposed control mechanisms are based on switches @xmath157 .
switch @xmath20 connects the small sd junction to a large capacitor , thus suppressing the tunneling . connecting @xmath118 for the duration @xmath158 creates an energy difference @xmath159 between @xmath147 and @xmath148 , because in the latter case we have a frustrated squid with @xmath156- and @xmath14- junctions , which generates @xmath160 spontaneous flux
this is a generalization of applying a @xmath161 operation to the qubit .
finally , if switch @xmath20 is open , the phase of the small junction can tunnel between @xmath156 and @xmath14 .
entanglement between different qubits is realized by connecting them through another josephson junction in a bigger squid loop.(the switches in question are in no way trivial , since they must operate without destroying quantum coherence of the system .
one possible solution is to use a frustrated dc - squid@xcite , that is , insert in the wire a small inductance loop with two equivalent josephson junctions in parallel .
the total supercurrent through the switch , @xmath162 , goes to zero if the phase difference ( tunable by external magnetic flux ) @xmath163 .
modifications of this scheme are discussed in @xcite .
another possibility is to use superconducting single - electron transistors ( ssets , `` parity keys'')@xcite . ) due to the absence of currents through the loop during tunneling between @xmath147 and @xmath148 the authors called it `` quiet '' , though as we have seen small currents and fluxes are still generated near sd boundaries .
another design based on the same bistability @xcite only requires one sd or dd boundary ( fig.[fig_az_ddqubit ] ) . here
a small island contacts a massive superconductor , and the angle between the orientation of the @xmath11-wave order parameter and the direction of the boundary can be arbitrary ( as long as it is compatible with bistability @xcite , fig.[fig_bistability_region ] ) .
the advantage of such a design is , that the potential barrier can be to a certain extent controlled and suppressed ; moreover , in general the two `` working '' minima ( @xmath164 ; the phase of the bulk superconductor across the boundary is zero ) will be separated from each other by a smaller barrier , than from the equivalent states differing by @xmath165 .
this allows us to disregard the `` leakage '' of the qubit state from the working space spanned by ( @xmath147,@xmath148 ) , which can not be done in the `` quiet '' design with the exact @xmath14-periodicity of the potential profile .
@xmath166 are magnetic force microscope tips , intended to read out spontaneous fluxes .
an alternative readout scheme relies on direct detection of the phase on the island , which affects the critical current between the bus and the ground @xcite . ]
qubit operations in this system are realized by connecting qubits to each other and to the ground electrode ( normal or superconducting ) through ssets ( or other kind of switches ) .
when isolated , a qubit undergoes natural evolution between @xmath147 ( state with phase @xmath167 ) and @xmath148 ( phase @xmath168 ) , which realizes the @xmath169 operation . the @xmath161 operation ( that is , adding a controllable phase shift to one of the states with respect to the other ) can be realized by e.g. connecting the island through a sset to the massive superconductor ( `` bus '' ) , the phase of which @xmath170 .
the same operation repeated periodically can be used to block unwanted tunneling ( so called `` bang - bang '' technique ) @xcite : physically , if we keep shifting the levels in the right and left well with respect to each other , the tunneling becomes suppressed , since they are practically never in resonance .
a better design was suggested in @xcite ( `` silent qubit '' ) .
here two small bistable @xmath11-wave grain boundary junctions with a small superconducting island between them are set in a squid loop ( fig .
[ fig_silent ] ) . as usual ,
`` small '' means that the total capacitance of the system allows phase tunneling : the coulomb energy term in ( [ eq_hamiltonian_phase_qubit ] ) is not negligible .
-wave qubit@xcite .
the structure is formed of a high - t@xmath0 film around the grain boundary aa and contains two grain boundary junction ( j , j ) . ] in the limit of negligible self - inductance of the loop , the quantization condition ( [ eq_fluxoid_quantization ] ) fixes the sum of the phases to the external flux , @xmath171 .
this leaves only one independent phase combination , the superconducting phase of the island , @xmath172 .
keeping for simplicity only the first two harmonics in the current - phase relation of the junctions , @xmath173 , \label{eq_current_phase_silent}\end{aligned}\ ] ] we find for the josephson potential of the qubit the expression ( fig .
[ fig_silent2 ] ) @xmath174.\end{aligned}\ ] ] here @xmath175 , and @xmath176 .
parameter @xmath177 $ ] provides a convenient parametrization for the current - phase dependence in a dd junction . ;
@xmath178 @xmath179 . in the limit of negligible self - inductance
the potential depends on the external magnetic flux , @xmath180 as on parameter ( a ) . in the absence of external flux
the potential has two degenerate minima ; degeneracy is lifted by finite external flux , which also affects the height of the potential barrier(b ) . ] in the absence of the external flux ( @xmath181 ) the qubit potential @xmath182 has two degenerate minima .
moreover , if the junctions only differ in the amplitude of critical current , but have the same @xmath183 , there is no spontaneous current in the loop in either minimum , which means that the qubit is decoupled from the external magnetic fields ( if we disregard spontaneous currents in the junctions themselves , which can be very small @xcite ) .
this justifies the moniker `` silent '' .
both the barrier and the bias between the wells can be controlled by the external flux .
it is noteworthy that the corrections are of at least second order in @xmath184 , which drastically reduces the influence of fluctuations in the external circuitry@xcite .
the mechanism of noise reduction is similar to that of the `` quantronium '' qubit@xcite , but the `` sweet spot '' ( an extremal point on the energy surface ) appears already on the classical level , and at zero external field .
finally , let us briefly mention two more proposals . a `` no tunneling '' design@xcite combines the ideas of cbjj ( current - based josephson junction ) qubits @xcite and the intrinsic bistability of d - wave junctions . in a single bistable josephson junction , the potential barrier between the degenerate levels corresponding to phase difference @xmath185 is too high to allow tunneling .
the transitions between the states are realized through rabi transitions via an auxiliary energy level , situated above the top of the barrier .
rabi transitions are induced by applying an external high - frequency field .
`` dot / antidot '' proposals are based on the spontaneous flux ( less than a flux quantum ) generated in a high - t@xmath0 island or around a hole in bulk high - t@xmath0 due to the presence of a subdominant order parameter@xcite ( see the end of section [ sec_1_josephson effect ] ) , but so far lack experimental support . due to fabrication difficulties , as well as
expected problems with decoherence from e.g. nodal quasiparticles ( see section [ sec_decoherence ] ) , experimental research on @xmath11-wave qubits is not as far advanced as on the devices with conventional superconductors . the experimental confirmation of quantum behaviour in these systems is still missing .
nevertheless several recent successes should be noted .
arrays of half - flux quanta were realized and manipulated ( classically ) in ybco - nb zigzag junctions @xcite .
each facet of the junction effectively constitutes a @xmath14-ring , which supports spontaneous flux of @xmath146 .
interaction between the fluxes is mainly due to the superconducting connection , which leads to robust antiferromagnetic ordering . in its absence , a weaker , magnetic interaction establishes ferromagnetic flux ordering .
the authors consider the possibility of using their structures in qubit design .
good quality submicron grain boundary ybco junctions were fabricated and bistable energy vs. phase dependence was demonstrated @xcite .
dc squids ybco ( 15@xmath18615 @xmath187 square loops with nominally @xmath188 m wide grain boundary junctions ) were fabricated and tested , and their classical behaviour is very well described by the existing theory @xcite ( fig .
[ fig_dc_squid ] ) . like in ( [ eq_current_phase_silent ] ) , only two harmonics of current - phase dependence were considered . from fitting the experimental data
, we had to conclude that the junctions in the same squid have not only different critical current amplitudes , but different ratios of first to second harmonic ( @xmath189 ) , probably due to the variation in the grain boundary properties over a distance of @xmath190 15 @xmath10 m .
this is one reason for putting the junctions in the silent qubit , fig .
[ fig_silent ] , close to each other , as it was done when fabricating its prototype@xcite .
@xmath191 @xmath192 @xmath193 ( b ) @xmath194 @xmath195 @xmath196 @xmath197 decoherence is the major concern for any qubit realization , especially for solid state qubits , due to abundance of low - energy degrees of freedom . in superconductors ,
this problem is mitigated by the exclusion of quasiparticle excitations due to the superconducting gap .
this explains also why the very fact of existence of gapless excitations in high - t@xmath0 superconductors long served as a deterrent against serious search for macroscopic quantum coherence in these systems .
an additional source of trouble may be zero - energy states in dd junctions .
nevertheless , recent theoretical analysis of dd junctions @xcite , all using quasiclassical eilenberger equations , shows that the detrimental role of nodal quasiparticles and zes could be exaggerated . before turning to these results ,
let us first do a simple estimate of dissipation due to nodal quasiparticles in bulk @xmath11-wave superconductors .
consider , for example , a three - junction ( `` mooij '' ) qubit with d - wave phase shifters .
the @xmath147 and @xmath148 states support , respectively , clockwise and counterclockwise persistent currents around the loop , with superfluid velocity @xmath198 .
tunnelling between these states leads to nonzero average @xmath199 in the bulk of the superconducting loop .
time - dependent superfluid velocity produces a local electric field @xmath200 and quasiparticle current @xmath201 the resulting average energy dissipation rate per unit volume is @xmath202 here @xmath203 is the quasiparticle lifetime , and @xmath204\ ] ] is the effective quasiparticle density .
the angle - averaged density of states inside the d - wave gap is @xcite @xmath205 where @xmath206 , and @xmath207 is the maximal value of the superconducting order parameter . substituting ( [ q7 ] ) in ( [ q6 ] ) , we obtain @xmath208 where li@xmath209 is the dilogarithm . expanding for small @xmath210 , and taking into account that li@xmath211 , li@xmath212 , and li@xmath213
we obtain @xmath214 the two terms in parentheses correspond to thermal activation of quasiparticles and their `` cherenkov '' generation by current - carrying state .
note that finite quasiparticle density does not lead by itself to any dissipation . in the opposite limit ( @xmath215 ) only the `` cherenkov '' contribution remains , @xmath216 ( since li@xmath217 for large negative @xmath218 , and @xmath219 for small @xmath218 ) .
the energy dissipation rate provides the upper limit @xmath220 for the decoherence time ( since dissipation is sufficient , but not necessary condition for decoherence ) .
denoting by @xmath221 the amplitude of the persistent current in the loop , by @xmath38 the inductance of the loop , and by @xmath222 the effective volume of d - wave superconductor , where it flows , we can write @xmath223 note that the thermal contribution to @xmath224 is independent on the absolute value of the supercurrent in the loop ( @xmath225 ) , while the other term scales as @xmath226 . both contributions are proportional to @xmath222 and ( via @xmath227 ) to @xmath228 , the characteristic frequency of current oscillations ( i.e. tunneling rate between clockwise and counterclockwise current states ) .
it follows from the above analysis that the intrinsic decoherence in a @xmath11-wave superconductor due to nodal quasiparticles can be minimized by decreasing the amplitude of the supercurrent through it , and the volume of the material where _ time - dependent _ supercurrents flow . from this point of view the designs with phase shifters , as well as the newns - tsuei and `` no tunneling '' designs are at a disadvantage ( the latter , because the microwave field necessary to produce rabi transitions will affect the whole sample ) .
now let us estimate dissipation in a dd junction .
first , following @xcite , consider a dnd model with ideally transmissive nd boundaries .
due to tunneling , the phase will fluctuate , creating a finite voltage on the junction , @xmath229 , and normal current @xmath230 .
the corresponding dissipative function and decay decrement are @xmath231 here @xmath232 , @xmath233 , @xmath234 are the coulomb energy , effective `` mass '' and number of quantum channels in the junction respectively .
the latter is related to the critical josephson current @xmath235 and spacing between andreev levels in the normal part of the system @xmath236 via @xmath237 we require , that @xmath238 where @xmath239 is the frequency of small phase oscillations near a local minimum .
this means , @xmath240 the above condition allows a straightforward physical interpretation . in the absence of thermal excitations ,
the only dissipation mechanism in the normal part of the system is through the transitions between andreev levels , induced by fluctuation voltage .
these transitions become possible , if @xmath241 which brings us back to ( [ eq_no_dissipation ] ) .
another interpretation of this criterion arises if we rewrite it as @xmath242 @xcite . on the right - hand side
we see time for a quasiparticle to traverse the normal part of the junction .
if it exceeds the period of phase oscillations ( on the left - hand side ) , andreev levels simply do nt have time to form . since they provide the only mechanism for coherent transport through the system , the latter is impossible , unless our `` no dissipation '' criterion holds .
for the thickness of the normal layer @xmath243 and @xmath244 cm / s this criterion limits @xmath245 , which is a comfortable two orders of magnitude above the usually obtained tunneling splitting in such qubits ( @xmath246 ghz ) and can be accommodated in the above designs .
nevertheless , while presenting a useful qualitative picture , the dnd model is not adequate for the task of extracting quantitative predictions .
for example , the coherence length @xmath67 in the normal metal can be very large , while in the high - t@xmath0 compound it is short
. therefore the estimates for crucial parameters ( like @xmath247 ) based on the assumption @xmath248 can be wrong .
moreover , the assumption of ideal nd boundaries is not realistic .
a calculation@xcite , which used a model of a dd junction interacting with a bosonic thermal bath , gave an optimistic estimate for the quality of the tricrystal qubit @xmath249 .
the role of size quantization of quasiparticles in small dd and snd structures was suggested in @xcite .
the importance of the effect is that it would exponentially suppress the quasiparticle density and therefore the dissipation below temperature of the quantization gap , estimated as @xmath250 k. recently this problem was investigated@xcite for a finite width dd junction .
contrary to the expectations , the size quantization as such turned out to be effectively absent on the scale exceeding @xmath15 ( that is , practically irrelevant ) . on the other hand
, the finite transverse size of the system imposed an effective band structure .
namely , due to the direction dependence of the order parameter , a quasiparticle travelling along the trajectory , bouncing between the sides of the junction , goes through a periodic 1d off - diagonal potential .
the influence of this band structure on dissipation in the system is not straightforward .
ironically , from the practical point of view this is a moot point , since the decoherence time from the quasiparticles in the junction , estimated in @xcite , already corresponds to the quality factor @xmath251 , which exceeds by two orders of magnitude the theoretical threshold allowing to run a quantum computer indefinitely . the expression for the decoherence time obtained in @xcite , @xmath252 where @xmath253 is the difference between equilibrium phases in degenerate minima of the junction ( i.e. @xmath254 in other notation ) , contains the expression for quasiparticle current in the junction at finite voltage @xmath255 ( where @xmath256 is the tunneling rate between the minima ) .
this agrees with our back - of - the - envelope analysis : phase tunneling leads to finite voltage in the system through the second josephson relation and with finite voltage comes quasiparticle current and decoherence .
the aforementioned quality factor is defined as @xmath257 , that is , we compare the decoherence time with the tunneling time .
this is a usual optimistic estimate , since it will certainly take several tunneling cycles to perform a quantum gate operation .
it turns out that a much bigger threat is posed by the contribution from zero energy bound states , which can be at least two orders of magnitude larger .
we can see this qualitatively from ( [ eq_fominov_decoherence ] ) : a large density of quasiparticle states close to zero energy ( i.e. , on fermi level ) means that even small voltages create large quasiparticle currents , which sit in the denominator of the expression for @xmath258 fortunately , this contribution is suppressed in the case of zes splitting , and such splitting is always present due to , e.g. , finite equilibrium phase difference across the junction@xcite .
a similar picture follows from the analysis presented in @xcite . a specific question addressed there is especially important
: it is known that @xmath259-constant measured in dd junctions is consistently 1 ps over a wide range of junction sizes @xcite , and it is tempting to accept this value as the dissipation rate in the system .
it would be a death knell for any quantum computing application of high - t@xmath0 structures , and nearly that for any hope to see there some quantum effects .
nevertheless , it is not quite that bad . indeed
, we saw that zes play a major role in dissipation in a dd junction , but are sensitive to phase differences across it@xcite . measurements of the @xmath259 constant are done in the resistive regime , when a finite voltage exists across the junction , so that the phase difference grows monotonously in time , forcing zes to approach the fermi surface repeatedly .
therefore @xmath260 reflects some averaged dissipation rate . on the other hand , in a free junction with not too high a tunneling rate phase differences obviously tend to oscillate around @xmath261 or @xmath262 , its equilibrium values , and do not spend much time near zero or @xmath14 ; therefore zes are usually shifted from the fermi level , and their contribution to dissipation is suppressed .
this qualitative picture is confirmed by a detailed calculation@xcite ( fig .
[ fig_aminsmirnov ] ) .
the decoherence time is related to the phase - dependent conductance via @xmath263 here @xmath264 is the dissipation coefficient , @xmath265 is interlevel spacing in the well , and @xmath266 for a realistic choice of parameters eq.([eq_aminsmirnov ] ) gives a conservative estimate @xmath267 ns , and quality factor @xmath268 .
this is , of course , too little for quantum computing , but quite enough for observation of quantum tunneling and coherence in such junctions .
since the @xmath11-wave character of superconducting pairing in high - t@xmath0 cuprates was established , our understanding of high - t@xmath0 josephson structures and ability to fabricate them progressed significantly .
now it enables us to ask the question in the title of this paper : can high - t@xmath0 cuprates play a role in quantum computing?- and to tentatively answer : yes .
we have seen that submicron junctions of sufficiently high quality are now fabricated .
several designs , which take advantage of the intrinsic bistability of @xmath11-wave structures , were developed , fabricated , and tested in classical regime .
if the decoherence time turns out to be large enough , their better scalability can be a decisive advantage for quantum computing applications .
this is , of course , a big if .
still , theoretical estimates tend to be rather optimistic .
even though they widely differ , all of them predict @xmath269 long enough to encourage experimental search for quantum coherence .
there are several compelling reasons to do that .
first , it would be a spectacular result .
second , it would clarify why different models give different answers .
third , it could indeed lead to practical application of high - t@xmath0 devices in quantum computing .
since this paper was published ( 2003 ) , macroscopic quantum tunneling was observed in ybco @xcite and biscco @xcite junctions .
this made the `` big if '' of the previous paragraph somewhat smaller .
i am grateful to m.h.s .
amin , a. blais , a. golubov , e. ilichev , a.n .
omelyanchouk , a. smirnov , and a. tzalenchuk for many illuminating discussions , and to b. wilson for the careful reading and helpful comments on the manuscript . 99 mineev , v.p . , and k.v .
samokhin , _ introduction to unconventional superconductivity _ ,
amsterdam , the netherlands : gordon and breach science publishers ( 1999 ) .
eilenberger , g . , z.phys .
* 214 * ( 1968 ) 195 .
tafuri , f . , and j.r .
kirtley , phys .
b * 62 * ( 2000 ) 13934 .
zagoskin , a.m . , _ quantum theory of many - body systems _ , new york berlin heidelberg : springer - verlag ( 1998 ) .
kulik , i.o . ,
phys . -jetp * 30 * ( 1970 ) 944 .
ishii , g .
* 44 * ( 1970 ) 1525 .
bardeen , j . , and j.l .
johnson , phys .
b * 5 * ( 1972 ) 72 .
furusaki , a .
, superlat . and microstruct .
* 25 * ( 1999 ) ( 1999 ) 809 .
barzykin , v . , and a.m. zagoskin , superlat . and microstruct .
* 25 * ( 1999 ) ( 1999 ) 797 .
zagoskin , a.m . , j. phys . : condens .
matter * 9 * ( 1997 ) l419 .
sigrist , m . , and t.m .
rice , rev .
mod . phys .
* 67 * ( 1995 ) 503 .
van harlingen , d.j .
* 67 * ( 1995 ) 515 .
tsuei , c.c .
, and j.r .
kirtley , rev .
* 72 * ( 2000 ) 969 .
wollman , d.a . , _
et al . _ ,
rev , lett .
* 71 * ( 1993 ) 2134 .
wollman , d.a . , _
et al . _ ,
rev , lett .
* 74 * ( 1993 ) 797 .
tsuei , c.c . , _
et al . _ ,
* 73 * ( 1994 ) 593 .
tinkham , m . ,
_ introduction to superconductivity _ , 2nd ed .
, new york etc .- mcgraw hill , inc .
huck , a . , a. van otterlo , and m. sigrist , phys .
b * 56 * ( 1997 ) 14163 .
zagoskin , a.m . , and m. oshikawa , j. phys . : condens .
matter * 10 * ( 1998 ) l105 .
hurd , m . , _
et al . _ ,
b * 59 * ( 1999 ) 4412 .
amin , m.h.s .
omelyanchouk , and a.m. zagoskin , phys .
b * 63 * ( 2001 ) 212502 .
amin , m.h.s . , _
et al . _ , physica b * 318 * ( 2002 ) 162 .
amin , m.h.s . , _
et al . _ ,
b * 66 * ( 2002 ) 174515 .
zagoskin , a.m . , cond - mat/9903170 .
tafuri , f . , _
et al . _ , phys .
b * 59 * , 11523 ( 1999 ) .
ilichev , e . , _ et al . _ ,
prl * 86 * ( 2001 ) 5369 .
lfwander , t .
shumeiko and g. wendin , supercond .
* 14 * ( 2001 ) r53 .
nakamura , y . , yu.a .
pashkin , and j.s .
tsai , nature * 398 * ( 1999 ) 786 .
friedman , j.r . , _
et al . _ ,
nature * 406 * ( 2000 ) 43 .
van der wal , c.h . , _
et al . _ , science * 290 * ( 2000 ) 773 .
blatter , g .
geshkenbein , and l.b .
ioffe , phys .
b * 63 * ( 2001 ) 174511 .
newns , d.m . , and
tsuei , us patent no . us 6,495,854 b1 ( dec.17 , 2002 ) .
ioffe , l.b . , _
et al . _ ,
nature * 398 * ( 1999 ) 679 .
blais , a . , and a.m. zagoskin , phys .
a * 61 * ( 2000 ) 042308 .
joyez , p . , _
* 72 * ( 1994 ) 2458 .
zagoskin , a.m .
, unpublished .
hilgenkamp , h . , _
et al . _ ,
nature * 422 * ( 2003 ) 50 .
amin , m.h.s . , _
et al . _ ,
b * 71*(2005)064516 .
vion , d. , et al .
, science * 296 * ( 2002 ) 886 .
amin , m.h.s .
smirnov , and a. maassen van den brink , prb * 67 * ( 2003 ) 100508(r ) .
martinis , j.m . , _
et al . _ ,
* 89 * ( 2002 ) 117901 .
yu , y . , _
_ , science * 296 * ( 2002 ) 889 .
tzalenchuk , a.ya . , _
et al . _ , phys .
b * 68 * ( 2003 ) 100501(r ) .
lindstrm , t . , _
* 90 * ( 2003 ) 117002 .
fominov , ya . , and a. golubov , to be published ( 2003 ) .
amin , m.h.s . , and a.yu.smirnov , phys .
rev . lett . * 92 * ( 2004 ) 017001 .
fominov , ya.v .
golubov , and m.yu .
kupriyanov , jetp lett .
* 77 * ( 2003 ) 587 .
, d.k . yip , and j.a .
sauls , phys .
b * 51 * ( 1995 ) 16233 . bauch , t . , _
et al . _ ,
* 94 * ( 2005 ) 087003 .
inomata , k . , _ et al .
_ , cond - mat/0412604 . | due to nontrivial orbital pairing symmetry , surfaces and interfaces of high - t@xmath0 superconductors support states which violate time - reversal ( @xmath1- ) symmetry . such naturally degenerate states , useful as working states of a qubit , are standard for atomic or molecular - size qubit prototypes ( e.g. based on nuclear spins ) , but exceptional for mesoscopic qubits .
( in particular , they hold promise of a better scalability . ) in these lectures i review the physics of @xmath2-breaking on surfaces and interfaces of high - t@xmath0 superconductors ; then describe existing proposals for high - t@xmath0 based qubits and the current state of experiments ; finally , i discuss the decoherence sources in the system , open questions , and future research directions . |
cosmic strings @xcite are one - dimensional topological defects which arise during phase transitions in the very early universe .
since they carry energy , they will lead to density fluctuations and cosmic microwave background ( cmb ) anisotropies ( see e.g. @xcite for reviews on cosmic strings and structure formation ) .
causality implies that the network of strings which forms during the phase transition contains infinite strings .
once formed in the early universe , the network of strings will approach a `` scaling solution '' which is characterized by of the order one infinite string segment in each hubble volume , and a distribution of cosmic string loops which are the remnants of the previous evolution .
in particular , this implies that in any theory which admits cosmic strings , a network of strings will be present during the time period relevant to the cmb , namely between the time of recombination @xmath0 and the present time @xmath1 . a network of cosmic strings will generate a scale - invariant spectrum of cosmological perturbations @xcite .
as first discussed by kaiser and stebbins ( ks ) @xcite , the non - gaussian nature of the density field produced by strings will lead to a distinctive signature ( which we will call ks signature ) in cmb anisotropy maps , namely line discontinuities .
these line discontinuities are a consequence of the non - trivial nature of the metric produced by a cosmic string : space perpendicular to a cosmic string is a cone with deficit angle given by = 8 g , where @xmath2 is the mass per unit length of the string and @xmath3 is newton s gravitational constant @xcite .
since cosmic strings have a tension comparable to their mass per unit length , they will typically be moving with a relativistic transverse velocity @xmath4 .
as illustrated in figure 1 , if we are looking at the cmb in direction of the string , we will see the photons passing on different sides of the string with a doppler shift [ kssig ] t t = 8 ( v ) v g , where @xmath5 is the relativistic gamma factor .
looking in direction of the string , we will see a line in the sky across which the cmb temperature jumps by the above amount .
any cosmic string which photons reaching us today pass on their way from the last scattering surface will produce an effect .
the most numerous strings , however , are those present close to the time of last scattering @xmath0 . at that time , the typical curvature radius of a long string segment is of the order of the hubble radius at @xmath0 .
the corresponding angular scale today is about @xmath6 . in order to be able to identify the cosmic signature as a line discontinuity , as good an angular resolution as possible
is required .
if probed with an angular resolution not significantly smaller than @xmath6 , the ks signature will be washed out . in this work
we will study the potential of surveys with an angular of several minutes of arc to detect the ks signature for strings .
less important than the angular resolution of the survey is the total sky coverage , as long as it includes a sufficiently large number of hubble patches at @xmath0 .
obviously , increasing the sky coverage will reduce the standard deviation of the results and thus lead to some better discriminating power .
there has been surprisingly little work devoted to detecting the ks signature . an early study @xcite showed that the angular resolution of the wmap satellite would not be adequate to pick out the ks signature , even for values of @xmath7 for which the cosmic strings would dominate the power spectrum of density fluctuations . in order to detect the ks signature ,
new statistical methods must be employed . in @xcite ,
a matched filtering method was applied to the wmap data to search for cosmic string signatures . in @xcite , new statistics ( such as a statistic measuring the connectedness of neighboring temperature steps or a decomposition of the temperature map into constant background temperature , gaussian noise plus straight string step discontinuities ) were introduced and applied to the wmap data to search for strings .
based on the null results of both studies , an upper bound on @xmath7 of @xmath8 could be set .
however , this bound is not competitive with other existing bounds on the string tension coming from the observed acoustic peak structure of the cmb angular power spectrum @xcite ( which yields @xmath9 ) and from pulsar timing measurements @xcite ( which yield limits of between @xmath10 and @xmath11 , the difference being due to conflicting ways @xcite in which the pulsar timing data is analyzed to obtain limits on the amplitude of the stochastic background of gravitational waves , and due to different assumptions about the distribution of cosmic string loops , the main source of gravitational radiation from a string network - see @xcite for a recent discussion ) . note that an advantage of using the ks signature over pulsar bounds is that the ks signature is more robust ( less dependent on the unknown distribution of cosmic string loops ) than the pulsar bounds ( which depend sensitively on the details of the cosmic string loop distribution ) as discussed above , we expect to be able to set much stronger limits on @xmath7 from searches for the ks signature by looking at small - scale anisotropy maps , provided we find a good statistic to identify line discontinuities .
the goal of this paper is to explore the potential of one specific edge detection technique , the canny algorithm @xcite to pick out these line discontinuities in cmb anisotropy maps .
one reason for the relative lack of previous work on detecting the ks signature of cosmic strings is the fact that interest in cosmic strings as a source for structure in the universe decreased dramatically after the discovery of the sharpness of the acoustic peaks in the cmb angular power spectrum @xcite .
since perturbations seeded by strings are not coherent @xcite , a scenario with only cosmic strings as the source of density perturbations would only produce one broad doppler peak rather than the observed narrow acoustic oscillations @xcite .
however , recent years have seen a resurgence of interest in cosmic strings .
this was fueled by two developments .
firstly , it was realized @xcite that in many supersymmetric particle physics models , cosmic strings are formed after inflation , and thus contribute to but do not completely replace inflationary perturbations as the seeds for structure formation . secondly , it has recently been realized that models with cosmic superstrings @xcite may well be viable @xcite .
they could , for example , be generated as the remnant of brane annihilation processes in brane inflation models @xcite , or they may play an important roles in inflationary models in warped backgrounds @xcite .
cosmic superstrings may also be left behind after the initial hagedorn phase in string gas cosmology @xcite , where they would add an additional component to the spectrum of fluctuations produced by thermal string gas fluctuations @xcite .
thus , it is of great interest to find new ways to search for signatures of strings in cosmological data .
the outline of this paper is as follows : in the following section we review the canny algorithm and describe its application to our problem .
section 3 discusses how the temperature maps with and without strings are produced , introduces the parameters chosen in the specific simulations , and presents our results .
we conclude with a summary and a discussion of the caveats of the current analysis and prospects for future work .
the canny algorithm was developed in 1986 as a technique to detect edges in images @xcite such as the two - dimensional images considered here . when applied to a map of raw data ,
the algorithm is intended to produce a map tracing the edges in the map , the lines across which the intensity contrast is largest .
the first step in the algorithm is to filter the data to eliminate point source noise .
the filtering is achieved using a convolution of the map with a gaussian filter .
the filtering length is a free parameter in the algorithm .
it must be chosen sufficiently large to eliminate unwanted point source noise , but must be smaller than the characteristic size of the structure in the maps which one is trying to identify .
let us denote the original map by @xmath12 , and the filter by @xmath13 .
the filtered map @xmath14 is then given by fm(i , j ) = _
k , l m(i - k , j - l ) f(k , l ) , where the pixel points of the map are denoted by the labels @xmath15 . in our simulations , the filtering length is taken to be @xmath16 grid units . the second step of the algorithm is to find the gradient of the filtered image , a vector obtained by taking the discrete derivative of the map in each of the two directions . the gradient vector at a grid point @xmath15 is denoted by @xmath17 , where @xmath18 and @xmath19 are the derivatives in the respective directions .
the _ edge strength _ @xmath20 at any given grid point is defined by |g| = , and the _ edge direction _ is given by the angle = arctan(g_y / g_x ) .
since we are working on a grid , it makes sense to replace the edge direction by one of the eight distinguished directions on a quadratic grid , namely the four directions along the coordinate axes and the four diagonal directions .
the average maximal gradient in maps without cosmic strings ( the average taken over all of our runs ) is denoted by @xmath21 . for a specific run ,
the maximal gradient is obtained by scanning the entire map and searching for the maximal value of the grid strength .
the value of @xmath21 is used to set the thresholds discussed below .
the next step of the canny algorithm is to find grid points with gradients which are maximal when we vary the point in direction of the gradient .
grid points which are not local maxima of the gradient are assigned a number @xmath22 . at this stage ,
two thresholds must be set , an upper and a lower threshold @xmath23 and @xmath24 .
the algorithm first finds local maxima in the gradient along one of the four directional axes on the grid .
next , it goes through each of the list of local maxima and determines whether the gradient is greater than the upper threshold fraction of @xmath25 , i.e. |g| > t_u g_m .
grid point thus selected are marked with a number @xmath26 .
if the gradient is larger than the lower threshold fraction of @xmath25 , |g| > t_l g_m , the grid point is assigned a value of @xmath27 . for points marked with @xmath26 or @xmath27 ,
the direction of the gradient is also stored . for each grid point with a value of @xmath27 , the algorithm next checks whether there is a grid point with value of @xmath26 or @xmath27 which is in a direction perpendicular to the gradient and whose gradient is parallel or next to parallel ( thus three possible directions in total ) to the initial gradient .
if such a grid point is found , the algorithm continues until it either does not find another point satisfying the criteria or else finds a point satisfying the criteria marked with a @xmath26 .
if it finds such a point , then all of the points ( initially marked @xmath27 ) found along the way are marked as @xmath26 .
neighboring ( in all eight directions ) grid points marked by @xmath26 are then said to belong to the same edge
. these are edges across which the gradient of the map is large .
the reason for using two thresholds is as follows : we want the algorithm to find lines with a large gradient . at the same time , we do not want noise to lead to points along a line of large gradients to be missed ( and thus the lines cut ) just because noise has reduced the amplitude of the edge strength at one point along the line . in this way , the algorithm produces a list of grid points belonging to edges , and this list can be read out as a map of edges . in our simulations ,
the upper and lower thresholds are @xmath28 and @xmath29 .
it is important to quantify the edge maps .
a simple way to do this is to produce a histogram of edge lengths .
thus , the implementation of the canny algorithm is arranged such as to output a list of edge lengths which can then be used to produce a histogram of edge lengths .
the histogram contains useful information about the presence of edges in the map .
eventually , we would like to apply the canny algorithm to actual cmb maps . for this initial feasibility study , however , we will apply the algorithm to theoretical maps produced in numerical simulations .
we wish to compare temperature maps for the `` standard '' gaussian @xmath30cdm model with maps in which a network of strings contributes a fraction @xmath31 of the total power .
the maps are characterized by the angular scale of the survey ( we take the survey area to be square ) and by the angular scale of the grid , i.e. the angular resolution of the survey .
the gaussian maps are produced in the following way : we start with the angular power spectrum @xmath32 of cosmic microwave anisotropies taken from the cmbfast simulation @xcite with the appropriate cosmological parameters .
values for which the cmbfast program generates @xmath32 values is limited to @xmath33 , we are currently unable to explore angular scales relevant to upcoming experiments .
we are at the moment completing an improved code which will allow us to obtain an improved angular resolution . ]
since we have in mind applications of our algorithm to small angular scale surveys , we perform a `` flat sky '' approximation @xcite .
we introduce a two - dimensional cartesian coordinate system covering the survey area which we take to be rectangular , choosing the upper right corner of the survey area to be the origin of the coordinates .
we need to compute the temperature field @xmath34 of the map at the coordinate values @xmath35 corresponding to the grid points .
this map is determined by an inverse fast fourier transform from the temperature values @xmath36 in fourier space . for each vector @xmath37 in fourier space , we find the integers @xmath38 and @xmath39 which bracket the @xmath40 value @xmath41 corresponding to @xmath37 and take a linear interpolation of the values of @xmath32 .
let us denote the result of this linear interpolation by @xmath42 , where @xmath43 .
the value of @xmath44 is then given as follows : ( ) = g ( ) , where @xmath45 is a random variable drawn from a probability distribution with variance @xmath26 . in the above analysis
, we are singling out one grid point to be special , namely the origin .
this introduces unwanted phase correlations which we overcome by superimposing the results from four separate simulations , indicated by @xmath46 : t_g(x , y ) & = & 1 2 ( t_1(x , y ) + t_2(x_m - x , y_m - y ) + & & + t_3(x_m - x , y ) + t_4(x , y_m - y ) ) , where @xmath47 and @xmath48 are the maximal @xmath49 and @xmath50 values of the survey volume ( the pre - factor of @xmath27 is required in order to maintain the original standard deviation ) .
a network of cosmic strings will produce a temperature map @xmath51 . in the following ,
we take a toy model for a temperature map produced by strings introduced by perivolaropoulos @xcite .
the toy model takes into account that at all times between @xmath0 ( the time of recombination ) and the present time @xmath1 , the network of strings is described by a scaling solution in which there are a fixed number of long string segments ( strings which are not loops with radius smaller than the hubble radius ) crossing each hubble volume .
each such string gives rise to a kaiser - stebbins line discontinuity ( [ kssig ] ) in the temperature map .
the cosmic string network is continuously evolving via motion of the strings and string interactions which lead to the production of string loops .
thus , in each hubble time step the string network can be taken to be uncorrelated . in the algorithm
, we divide the time interval @xmath52 into 15 hubble time steps . in each time step @xmath53
we lay down a network of strings at random , uncorrelated with the network at the previous time step .
we take the network to consist of straight string segments of length @xmath54 . in order to avoid missing strings at the edges of the survey area , the survey area
is extended in each direction by a hubble distance .
the program runs through all points of the simulation volume and picks points to be midpoints of a string segment with a probability chosen such that the average number of strings in the hubble volume equals the number @xmath55 of the scaling solution .
the directions of the strings are chosen at random , as are the string transverse velocities . to take into account also the projection onto the last scattering surface , we add a temperature = 1 2 v r 8 g , to one side of the string projection onto the last scattering surface , and subtract it from the other ( so as to maintain the average temperature of the cmb ) . here
, @xmath56 stands for the maximal value of @xmath57 , where @xmath4 is the transverse velocity of the string , and @xmath5 is the relativistic @xmath58 factor associated with @xmath4 . also , @xmath59 is a random number between @xmath22 and @xmath26 , to take into account both the distribution of velocities , and also projection effects ( the formula ( [ kssig ] ) for the line discontinuity in temperature is modulated by an angular factor if the string velocity is not perpendicular to our line of sight to the string ) .
the regions affected by the temperature fluctuation are rectangles on either side of the string .
the depth of these rectangles in direction transverse to the direction of the string is taken to be a fraction @xmath31 of the hubble radius .
the length of the string segments is taken to be @xmath60 times the hubble radius .
the survey area is always taken to correspond to the central region of the simulation area .
the free parameters of the string simulation are @xmath7 , the number of strings @xmath55 per hubble volume , the length coefficients @xmath60 and @xmath31 , as well as @xmath56 . in our work , we will fix @xmath61 , @xmath62 , @xmath63 and @xmath64 .
these values are representative of the results of numerical simulations of cosmic string evolution ( see @xcite for reviews ) .
however , it should be kept in mind that the results for these parameters obtained using different numerical codes differ significantly .
let us make a couple of comments on the expected dependence of the results on the value of @xmath55 .
as @xmath55 initially increases from zero , the signature of the strings should increase since there are more strings .
however , once @xmath55 rises above a certain critical value , then further increasing the number @xmath55 will render the string distribution more gaussian by the central limit theorem and will hence decrease the sensitivity of the canny algorithm .
we expect that the critical value of @xmath55 will be the one for which the regions of the sky effected by individual string segments begin to overlap .
if the strings have length comparable to the hubble radius , the the critical value of @xmath55 is expected to be less than @xmath61 . increasing the string velocity @xmath56 will boost the line discontinuites in the temerature maps and will make our algorithm more effective . since the weight in our statistical analysis of the edge maps comes from fairly short segments ,
our results are not very sensitive to the specific values of @xmath31 and @xmath60 .
we are interested in using the canny algorithm to find signatures for cosmic strings in maps which contain both gaussian fluctuations from inflation and a cosmic string component ( which is sub - dominant in terms of its contribution to the power spectrum ) . the total map @xmath65 is given by the superposition of a pure gaussian noise map and the map produced by a distribution of strings : t_t(x , y ) = t_g(x , y ) + t_cs(x , y ) , and where the amplitude of the gaussian term is adjusted such that the total angular power spectrum agrees with the cobe results @xcite .
the output of the canny algorithm is a map of edges which have been picked out . in the presence of cosmic strings
, we would expect more longer edges than in the absence of strings . in order to test for this effect , we produced histograms of the distribution of edge lengths for each simulation .
since both the gaussian maps and the maps with cosmic strings are produced by a gaussian random process ( the phases of the fourier modes for the gaussian maps are picked at random , and for the string maps the locations of the string centers , their directions and their transverse velocity vectors are all random ) , we ran many ( @xmath66 ) simulations for both the gaussian map and the maps with strings .
this gave us statistical error bars on the histograms of edge lengths .
in turn , this allows us to assign statistical weight to the difference in the histograms .
our results are based on this analysis . in our simulations ,
we consider the `` gaussian noise '' maps to be given by a standard @xmath30cdm model with adiabatic fluctuations with a spectral index of @xmath67 , and with parameters @xmath68 ( baryon fraction ) , @xmath69 ( cold dark matter fraction ) , @xmath70 ( cosmological constant contribution ) , and @xmath71 ( hubble expansion rate ) .
no massless neutrinos , standard recombination history and a helium fraction of @xmath72 were assumed ( in the @xmath32 spectra which were used to construct the maps ) .
the parameters of the string simulation are indicated above .
our maps have angular extent @xmath73 degrees by @xmath73 degrees .
we fix the angular resolution and search for the minimal value of the cosmic string mass parameter @xmath7 for which the maps with the strings can be distinguished from the pure gaussian maps at a statistically significant level .
we demand significance at the @xmath74 level .
figure 2 shows a cmb temperature fluctuation map of a simulation without strings , and figure 3 is the corresponding map of a simulation which includes strings according to the prescription described above .
square of the sky for a @xmath30cdm simulation with the parameters described in the text.,height=340 ] ( and other parameters as described in the text.,height=340 ] the output of the canny algorithm applied to the above maps is shown in figures 4 and 5 , respectively .
cdm map of figure 2.,height=340 ] based on these results , our algorithm evaluates the distribution of edge lengths and checks whether the distributions of the histograms with and without strings are statistically different .
maps with cosmic strings showed , as expected , a slight excess of edges of all lengths .
this excess is more pronounced for edge lengths which lie in the non - gaussian tail .
the results for an angular resolution of @xmath75 are presented in table 1 . in this table ,
the rows indicate the length @xmath76 of the edge .
the entries in the columns are the mean number [ standard deviation of the mean ] of edges of the maps which have the respective length .
the first column of numbers is for simulations without strings , the next two are for simulations including a network of strings with the indicated mass per unit length parameter . each histogram was produced from 50 runs . from the histogram
it follows that the difference between the distributions is significant ( at the @xmath74 level ) if @xmath77 , but not if @xmath78 . for a given angular resolution , we varied the string mass parameter @xmath7 to find the limiting value below which the differences in the histograms ceased to be statistically significant ( using fisher s combined probability test ) .
our results are summarized in table 2 .
based on the above results , prospects for being able to use the canny algorithm to significantly improve the limits on the cosmic string mass parameter @xmath7 ( or detect cosmic strings ) are excellent .
the angular resolution of the planck satellite experiment will be about @xmath79 , the south pole telescope is aiming for a resolution of @xmath80 with a field of view of @xmath81 square degrees , and the resolution of the act telescope in chile will be @xmath82 .
.histograms of the distribution of edge lengths @xmath76 . [ cols="^,^,^,^",options="header " , ]
we have suggested a new way of looking for the specific signature of cosmic strings , namely the kaiser - stebbins line discontinuities , in small - scale cosmic microwave anisotropy maps .
our method makes use of the canny algorithm , an edge - detection technique used often in pattern recognition .
since the cosmic microwave temperature anisotropy maps induced by a network of cosmic strings are dominated by the strings present at the time of last scattering ( when the hubble radius , which is comparable to the correlation length of the string network at that time , is of the order of one degree ) , good small - scale angular resolution is essential in order to be able to detect strings .
an angular resolution substantially less than @xmath83 is required , otherwise the signals from the line discontinuities are washed out .
we have constructed cmb anisotropy maps which correspond to having both gaussian `` noise '' with a nearly scale - invariant power spectrum from inflation and anisotropies produced by a distribution of straight string segments . based on our numerical simulations
, we find that for an angular resolution of cmb maps of @xmath75 the canny algorithm has the potential to detect strings with a mass per unit length @xmath2 above a value of @xmath84 , close to an order of magnitude better than current limits based on the cmb .
a drawback of our work is that it is based on toy model cosmic string cmb maps which are obtained by superimposing idealized line discontinuities of straight string segments , not from actual string networks .
actual string networks contain both infinite strings and string loops .
the infinite strings are not completely straight , but have a curvature radius comparable to the hubble radius .
an improved analysis should start from a numerical simulation of the distribution of cosmic strings , calculate the induced temperature anisotropies taking into account of all effects following the formalism set out in @xcite , and then apply the canny algorithm to the resulting maps .
after completion of this work , a paper appeared @xcite in which cmb temperature maps on scales similar to the ones we are considering were constructed based on full cosmic string simulations .
this work confirmed that the kaiser - stebbins from long string segments is the dominating visible effect , and suggested , in agreement with the message of our work , that strings should be visible in high resolution small - scale cmb anisotropy maps .
however , no string - specific statistical analyses of the maps like the one we are proposing were performed in @xcite .
another important issue which remains to be analyzed is the effect of instrumental noise .
we have done preliminary work on this topic and modelled instrumental noise by a component to the @xmath32 spectrum which rises rapidly as a function of @xmath40 at the scale of the angular resolution of the survey , becoming dominant at the angular resolution scale .
initial results show that the efficiency of the canny algorithm is not reduced .
we plan to study this question in more detail .
this work is supported in part by a nserc discovery grant , by funds from the crc program , and by a fqrnt team grant .
we wish to thank matt dobbs , christophe ringeval , and in particular gil holder for useful discussions .
t. w. b. kibble , `` topology of cosmic domains and strings , '' j. phys . a * 9 * , 1387 ( 1976 ) ; + y. b. zeldovich , `` cosmological fluctuations produced near a singularity , '' mon . not .
soc . * 192 * , 663 ( 1980 ) ; + a. vilenkin , `` cosmological density fluctuations produced by vacuum strings , '' phys .
* 46 * , 1169 ( 1981 ) [ erratum - ibid .
* 46 * , 1496 ( 1981 ) ] .
a. vilenkin and e.p.s .
shellard ; _ cosmic strings and other topological defects _ , ( cambridge univ .
press , cambridge , 1994 ) ; + m. b. hindmarsh and t. w. kibble , `` cosmic strings , '' rept .
phys . * 58 * , 477 ( 1995 ) [ arxiv : hep - ph/9411342 ] ; + r. h. brandenberger , `` topological defects and structure formation , '' int .
j. mod .
phys . a * 9 * , 2117 ( 1994 ) [ arxiv : astro - ph/9310041 ] . n. turok and r. h. brandenberger , `` cosmic strings and the formation of galaxies and clusters of galaxies , '' phys .
d * 33 * , 2175 ( 1986 ) ; + h. sato , `` galaxy formation by cosmic strings , '' prog .
phys . * 75 * , 1342 ( 1986 ) ; + a. stebbins , `` cosmic strings and cold matter '' , ap .
j. ( lett . ) * 303 * , l21 ( 1986 ) .
n. kaiser and a. stebbins , `` microwave anisotropy due to cosmic strings , '' nature * 310 * , 391 ( 1984 ) .
a. vilenkin , `` gravitational field of vacuum domain walls and strings , '' phys
. rev .
d * 23 * , 852 ( 1981 ) .
r. moessner , l. perivolaropoulos and r. h. brandenberger , `` a cosmic string specific signature on the cosmic microwave background , '' astrophys .
j. * 425 * , 365 ( 1994 ) [ arxiv : astro - ph/9310001 ] .
a. s. lo and e. l. wright , `` signatures of cosmic strings in the cosmic microwave background , '' arxiv : astro - ph/0503120 .
e. jeong and g. f. smoot , `` search for cosmic strings in cmb anisotropies , '' astrophys .
j. * 624 * , 21 ( 2005 ) [ arxiv : astro - ph/0406432 ] ; + e. jeong and g. f. smoot , `` the validity of the cosmic string pattern search with the cosmic microwave background , '' arxiv : astro - ph/0612706 . l. pogosian , s. h. h. tye , i. wasserman and m. wyman , `` observational constraints on cosmic string production during brane inflation , '' phys .
d * 68 * , 023506 ( 2003 ) [ erratum - ibid .
d * 73 * , 089904 ( 2006 ) ] [ arxiv : hep - th/0304188 ] ; + m. wyman , l. pogosian and i. wasserman , `` bounds on cosmic strings from wmap and sdss , '' phys .
d * 72 * , 023513 ( 2005 ) [ erratum - ibid .
d * 73 * , 089905 ( 2006 ) ] [ arxiv : astro - ph/0503364 ] ; + a. a. fraisse , `` limits on defects formation and hybrid inflationary models with three - year wmap observations , '' jcap * 0703 * , 008 ( 2007 ) [ arxiv : astro - ph/0603589 ] ; + u. seljak , a. slosar and p. mcdonald , `` cosmological parameters from combining the lyman - alpha forest with cmb , galaxy clustering and sn constraints , '' jcap * 0610 * , 014 ( 2006 ) [ arxiv : astro - ph/0604335 ] ; + n. bevis , m. hindmarsh , m. kunz and j. urrestilla , `` cmb power spectrum contribution from cosmic strings using field - evolution simulations of the abelian higgs model , '' phys .
d * 75 * , 065015 ( 2007 ) [ arxiv : astro - ph/0605018 ] ; + n. bevis , m. hindmarsh , m. kunz and j. urrestilla , `` fitting cmb data with cosmic strings and inflation , '' arxiv : astro - ph/0702223 .
f. r. bouchet and d. p. bennett , `` does the millisecond pulsar constrain cosmic strings ? , '' phys .
d * 41 * , 720 ( 1990 ) . v. m. kaspi , j. h. taylor and m. f. ryba , `` high - precision timing of millisecond pulsars .
3 : long - term monitoring of psrs b1855 + 09 and b1937 + 21 , '' astrophys . j. * 428 * , 713 ( 1994 ) ; + s. e. thorsett and r. j. dewey , `` pulsar timing limits on very low frequency stochastic gravitational radiation , '' phys .
d * 53 * , 3468 ( 1996 ) ; + m. p. mchugh , g. zalamansky , f. vernotte and e. lantz , `` pulsar timing and the upper limits on a gravitational wave background : a bayesian approach , '' phys .
d * 54 * , 5993 ( 1996 ) ; + a. n. lommen , `` new limits on gravitational radiation using pulsars , '' arxiv : astro - ph/0208572 ; + f. a. jenet _ et al .
_ , `` upper bounds on the low - frequency stochastic gravitational wave background from pulsar timing observations : current limits and future prospects , '' astrophys . j. * 653 * , 1571 ( 2006 ) [ arxiv : astro - ph/0609013 ] .
j. polchinski , `` introduction to cosmic f- and d - strings , '' arxiv : hep - th/0412244 ; + j. polchinski , `` cosmic string loops and gravitational radiation , '' arxiv:0707.0888 [ astro - ph ] . j. canny , `` a computational approach to edge detection '' , ieee trans .
pattern analysis and machine intelligence * 8 * , 679 ( 1986 ) .
d. mauskopf _ et al .
_ [ boomerang collaboration ] , `` measurement of a peak in the cosmic microwave background power spectrum from the north american test flight of boomerang , '' astrophys . j. * 536 * , l59 ( 2000 ) [ arxiv : astro - ph/9911444 ] .
c. l. bennett _ et al .
_ , `` first year wilkinson microwave anisotropy probe ( wmap ) observations : preliminary maps and basic results , '' astrophys .
j. suppl .
* 148 * , 1 ( 2003 ) [ arxiv : astro - ph/0302207 ] .
a. albrecht , d. coulson , p. ferreira and j. magueijo , `` causality and the microwave background , '' phys .
lett . * 76 * , 1413 ( 1996 ) [ arxiv : astro - ph/9505030 ] .
l. perivolaropoulos , `` spectral analysis of microwave background perturbations induced by cosmic strings , '' astrophys .
j. * 451 * , 429 ( 1995 ) [ arxiv : astro - ph/9402024 ] . j. magueijo , a. albrecht , d. coulson and p. ferreira , `` doppler peaks from active perturbations , '' phys .
* 76 * , 2617 ( 1996 ) [ arxiv : astro - ph/9511042 ] .
u. l. pen , u. seljak and n. turok , `` power spectra in global defect theories of cosmic structure formation , '' phys .
lett . * 79 * , 1611 ( 1997 ) [ arxiv : astro - ph/9704165 ] .
r. jeannerot , `` a supersymmetric so(10 ) model with inflation and cosmic strings , '' phys .
d * 53 * , 5426 ( 1996 ) [ arxiv : hep - ph/9509365 ] ; + r. jeannerot , j. rocher and m. sakellariadou , `` how generic is cosmic string formation in susy guts , '' phys .
d * 68 * , 103514 ( 2003 ) [ arxiv : hep - ph/0308134 ] .
e. witten , `` cosmic superstrings , '' phys .
b * 153 * , 243 ( 1985 ) .
e. j. copeland , r. c. myers and j. polchinski , `` cosmic f- and d - strings , '' jhep * 0406 * , 013 ( 2004 ) [ arxiv : hep - th/0312067 ] .
s. sarangi and s. h. h. tye , `` cosmic string production towards the end of brane inflation , '' phys .
b * 536 * , 185 ( 2002 ) [ arxiv : hep - th/0204074 ] .
c. p. burgess , `` inflatable string theory ?
, '' pramana * 63 * , 1269 ( 2004 ) [ arxiv : hep - th/0408037 ] ; + j. m. cline , `` inflation from string theory , '' arxiv : hep - th/0501179 ; + a. linde , `` inflation and string cosmology , '' econf * c040802 * , l024 ( 2004 ) [ arxiv : hep - th/0503195 ] .
r. h. brandenberger and c. vafa , `` superstrings in the early universe , '' nucl .
b * 316 * , 391 ( 1989 ) .
a. nayeri , r. h. brandenberger and c. vafa , `` producing a scale - invariant spectrum of perturbations in a hagedorn phase of string cosmology , '' phys .
lett . * 97 * , 021302 ( 2006 ) [ arxiv : hep - th/0511140 ] ; + r. h. brandenberger , a. nayeri , s. p. patil and c. vafa , `` string gas cosmology and structure formation , '' arxiv : hep - th/0608121 .
u. seljak and m. zaldarriaga , `` a line of sight approach to cosmic microwave background anisotropies , '' astrophys . j. * 469 * , 437 ( 1996 ) [ arxiv : astro - ph/9603033 ] .
m. j. white , j. e. carlstrom and m. dragovan , `` interferometric observation of cosmic microwave background anisotropies , '' astrophys .
j. * 514 * , 12 ( 1999 ) [ arxiv : astro - ph/9712195 ] . l. perivolaropoulos , `` cobe versus cosmic strings : an analytical model , '' phys .
b * 298 * , 305 ( 1993 ) [ arxiv : hep - ph/9208247 ] ; + l. perivolaropoulos , `` statistics of microwave fluctuations induced by topological defects , '' phys
. rev .
d * 48 * , 1530 ( 1993 ) [ arxiv : hep - ph/9212228 ] .
g. f. smoot _ et al .
_ , `` structure in the cobe differential microwave radiometer first year maps , '' astrophys .
j. * 396 * , l1 ( 1992 ) .
a. a. fraisse , c. ringeval , d. n. spergel and f. r. bouchet , `` small - angle cmb temperature anisotropies induced by cosmic strings , '' arxiv:0708.1162 [ astro - ph ] . | line discontinuities in cosmic microwave background anisotropy maps are a distinctive prediction of models with cosmic strings .
these signatures are visible in anisotropy maps with good angular resolution and should be identifiable using edge detection algorithms .
one such algorithm is the canny algorithm .
we study the potential of this algorithm to pick out the line discontinuities generated by cosmic strings . by applying the algorithm to small - scale microwave anisotropy maps generated from theoretical models with and without cosmic strings , we find that , given an angular resolution of several minutes of arc , cosmic strings can be detected down to a limit of the mass per unit length of the string which is one order of magnitude lower than the current upper bounds . |
before reviewing the prospects for spin physics at the bnl - rhic we briefly turn to longitudinally polarized deep - inelastic scattering ( dis ) and what we have learned from more than twenty years of beautiful data @xcite .
to next - to - leading order ( nlo ) in the strong coupling @xmath0 , the dis structure function @xmath1 , which parametrizes our ignorance about the nucleon spin structure , can be expressed as @xmath2 ( x , q^2)\;\;\;.\ ] ] the symbol @xmath3 denotes the usual convolution , and @xmath4 are the perturbatively calculable coefficient functions , which are known even up to next - to - next - to - leading order @xcite .
the @xmath5 , @xmath6 , are the spin - dependent parton distributions , defined as @xmath7 where @xmath8 @xmath9 is the number density of a parton type @xmath10 with helicity `` @xmath11 '' ( `` @xmath12 '' ) in a proton with positive helicity , carrying a fraction @xmath13 of the proton s momentum .
once they are known at some initial scale @xmath14 , their scale @xmath15 dependence is governed by a set of evolution equations @xmath16 so far , the spin - dependent @xmath17 splitting functions entering these evolution equations have been calculated up to nlo accuracy @xcite .
= 0.59 figure [ fig : fig1 ] compares the available information on @xmath18 for dis off a proton target to results of a typical nlo qcd fit . from such types of analyses
a pretty good knowledge of certain combinations of different quark flavors has emerged , and it became clear that quarks contribute only a small fraction to the proton s spin .
however , there is still considerable lack of knowledge regarding the polarized gluon density @xmath19 , which is basically unconstrained by present data , the separation of quark and antiquark densities and of different flavors , and the orbital angular momentum of quarks and gluons inside a nucleon . with the exception of orbital angular momentum rhic
can address all of these questions as will be demonstrated in the following @xcite .
there is also an important difficulty when analyzing polarized dis data in terms of spin - dependent parton densities : compared to the unpolarized case the presently available kinematical coverage in @xmath13 and @xmath20 and the statistical precision of polarized dis data are much more limited @xcite . as a consequence ,
one is forced to include data into the fits from @xmath21-regions where corresponding fits of unpolarized leading - twist parton densities start to break down , see fig .
[ fig : fig1 ] . data from rhic , taken at `` resolution '' scales @xmath15 where perturbative qcd and the leading - twist approximation are supposed to work , can shed light on the possible size of unwanted higher - twist contributions in presently available sets of polarized parton distributions .
the qcd - improved parton model has been successfully applied to many high energy scattering processes .
the predictive power of perturbative qcd follows from the universality of the parton distribution and fragmentation functions which is based on the factorization theorem . to be specific ,
let us consider the inclusive production of a hadron @xmath22 , e.g. , a pion , in longitudinally polarized @xmath23 collisions at a c.m.s .
energy @xmath24 .
the cross section can be written in a factorized form as a convolution of perturbatively calculable partonic cross sections @xmath25 describing the hard scattering @xmath26 and appropriate combinations of parton densities @xmath27 and fragmentation functions @xmath28 embodying the non - perturbative physics : @xmath29 here , @xmath30 stands for any appropriate set of kinematical variables like the transverse momentum @xmath31 and/or rapidity @xmath32 of the observed hadron .
the @xmath28 are the parton - to - hadron fragmentation functions .
their scale @xmath15-dependence is governed by a set of equations very similar to ( [ eq : evol ] ) .
the factorization scale @xmath15 , introduced on the r.h.s . of ( [ eq : eq1 ] ) , separates long- and short - distance phenomena .
@xmath15 is completely arbitrary but usually chosen to be of the order of the scale characterizing the hard interaction , for instance @xmath31 . since the l.h.s
. of ( [ eq : eq1 ] ) has to be independent of @xmath15 ( and other theoretical conventions ) , any residual dependence of the r.h.s . on the actual choice of @xmath15
gives an indication of how well the theoretical calculation is under control and can be trusted .
in particular , leading order ( lo ) estimates suffer from a strong , uncontrollable scale dependence and hence are not sufficient for comparing theory with data .
figure [ fig : fig2 ] shows a typical factorization scale dependence for various `` high-@xmath31 '' processes and experiments as a function of @xmath31 . clearly , the situation is only acceptable at collider experiments where one can easily reach @xmath31 values in excess of 5 gev .
@xmath31 values of the order of 1 - 2 gev , accessible at fixed - target experiments , are not sufficient to provide a large enough scale @xmath15 for performing perturbative calculations reliably .
for simplicity we have not distinguished between renormalization and initial / final - state factorization scales in ( [ eq : eq1 ] ) which can be chosen differently . in practice , spin experiments do not measure the polarized cross section @xmath33 itself , but the longitudinal spin asymmetry , which is given by the ratio of the polarized and unpolarized cross sections , i.e. , for our example above , eq .
( [ eq : eq1 ] ) , it reads @xmath34 the unpolarized cross section @xmath35 is given by eq .
( [ eq : eq1 ] ) with all polarized quantities replaced by their unpolarized counterparts . at rhic
one can also study doubly transverse spin asymmetries @xcite but here we will focus on longitudinal polarization only .
the main thrust of the rhic spin program @xcite is to pin down the so far elusive gluon helicity distributions @xmath36 .
the strength of rhic is the possibility to probe @xmath36 in a variety of hard processes @xcite , in each case at sufficiently large @xmath31 where perturbative qcd is expected to work .
this not only allows to determine the @xmath13-shape of @xmath36 for @xmath37@xmath38 but also verifies the universality property of polarized parton densities for the first time . in the following
we review the status of theoretical calculations for processes sensitive to @xmath19 , experimental aspects can be found , e.g. , in @xcite .
the `` classical '' tool for determining the gluon density is high-@xmath31 prompt photon production due to the dominance of the lo compton process , @xmath39 . exploiting this feature , both rhic experiments , phenix and star , intend to use this process for a measurement of @xmath19 .
apart from `` direct '' mechanisms like @xmath39 , the photon can also be produced by a parton , scattered or created in a hard qcd reaction , which fragments into the photon .
such a contribution naturally arises in a qcd calculation from the necessity of factorizing final - state collinear singularities into a photon fragmentation function @xmath40 .
however , since photons produced through fragmentation are always accompanied by hadronic debris , an `` isolation cut '' imposed on the photon signal in experiment , e.g. , a `` cone '' , strongly reduces such contributions to the cross section .
= 3.15 in the nlo qcd corrections to the direct ( non - fragmentation ) processes have been calculated in @xcite and lead to a much reduced factorization scale dependence as compared to lo estimates .
in addition , monte carlo codes have been developed @xcite , which allow to include various isolation criteria and to study also photon - plus - jet observables .
the latter are relevant for @xmath19 measurements planned at star @xcite . since present comparisons between experiment and theory are not fully satisfactory in the unpolarized case , in particular in the fixed - target regime , considerable efforts have been made to push calculations beyond the nlo of qcd by including resummations of large logarithms @xcite .
it is hence not unlikely that a better understanding of prompt photon production can be achieved soon .
figure [ fig : fig3 ] shows @xmath41 as predicted by a nlo qcd calculation @xcite as a function of the photon s transverse momentum @xmath31 .
the applied rapidity cut @xmath42 matches the acceptance of the phenix detector .
the important result is that the expected statistical errors @xmath43 are considerably smaller than the changes in @xmath41 due to different spin - dependent gluon densities over a wide range of @xmath31 .
hence rhic should be able to probe @xmath19 in prompt photon production .
@xmath43 may be estimated by the formula @xmath44 where @xmath45 is the polarization of one beam , @xmath46 the integrated luminosity of the @xmath23 collisions , and @xmath47 the unpolarized cross section integrated over the @xmath31-bin for which the error is to be determined .
unless stated otherwise , @xmath48 and @xmath49 is used in eq .
( [ eq : error ] ) for @xmath23 collisions at @xmath50 @xcite
. jets are another key - process to pin down @xmath19 at rhic : they are copiously produced at @xmath51 , even at high @xmath31 , @xmath52@xmath53@xmath54 , and gluon - induced @xmath55 and @xmath56 processes are expected to dominate in accessible kinematical regimes .
= 3.05 in due to limitations in the angular coverage , jet studies will be performed by star only . as jet surrogates ,
phenix can look for high-@xmath31 leading hadrons , such as pions , whose production proceeds through the same partonic subprocesses as jet production .
hadrons have the advantage that they can be studied also at @xmath57 and down to lower values in @xmath31 than jets as they do not require the observation of clearly structured `` clusters '' of particles ( jets ) .
contrary to jet production , a fragmentation function has to be introduced into the theoretical framework , cf .
( [ eq : eq1 ] ) , to take care of final - state collinear singularities .
in case of pion production , the @xmath58 are , however , fairly well constrained by @xmath59 data .
it should be also emphasized that in the unpolarized case , the comparison between nlo theory predictions with jet production data from the tevatron is extremely successful .
the same is true for first preliminary data on the @xmath31-spectrum for pions at @xmath57 from phenix @xcite .
the nlo qcd corrections to polarized jet production are available as a monte carlo code @xcite . apart from a significant reduction of the scale dependence ,
they are also mandatory for realistically matching the procedures used in experiment in order to group final - state particles into jets . for single - inclusive high-@xmath31 hadron production
the task of computing the nlo corrections has been completed only very recently @xcite .
figure [ fig : fig4 ] shows @xmath60 for single - inclusive jet production at the nlo level as a function of the jet @xmath31 .
a cut in rapidity , @xmath61 , has been applied in order to match the acceptance of star .
the asymmetries turn out to be smaller than for prompt photon production , but thanks to the much higher statistics one can again easily distinguish between different spin - dependent gluon densities .
results for single - inclusive @xmath62 production are presented in fig .
[ fig : fig5a ] . note that here the expected statistical accuracy refers to only a very moderate integrated luminosity and beam polarization as targeted for the upcoming run of rhic .
even under these assumptions a first determination of @xmath19 can be achieved .
= 3.00 in the last process which exhibits a strong sensitivity to @xmath19 is heavy flavor production . here ,
the lo gluon - gluon fusion mechanism , @xmath63 , dominates unless @xmath31 becomes rather large .
unpolarized calculations have revealed that nlo qcd corrections are mandatory for a meaningful quantitative analysis . in the polarized case
they have been computed recently in case of single - inclusive heavy quark production @xcite .
again , one observes a strongly reduced scale dependence for charm and bottom production at rhic energies .
it turns out that the major theoretical uncertainty stems from the unknown precise values for the heavy quark masses @xcite .
since the heavy quark mass already sets a large scale , one can perform calculations for small transverse momenta or even for total cross sections which , in principle , give access to the gluon density at smaller @xmath13-values than relevant for high-@xmath31 jet , hadron or prompt photon production .
heavy flavors are not observed directly at rhic but only through their decay products .
possible signatures for charm / bottom quarks at phenix are inclusive - muon or electron tags or @xmath64-coincidences .
the latter provide a much better @xmath65-separation which is an experimental problem .
in addition , lepton detection at phenix is limited to @xmath66 and @xmath67 for electrons and muons , respectively .
since heavy quark decays to leptons proceed through different channels and have multi - body kinematics , it is a non - trivial task to relate , e.g. , experimentally observed @xmath31-distributions of decay muons to the calculated @xmath31-spectrum of the produced heavy quark .
one possibility is to model the decay with the help of standard event generators like pythia @xcite by computing probabilities that a heavy quark with a certain @xmath68 is actually seen within the phenix acceptance for a given decay mode .
figure [ fig : fig5 ] shows a prediction for the charm production asymmetry @xmath60 at phenix in nlo qcd for the inclusive - electron tag .
the sensitivity to @xmath19 is less pronounced than for the processes discussed above .
it remains to be checked if heavy flavor production at rhic can be used to extend the measurement of @xmath19 towards smaller @xmath13-values .
= 2.60 in inclusive dis data only provide information on the sum of quarks and antiquarks for each flavor , i.e. , @xmath71 . at rhic
a separation of @xmath72 , @xmath73 , @xmath74 , and @xmath75 can be achieved by studying @xmath76-boson production . exploiting the parity - violating properties of @xmath76-bosons
, it is sufficient to measure a single spin asymmetry , @xmath77 , with only one of the colliding protons being longitudinally polarized .
the idea is to study @xmath77 as a function of the rapidity of the @xmath78 , @xmath79 , relative to the polarized proton @xcite . in lo
it is then easy to show @xcite that for @xmath80-production , @xmath81 , and large and positive ( negative ) @xmath79 , @xmath77 is sensitive to @xmath82 ( @xmath83 ) . similarly , @xmath84-production probes @xmath85 and @xmath86 .
the nlo qcd corrections for @xmath87 as well as the factorization scale dependence are small @xcite .
experimental complications @xcite arise , however , from the fact that neither phenix nor star are hermetic , which considerably complicates the reconstruction of @xmath79 .
therefore it is important to understand @xmath87 on the decay - lepton level . here , fully differential nlo cross sections are available as a mc code @xcite .
the anticipated sensitivity of phenix on the flavor decomposed quark and antiquark densities is illustrated in fig .
[ fig : fig6 ] .
semi - inclusive dis measurements , @xmath88 , are another probe to separate quark and antiquark densities .
hermes has recently published first preliminary results @xcite .
the accessible @xmath13-range for the @xmath69 and @xmath70 densities is comparable to that of rhic , see fig .
[ fig : fig6 ] , but at scales @xmath89 rather than @xmath90
. the combination of both measurements can provide an important test of the qcd scale evolution for polarized parton densities and of the possible relevance of higher twist contributions at low scales . having available at some point in the near future data on various different reactions
, one needs to tackle the question of how to set up a `` global qcd analysis '' for spin - dependent parton densities .
the strategy is in principle clear from the unpolarized case : an ansatz for the densities , eq .
( [ eq : pdfs ] ) , at some initial scale @xmath14 , given in terms of some functional form with a set of free parameters , is evolved , eq .
( [ eq : evol ] ) , to a scale @xmath15 relevant for a certain data point .
a @xmath91-value is assigned that represents the quality of the comparison of the theoretical calculation to the experimental point .
the parameters are varied until eventually a global minimum in @xmath91 is reached mutually for all data points .
in practice , this approach is not fully viable since the numerical evaluations of the cross sections in nlo qcd are usually time - consuming as they require several tedious integrations .
hence the computing time for a qcd fit easily becomes excessive .
= 0.35 in the unpolarized case , the wealth of dis data already provides a pretty good knowledge of the parton densities , and reasonable approximations can be made for the most time - consuming processes .
for instance , one can absorb all nlo corrections into some pre - calculated `` correction factors '' @xmath92 , and simply multiply them in each step of the fit to the lo approximation for the cross sections which can be evaluated much faster . in the polarized case ,
it is in general not at all clear whether such a strategy will work . here
, parton densities are known with _
much _ less accuracy so far .
it is therefore not possible to use @xmath92-factors reliably .
in addition , spin - dependent parton densities as well as partonic cross sections may oscillate , i.e. , have zeros , in the kinematical regions of interest such that predictions at lo and the nlo can show marked differences .
clearly , in the polarized case the goal _ must _ be to find a way of implementing efficiently , and without approximations , the _ exact _ nlo expressions for all relevant hadronic cross sections .
a very simple and straightforward method based on `` double mellin transformations '' was proposed in @xcite .
recently , its actual practicability and usefulness in a global qcd analysis has been demonstrated @xcite in a case study based on fictitious prompt-@xmath93 data .
spin observables are also an interesting tool to uncover important new physics .
one idea is to study single spin asymmetries @xmath94 for large-@xmath31 jets . in the standard model @xmath94
can be only non - zero for parity - violating interactions , i.e. , qcd - electroweak interference contributions , which are fairly small .
the existence of new parity - violating interactions could lead to sizable modifications @xcite of @xmath94 .
possible candidates are new quark - quark contact interactions , characterized by a compositeness scale @xmath95 .
rhic is surprisingly sensitive to quark substructure at the 2 tev scale , and is competitive with the tevatron despite the much lower c.m.s .
energy @xcite .
other candidates for new physics are possible new gauge bosons , e.g. , a leptophobic @xmath96 .
of course , high luminosity and precision as well as a good knowledge of polarized and unpolarized parton densities and of the standard model `` background '' are mandatory . for details , see @xcite .
with first data from rhic hopefully starting to roll in soon , we can address many open , long - standing questions in spin physics like the longitudinally polarized gluon density . with data from many different processes taken at high energies , where perturbative qcd should be at work , a first global analysis of spin - dependent parton densities
will be possible . for a long time to come rhic will provide the best source of information on polarized parton densities , certainly much improving our knowledge of the spin structure of the nucleon , and , perhaps , the next `` spin surprise '' is just round the corner . future projects like the eic @xcite , which is currently under scrutiny , would help to further deepen our understanding by probing aspects of spin physics not accessible in hadron - hadron collisions .
the structure function @xmath1 at small @xmath13 or the spin content of circularly polarized photons are just two examples .
99 for a review on experimental efforts and results in polarized deep - inelastic scattering , see : e. hughes and r. voss , _ annu .
nucl . part .
_ * 49 * , 303 ( 1999 ) .
van neerven and e.b .
zijlstra , _ nucl .
phys . _ * b417 * , 61 ( 1994 ) , ( e ) * b426 * , 245 ( 1994 ) .
r. mertig and w.l .
van neerven ,
_ z. phys . _
* c70 * , 637 ( 1996 ) ; w. vogelsang , _ phys .
rev . _ * d54 * , 2023 ( 1996 ) ; _ nucl .
phys . _ * b475 * , 47 ( 1996 ) . for a review on rhic spin ,
see : g. bunce , n. saito , j. soffer , and w. vogelsang , _ annu .
nucl . part .
sci . _ * 50 * , 525 ( 2000 ) .
bland , these proceedings . l.e .
gordon and w. vogelsang , _ phys .
* d48 * , 3136 ( 1993 ) ; a.p .
contogouris et al . , _ phys .
* b304 * , 329 ( 1993 ) ; _ phys .
rev . _ * d48 * , 4092 ( 1993 ) .
gordon , _ nucl .
phys . _ * b501 * , 197 ( 1997 ) ; l.e .
gordon and g.p .
ramsey , _ phys .
rev . _ * d59 * , 074018 ( 1999 ) ; s. chang et al .
* d58 * , 074002 ( 1998 ) .
s. frixione and w. vogelsang , _ nucl .
phys . _ * b568 * , 60 ( 2000 ) .
e. laenen et al .
_ * b438 * , 173 ( 1998 ) ; _ phys .
lett . _ * 84 * , 4296 ( 2000 ) ; s. catani et al . , _ jhep _ * 9807 * , 024 ( 1998 ) ; * 9903 * , 025 ( 1999 ) .
h. torii , talk presented at quark matter 2002 " , nantes , france , 2002 .
d. de florian et al .
phys . _ * b539 * , 455 ( 1999 ) .
d. de florian , hep - ph/0210442 .
b. jger , a. schfer , m. stratmann , and w. vogelsang , hep - ph/0211007 . i. bojak and m. stratmann , hep - ph/0112276 and work in progress
. t. sjstrand et al . , hep - ph/0108264 .
p. chiappetta and j. soffer , _ phys .
_ * b152 * , 126 ( 1985 ) ; c. bourrely and j. soffer , _ phys .
* b314 * , 132 ( 1993 ) ; _ nucl
. phys . _ * b423 * , 329 ( 1994 ) ; * b445 * , 341 ( 1995 ) . a. weber , _ nucl . phys .
_ * b403 * , 545 ( 1993 ) ; b. kamal , _ phys .
rev . _ * d57 * , 6663 ( 1998 ) ; t. gehrmann , _ nucl .
* b534 * , 21 ( 1998 ) .
p.m. nadolsky , these proceedings [ hep - ph/0210190 ] .
m. beckmann , hep - ex/0210049 ; a. miller , these proceedings .
kosower , _ nucl .
phys . _ * b520 * , 263 ( 1998 ) . m. stratmann and w. vogelsang , _ phys .
rev . _ * d64 * , 114007 ( 2001 ) .
p. taxil and j .- m .
virey , _ phys .
lett . _ * b364 * , 181 ( 1995 ) ; * b383 * , 355 ( 1996 ) ; * b441 * , 376 ( 1998 ) ; _ phys . rev . _ * d55 * , 4480 ( 1997 ) . for information concerning the eic project , see : http://www.bnl.gov/eic . | we review how rhic is expected to deepen our understanding of the spin structure of longitudinally polarized nucleons . after briefly outlining the current status of spin - dependent parton densities and pointing out open questions , we focus on theoretical calculations and predictions relevant for the rhic spin program
. estimates of the expected statistical accuracy for such measurements are presented , taking into account the acceptance of the rhic detectors .
address = inst . for theor .
physics , univ . of regensburg ,
d-93040 regensburg , germany + e - mail : [email protected] |
as discussed in mathur et al .
( these proceedings ) and in @xcite , the only way to detect low - mass supermassive black holes ( smbhs ) may be by their accretion activity .
we choose to use x - ray selection to identify candidate agns for the following reasons : first , x - ray emission is a hallmark of agns .
second , x - rays can penetrate obscuring material which may be hiding the line emitting regions .
third , there are fewer sources of x - rays in a galaxy than there are of optical and uv emission and so dilution of the agn signature by host galaxy light is less of a problem , even if the agn is moderately obscured .
fourth , even if , as expected in some theories @xcite agns that have luminosities or accretion rates below a cut - off value do not have broad - line regions , they should still be detectable in x - rays .
the disadvantage is that x - ray observations by themselves can not always distinguish between agns and other x - ray sources , such as x - ray binaries ( xrbs ) and ultraluminous x - ray sources ( ulxs ) .
multi - wavelength data are needed to determine the type of source .
as examples , we present here two late - type spiral galaxies , ngc 3184 and ngc 5457 ( m101 ) .
their nuclear optical spectra show no signs of agns @xcite . but using archival _
chandra _ data and multi - wavelength data from the literature ,
we show that the sum total of evidence strongly suggests that the galaxies host agns .
this galaxy is of type scd , classified as having an hii nucleus by @xcite , and is at a distance of 8.7 mpc @xcite .
the argument for the presence of an agn is based on the following : ( a ) _ the nucleus is an x - ray source_. observations with _
chandra _ totaling 60 ks detect a point - like source with 36 counts and a more extended component with 117 counts ( fig .
[ fig : n31n54 ] ) . assuming a power - law spectrum with slope @xmath0 and correcting for galactic absorption only , the point source has luminosity @xmath1 erg s@xmath2 .
luminosities in this range are typically associated with xrbs rather than agns , but does not rule out an agn , especially since intrinsic absorption is unknown .
( b ) _ infrared line ratios are not that of an hii nucleus but instead suggest the presence of an agn component_. a stronger argument that the source is an agn derives from infrared data .
this galaxy is part of the _ spitzer _ infrared nearby galaxies survey @xcite .
@xcite have used the equivalent width of the pah feature at @xmath3 m and the fluxes in a mix of high- and low - ionization ir lines of s , ne , and o to create diagnostic diagrams that distinguish between agn and star - forming galaxies .
this nucleus falls into the `` transition '' region between agns and hii regions .
this suggests that the emission has an agn component that was diluted because of the large aperture used ( @xmath4 ) to extract the fluxes .
( c ) _ mid - infrared colors are redder than those of normal galaxies_. the observed irac colors , @xmath5-[4.5 ] = + 0.20 $ ] @xmath6 and @xmath7-[8.0 ] = + 1.24 $ ] @xmath6 ( magnitudes in vega system ) ( d. a. dale , priv .
comm . ) , are redder than more than 80% of normal late - type galaxies ( fig . [
fig : n3184 ] ; also see @xcite ) .
this is expected if there is an agn , as agn power - law emission falls off more slowly than galactic emission in the mir .
thus , the ir line ratios and irac colors strongly argue that the source is an agn . on a side and
have north up and east to the left . _
ngc 3184 _ : the nucleus consists of a point - like source and an extended component to the south of it .
the bar on the lower left corresponds to a distance of @xmath8 pc . _
ngc 5457 _ : the nucleus is resolved into two sources , marked `` n '' and `` s '' in the figure .
source n is the active nucleus and source s is a star cluster .
the bar on the lower left corresponds to a distance of @xmath9 pc . ]
galaxies , separated by factors of 1.8 in galaxy density .
c4 stand for the 3.6 @xmath10 m , 4.5 @xmath10 m , 5.8 @xmath10 m , 8.0 @xmath10 m bands respectively .
the left- and right - hand concentrations are spiral and elliptical galaxies respectively .
a typical error bar is shown in the lower right .
( r. j. assef 2007 , priv . comm . also see @xcite ) .
_ _ reprinted with permission of the author . _ _ ]
ngc 5457 is a galaxy of type scd at a distance of approximately 7 mpc @xcite .
the nucleus was classified as hii by @xcite .
ngc 5457 was observed multiple times by _ chandra _ for a total observing time of approximately 1 ms ( fig .
[ fig : n31n54 ] ) . for this galaxy
the x - ray data by themselves provide evidence suggesting the presence of an agn .
( a ) _ the nucleus is an x - ray source_. ( b ) even more importantly , _ the nuclear source is variable_. the nucleus varies by about a factor of 10 over nine months ( fig . [ fig : n5457 ] ) , with luminosity ranging from @xmath11 to @xmath12 erg s@xmath2 .
the variability rules out multiple xrbs as the origin of the x - ray emission , as they would have to be varying in phase .
it is possible for a single xrb to produce the observed x - ray emission , but xrbs with luminosities this high are uncommon .
( c ) _ the x - ray spectrum _ shows an absorbed power - law with slope @xmath13 , similar to what is seen in agns .
the best - fit intrinsic absorption is consistent with zero when the source is faint .
( d ) _ the ratios of 0.32 kev , 25 kev , and 58 kev counts _ for this nucleus are similar to those seen in compton - thick agns @xcite .
points ( c ) and ( d ) together suggest a highly obscured agn where , like in the seyfert galaxy ngc 1068 , the dominant observed emission is radiation that has been scattered into our line of sight . finally , ( e ) as can be seen in fig . [
fig : n5457 ] , _ the behavior of the nucleus is markedly different from that of a known star cluster _ of comparable x - ray luminosity which is @xmath14 pc away from the nucleus .
this suggests the nuclear source is not simply another star cluster .
there are compelling , though not conclusive , arguments that ngc 3184 and ngc 5457 actually host agns even though neither shows signs of agns in their optical spectra .
these agns are excellent candidates for being low - mass smbhs , as they reside in low - mass bulges .
a similar approach can be used to uncover more candidate low - mass smbhs .
we are grateful to d. a. dale for kindly providing _ spitzer _
fluxes for the nucleus of ngc 3184 prior to publication .
support for this work was provided by the national aeronautics and space administration through chandra award number go7 - 8111x issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of the national aeronautics space administration under contract nas8 - 03060 . | we demonstrate the feasibility of uncovering supermassive black holes in late - type , `` quiescent '' spiral galaxies by detecting signs of very low - level nuclear activity .
we use a combination of x - ray selection and multi - wavelength follow - up . here
, we apply this technique to ngc 3184 and ngc 5457 , both of type scd , and show that strong arguments can be made that both host agns .
address = department of astronomy , the ohio state university , 140 w 18th ave , columbus , oh 43210 , usa address = department of astronomy , the ohio state university , 140 w 18th ave , columbus , oh 43210 , usa address = inaf - osservatorio astronomico di roma , via frascati 33 , 00040 monteporzio catone ( roma ) , italy address = herzberg institute of astrophysics , 5071 west saanich road , victoria , bc v8x 4m6 , canada |
the author thanks to bumseok kyae and fuminobu takahashi for useful discussions .
this work is supported by the national research foundation of korea ( nrf ) grant funded by the korea government ( msip ) ( nrf-2015r1d1a3a01019746 ) .
b. batell , g. f. giudice and m. mccullough , `` natural heavy supersymmetry , '' jhep * 1512 * , 162 ( 2015 ) [ arxiv:1509.00834 [ hep - ph ] ] .
k. choi and s. h. i m , `` realizing the relaxion from multiple axions and its uv completion with high scale supersymmetry , '' jhep * 1601 * , 149 ( 2016 ) [ arxiv:1511.00132 [ hep - ph ] ] .
j. l. evans , t. gherghetta , n. nagata and z. thomas , `` naturalizing supersymmetry with a two - field relaxion mechanism , '' jhep * 1609 * , 150 ( 2016 ) [ arxiv:1602.04812 [ hep - ph ] ] .
n. arkani - hamed , s. dimopoulos and s. kachru , `` predictive landscapes and new physics at a tev , '' hep - th/0501082 .
l. j. hall and y. nomura , `` spread supersymmetry , '' jhep * 1201 * , 082 ( 2012 ) [ arxiv:1111.4519 [ hep - ph ] ] .
k. cheung , c. w. chiang and j. song , `` a minimal supersymmetric scenario with only mu at the weak scale , '' jhep * 0604 * , 047 ( 2006 ) [ hep - ph/0512192 ] .
v. a. beylin , v. i. kuksa , g. m. vereshkov and r. s. pasechnik , `` neutralino - nucleon interaction in the split higgsino scenario of the dark matter , '' int .
j. mod .
a * 24 * , 6051 ( 2009 ) [ arxiv:0903.4201 [ hep - ph ] ] .
p. j. fox , g. d. kribs and a. martin , `` split dirac supersymmetry : an ultraviolet completion of higgsino dark matter , '' phys .
d * 90 * , no . 7 , 075006 ( 2014 ) [ arxiv:1405.3692 [ hep - ph ] ] .
n. nagata and s. shirai , `` higgsino dark matter in high - scale supersymmetry , '' jhep * 1501 * , 029 ( 2015 ) [ arxiv:1410.4549 [ hep - ph ] ] .
e. j. chun , s. jung and j. c. park , `` very degenerate higgsino dark matter , '' jhep * 1701 * , 009 ( 2017 ) [ arxiv:1607.04288 [ hep - ph ] ] .
j. hisano , d. kobayashi and n. nagata , `` enhancement of proton decay rates in supersymmetric su(5 ) grand unified models , '' phys .
b * 716 * , 406 ( 2012 ) [ arxiv:1204.6274 [ hep - ph ] ] .
j. hisano , t. kuwahara and y. omura , `` threshold corrections to baryon number violating operators in supersymmetric @xmath81 guts , '' nucl .
b * 898 * , 1 ( 2015 ) erratum : [ nucl .
b * 907 * , 476 ( 2016 ) ] [ arxiv:1503.08561 [ hep - ph ] ] .
h. murayama and a. pierce , `` not even decoupling can save minimal supersymmetric su(5 ) , '' phys .
d * 65 * , 055009 ( 2002 ) [ hep - ph/0108104 ] .
j. hisano , d. kobayashi , t. kuwahara and n. nagata , `` decoupling can revive minimal supersymmetric su(5 ) , '' jhep * 1307 * , 038 ( 2013 ) [ arxiv:1304.3651 [ hep - ph ] ] . | we explore gauge coupling unification and dark matter in high scale supersymmetry where the scale of supersymmetry breaking is much above the weak scale .
the gauge couplings unify as precisely as in low energy supersymmetry if the higgsinos , whose mass does not break supersymmetry , are much lighter than those obtaining masses from supersymmetry breaking .
the dark matter of the universe can then be explained by the neutral higgsino or the gravitino .
high scale supersymmetry with light higgsinos requires a large higgs mixing parameter for electroweak symmetry breaking to take place .
it is thus naturally realized in models where superparticle masses are generated at loop level while the higgs mixing parameter is induced at tree level , like in anomaly and gauge mediation of supersymmetry breaking .
supersymmetry ( susy ) is a theoretically well - motivated framework for extending the standard model ( sm ) , and can be realized in various ways . taken as a theoretical guiding principle to physics beyond the sm
, naturalness suggests low energy susy spontaneously broken at a tev scale or below , for which the minimal supersymmetric sm ( mssm ) becomes compatible with gauge unification @xcite and provides a viable candidate for cold dark matter @xcite .
however , the observation of the sm - like higgs boson with mass near 125 gev and non - observation of new physics signals at the lhc seem to indicate that the scale of susy breaking is around multi tev or higher
. one may then need a different approach to the naturalness problem , like for instance the dynamical relaxation of the weak scale @xcite .
if realized at a high energy scale , susy would play only a partial role in stabilizing the weak scale against quantum corrections .
receives two or more contributions whose relative phase is determined by the relaxion , generating the weak scale via cosmological evolution .
this mechanism would work also in high scale susy with light higgsinos because @xmath0 , on which the determinant of the higgs mass matrix depends , still varies along the relaxion direction . here
@xmath1 denotes the higgs mixing parameter . ]
nonetheless it still remains attractive , in particular from the viewpoint of grand unification and dark matter .
in this letter we examine if susy broken much above the weak scale can account for the dark matter of the universe , and retain the unification of gauge couplings in the absence of high scale threshold corrections so that extension towards a grand unified theory ( gut ) is possible .
interestingly , gauge couplings unify as precisely as in low energy susy if the higgsinos , whose mass is not directly connected to susy breaking , are much lighter than other sparticles and the heavy higgs doublet @xcite .
that is , if || m_susy , where @xmath2 is the higgsino mass parameter , and @xmath3 denotes the scale of susy breaking .
the lightest sparticle is then the neutral higgsino or the gravitino , which can make up all the dark matter if produced non - thermally or from thermal scatterings of sparticles .
high scale susy with light higgsinos can have interesting phenomenology @xcite , but needs to address how to achieve electroweak symmetry breaking ( ewsb ) that requires a higgs mixing parameter @xmath1 much larger than @xmath3 .
we stress that a natural framework for large @xmath1 is provided by models where sparticle masses are generated at loop level while @xmath1 is induced at tree level , like in anomaly @xcite or gauge mediation @xcite .
we examine the relation between @xmath3 and @xmath2 required for ewsb and gauge unification in the scenario where anomaly or gauge mediation is sizable , and estimate the value and scale at which the gauge couplings unify .
how light can the higgsinos be compared to other sparticles ?
the most important theoretical constraint on @xmath2 comes from ewsb . in the mssm scalar potential ,
the squared - mass terms of the neutral higgs fields read & & ( m_h_u^2+||^2)|h_u^0|^2 + ( m_h_d^2+||^2 ) |h_d^0|^2 + & & - ( bh_u^0h_d^0 + h.c . ) , including susy breaking terms associated with the higgs doublet fields .
thus ewsb occurs if |b|^2 > ( m^2_h_u + ||^2 ) ( m^2_h_d + ||^2 ) , where the involved parameters should satisfy 2 |b| & < & m^2_h_u + m^2_h_d + 2 ||^2 , for the scalar potential to be stable along the @xmath4-flat direction , @xmath5 .
it is obvious that the above condition can not be satisfied if @xmath2 is much larger than @xmath3 as inferred from that it gives a supersymmetric mass to the higgs bosons . for the conventional scenario with [ ewsb-1 ]
|| ~|b| ~ ~m_susy , ewsb is triggered when renormalization group ( rg ) flow drives @xmath6 to negative or small values at low energy , which is possible as it is considerably affected by loop effects associated with the large top yukawa coupling .
it is remarkable that ewsb is possible also for @xmath7 if the higgs mixing parameter @xmath1 has a large value around @xmath8 . to be in more detail , the minimization condition 2&= & , + m^2_z & = & - ||^2 +
, shows that ewsb can be achieved when the higgsinos are relatively light and the higgs sector parameters have the hierarchy @xcite , [ higgs - hierarchy ] || m_susy |b| , for moderate and large @xmath9 , while @xmath6 is negative or much smaller than @xmath10 in size as in the conventional scenario
. here @xmath11 , and @xmath12 is the @xmath13-boson mass .
the hierarchy between the higgs mixing parameter and sparticle masses is naturally realized in a class of susy breaking models where sparticles obtain masses at loop level , that is , for instance , models in which anomaly or gauge mediation is sizable .
a large @xmath1 in such models has been regarded as problematic in low energy susy because it makes ewsb difficult unless @xmath2 is below 100 gev , which is in conflict with the lep bound on the chargino mass .
the situation however changes in high scale susy . in supergravity
, anomaly mediation always induces sparticle masses radiatively , but the higgs mixing parameter associated with @xmath2 arises at tree - level : m^2_h_d|_am & ~ & , + b|_am & ~ & m_3/2 , where @xmath14 is the gravitino mass . hence @xmath15 is obtained in models where anomaly mediation is sizable to other mediations in size . in such a case ,
the gravitino is quite heavy as would be required to make the gravitino decay before big bang nucleosynthesis ( bbn ) .
note also that the higgsino is likely the lightest supersymmetric particle ( lsp ) because ewsb requires @xmath7 .
another possible way to get a large higgs mixing parameter is to consider models with sizable gauge mediation where @xmath2 is dynamically generated from the superpotential term [ gm - mu ] f(x ) h_u h_d , for some function @xmath16 . here
@xmath17 is the susy breaking field that provides masses to messengers .
one may instead consider @xmath18 in the k@xmath19hler potential .
the sparticles then obtain masses at loop level , while the higgs mixing parameter is generated at tree level : m^2_h_d|_gm & ~ & , + b|_gm & ~ & f / m , for @xmath20 .
on the other hand , the vanishing cosmological constant puts a lower bound on the gravitino mass m_3/2 f / m , with @xmath21 being the reduced planck mass .
the gravitino can be the lsp if the messenger scale @xmath22 is intermediate or low .
to summarize , there naturally arises a hierarchy between the higgs mixing parameter and the sparticle masses : |b| 8 ^ 2 m_susy . in models where anomaly or gauge mediation gives sizable contributions to sparticle masses . for large @xmath1
, we need @xmath7 to trigger ewsb . in this class of models , therefore , the lightest ordinary supersymmetric particle would be a nearly pure higgsino , while the gravitino can be very heavy or light .
let us now turn to the issue of gauge coupling unification .
the mssm leads to quantitative unification of gauge couplings for low scale susy where the sparticles are around tev . the unified gauge coupling and unification scale are approximately given by [ gut - tev ] & & 2 , + m_gut|_tev & & 210 ^ 16 gev .
let us consider a simple case in which all the sparticles and the heavy higgs doublet have a common mass .
then , there exists a value of the common sparticle mass m_at which the three gauge couplings meet exactly at one point within the mssm . the value of @xmath23 is estimated to be m_1tev , where the dependence on @xmath9 , which arises at the two - loop level , is very mild . to see how gauge coupling unification is affected by sparticle masses , we use one - loop rg evolution of gauge couplings in the dimensional reduction scheme under the assumption , for simplicity , that the sleptons ( squarks ) have a universal mass @xmath24 ( @xmath25 ) . from the fact that gauge unification works in low scale susy
, the unified gauge coupling is found to be & = & + ( ) + & & + _ ( ) , for @xmath26 and summing over the sparticles and heavy higgs doublet .
here @xmath27 denotes the @xmath28 function coefficient in the mssm , and @xmath29 is the contribution of the particle @xmath30 to it . using the above relations , one finds that successful unification is maintained for a sparticle spectrum satisfying [ unification - condition ] ( ) ^ ( ) ^ ( ) ^ ( ) ^ = 1 , with @xmath31 tev @xcite . here
@xmath32 and @xmath33 are the wino and gluino mass , respectively , and @xmath34 is the heavy higgs doublet mass .
the unification condition is sensitive to the masses of the higgsino , wino and gluino .
but it is relatively insensitive to the masses of the heavy higgs doublet and sfermions . in particular ,
gauge unification becomes independent of the sfermion spectrum if the squarks and sleptons have a common mass , which reflects the fact that they form complete su@xmath35 multiplets .
this implies that unification works even when the sfermions are very heavy while other sparticles are around @xmath23 as in the split susy @xcite .
it is also worth noting that the inclusion of two - loop corrections leads to small shift of the scale @xmath23 , and does not change the qualitative features . under the condition eq .
( [ unification - condition ] ) , the three gauge couplings converge to a common value [ gut - value - scale1 ] & & = + & & + + & & + , at a gut scale [ gut - value - scale2 ] m_gut & = & m_gut|_tev ( ) ^ ( ) ^ + & & ( ) ^ ( ) ^ . for higgsinos
much lighter than other sparticles , the gut scale and the value of unified gauge coupling will get smaller .
if the mssm is embedded in a gut , operators mediated by heavy gauge bosons around @xmath36 would induce proton decay mainly via @xmath37 @xcite .
the experimental limits on proton lifetime are evaded for ( ) ^-1 ( ) ^2 0.1 , which is combined with the relations , eqs .
( [ gut - value - scale1 ] ) and ( [ gut - value - scale2 ] ) , to put an upper bound on the ratio between @xmath3 and @xmath2 : 0.610 ^ 4 ( ) ^- , for a simplified case where the wino , sfermions and heavy higgs doublet have a similar mass about @xmath3 .
the constraint is satisfied in most of the parameter region of our interest . the gauge coupling at @xmath36 increases if there are gauge messengers , enhancing the proton decay rate @xcite . as a consequence ,
the upper bound on @xmath38 presented above is reduced , for instance , by a factor of about 0.6 ( 0.2 ) for @xmath39 ( 3 ) and @xmath40 gev . here
@xmath41 is the number of gauge messengers in @xmath42 of su@xmath35 , and @xmath43 is their mass .
note that there are also model - dependent dimension - five operators leading to proton decay mainly via @xmath44 , which are induced by the exchange of colored higgs multiplets @xcite .
the lifetime is approximately proportional to @xmath45 for sfermions around @xmath3 , and it follows that high scale susy with @xmath3 above 10 tev and low @xmath9 is favored to evade the experimental bound @xcite . for gauginos , sfermions and the heavy higgs doublet around @xmath3 , the unification condition eq .
( [ unification - condition ] ) is reduced roughly to m_susy ( ) ^ ( )
^4 m _ , where we have kept the dependence on the ratio between the wino and gluino mass as it can be important .
the above shows that the sparticle masses @xmath3 can be larger than @xmath23 , i.e. above tev , if the higgsinos are much lighter than other sparticles .
the value of @xmath3 consistent with unification is further pushed up if the gluino is heavier than the wino @xcite .
let us examine if gauge unification works for the higgs sector with the hierarchy of eq .
( [ higgs - hierarchy ] ) . combined with gauge unification ,
the low energy gaugino masses have a distinctive and robust pattern @xcite in many mediation schemes including anomaly , gauge , gravity mediation and mixed mediations , .|_m_susy 1 + b_a , at the one - loop level , whereas scalars have a quite model - dependent mass pattern . here
@xmath46 represents the relative importance of anomaly - mediated contributions , and it is positive in many mixed mediation models @xcite .
negative @xmath46 is also possible but in rather involved models @xcite . from the gaugino mass pattern ,
it follows = , where one would need || ( 1 ) , because pure anomaly mediation suffers from the tachyonic slepton problem . the gauge coupling ratio @xmath47 is equal to 3 around tev , and it decreases to 1 as energy scale increases . for other soft susy breaking parameters , one obtains = , for @xmath48 lying in the range 0 < 8 ^ 2 .
the parameter @xmath48 is much larger than order unity if anomaly and/or gauge mediation provide sizable masses to sparticles while generating @xmath1 at tree level .
finally , from the minimization condition , the value of @xmath2 appropriate for ewsb reads [ ewsb - mu ] || , and it should not exceed much @xmath3 . as noticed above , the value of @xmath3 consistent with gauge unification becomes higher than tev for @xmath7 , and is further pushed up if the gluino is heavy relative to the wino .
l we now perform a simple numerical discussion of the sparticle spectrum required for ewsb and gauge unification .
the ewsb relation eq .
( [ ewsb - mu ] ) is combined with the unification condition eq .
( [ unification - condition ] ) to uniquely fix the values of @xmath2 and @xmath3 : [ msusy - mu ] & & 0.1 10 ^ 3 ( ) ^ ( ) ^ , + & & 0.4 ( ) ^ ( ) ^- , for @xmath49 , and @xmath46 of order unity .
here we have taken into account that low @xmath9 is favored to accommodate the 125-gev higgs boson in high scale susy , for instance , @xmath9 smaller than @xmath50 for @xmath3 above a few tens tev @xcite , and that one has @xmath51 around 100 tev .
we emphasize that high scale susy can be reconciled with gauge unification when the higgsinos are much lighter than other sparticles .
for instance , ewsb and unification are achieved for @xmath2 around a few hundred gev and @xmath52@xmath53 tev .
if @xmath54 gets larger , the required value of @xmath2 and @xmath3 increase by the same factor .
fig .
[ fig : unification ] shows rg flow of gauge couplings in the mssm .
high scale susy with light higgsinos can lead to gauge coupling unification , where the three gauge coupling unify as precisely as in the conventional tev susy . in fig .
[ fig : region ] , the shaded region is compatible with gauge unification .
here we have used the fact that the susy particle mass @xmath3 and the higgsino mass parameter @xmath2 are fixed according to eq .
( [ msusy - mu ] ) once the ewsb and unification conditions are imposed .
l finally we discuss dark matter and collider signs in high scale susy under consideration .
the lsp , which is stable under @xmath55-parity conservation , is the higgsino or the gravitino . for @xmath7 ,
the lightest neutralino and chargino are mostly pure higgsino , and are nearly mass degenerate : m m_^+_1 - m_^0_1 = m_tree + m_loop , where the tree - level contribution is due to mixing with the bino and wino , and is positive unless the bino and wino mass have a different sign |m_tree| 30mev ( ) | 1 + 0.3 | , while the radiative mass difference comes mainly from gauge boson loops @xcite , and is approximated by m_loop 260mev ( ) ^0.15 , for @xmath2 below about tev .
hence , the mass difference is expected to be positive and larger than the pion mass , for which the lightest chargino dominantly decays to the lightest neutralino and the charged pion .
the decay time of the lightest chargino is about @xmath56 for @xmath57 not close to the pion mass
. it would thus be difficult to probe at the lhc , but @xmath58 or @xmath59 mediated by virtual @xmath13 boson may provide a visible signal in future linear colliders .
we first examine the gravitino lsp case , @xmath60 . in this case , the gravitino production from thermal scatterings can generate the right dark matter density . if the freeze - out temperature of the gravitino @xcite t_f ( ) ^2 ( ) ^2 is higher than the reheating temperature @xmath61 after inflation , the gravitino relic abundance is determined by _ 3/2 h^2 & & 0.3 ( ) + & & ( ) ( ) ^2 , where the gravitino should be heavier than about @xmath62 kev to be a cold dark matter @xcite . the reheating temperature producing
the observed dark matter density can be higher than about @xmath63 gev as required for standard thermal leptogenesis @xcite .
one should however note that the next to lightest supersymmetric particle ( nlsp ) , which is the neutral higgsino , decays into the gravitino and ordinary particles with a width _
~ , and so it occurs during or after the bbn epoch in the parameter region of our interest while producing high energetic electromagnetic and hadronic showers
. such late - time decay of the nlsp can significantly alter the abundances of light elements , spoiling the success of bbn . a simple way to avoid
this cosmological difficulty is to consider @xmath55-parity violation .
one possibility is to add @xmath55-parity violating terms that violate the lepton number as well but preserve the baryon number to forbid dangerous proton decay operators @xcite .
for instance , @xmath64 with @xmath65 larger than about @xmath66 can allow the nlsp to decay very shortly into ordinary particles while making the gravitino live long enough to make up the dark matter of the universe .
note that one needs @xmath67 to avoid washout of baryon asymmetry before the electroweak phase transition @xcite .
let us move on to the case of @xmath68 with @xmath55-parity conservation , for which the lsp is the neutral higgsino .
the lsp thermal relic density is significant only for @xmath2 above 1 tev . for smaller @xmath2 ,
the right dark matter density can be generated via non - thermal lsp production .
here we consider a scenario where gravitinos are abundantly produced from the decay of a heavy scalar field such as inflaton , polonyi field , or string moduli .
the gravitino decay width reads _
3/2 = , for @xmath69 , and it corresponds to the decay temperature t_3/2 0.25gev ( ) ^ ( ) ^ , where @xmath70 counts the effective number of relativistic degrees of freedom .
thus , lsps are produced below the lsp freeze - out temperature @xmath71 .
if the gravitino abundance is large enough to make annihilation among produced lsps effective , the lsp relic density becomes independent of the initial gravitino abundance @xcite : _
^0_1 h^2 & & ( ) ^ + & & ( ) ^3 ( ) ^ , where @xmath72 , and @xmath73 is similarly defined by taking @xmath74 .
note that the gravitino should be heavier than about 40 tev to decay before nucleosynthesis , and its mass is bounded from above , @xmath75 , because anomaly mediation is a model - independent source of sparticle masses .
the direct detection of the higgsino dark matter would be challenging since it has only small couplings to the sm higgs boson and @xmath13 boson suppressed by @xmath76 , making both spin - dependent and independent scatterings too small to be detected by current experiments .
however , indirect detection via the process @xmath77 or @xmath78 may be possible in future experiments .
l fig .
[ fig : dm ] shows the region where the right dark matter density @xmath79 is obtained .
if the neutral higgsino is the lsp , one can consider non - thermal production of higgsinos from heavy gravitino decays . on the other hand , in the case of a gravitino lsp ,
gravitino production by thermal scatterings can yield the right relic density .
we have studied gauge coupling unification and dark matter in high scale susy where sparticles have masses much higher than the weak scale as would be indicated by tensions between low energy susy and the lhc results .
high scale susy admits gauge coupling unification as exactly as in low energy susy if the higgsinos are much lighter than those obtaining susy breaking masses . in this scenario ,
the lsp is the neutral higgsino or the gravitino , which can make up all the dark matter of the universe if produced non - thermally in the former case and thermally from sparticle scatterings in the latter case .
a large higgs mixing parameter required for ewsb is naturally obtained in models where sparticle masses are generated at loop level while the higgs mixing parameter arises at tree level , as is the case in anomaly and gauge mediation .
then , gauge coupling unification and dark matter would indicate high scale susy around 10 tev a pev and light higgsinos below a few tev . |
a multi - element radio telescope is a spatially spread array of antennas ( or antenna elements ) whose noise - like responses are required to be time aligned , dynamically calibrated and combined or correlated in real time .
the resulting estimates of spatio - temporal and spectral correlations between responses of pairs of elements can be used to recover the desired information on the strength and distribution of radio emission within the common field of view @xcite using standard post - processing software .
thus the signal of interest is statistical in nature , resulting from a minute level of mutual coherence arising from weak celestial signals buried in noise . because of the large number of elements and the high sampling rates necessary for bandwidths exceeding several tens of mhz in recent arrays , real - time statistical estimation is essential to achieve practical data rates and volumes for recording and post processing .
for instance , the recently initiated upgrade for the ooty radio telescope ( ort ) aims at treating the 30 m x 506 m parabolic cylindrical antenna of the ort as 264 independent sets of elements , each of which is required to be sampled at 80 ms / s , leading to a data generation rate of 21 gigasamples per second , exceeding 80 terabytes per hour . till recently
, computing requirements of this scale forced a choice of custom hardware to be the most favored platform .
however , rapid developments in the fields of digital technology , communication and computing have led to a changing trend towards alternative approaches for upcoming telescopes .
such approaches range between a customized and reusable hardware library of components on an fpga platform , e.g. , the casper project @xcite , and a software - only approach , e.g. , at the gmrt @xcite .
the gmrt case is an example of a recent transition from custom hardware to a software - only approach . in this paper , we have taken a middle path , where the real - time processing of a multi - element radio telescope is abstracted as a multi - sensor , data fusion problem and addressed in a new platform called the networked signal processing system ( nsps ) in terms of packetized , heterogeneous , distributed and real - time signal processing .
it is a co - operative of two kinds of networks , among which one is a custom peer - to - peer network while the other is a part of a commodity processor network .
the custom network includes subsystems related to the digitization and all intermediate routing and preprocessing blocks as the network nodes , in which the emphasis is on traffic shaping , on - the - fly processing and load balancing for effective distributed computing .
however , all customized protocols are absorbed while crossing over the last mile to interface to the commodity processor network using a common industry - standard network protocol .
the actual estimation of the correlations is carried out by nodes on the commodity network .
in contrast to the traditional use of a packetized network as merely a data transport fabric between processing entities , we have the notion of `` a logical packet '' based on an application specific `` transaction unit '' , which itself may be composed of a large number of physical packets whose sizes are network - specific this unit refers to a time stretch long enough to facilitate a dynamical flagging mechanism or / and to relax the constraint on timing , synchronization and scheduling of workloads on the commodity operating systems on which processing is expected to be carried out . both these requirements necessitate lower level nsps nodes to be equipped with large memories , which are also used to route traffic selectively ( traffic shaping ) to higher levels in the nsps .
two independent considerations have led us towards stretching the transaction unit to a good fraction of a second .
one of these , as explained above , is to provide latency tolerance in order to simplify software on standard computing platforms , while the other arises from a desire to make explicit provision for preprocessing using concepts related to modern information theory . from this point of view and to attract the attention of experts from outside the field of radio astronomy , we have given a somewhat unconventional description of the signal path and analysis of processing requirements in sections [ sec : multielementradiotelescope ] and [ sec : realtimeprocessing ] in an attempt to illustrate the connection of the problem to information theory .
another concept in literature which we find useful in the present context is that of `` multi - sensor data fusion '' , defined by @xcite , as a system model where `` spatially and temporally indexed data provided by different sources are combined ( fused ) in order to improve the processing and interpretation of these data '' .
this model , widely used in applications like military target tracking , weather forecasting etc , has many features relevant for describing the control flow and pre - processing required in a multi - element radio telescope before correlation . in a sense ,
the nsps is an adaptation of data fusion architecture to our domain .
our analysis of the nature of the real - time problem results in a natural partitioning into two broad categories as elaborated in section [ sec : realtimeprocessing ] .
this is our primary motivation for defining the architecture , described as a fusion tree in section [ sec : fusiontree ] .
an illustration of the feasibility of its implementation is presented in section [ sec : an - implementation - example : ] by considering the case of the ort upgrade .
we present here an abstract picture of a multi - element radio telescope , in which the primary beams of individual elements are viewed as virtual communication channels carrying different combinations of radiation from a set of independent celestial `` radio emitters '' .
these are located in different directions within the common celestial region intercepted by the element primary beams .
each individual radio emitter from this set is a source of stationary ( within the observing time ) random process .
since signals received in any finite bandwidth can be spectrally decomposed , each primary beam can be considered equivalent to a set of independent communication channels of identical bandwidths corresponding to the spectral resolution .
each such channel is characterized by a correlation timescale equal to reciprocal of its bandwidth .
thus , spectral decomposition can be used to enable the propagation time differences for noise from different radio emitters to be within their correlation timescale .
this implies that the corresponding channels of different primary beams are carriers of noise arising from different superpositions of the same set of random processes , but with different weights .
information about the strength and distribution of radio emitters in the field of view can be considered to be coded in the correlation between corresponding communication channels of different elements .
hence , we consider the real - time spectral correlation to be a fusion process for compressing the information conveyed by the responses of different elements without affecting decoding to be carried out in post - processing , say in the form of an image of a celestial region
. however , in practice , the propagation medium introduces a variety of correlated and uncorrelated noise into these virtual channels which can erase or distort the information related to celestial emitters .
this is represented schematically in fig .
[ fig : antenna - beams - viewed ] . often , some of these distortions are characterized by a combination of dispersive and non - dispersive components which are _ localized _ in time and/or frequency , unlike the celestial signal which is more like a random noise . such a localization can be exploited by a pre - processing algorithm to enable suitable recognition and characterization of deviant ( non - random ) data and to tag them suitably .
such data can then be segregated from those passed on to an irreversible fusion operation , e.g. , to minimize the biases in the correlations . on the other hand , in situations like observing fast transients with low duty cycle , or
when one wants to find the direction in which such a non - random signal is present , the segregated data can be passed to an independent processing or recording stage for later use .
we refer the interested reader for an analogous situation in a re - visit of maxwell s demon by @xcite to connect algorithmic randomness to physical entropy . for the present purpose
, it suffices to note that such a segregation results effectively from an algorithmic feedback , requiring multiple passes through the data before fusion .
the output of this feedback via multiple passes on stored data is also analogous to the concept of a relay network with `` side '' or meta information , described in @xcite , where this side information is used by other entities for selective processing of the data reaching them . in our abstraction , a provision for such an algorithmic feedback should be an essential part of the signal processing architecture , and should be present in the path between the digitizer and the fusion operator .
this is only possible if the real - time system is equipped with adequate memory and preprocessing for characterization of data before they are sent to a correlator or beam forming system . _ _ by excluding such a provision , digital receivers in existing large radio telescopes are a potential source of irreversible biases in the recorded correlations , and suffer from inefficiency for requirements like detection of short term transients .
antenna beams viewed as noisy virtual communication channels connecting a sky region and an antenna element .
without getting into specific details of the real - time processing required for a large multi - element radio telescope , we abstract them into a combination of three broad categories : * _ embarrassingly parallel processing _ , e.g. , spectral decomposition of the incoming time series ( say via fft or polyphase filter banks ) and the recognition and management of path - induced distortions / interference on timescales significantly smaller than the integration time chosen for the correlations . * _ pipelined processing _ , e.g. , multi - beam formation ( say _ k _ beams ) by phasing n elements requires @xmath0 pipelined operations for each spectral band . *
_ data fusion _ operations , in which the data originating from different antenna elements are hierarchically fused ( combined via routing and application - dependent processing ) along a chosen set of dimensions which include time , frequency and spatial spread .
the most important fusion operation for an antenna array is the real - time correlation of signals from every possible pair of antenna elements in different frequency sub - bands . apart from being an @xmath1 process ( for an n element array ) from a computational point of view
, this brings in the additional complication of routing large volumes of distributed data to appropriate data processing elements to provide a complete - graph connectivity between the sources of data and processing elements .
significantly , cross - correlation between all possible pairs of signals is also essential for using self - calibration techniques to enable dynamic calibration of instrumental and atmospheric contributions to the data corresponding to different elements , before they are subjected to an irreversible fusing operation in a phased array .
this makes a spectral correlator an implicit requirement , even for a phased array , for minimizing the irreversible loss of information resulting from distortions induced in the the path or the local environment . in our approach ,
we bifurcate the requirements of a real - time system into _ commodity _ and _ custom _ segments . in the current state of technology ,
the commodity segment can be fulfilled by subsystems available in the market while the custom segment can generally be realized on the basis of customized hardware and/or firmware layers based on cots technology .
such a bifurcation is explained below for different functional categories of the nsps : * _ cots segment _ : computationally complex and/or latency tolerant processing , typically realized on a programmable platform ranging from workstations to a high performance cluster . * _ custom segment _ : latency critical , logic intensive and repetitive pattern of deterministic processing , well suited for a configurable platform , typically fpga - based . for efficient computation
, we pay special attention to reducing coupling between data in order to target explicit parallelism at all levels of processing .
multiple parallel circuits are implemented in the fpga - based custom segment , while the current trend of multicore processors with access to shared memory is exploited in high level software .
further , the desired high signal bandwidth and large number of antenna elements make the processing complex and compute intensive .
this aspect , and the advantage of quickly implementing exploratory algorithms , make a commodity compute cluster an attractive choice for the central computing .
this is recognized by explicitly including the cluster in the cots segment mentioned above . * _ cots segment _ :
commercial switches with all - to - all connectivity are used for data routing to commodity processors and broadcasting in the last mile , as well as for load balancing .
the routing is controlled by manipulating the destination addresses on data packets .
connection - less protocols like user datagram protocol ( udp ) are adequate for high - speed , streaming applications where a small fraction of lost packets does not affect performance adversely .
packet collisions are minimized in full duplex , point - to - point connections between network partners , and also because data flow is extremely asymmetric .
further , the criteria for load partitioning discussed in section [ sub : load - partitioning - and ] very often result in under - utilization of link speeds to match them to sustainable processing bandwidths . * _ custom segment _ : customized switches with static routes for traffic shaping are relevant when only a subset of the network data needs to flow to a subset of the nodes based on certain conditions . they are generally implemented in configurable logic . since many fpgas support gigabit ethernet mac as a hard ( or publicly accessible soft ) ip , this feature is useful while introducing a bridge between the peer - to - peer network and the commodity network in the last stage of the custom segment .
in addition , some or all the major subsystems may have management support from an embedded or explicit on - board processor . * _ cots segment _ : a commodity network compatible with a typical high performance compute cluster , which includes gigabit ethernet as a _ de facto _ standard for interfacing with external systems . * _ custom segment _ : a peer - to - peer network which may include significant on - the - fly application specific operations , suitable for implementing on a standard fpga platform . a data pooler node of the nsps shown carrying out data fusion and traffic shaping .
_ ] an implementation of the actual processing on a dedicated set of identical hardware circuits or parallel processors can take advantage of an intelligent network capable of elementary on - the - fly operations to achieve a balance on the dataflow and computing requirements .
this is depicted in fig .
[ fig : a - data - pooler , ] , where a `` data pooler '' node , as defined in section [ sec : fusiontree ] , is shown carrying out real - time fusion of the streaming data by using the side information made available by external sources . at the same time , the pooler is seen generating side information out of the fused data set by way of multiple processing passes on the stored data .
the pooler can then segregate the data , and route the segregated components to different data sinks using the routing information available with the peer - to - peer link nodes .
we use `` traffic shaping '' here in a more general sense than in internet traffic shaping ( which delays lower priority packets in favour of better network performance of higher priority packets ) to refer to both segregation of the incoming stream , as well as the specific routing of segregated data to different sinks .
further , the efficiency of hierarchical computation can be significantly improved by accommodating some degree of pre - processing and/or partitioning of data in each level of the custom segment to facilitate the next level .
* _ cots segment : _ the overall task supervision , command and monitor , user interface and the dynamic system monitoring are tasks whose complexity is best left to the commodity segment to handle , where a variety of tools ranging from mpi , compiler resources and advanced operating systems like linux or vxworks are available . * _ custom segment _ : event driven scheduling with periodic or quasi - periodic events generated conveniently in a low latency logic implementation suitable for an fpga .
the interval between the events is stretched to handle an application - specific transaction unit to the extent permissible within the available resources .
in this section , we present the nsps architecture as a _ multi - sensor data fusion tree , _ in which both conventional and `` virtual '' sensors play a role . while entities like antenna elements , round - trip phase / delay monitors , noise calibration etc can be treated as `` conventional '' sensors , `` virtual '' sensors result from processing blocks at various levels .
for instance , pre - processing can result in a flagging mechanism to improve the reliability of fusion systems like correlators , in which the original data are erased while compressing their information into a statistical estimate to be passed to the next level .
the signal processing system proposed in this paper is a set of spatially separated nodes of varying communication and processing capabilities , which are interconnected by a customized high speed tree - like packet switched network interfaced to master commodity nodes .
this is equivalent to a _ data fusion tree _
, with the nodes of the tree performing operations like traffic shaping , packet routing , or pre - processing before data fusion .
accordingly , we have described the overall architecture of nsps in the form of a _ fusion tree _ ,
schematically represented in fig .
[ fig : a - conceptual - layout ] .
however , each level provides a different mixture of functional capabilities .
this has resulted from our recognition of the following features of the functional requirements : * distributed computing across many nodes of different processing capabilities , with local parameters guided by a processor capable of seeing a subset of all data . * data routing nodes with configurable routes for routing preprocessed data subsets . * nodes with large memory buffers to enable multiple passes on data , also for enabling memory based data transposition . * a high speed network interconnecting all nodes , with ability of cots nodes to tap into the network .
* interface to a master commodity node through a standard interface like gigabit ethernet ( gige ) .
thus , from a functional point of view , we classify the nodes in nsps into the following three categories : 1 .
data poolers / fusers : nodes with sufficient on - board memory to allow packetizing and multiple processing passes on incoming data .
these break the need for many - to - many connectivity in the correlation process by transposing data via memory based switches .
the transposition , based on different parameters , ultimately serves a packet of data suitable for processing by a single element of a distributed system in an embarrassingly parallel manner .
data routers : these elements are endowed with high speed links to either peers or more powerful processors to whom incoming packets are routed based on statically configured routes .
these form an integral part of our architecture , helping in the load balancing by directing appropriate subsets of preprocessed data to different processing elements .
we use commodity switches for routing data to multiple external sinks by forming many - to - many connections between data sources and data processors .
data processors : these elements have high compute density and can be used for preprocessing , as well as for data rate reduction .
we classify processors into two groups as mentioned earlier : 1 .
those catering to computationally complex and/or latency tolerant processing , an example of which is the estimation of system calibration parameters based on a long ( few minutes ) history of data , and its dynamic updation .
this processing is generally carried out by sending a subset of the data to a central processor .
2 . latency critical , logic intensive and repetitive pattern of deterministic processing .
an example of this class is the real - time block - level data encoding process requiring the estimation of block level statistics .
this can be carried out by multiple passes on small segments of data .
thus , we visualize the nsps as a _ restricted _ distributed system , depending primarily on stripped down lightweight networking protocols and the static routes set up during system configuration .
data routers , both customized and commodity , play an important role in reorganizing data to be computationally palatable to processing nodes in this scheme .
a master node is in charge of command , configuration and control , and is almost always a commodity node like a pc .
it may be noted that custom processing is spread across the nsps tree by explicitly advocating local intelligence in every node . as an illustration of the inherent facilitation of distributed processing in the nsps ,
some important aspects of the interconnection mechanism are elaborated in the following subsections .
interconnects in the data fusion tree consist of the following four essential graphs , which can either be logical or explicitly physical manifestations . 1 . the _ data _ network is a simplex , high bandwidth net connecting the leaf nodes of the fusion tree to a central processor , possibly passing through several collation levels of the signal processing tree .
the _ control and monitor _ network is a full duplex , low bandwidth network and interconnects all nodes hierarchically through management processors to a central monitoring station .
the _ calib _ network is a full duplex , low bandwidth network .
this allows calibration information to reach the data fusing nodes before the sequence of irreversible fusion operations take place .
the _ clock _ network is a full duplex network , providing the distribution of clocking signals to the various nodes , as well as allowing a round trip clock phase measurement . in actual implementations , it may be simpler to realize these in terms of a set of simplex networks , among which clock , control , monitor and calib are directed towards the leaf nodes while the data and status ( including response to monitor queries ) belong to simplex networks which flow from different levels of the fusion tree into the master node .
our network implementation can be bifurcated into the following sets : 1 .
customized high speed serial peer - to - peer links terminating into peer - to - peer switches which implement a subset of the complete graph connectivity .
commodity high speed serial links terminating into commodity networking equipment , with ability to interface to standard processing nodes . in typical implementations
, we expect custom links to be on a passive optical network ( pon ) based communication stack .
gigabit ethernet is the preferred choice for the backbone of commodity segment to connect to the custom network .
we exploit the high speed serializing ability of modern fpgas to dispatch data on high bandwidth copper links and use pon components to meet the spatial spread required to reach remote nodes over fibre links . as long as the bandwidth requirements are met
, no specific preference is implied for a choice among different networking technologies .
thus , some implementations may utilize the embedded multi - gigabit serializers in fpgas for peer - to - peer links while others may refer the cross dispersion of data to a compute cluster s high bandwidth infiniband network , or use commodity gigabit ethernet .
this leads to the need for a flexible bridging mechanism which can be exploited by implementations .
for instance , data can be conveniently transmitted over a peer - to - peer or a commodity link due to the maintained commonality of their interfaces . the high speed network internal to the nsps tree has features which are restricted and stripped down versions of those found in commodity networks .
this is an optimization due to the highly controlled network which exists within the telescope receiver environment .
our network differs from regular networks in several aspects : * a controller node is assigned for every sub - tree at a given level . this entity forwards command and status information between controller ( level-0 ) and upstream levels .
thus , it is not a typical peer - to - peer network which does not have such a hierarchical control structure .
broadcast and multicast domains available in commodity networks are used to implement control and monitor mechanisms , while explicit `` pull '' mechanisms are implemented on the custom network . here , the `` pull '' refers to the explicit request for data made by a downstream node to an upstream node .
the advantage of a `` pull '' mechanism is that data is made available to a downstream node only when it is ready to handle the data , as inferred from the downstream node s request .
also , if a downstream node is busy , then the upstream node loses data in integral packets , thus maintaining timing information .
* the network configuration and routes are fixed statically in an application dependent manner at configure time .
there is no node discovery , and data routing does not have an explicit mechanism for handling node failure .
since all data flows towards a logical sink , there is also no destination address , although source addresses can be preserved .
this simplifies network management to a large extent at the cost of non - redundancy of nsps entities . *
communication protocols : all elements in our network , including master nodes , generate similar kinds of packets which contain an 8 byte header with fixed fields .
these are typically indicative of the nature of accompanying data , as well as its timestamp , source and other meta information .
system state can also be propagated through these packets , or by forming special status packets .
the restricted meta information processing makes it simpler to realize packet formation in hardware with simple state machines . * high speed
serial interconnects : all entities in the internal network communicate via high speed serial interconnects with a clock recoverable from the encoded data .
this approach allows us to transmit data long - haul over fibre , or short - haul over copper without any changes .
in particular , we discourage bus - based interconnection between physically separated nodes .
once preprocessed data is ready within the nsps , it needs to be transferred onto a commodity network for reaching commodity nodes for post processing or archiving .
local intelligence in the peer network can be used to partition the data such that the interfaces to commodity nodes use link speeds commensurate with their processing ability . for simplicity ,
we have used gigabit ethernet as a typical standard external interface .
this is a popular high speed serial interconnect with a vast amount of infrastructure available in the commodity market .
it also allows transmission over copper ( utp ) to interface directly with commodity servers , or fibre ( via conversion to 1000basex ) for long - haul transmission .
commodity servers of moderate ability can then be used as data sinks with minimal customization .
this is also motivated by the fact that many modern fpgas have embedded high speed serial interconnects on chip , with complete gigabit ethernet support in the form of on - chip gigabit macs or as publicly available libraries . for the last mile connectivity ,
udp can be used since it is a simple connectionless protocol with minimal overheads on top of ip .
it is also possible to fill the relevant udp fields during system configuration , and hold them static for the duration of an observation .
each of the internal network types carrying data ( _ data , calib _ and _ control _ ) can then be easily made available over a different udp port as part of the design .
this allows an application program to associate independent threads to service these streams .
the distributed nature of our architecture requires status monitoring of all nodes and links , which can be handled by the individual sub - tree roots and communicated to the master controller .
this is implemented by a status `` pull '' scheme by which controlling entities periodically query the status of all nodes in the nsps tree rooted with them by way of a special aya ( `` areyoualive '' ) packet .
the nodes respond with an iaa ( `` iamalive '' ) packet containing selected status information .
similarly , control packets contain command and configuration data .
each command packet typically results in a status reply from the targeted entity , which confirms the receipt of the command , and regenerates a control packet for the entities controlled by it .
the master node can use this in an appropriately scheduled housekeeping operation to discover failure of nodes .
this network is meant to carry data from the nsps tree which is relevant to forming calibration solutions for the array .
the calibration mechanism is to be applied differently for the two main modes of observation with the nsps : * in the interferometric mode , the correlator can work independently of the actual gains and phases of the sensor elements , since the observations include calibration scans at reasonable time intervals .
off - line processing can infer intermediate variations by supplementing interpolation between calibration scans with dynamic calibration schemes like self - calibration based on the partial , low bandwidth dataset available over the calibration network .
* on the other hand , real - time beam formation includes an irreversible fusing operation which requires dynamic calibration to be part of data fusion .
fortunately , it is often possible to use a relatively small subset of the data ( non - contiguous timeslices or a chosen frequency sub - band ) for this purpose to enable short term predictions of gain variations . these can be fed to the fusing nodes well in time before irreversible fusing operations are performed .
since the complexity of the actual algorithm used for calibration makes it better suited for a general purpose computer , the _ calibration network _ can be used to route the relevant subset of data to a commodity switch and deliver the calibration parameters to the appropriate nsps tree level . in the spatially distributed , direct rf sampling nsps architecture
, clocks passed to samplers have very stringent signal quality constraints in terms of net jitter and stability .
the alignment of the multiple data sources before fusion requires high relative stability of the sampling clocks with random jitters much smaller than the reciprocal of the highest frequency in the sampled signal .
clock distribution should also include a mechanism for ensuring the traceability of timekeeping at all digitizing blocks to a centrally maintained time standard to a very high accuracy .
the implementation can benefit from commercial clock distributors which have embedded phase - lock loop clock synthesizer with on - chip voltage controlled oscillator ( vco ) and a per port delay tuning for the distributed clocks .
all data flowing in the nsps is packetized with a custom , low overhead header .
all subsystems accept and generate data in a packetized fashion .
this reflects the inherent asynchronous nature of our system .
packets traversing our platform are atomic and capable of independent existence .
the packetizing of data means that data loss due to network congestion or buffer over - runs is never arbitrary , but always in units of packets . at any instant ,
our network can have different kinds of packets traversing it , corresponding to different stages in the processing .
the basic unit of packet size is maintained as 8 bytes , which is a natural unit or sub - unit for different memory and processing hierarchies .
adequate padding is used if necessary to maintain this condition .
the header is mandated to have a few fixed fields which are common in size and layout across packet types , allowing processing entities to examine packets which can be processed by them , while discarding the others . in a broadcast network
, this approach can waste bandwidth when packets are discarded , but the wastage can be minimised by setting up static routes between partner nodes .
this is possible in both the custom and commodity peer - to - peer link nodes .
the command network , on the other hand , is a broadcast network , with nodes passing on commands not addressed to them to all other nodes downstream of themselves .
data sources can include packet specific extensions to the packet headers generated by them .
the following fields are suggested as a mandatory part of the packet header : * source identifier : at every level of the tree , nodes are endowed with a unique identifier which supplants the existing upstream source i d , if any processing is carried out on the packet . *
datatype : this field allows processing entities to recognize which packets are palatable to them and to reject others .
* data pixel descriptor : this field lays out the size and description of the smallest unit of data transfer to be one of an allowed set , which is implementation dependent .
* streams : this field records the number of independent signal sources present in each packet .
* packet size : the size of a packet is expressed in units of words as specified by the datatype field . *
timestamp : this field is populated as early as possible in the data generation path and maintained across data processing .
this field is generally populated by a timestamp counter running on either a reference clock or on the sampling clock itself and is traceable to the centrally maintained time standard .
this specification is efficient for real - time streaming data description with minimum overhead . for archival of processed data , a standard format which allows multiple binary streams to maintain their identity , like the fits or vlbi data interchange format ( as proposed by the vdif task force ( 2009 ) )
can be used .
no subsystem in our scheme is source synchronous , be it at the hardware or the software level .
all subsystems have enough memory for a store - and - forward of several packets .
this allows the processing to happen at the packet level , on a faster clock than the sampling clock of front end adcs .
it also eases the timing requirements of designs implemented in fpgas and makes them more tolerant of clocking errors .
sequencers play an important role in our architecture , generating events on which the processing progresses .
the sequencers generate necessary globally aligned events to which any action taken on the basis of commands from commodity network will get aligned .
a simple example is an implementation where all processes in the peer network operate at a block level , with a periodic event signifying the need for scheduling a new process as a result of the arrival of a new block of data .
it is important to match the communication bandwidth to processing abilities at every level in the signal processing tree .
more specifically , event markers generated by the sequencers should facilitate a partitioning of processing in each level into abstract transactions , where each transaction deals with the entire data collected over a convenient timeslice * and * the relevant data are locally available _ on demand_. in particular , it is desirable that processing at the central commodity segment is facilitated at a cadence suited for general purpose operating system scheduling to achieve latency tolerance .
for instance , to be commensurate with a housekeeping tick of 10 milliseconds in typical linux configurations , the transaction timeslices should be several times longer . among the available axes in processing space along which the data can be partitioned and distributed to parallel processors ,
the time axis is often the most convenient for slicing , as individual timeslices can be considered independent . for a large network ,
we realize this from a hierarchical set of data poolers _ _ populating different levels of the processing tree , which can collate data from different sources and partition them along the time dimension at each level .
in this section , we provide an example of implementation of nsps by giving an outline of the system being planned for modernizing the ort @xcite .
the ort is a 506 m x 30 m equatorially mounted cylindrical telescope with an equispaced linear array of 1056 dipoles along the focal line .
each dipole has a tuned low noise amplifier @xcite with about 40 mhz bandwidth centered at 327mhz , although the existing analog phasing network restricts the bandwidth to about 10 mhz .
a collaborative program for upgrading the ort has been undertaken jointly by the raman research institute ( rri ) and the national centre for radio astrophysics ( ncra ) , which operates the ort . in this program , the feed array is logically divided into 264 identical segments , where each segment represents an independent antenna element of size 1.93 m x 30 m . the aim of this upgrade is to reconfigure the ort into a programmable 264-element array .
when completed , the reconfigured ort will have an instantaneous field of view of @xmath2 , bandwidth of @xmath3 and will be equipped with an nsps - based digital receiver .
currently , prototypes have been tested for a large fraction of the custom segment of the nsps and the analog signal conditioning subsystems . the final production and integration
is expected to be completed in 2011 .
the digitizers are organized in 22 digitization blocks located below the reflector at a spacing of about 23 m , where each digitization block includes a 12-channel digitizer capable of operating at 100 ms / s .
all the 22 digitization blocks are connected in a star topology with a central system using a peer - to - peer network on optical fibres with multiple links operating at speeds of 2.5 gigabits / sec from each block .
proposed nsps implementation for configuring ort as a 264 element programmable telescope .
_ ] the proposed system consists of four hierarchical levels as illustrated in fig .
[ fig : proposed - nsps - implementation ] , the high speed peer network uses the light weight aurora protocol simplex links for data uplink , with last mile via gige .
currently , bandwidths of upto 100mbps per gige link into level-0 memory have been sustained .
* _ level 3 : _ this level is composed of the distributed digitization infrastructure and is installed at the antenna base .
the prime components of this level are the digitizers , the sampling clock derivation and conditioning circuitry , the first level data organizer and peer - to - peer link handler .
all these components required for handling 12 sensor elements from a 23 m section of the ort are implemented as a single board .
* * the adcs ( dual channel ad9600 ) are capable of a 100ms / s sampling for signals with frequencies upto @xmath4 .
the @xmath5 dynamic range at 10 bit resolution allows us to implement an agc in software .
more importantly , the implementation can benefit from the on - chip sampling clock conditioner , divider and duty cycle stabilizer .
for instance it is possible to provide a sine - wave with frequency 2 - 8 times the sampling clock , and use the on - chip features to convert to square - wave , enable duty - cycle stabilizer and divide by suitable integer to get the sampling clock , thus reducing the overall sampling clock jitter and hence the phase noise in the sampled data .
this feature is useful in direct ( harmonic ) sampling of the incoming @xmath6 rf since the nyquist sampling interval in a band - limited rf is decided only by the bandwidth while jitter tolerance depends on the highest frequency content . *
* this level has an embedded clock synthesizer and distributor for the on - board 12-channel adc based on a reference distributed on fibre by the central high stability clock distributor .
it is received using a digital fibre optic receiver .
the embedded clock distributor is based on a clock buffer and distributor ( lmk03020 ) which has an on - chip vco and per - port delay tuning . * * the first data processing block is implemented in a xilinx spartan6 ( lx45 t ) fpga to perform a conversion of the adc 10 bit resolution data to @xmath7bit via configurable , table - driven logic , where the look - up - table ( lut ) is dynamically updated to accommodate innovative schemes of compression and segregation . for instance , let us assume that a choice among a pre - determined set of luts is best suited for coding / compressing the data in a set of physical packets associated with a transaction unit . here
, each table implements a different encoding of the input word to an output with lesser number of bits per word .
further , we assume that every word of each physical packet is encoded by a choice between two luts out of the set , to represent normal and segregated(flagged ) data . for decompression by downstream nodes ,
the tags of these two luts can be accommodated in the packet header , while a one - bit selection between the two can be associated with each data word - thus achieving the dual purpose of flagging and scaling at the word level .
such a scheme can accommodate a wide range of scale factors and hence a large dynamic range within a logical packet .
suitable thresholds for packet - level choice of luts can be generated on the basis of integrated power over a reasonable time stretch as part of the pre - processing . * * the data router node buffers data from all 12 sensors in internal memory and reorganizes them to form packets containing identical time - stretches from all antenna elements .
the peer - to - peer link out of each 23 m section which connects level-3 to level-2 is implemented using the 4 available rocketio multi - gigabit onboard serializers on the spartan6 .
our choice of sfp for the current implementation can sustain link speeds of upto 2.6gbps on single mode fibre , while we use the light - weight aurora protocol at a wire speed of 2.5gbps for communication between level 3 and level 2 .
* _ level 2 : _ this level is implemented using an fpga ( virtex5 lx50 t ) board whose on - board resources include 2 gb of memory and 8 multi - gigabit transceivers and expansion connectors . this board can sustain the following level-2 functionalities : * * fft block : here , the data processor block first decodes data from a pair of sensors , packs them into the real and imaginary parts of a 32-bit complex integer word , and implements a pipeline stage ( e.g. , radix-64 ) of a split radix fft for all pairs of incoming channels .
the processing resources are enough to handle upto 24 channels ( 2 level-3 entities ) at the maximum sampling rate . * * the data pooler block operates using the large local ddr2 ram and the interconnection with other level-2 cards to pool subsets of both local and remote level-2 data into transaction oriented packets by partitioning data along the time axis . here
, each transaction refers to the processing of a specific timeslice for the entire array . * * the data router block collects data from the local ram in units appropriate for transfer to each outgoing port , packetizes them and sends out selected time slices to level-1 entities .
provision for computation offloading is provided in the form of spare peer - to - peer links which can add on more level-2 cards . * _
level 1 : _ at this level , a memory - based , nsps to commodity network bridge is implemented .
large bursts of continuous time slices are first buffered in ram , and then sent out over gige as properly timestamped packets .
the level implements load partitioning as configured by the root level by manipulating ethernet destination addresses of streams going into the data gige switch .
there is possibility of implementing a data processor block for the remaining 16 point fft operation pending from the split radix fft .
* _ level 0 : _ at the root of the nsps tree , a medium level cluster is proposed to handle both the communication and processing requirements for forming correlations between all sensors .
it is important to note that the cluster inter - node traffic is significantly reduced due to the data routing and transposition carried out using the data pooling nodes at the various levels .
the formation of actual correlations and the calibration parameter estimation is carried out by this level .
the above mentioned partitioning of the load into the 3 levels can be used to bring a subset of data from all 264 elements into one node via a quad - gige card .
the control network is a simplex , one - way channel from a master with unique i d ( in the commodity segment ) to the peer network through the bridge node .
thus , while data can be routed to arbitrary nodes in the cluster depending on the udp destination addresses set during configuration , commands are accepted by the bridge node only through a privileged link from the master .
this simplifies assigning privileges to operations related to starting , stopping or resetting the acquisition state machines , configuring the network routes on customized hardware switches , changing ethernet destination addresses , or changing the contents of the luts used in earlier nodes like the digitizers .
while asynchronous , packetized processing over standard networks is a relatively new concept in radio telescopes , it is being embraced enthusiastically due to the many benefits it offers to the system designer .
even among this class of telescope data processors , contemporary architectures usually have a direct link from the samplers to the central processor . some operations like a digital filter bank or fft
are carried out remotely , while others like cross - correlation is done centrally .
the memory - rich architectures of modern fpgas help in distributing computing to remote nodes and enables buffering to allow multiple passes on the streaming data .
as the data volume grows , e.g. , in the central pooling stations , the processing can be supported by large off - chip memory using commercial memory modules , routinely supported by modern fpgas .
this provides substantial enhancement to buffering for transaction level operations and data partitioning . a peak data rate of 100 ms / s x 4bits for 264 elements for
the ort would correspond to about 13.2gb / s for which the level-1 buffering of 22 gb in 11 virtex-5 cards shown in fig .
[ fig : proposed - nsps - implementation ] is comfortable to sustain transactions of upto a fraction of a second duration .
the use of standard software stacks also allows us to leverage the various high performance modes being worked upon by system optimizers , e.g. , the zero copy mechanism on linux .
another challenging problem with large arrays is the so - called _ corner turning _ problem _ _ , _ _ which refers to the transposition of the input signal matrix needed to achieve the all - to - all communication necessary for correlation .
earlier approaches have looked at either commercial switches or entirely customized switches for routing data @xcite .
we break this problem down into levels , and apply a hybrid of commercial as well as custom routing .
the data pooler element is utilized to implement a memory based switch , while the cots ( gige ) network controller manages another level of redirection by manipulating udp destination addresses .
we have presented a packetized , heterogeneous and distributed signal processing architecture for radio interferometric signal processing which elevates the network to a core system component .
the architecture addresses some of the core issues pertaining to interferometric signal processing .
we visualize this problem as that of an appropriate workload creation and scheduled dispatch to matched processors over a data flow tree . here , the leaf nodes are sources of data , with data processors handling a managed slice of the processing at the intermediate nodes of this tree .
we emphasize the use of cots components , both hardware and software , for rapid deployment , ease of maintenance , and lowering the cost of the architecture implementation .
the goal of realizing a programmable telescope with nsps is facilitated by defining rigid interfaces between both hardware and software components .
this can allow exchange of a variety of data with varying communication and computing requirements between levels in the network .
most of the individual nodes in the nsps can change the nature of their processing within the limits specified by their designed personality and available resources at the node .
this allows offloading of computing requirements in a hierarchical manner up the nsps tree , trading off implementation time with hardware capability of an application mode . due to the rigidity of interfacing protocols as well as the standardized networks making up the system
, we can comfortably add nodes which can tap into the nsps in order to carry out a different processing chain .
data duplication , if required , can be carried out by cots components ( e.g. , by switches operating in broadcast mode ) thus reducing development load .
we have presented an outline of the nsps implementation being planned for configuring the ort as a programmable 264 element telescope .
our architecture is optimally tuned to service the needs of medium sized arrays .
we advocate full software processing for smaller arrays , with an increasing factor of hardware offload as the array size grows .
this approach has being taken by us in building a 44 element demonstrator as a precursor to the receiver for the full 264 element ort array .
this receiver exploits all nsps aspects we have dwelt on , and is in an advanced stage of completion .
the 44-element demonstrator for the ort is being built as a collaborative effort between radio astronomy laboratory at raman research institute and the observatory staff at the ooty radio telescope .
many of the ideas presented here evolved during the trials of demonstrator subsystems for which we specially thank colleagues both at rri and ort .
crs would like to thank madan rao and dwarakanath whose comments have helped improving the manuscript .
we thank the anonymous referee whose comments have been very helpful in improving the clarity of presentation in the manuscript | a new architecture is presented for a networked signal processing system ( nsps ) suitable for handling the real - time signal processing of multi - element radio telescopes . in this system
, a multi - element radio telescope is viewed as an application of a _ multi - sensor , data fusion _ problem which can be decomposed into a general set of computing and network components for which a practical and scalable architecture is enabled by current technology .
the need for such a system arose in the context of an ongoing program for reconfiguring the ooty radio telescope ( ort ) as a programmable 264-element array , which will enable several new observing capabilities for large scale surveys on this mature telescope . for this application ,
it is necessary to manage , route and combine large volumes of data whose real - time collation requires large i / o bandwidths to be sustained . since these are general requirements of many multi - sensor fusion applications , we first describe the basic architecture of the nsps in terms of a _ fusion tree _ before elaborating on its application for the ort .
the paper addresses issues relating to high speed distributed data acquisition , field programmable gate array ( fpga ) based peer - to - peer networks supporting significant on - the fly processing while routing , and providing a last mile interface to a typical commodity network like gigabit ethernet .
the system is fundamentally a pair of two co - operative networks , among which one is part of a commodity high performance computer cluster and the other is based on commercial - off the - shelf ( cots ) technology with support from software / firmware components in the public domain . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.