article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the methane - bearing t dwarfs comprise the coolest known class of brown dwarfs ( * ? ? ?
* references therein ) .
these sources have atmospheres that are sufficiently cool for h@xmath8o and ch@xmath9 gases to form in abundance , resulting in distinct , planet - like spectra @xcite .
the similar masses and effective temperatures of t dwarfs as compared to
warm " exoplanets make them critical benchmarks for theoretical atmospheric and interior models of substellar objects @xcite .
although more than 200 t dwarfs have been discovered over the last decade , there remains a large gap in temperature between the coolest known t dwarfs ( @xmath10 k * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the jovian planets ( @xmath11 k ) . based on theoretical spectra ,
objects in this temperature range may exhibit spectroscopic characteristics that are distinct from those of t dwarfs , including strong near - ir nh@xmath12 absorption and scattering from water ice clouds @xcite .
these sources may therefore represent a new spectral class tentatively designated as y dwarfs @xcite .
in addition , the coolest brown dwarfs encompass the oldest and lowest mass stars " in the galaxy , probing the history and efficiency of star formation at the planetary - mass limit .
wide - field imaging surveys over the last decade have been successful in uncovering free - floating brown dwarfs at progressively cooler temperatures .
however , the relatively low surface density of field brown dwarfs on the sky @xcite necessitates imaging of very large areas and spectroscopy of samples of candidates that can be highly contaminated . a survey for companions to stars in the solar neighborhood ( @xmath13 pc ) is an appealing alternative for identifying the coolest brown dwarfs because of the smaller search volume that is required .
in addition , the age , distance , and metallicity measured for a primary can be adopted for its companion , parameters which are usually difficult or impossible to measure for isolated brown dwarfs .
several of these benchmark t dwarfs , including the prototype gliese 229b , have been discovered as companions to nearby main sequence stars @xcite as well as a white dwarf @xcite .
the infrared array camera ( irac ; * ? ? ?
* ) on board the _ spitzer space telescope _ @xcite offers the best available sensitivity to cool companions in wide orbits ( @xmath14 k , @xmath15 au )
. we have used multi - epoch images from irac to search for substellar companions in the solar neighborhood based on common proper motions . in this letter , we present the discovery of a new companion that is a candidate for the coolest and faintest known brown dwarf .
to search for substellar companions around nearby stars , we have considered nearly all stars that have been observed with irac in multiple epochs and that should show detectable motions in these data ( @xmath16 , see figure [ fig : pm ] ) based on their known proper motions .
this sample consists of @xmath17600 stars , brown dwarfs , and white dwarfs .
the plate scale and field of view of irac are @xmath18 pixel@xmath19 and @xmath20 , respectively .
the camera produces images with fwhm@xmath21-@xmath22 from 3.6 to 8.0 . among the four irac filters ,
we have analyzed the images collected at 3.6 and 4.5 since they offer the best spatial resolution and sensitivity to cool brown dwarfs .
initial processing of the images was performed by the pipeline at the spitzer science center and the resulting data were combined into the final images using r. gutermuth s wcsmosaic idl package @xcite .
pipeline versions s18.7.0 and s18.12.0 were used for the data discussed in this paper .
for each nearby star , we used the task _ starfind _ within iraf to measure positions for all point sources in the images from the multiple epochs .
we converted the pixel coordinates to equatorial coordinates using the world coordinate systems produced by the pipeline and matched the source lists between the epochs .
we then measured a new world coordinate system for the first epoch image using the pixel coordinates from the first epoch and the equatorial coordinates from the second epoch , which enabled more accurate relative astrometry between the two epochs . to identify possible companions , we checked for sources with motions that were consistent with the motions measured for the primaries from these data .
if accurate positions were unavailable for the primary because of saturation in the irac data , we instead computed the motion expected for the primary based on its published proper motion and the elapsed time between the two epochs .
we have identified a new companion to the dq white dwarf wd 0806 - 661 ( l97 - 3 ) , which we refer to as wd 0806 - 661 b. the primary was observed by irac on 2004 december 15 and 2009 august 24 through programs 2313 ( m. kuchner ) and 60161 ( m. burleigh ) with total exposure times of 134 and 536 s , respectively .
the companion has a projected separation of @xmath0 , corresponding to 2500 au at the distance of the white dwarf ( 19.2 pc , * ? ? ?
the motion of wd 0806 - 661 b between the two epochs of irac data agrees closely with that of the primary , and is significantly greater than the zero motion expected for a background source .
this is demonstrated in figure [ fig : pm ] , which shows the differences in equatorial coordinates between the two epochs for all stars with photometric errors less than 0.1 mag in the second epoch . for stars that have fluxes within @xmath23 mag of wd 0806 - 661 b , the standard deviations of the differences in right ascension and declination in figure [ fig : pm ] are @xmath24 , whereas wd 0806 - 661 b exhibits a motion of @xmath25 .
the astrometric accuracy for the primary is higher since it is 3 mag brighter than wd 0806 - 661 b. we measured photometry at 4.5 for wd 0806 - 661 b with the methods described by @xcite .
we present the astrometry and photometry for this object as well as other properties of the system in table [ tab : data ] .
the 4.5 image of wd 0806 - 661 a and b from the second epoch is shown in figure [ fig : image ] .
to investigate the nature of wd 0806 - 661 b , we have examined the available constraints on its colors .
it was not observed at 3.6 or 5.8 by irac in either epoch .
it was encompassed by the 8.0 images from the first epoch but was not detected . the resulting limit of @xmath26>15 $ ] implies @xmath27-[8.0]<1.7 $ ] , which is consistent with a stellar object at any spectral type , including a t dwarf @xcite .
we have not found any detections or useful flux limits for this object in any publicly available images . on 2010 december 23 , we sought to obtain a spectrum of wd 0806 - 661 b with the folded - port infrared echellette ( fire , * ? ? ?
* ) at magellan 6.5 m baade telescope
. however , it was not detected in the @xmath28-band acquisition images , which had a total exposure time of 45 s. by flux calibrating these images with photometry of unsaturated sources from the point source catalog of the two - micron all - sky survey @xcite , we estimate a magnitude limit of @xmath29 for wd 0806 - 661 b. the combination of this limit with the irac measurement at 4.5 indicates a color of @xmath30>3.3 $ ] , which is consistent with a spectral type of @xmath31t8 @xcite .
a substellar object is the only type of companion that could be as red and faint as wd 0806 - 661 b.
to estimate the physical properties of wd 0806 - 661 b , we begin by assuming that it has the distance of the primary and plotting the resulting absolute magnitude at 4.5 versus @xmath30 $ ] and spectral type in figure [ fig : cmd ] . for comparison , we include t dwarfs that have measured distances and irac photometry .
these data indicate that wd 0806 - 661 b should be the coolest and faintest brown dwarf identified to date by a fairly large margin . for perspective
, the sequence of known t dwarfs spans 2.5 magnitudes while wd 0806 - 661 b is one magnitude fainter than the bottom of that sequence . as a likely companion , wd 0806 - 661
b should have the same age as its primary . by combining the mass of @xmath32 @xmath7 for wd 0806 - 661 @xcite with initial - final mass relations for white dwarfs
@xcite , we compute a mass of @xmath33 @xmath7 for the progenitor . a star born with this mass should have a main sequence lifetime of @xmath34 myr @xcite .
the sum of that lifetime and the cooling age of 670@xmath3540 myr from @xcite produces a total age of @xmath36 gyr .
we can estimate the mass and temperature of wd 0806 - 661 b by comparing its photometry to the fluxes predicted by theoretical evolutionary models . in figure
[ fig : mag ] , we plot it on a diagram of @xmath2 versus age under the assumption that it has the age and distance of the primary .
@xcite used their model interiors and atmospheres to compute effective temperatures , bolometric luminosities , and spectra for cool brown dwarfs across a range of masses and ages .
we have convolved those synthetic spectra with the system throughput for the 4.5 irac filter .
the resulting magnitudes are plotted at constant values of mass in figure [ fig : mag ] .
they indicate that wd 0806 - 661 b should have a mass of @xmath4 @xmath5 .
the models that best match the absolute magnitude and age of wd 0806 - 661 b have effective temperatures of @xmath3 k.
wd 0806 - 661 b is a strong contender for the faintest known brown dwarf based on the available photometric and astrometric measurements . with an estimated @xmath37
k , it would be significantly cooler than the latest t dwarf currently known ( @xmath10 k * ? ? ? * ) . moreover , its photosphere would be sufficiently cool to harbor water ice clouds , making it a likely prototype for the y dwarf spectral class .
we also estimate a mass of @xmath4 @xmath5 for wd 0806 - 661 b from the evolutionary models . given that cloud core fragmentation appears capable of making binary companions near this mass ( e.g. , * ? ? ?
* ) , it seems likely that a companion of this kind in a very large orbit ( 2500 au ) would have formed in this manner .
however , the mass estimate also falls within the range of masses measured for close - in extrasolar planets ( @xmath38 @xmath5 , * ? ? ?
* ; * ? ? ?
because the progenitor of the primary was fairly massive ( @xmath6 @xmath7 ) , its circumstellar disk at birth could have been massive enough to give rise to a companion at this mass .
thus , wd 0806 - 661 b could be a giant planet that has been dynamically scattered to a larger orbit .
spectroscopy and multi - band photometry are necessary to verify the substellar nature of wd 0806 - 661 b and to better estimate its physical properties . if confirmed as the coolest known brown dwarf , it will represent a valuable laboratory for studying atmospheres in a new temperature regime , and its colors will help to guide searches for the coldest brown dwarfs with facilities like the wide - field infrared survey explorer and the james webb space telescope .
we thank robin ciardullo for advice regarding age estimates for white dwarfs , telescope operator mauricio martinez at magellan for his assistance with the fire observations , and nigel hambly for his helpful referee report .
we acknowledge support from grant ast-0544588 from the national science foundation ( k. l. , j. b. ) and the chris and warren hellman fellowship program ( a. b. ) .
the center for exoplanets and habitable worlds is supported by the pennsylvania state university , the eberly college of science , and the pennsylvania space grant consortium . ll distance & [email protected] pc + @xmath39 & + 340.3@xmath40 mas yr@xmath19 + @xmath41 & @xmath42289.6@xmath40 mas yr@xmath19 + mass & @xmath32 @xmath7 + progenitor mass & @xmath33 @xmath7 + age & @xmath36 gyr + separation & [email protected]@xmath43 ( 2500 au ) + position angle & 104.2@xmath44 + @xmath28 & @xmath1720 mag + @xmath27 $ ] & [email protected] mag + est .
@xmath45 & @xmath3 k + est . mass & @xmath4 @xmath5 +
|
we have used multi - epoch images from the infrared array camera on board the _ spitzer space telescope _ to search for substellar companions to stars in the solar neighborhood based on common proper motions . through this work
, we have discovered a faint companion to the white dwarf wd 0806 - 661 .
the comoving source has a projected separation of @xmath0 , corresponding to 2500 au at the distance of the primary ( 19.2 pc ) . if it is physically associated , then its absolute magnitude at 4.5 is @xmath1 mag fainter than the faintest known t dwarfs , making it a strong candidate for the coolest known brown dwarf .
the combination of @xmath2 and the age of the primary ( 1.5 gyr ) implies an effective temperature of @xmath3 k and a mass of @xmath4 @xmath5 according to theoretical evolutionary models .
the white dwarf s progenitor likely had a mass of @xmath6 @xmath7 , and thus could have been born with a circumstellar disk that was sufficiently massive to produce a companion with this mass .
therefore , the companion could be either a brown dwarf that formed like a binary star or a giant planet that was born within a disk and has been dynamically scattered to a larger orbit .
|
it is a pleasure to acknowledge helpful discussions with winfried zimdahl .
this project was in part supported by national basic research program of china under grant no.2003cb716300 and also by nnsfc under grant no.10175070 .
99 b.ratra and p.j.e.peebles , phys.rev . *
d37 * ( 1988 ) 3406 ; c.wetterich , nucl.phys . *
b302 * ( 1988 ) 668 .
i.zlatev , l.wang and p.j.steinhardt , phys.rev.lett , * 82 * ( 1999 ) 896 ; p.j.steinhardt , l.wang and i.zlatev , phys.rev . * d59 * ( 1999 ) 123504 .
u.alam , v.sahni , t.d.saini and a.a.starobinsky , astro - ph/0311364 ; u.alam , v.sahni and a.a.starobinsky , astro - ph/0403687 ; d.huterer and a.cooray , astro - ph/0404062 ; b.feng , x.l.wang and x.zhang , astro - ph/0404224 .
r.r.caldwell , phys.lett . *
b545 * ( 2002 ) 23 .
l.mersini , m.bastero-gil and p.kanti , phys.rev . *
d64 * ( 2001 ) 043508 ; m.bastero-gil , p.h.frampton and l.mersini , phys.rev . *
d65 * ( 2002 ) 106002 ; p.h.frampton , phys.lett .
* b555 * ( 2003 ) 139 ; f.piazza and s.tsujikawa , jcap 0407 ( 2004 ) 004 .
j.d.barrow , nucl.phys . *
b310 * ( 1988 ) 743 .
m.d.pollock , phys.lett . *
b215 * ( 1988 ) 635 ; g.calcagni , gr - qc/0410027 .
d.f.torres , phys.rev . *
d66 * ( 2002 ) 043522 ; e.elizalde , s.nojiri and s.d.odintsov , phys.rev . *
d70 * ( 2004 ) 043539 .
a.kehagias and e.kiritsis , jhep * 9911 * ( 1999 ) 022 .
t.chiba , t.okabe and m.yamaguchi , phys.rev .
* d62 * ( 2000 ) 023511 ; j.m.aguirregabiria , l.p.chimento and r.lazkoz , phys.rev . *
d70 * ( 2004 ) 023509 .
g.w.gibbons , hep - th/0302199 ; s.nojiri and s.d.odintsov , phys.lett .
* b562 * ( 2003 ) 147 ; s.nojiri and s.d.odintsov , phys.lett .
* b565 * ( 2003 ) 1 .
a.e.schulz and m.white , phys.rev .
* d64 * ( 2001 ) 043514 ; m.p.dabrowski , t.stachowiak and m.szydlowski , phys.rev .
* d68 * ( 2003 ) 103519 ; y.s.piao and e.zhou , phys.rev . *
d68 * ( 2003 ) 083515 ; v.b.johri , phys.rev . *
d70 * ( 2004 ) 041303 ; x.h.meng and p.wang , hep - ph/0311070 ; y.s.piao and y.z.zhang , astro - ph/0401231 ; j.lima and j.s.alcaniz , astro - ph/0402265 ; m.bouhmadi-lopez and j.j.madrid , astro - ph/0404540 ; y.h.wei and y.tian , gr - qc/0405038 ; l.p.chimento and r.lazkoz , astro - ph/0405518 ; b.feng , m.li , y.s.piao and x.zhang , astro - ph/0407432 ; p.f.gonzalez-diaz , astro - ph/0408450 .
r.r.caldwell , m.kamionkowski and n.n.weinberg , phys.rev.lett . *
91 * ( 2003 ) 071301 ; m.sami and a.toporensky , mod.phys.lett .
* a19 * ( 2004 ) 1509 ; p.f.gonzalez-diaz , phys.lett . *
b586 * ( 2004 ) 1 ; l.p.chimento and r.lazkoz , gr - qc/0405020 .
s.m.carroll , m.hoffman and m.trodden , phys.rev . *
d68 * ( 2003 ) 023509 ; p.singh , m.sami and n.dadhich , phys.rev .
* d68 * ( 2003 ) 023522 ; z.k.guo , y.s.piao and y.z.zhang , phys.lett . *
b594 * ( 2004 ) 247 .
w.zimdahl , d.pavon and l.p.chimento , phys.lett .
* b521 * ( 2001 ) 133 .
l.p.chimento , a.s.jakubi , d.pavon and w.zimdahl , phys.rev . *
d67 * ( 2003 ) 083513 .
c.wetterich , astron.astrophys .
* 301 * ( 1995 ) 321 .
z.k.guo , y.s.piao , x.zhang and y.z.zhang , astro - ph/0410654 . m.doran and j.jaeckel , phys.rev .
* d66 * ( 2002 ) 043519 .
r.g.cai and a.wang , hep - th/0411025 .
|
we investigate the role of a suitable interaction between a matter fluid and a phantom field for the coincidence problem .
there exists a stationary scaling solution which is a stable attractor at late times .
furthermore , the cosmic doomsday is avoided in one region of the parameter space .
scalar field plays an important role in modern cosmology .
dark energy can be attributed to the dynamics of a scalar field , called quintessence @xcite , which convincingly realize the present accelerated expansion of the universe by using late - time attractor solutions in which the scalar field mimics the perfect fluid in a wide range of parameters . but regarded as dark energy , quintessence field with the state equation parameter @xmath0 may be not consistent with recent observation @xcite . in order to obtain @xmath1 , phantom field with reverse sign in its dynamical term
may be a simplest implementing and can be regarded as one of interesting possibilities describing dark energy @xcite .
the physical background for phantom type of matter with strongly negative pressure would be found in string theory @xcite .
phantom field may also arise from a bulk viscous stress due to particle production @xcite or in higher - order theories of gravity @xcite , brans - dicke and non - minimally coupled scalar field theories @xcite .
the cosmological models which allow for phantom matter appear naturally in the mirage cosmology of the braneworld scenario @xcite and in k - essence models @xcite . in spite of the fact that the field theory of phantom fields encounters the problem of stability which one could try to bypass by assuming them to be effective fields @xcite , it is nevertheless interesting to study their cosmological implication .
recently , there are many relevant studies of phantom energy @xcite .
the physical properties of phantom energy are rather weird , as they include violation of the dominant - energy condition , naive superluminal sound speed and increasing energy density with time .
the latter property ultimately leads to unwanted future singularity called big rip .
this singularity is characterized by the divergence of the scale factor in a finite time in future @xcite . to avoid the cosmic doomsday ,
specific scalar field models were proposed @xcite .
it requires a special class of phantom field potentials with local maximum .
moreover , the energy density of the phantom field increases with time , while the energy density of the matter fluid decreases as the universe evolves . why the energy densities of dark matter and phantom energy are of the same order just at the present epoch ? in this paper
we investigate the role of a possible coupling of dark matter and phantom field for the coincidence problem . with the help of a suitable coupling @xcite
, we find that there exists a stationary scaling solution and demonstrate numerically that it is a stable attractor at late times .
furthermore , the cosmic doomsday is avoided in one region of the parameter space .
we consider the case in which both the phantom energy with constant @xmath1 and the cold dark matter are present .
then , if the universe ceases to be matter - dominated at cosmological time @xmath2 , the scale factor can be written as @xmath3^{2/3(1+w)}\ ] ] at @xmath4 .
it is easy to see that the scale factor blows up at @xmath5 .
this occurs because , even though the energy densities in ordinary types of matter are redshifting away , the energy density of phantom energy increases in an expanding universe .
it is possible that both components decrease with time if there is a transfer of energy from the phantom field to the matter fluid .
then the cosmic doomsday may be avoided . for a spatially flat frw universe with the matter fluid @xmath6 and the phantom field @xmath7 ,
friedmann equation can be written as @xmath8 where @xmath9 is the gravitational coupling and the energy density and pressure , @xmath10 and @xmath11 , of the homogeneous phantom field @xmath7 are given by @xmath12 respectively , in which @xmath13 is the phantom field potential .
we postulate that the two components , @xmath10 and @xmath6 , interact through the interaction term @xmath14 according to @xmath15 where @xmath16 satisfy @xmath17 and @xmath18 .
the interaction term @xmath14 represents an additional degree of freedom which can be specified by the existence of solution with a stationary energy density ratio @xmath19 at late times . using eqs.([em1 ] ) and ( [ em2 ] ) we obtain the evolution equation of the ratio @xmath20
: @xmath21 obviously , the suitable interaction @xmath22 guarantees the existence of the stationary solution .
note that @xmath23 , which implies there is a transfer of energy from the phantom field to the matter fluid .
we assume an interaction characterized by @xmath24 where @xmath25 denotes the transfer strength , which has already been discussed in ref .
@xcite . when @xmath26 and @xmath27 are assumed to be constants , the two stationary solutions to eq.([qe ] ) are @xmath28 which imply @xmath29 and @xmath30 when @xmath31 .
a stability analysis of the stationary solution indicates that the matter - dominated scaling solution @xmath32 is unstable , while the phantom - dominated scaling solution @xmath33 is stable @xcite .
the energy density ratio @xmath20 evolves from the unstable stationary value @xmath34 to the stable stationary solution @xmath35 , which is clearly compatible with the presently favored observational data @xmath36 and @xmath37 .
this may provide us with the dynamics of the density ratio that is relevant to the solution of the coincidence problem .
now consider the case of @xmath38 ( i.e. cold dark matter ) .
we will investigate the cosmological evolution with the stationary density ratio @xmath20 .
substituting eq.([qe ] ) into eq.([em1 ] ) , it follows that @xmath39 which yields @xmath40 where @xmath41 .
if the expansion is matter dominated until the time @xmath2 , then we can write the scale factor as @xmath42 using the definitions of @xmath20 and @xmath26 , the friedmann equation ( [ fe ] ) gives @xmath43 in terms of @xmath7 , ( [ rp ] ) becomes @xmath44 since @xmath26 is a constant , that is , potential and kinetic energies of the phantom field remain proportional , it follows that @xmath45 it is reassuring to find an exponential potential because this type potentials arise very naturally in the models of unification , such as kaluza - klein theories , supergravity theories and string theories .
similar result has been obtained in refs .
@xcite .
these authors started with an exponential potential in which there is a free parameter and then investigated the parameter range for which there exists a stable scaling solution .
we have first constructed a solution with expected properties and then derived the corresponding potential . in the coupled phantom scenario , it is surprising that the signs of the exponent in eq.([ve ] ) are different in different regions of the parameter space , which are determined by the signs of the parameter @xmath46 .
we will find that the increasing potential corresponds to a climbing - up phantom field while the decreasing potential corresponds to a rolling - down phantom field .
the recent observations indicate that approximately 0.3 and 0.7 of the total energy density of the universe attribute to dark matter and dark energy , respectively . to solve the coincidence problem of our present universe ,
the stable stationary ratio of the energy densities should satisfy @xmath47 , which corresponds to @xmath48 ( i.e. the regions i and ii in figure 1 ) . in the region
i @xmath49 for @xmath50 , since @xmath51 , the energy density of the phantom field decreases in an expanding universe and the universe accelerates without a big rip in the future .
however , in the region ii @xmath52 for @xmath50 and @xmath53 for @xmath54 , @xmath55 indicates that the universe accelerates until a big rip occurs at @xmath56 .
we see that the energy transfer from the phantom field to the cold dark matter prolongs the lifetime of the universe .
, @xmath25 ) parameter space with @xmath38 . in the regions i and ii , the scaling solution is a stable attractor . in the region iii ,
there exist no stable scaling solutions with @xmath47 .
the universe accelerates for ever in the region i , while the universe leads to a big rip in the future in the region ii . ] to study an explicit numerical evolution of the energy density ratio @xmath20 and the phantom energy density ( or the matter energy density ) , it is most convenient to rewrite the evolution equations ( [ re ] ) and ( [ em1 ] ) ( or eq.([em2 ] ) ) as a set of two first - order differential equations with two independent variables @xmath20 and @xmath10 ( or @xmath6 ) @xmath57 where the prime denotes a derivative with respect to the logarithm of the scalar factor , @xmath58
. let us consider two points , ( @xmath59 , @xmath60 ) and ( @xmath61 , @xmath60 ) in the regions i and ii of the ( @xmath26 , @xmath25 ) parameter space in figure 1 , respectively .
we choose different initial conditions and follow the evolution .
we find that an initial energy density ratio decreases and there exist an attractor curve , which corresponds to the stationary scaling solution .
it implies that the universe evolves from a dark matter dominance to a phantom dominance and the energy density ratio of the two components becomes a constant ultimately .
furthermore , we find that the stationary energy density ratio increases as the transfer strength @xmath25 increases when @xmath26 is fixed . in figures 2 and 3 ,
the energy densities of the two components redshift away . from eqs.([rpe ] ) and ( [ ve ] ) , we note that the phantom field rolls down the exponent potential . in figures 4 and 5 ,
the energy density of the phantom field increases with time , while the energy density of the cold dark matter decreases initially and then increases ultimately with time . from eqs.([rpe ] ) and ( [ ve ] ) , we note that the phantom field climbs up the exponent potential .
we have considered above that the present accelerated expansion of our flat frw universe is driven by an interacting mixture of a matter fluid and a phantom field with @xmath1 . in the absence of interaction , as shown in ref .
@xcite , there exist no scaling solutions because the phantom energy increases while the matter energy decreases with time . with the help of a suitable coupling
, there exists a stable , stationary scaling solution , which requires a transfer of energy from the phantom field to the matter fluid .
furthermore , we demonstrate numerically that an interaction between cold dark matter and phantom field can drive the transition from a matter dominance to a phase of accelerated expansion with a stationary ratio of the energy densities of the two components .
this interacting phantom approach indicates a phenomenological solution of the coincidence problem .
the different regions in the ( @xmath26 , @xmath25 ) parameter space lead to different fates of the universe . in the region
i , the phantom field rolls down the exponent potential and the universe accelerates without the cosmic doomsday .
however , in the region ii , the phantom energy increases as the phantom field climbs up the potential , which lead to the divergence of scale factor in the future .
coupling of the quintessence field to dark matter may be worrisome because of quantum corrections to the quintessence potential @xcite .
it is valuable to study the stability of the phantom potential under quantum fluctuations in the case with a coupling to dark matter . * note added .
* while this letter is under review , ref .
@xcite presents a similar idea .
|
one of the important areas of investigation at the proposed high energy electron - positron linear collider will be the study of photon - photon collisions . since photons couple directly to all fundamental fields carrying the electromagnetic current
leptons , quarks , @xmath13 supersymmetric particles , etc.high energy @xmath14 collisions will provide a comprehensive laboratory for exploring virtually every aspect of the standard model and its extensions @xcite .
effective photon beams from virtual bremsstrahlung provide access to low energy quasi - real @xmath15 collisions @xcite .
a large number of studies have been performed at the 4 detectors at lep , at cesr , babar , belle , vepp-4 , vepp-2 , adone , and other @xmath16 storage rings .
the qed processes @xmath17 and @xmath18 have been studied at the l3 detector at lep at 161 gev @xmath19 209 gev .
the muon pair invariant mass was measured in the range 3 gev @xmath20 40 gev .
good agreement was found with @xmath21 qed expectations . in addition , limits on the anomalous magnetic and electric dipole moments of the tau lepton were obtained @xcite . it should be noted that significant deviations from qcd predictions have been reported by the l3 and opal collaborations at lep for the cross sections of inclusive single - particle , jet , and beauty production @xcite in @xmath15 collisions .
a recent review of these measurements has been given by braccini @xcite .
for example , the production of @xmath22 and @xmath23 quarks in @xmath24 collisions has been studied by l3 at @xmath25 from @xmath26 gev to @xmath27 gev .
the measured cross sections are in excess of the next - to - leading order perturbative qcd predictions by a factor of 3 @xcite .
[ fig : sigma_ccbb ] .
[ fig : sigma_ccbb ] the cross section for @xmath28 , @xmath29 and @xmath30 inclusive production in the reaction @xmath31 hadrons has been measured as a function of the transverse momentum and the pseudorapidity by l3 @xcite and opal @xcite .
agreement with respect to qcd predictions is found for all particles at @xmath32 gev ; however the data are significantly higher than the qcd predictions for @xmath28 and @xmath29 at high @xmath33 , as shown in fig .
[ fig02 ] .
the advent of back - scattered laser beams for @xmath34 colliders will allow the efficient conversion of a substantial fraction of the incident lepton energy into high energy photons @xcite . when a polarized laser beam compton - scatters on a polarized electron beam , each electron is effectively converted into a polarized photon with a high fraction of its energy .
the technology for the required high - powered lasers is well along in development @xcite .
the effective luminosity and energy of photon - photon collisions from back - scattered laser beams is expected to be comparable to that of the primary electron - positron collisions .
polarized electron - photon collisions are also an important by - product of this program the high energy luminosity , and polarization of back - scattered laser beams thus has the potential to make photon - photon collisions a key component of the physics program of the next linear collider @xcite .
this capability will allow detailed studies of a large array of high energy @xmath35 and @xmath36 collision processes , including polarized beams .
the physics program includes tests of electroweak theory in photon - photon annihilation such as @xmath37 , @xmath38 neutral and charged higgs bosons , and higher - order loop processes , such as @xmath39 and @xmath40 since each photon can be resolved into a @xmath4 pair , high energy photon - photon collisions can also provide a remarkably background - free laboratory for studying @xmath41 collisions and annihilation .
there are also important high energy @xmath15 and @xmath42 tests of quantum chromodynamics , including the production of two gluon jets in photon - photon collisions , deeply virtual compton scattering on a photon target , and leading - twist single - spin asymmetries for a photon polarized normal to a production plane .
exclusive hadron production processes in photon - photon collisions provide important tests of qcd at the amplitude level , particularly as measures of hadron distribution amplitudes which are also important for the analysis of exclusive semi - leptonic and two - body hadronic @xmath12-decays .
some of these processes are illustrated in fig .
[ fig : feynmans ] .
the photon - photon collider is also compatible with @xmath43 collisions which are interesting in their own right @xcite .
some of the interesting processes are the production of single and double - charged higgs bosons via @xmath44 and @xmath45 studies of @xmath46 fusion in @xmath47 measurements of elastic vector boson scattering in @xmath48 including tests of the 3- and 4-point @xmath49 @xmath50 and @xmath51 couplings via @xmath52 and @xmath53 studies of the @xmath54 qcd structure function in @xmath55 as well as measurements of high momentum transfer @xmath43 elastic scattering .
some of these processes are illustrated in fig .
[ fig : eeprocesses ] .
photon - photon collisions can be classified as follows : ( a ) the photons can annihilate into a charged pair such as @xmath56 , lepton pairs or charged higgs ; ( b ) the photons can produce neutral pairs via loop diagrams such as @xmath57 and @xmath58 ; or ( c ) the photons can each couple to separate charged pairs which scatter by a gauge particle exchange : @xmath59 ( d ) the photons can fuse to produce a single @xmath1 resonance such as a neutral higgs , an @xmath60 or @xmath61 higher orbital state .
exclusive hadronic final states such as meson or baryon pairs can be formed . in each case , a state of even charge conjugation @xmath62 is produced in a general partial wave .
a recent survey of the physics potential of @xmath63 @xmath42 and @xmath14 colliders has been given by de roeck @xcite .
a detailed study of higgs production in the standard model and the minimal supersymmetric standard model ( mssm ) has been summarized by krawczyk @xcite and asner @xcite . probes of the physics of alternative models , such as the little higgs " model @xcite is discussed by asner et al .
a review of recent experimental results in two - photon interactions is given by urner @xcite .
table i summarizes some of the important processes accessible at a photon collider , as itemized by boos _
et al . _
@xcite one can add processes such as @xmath64 , leptoquark production , strong @xmath65 scattering , and @xmath66 .
.update of the gold
plated processes at photon colliders . [ cols="<,^",options="header " , ] a unique advantage of a photon - photon collider is its potential to produce and determine the properties of fundamental @xmath1 resonances such as the higgs boson . a simulation of events for @xmath67 is shown in fig . [
fig : higgs1 ] .
one can also use the transverse polarization of the colliding photons to distinguish the parity of the resonance : the coupling for a scalar resonance is @xmath68 versus @xmath69 for the pseudoscalar .
more generally , one can use polarized photon - photon scattering to study cp violation in the fundamental higgs to two - photon couplings @xcite . in the case of electron - photon collisions
, one can use the transverse momentum fall - off of the recoil electron in @xmath70 to measure the fall - off of the @xmath71 higgs transition form factor and thus check the mass scale of the internal massive quark and @xmath54 loops coupling to the higgs @xcite .
the cross sections for pairs of scalars , fermions or vectors particles are all significantly larger ( by about one order of magnitude ) in @xmath15 collisions than in @xmath72 collisions , as demonstrated in fig .
[ charged ] . unlike the @xmath16 annihilation cross section , which falls at least as fast as @xmath73 many of the @xmath15 cross sections increase with energy .
the energy dependence of a cross section follows from the spin of the exchanged quanta . using regge analysis , a two - body cross section @xmath74 at fixed @xmath75 where @xmath76 is the spin of the exchanged particle or effective trajectory .
for example , the @xmath37 differential cross section is constant at high energies since the spin of the exchanged @xmath54 is @xmath77 in fact , after integration over phase space , the cross section for pairs of vector bosons in photon - photon collisions increases logarithmically with energy .
this is in contrast to @xmath78 which produces a single @xmath79 pair in one partial wave and falls as @xmath80 it is also interesting that a dominant two - jet high @xmath81 reaction in photon - photon collisions at high energies @xmath82 is @xmath83 which proceeds via two quark loops coupling via gluon exchange in the @xmath75 channel @xcite .
since each photon can be resolved into a @xmath4 pair , high energy photon - photon collisions produce equivalent effective @xmath79 beams , thus providing a remarkably background - free laboratory for studying @xmath41 interactions and testing for any anomalous magnetic and quadruple couplings .
the interacting vector bosons can scatter pair - wise or annihilate ; _
e.g. _ , they can annihilate into a standard model higgs boson or a pair of top quarks .
there is thus a large array of tests of electroweak theory possible in photon - photon collisions . the splitting function for @xmath84 can be relatively flat for some @xmath54 helicities , so that one has a high probability for the @xmath54 s to scatter with a high fraction of the energy of the photon .
one can thus study tree graphs contributions derived from photon , z , or higgs exchange in the @xmath75-channel , and in the case of identical @xmath54 s , the additional u - channel amplitudes . in the case of oppositely - charged @xmath54 s , @xmath85-channel annihilation processes such as @xmath86 contribute .
the largest cross sections will arise if the @xmath54 s obey a strongly coupled theory ; in this case the longitudinal @xmath54 s scattering amplitude saturates unitarity and the corresponding @xmath87 cross section will be maximal .
the cross sections of many standard model processes are illustrated in fig .
[ fig : cs ] . reviews of this physics are given in the references .
@xcite one of the most important applications of two - photon physics is the direct production of @xmath79 pairs . by using polarized back - scattered laser beams
, one can in principle study @xmath37 production as a function of initial photon helicities as well as resolve the @xmath54 helicities through their decays .
the study of @xmath88 is complimentary to the corresponding @xmath89 channel , but it also can check for the presence of anomalous four - point @xmath90 interactions not already constrained by electromagnetic gauge invariance , such as the effects due to @xmath91 exchange .
a main focus of the pair production measurements are the values of the @xmath54 magnetic moment @xmath92 and quadruple moment @xmath93 the standard model predicts @xmath94 and @xmath95 up to radiative corrections analogous to the schwinger corrections to the electron anomalous moment .
the anomalous moments are thus defined as @xmath96 and @xmath97 the fact that @xmath98 and @xmath99 are close to zero is actually a general property of any spin - one system if its size is small compared to its compton scale . for example , consider the drell - hearn - gerasimov sum rule @xcite for the @xmath54 magnetic moment : @xmath100
. $ ] here @xmath101 is the total photoabsorption cross section for photons on a @xmath54 with ( anti- ) parallel helicities . as the radius of the @xmath54 becomes small , or its threshold energy for inelastic excitation becomes large , the dhg integral and
hence @xmath102 vanishes .
hiller and i have shown @xcite that this argument can be generalized to the spin - one anomalous quadruple moment as well , by considering one of the unsubtracted dispersion relations for near - forward @xmath103 spin - one compton scattering @xcite : @xmath104 here @xmath105 .
one again sees that in the point - like or high threshold energy limit , both @xmath106 and @xmath107 this result applies to any spin - one system , even to the deuteron or the @xmath108 the essential assumption is the existence of the unsubtracted dispersion relations ; _ i.e. _ , that the anomalous moments are in principle computable quantities . in the case of the @xmath54 ,
the finite size correction is expected to be order @xmath109 , since the underlying composite theory should be chiral to keep the @xmath54 mass finite as the composite scale @xmath110 becomes large @xcite .
thus the fact that a spin - one system has nearly canonical values for its moments signals that it has a small internal size ; however , it does not necessarily imply that it is a gauge field .
yehudai @xcite has made extensive studies of the effect of anomalous moments on different helicity amplitude contributing to @xmath111 cross section .
the empirical sensitivity to anomalous couplings from @xmath112 reactions is comparable and complimentary to that of @xmath113 a comprehensive analysis of fermion processes in photon - photon collisions is given by layssac and renard @xcite . as emphasized by jikia and tkabladze @xcite , pairs of neutral gauge bosons
can be produced in @xmath112 reactions through one loop amplitudes in the standard model at a rate which should be accessible to the nlc .
leptons , quarks , and @xmath54 all contribute to the box graphs .
the fermion and spin - one exchange contributions to the @xmath114 scattering amplitude have the characteristic behavior @xmath115 and @xmath116 respectively .
the latter is the dominant contribution at high energies , so one can use the optical theorem to relate the forward imaginary part of the scattering amplitude to the total @xmath37 cross section .
the resulting cross section @xmath117 is of order 20 fb at @xmath118 , corresponding to 200 events / year at an nlc with luminosity 10 fb@xmath119 @xcite .
the corresponding @xmath120 process has been analyzed by gounaris , porfyriadis and renard @xcite .
a single top quark can be produced in electron - photon collisions at an nlc through the process @xmath121
@xcite . see fig .
[ fig : f9 ] .
@xcite this process can be identified through the @xmath122 decay with @xmath123 the rate is strongly polarization dependent and is sensitive to the structure of the @xmath124 matrix element , possible fourth generation quarks , and anomalous couplings .
an interesting background is the virtual @xmath54 process @xmath125 where the higgs boson decays to @xmath126 and @xmath127 schmidt , rizzo , and i @xcite have shown that one can use the sign change of the integrand of the dhg sum rule to test the canonical couplings of the standard model and to isolate the higher order radiative corrections .
for example , consider the reactions @xmath128 , @xmath129 and @xmath130 which can be studied with back - scattered laser beams .
in contrast to the time - like process @xmath131 , the @xmath15 and @xmath36 reactions are sensitive to the anomalous moments of the gauge bosons at @xmath132 the vanishing of the logarithmic integral of @xmath133 in the born approximation implies that there must be a center - of - mass energy , @xmath134 , where the polarization asymmetry @xmath135 possesses a zero , _
i.e. _ , where @xmath136 reverses sign .
the cancellation of the positive and negative contributions @xcite of @xmath137 to the dhg integral is evident in fig .
we find strong sensitivity of the position of this zero or `` crossing point '' ( which occurs at @xmath138 gev in the sm ) to modifications of the sm trilinear @xmath139 coupling and thus can lead to high precision constraints .
in addition to the fact that only a limited range of energy is required , the polarization asymmetry measurements have the advantage that many of the systematic errors cancel in taking cross section ratios . this technique can clearly be generalized to other higher order tree - graph processes in the standard model and supersymmetric gauge theory .
the position of the zero in the photoabsorption asymmetry thus provides an additional weapon in the arsenal used to probe anomalous trilinear gauge couplings .
because of the simplicity of its initial state , two - photon collisions provide an important laboratory for testing coherent and incoherent effects in quantum chromodynamics . in qcd events where each photon is resolved @xcite in terms of its intermediate quark and gluon states , @xmath35 collisions resemble point - like meson - meson collisions .
one can study detailed features of @xmath140 at threshold and its final state evolution . in the case of single or double diffractive two - photon events , one can study fundamental aspects of pomeron and odderon @xmath75-channel physics . for example , the energy asymmetry of charm quark versus anti - charm jets in four - jet reactions can measure the interference of the pomeron and odderon @xcite .
consider , the cross section for producing two quark pairs .
@xmath141 , which can be described as the interaction of two color dipoles @xcite .
if one of the heavy quark pairs is a charm or bottom quark pair with a small color dipole moment , then the multiple gluonic exchange graphs will reggeize into a hard pomeron , and one predict a strong energy growth of the production cross section similar to that observed at hera in diffractive charm electroproduction @xcite .
the qcd description of the hard qcd pomeron is given by the bfkl analysis @xcite .
there has been progress in stabilizing the bfkl predictions at next - to - leading - order by using blm scale fixing @xcite .
the analogous qcd physics of the gluon rescattering in diffractive deep inelastic scattering is discussed in the references @xcite .
one can also utilize electron - photon collisions at a linear collider to test the shape and growth of the photon structure functions @xcite the back - scattered laser beam provides a high energy polarized target photon , and the neutral current probe is obtained by tagging the scattered electron at momentum transfer squared @xmath142 one can also reconstruct the charged current contributions where the electron scatters into a neutrino from calorimetric measurements of the recoiling system .
it also should be possible to identify the separate charm , bottom , top and @xmath54 contributions to the photon structure functions .
the photon structure functions receive hadron - like contributions from the photon s resolved fock components as well as its direct component derived from the @xmath143 time - like qcd compton amplitude . because of the direct contributions , the photon structure functions obey an inhomogeneous evolution equation . the result ,
as first shown by witten @xcite is that the leading order qcd structure functions of the photon have a unique scaling behavior : @xmath144 , @xmath145 , @xmath146 .
the most characteristic behavior of the photon structure function @xmath147 in qcd is its continuous linear rise of with @xmath148 at fixed @xmath149 . as emphasized by peterson , walsh and zerwas @xcite ,
the fact that this tree graph behavior is preserved to all orders in perturbation theory is due to the balance in qcd between the increase of the phase space for gluon emission in the scattering processes versus the decreasing strength of the gluon coupling due to asymptotic freedom .
although the logarithmic rise of the born approximation result is preserved , the shape of @xmath150 is modified by the qcd radiation . if the running coupling constant were to freeze to a constant value at large momentum transfer , the photon structure function stops rising at high @xmath151 due to the increased phase space for gluon radiation . thus probing the qcd photon structure functions at the high momentum transfers available at the nlc
will provide a valuable test of asymptotic freedom .
hoyer , marchal , peigne , and sannino and i @xcite have challenged the common view that structure functions measured in deep inelastic lepton scattering are determined by the probability of finding quarks and gluons in the target hadron .
we show that this is not correct in gauge theory .
gluon exchange between the fast , outgoing partons and target spectators , which is usually assumed to be an irrelevant gauge artifact , affects the leading twist structure functions in a profound way .
this observation removes the apparent contradiction between the projectile ( eikonal ) and target ( parton model ) views of diffractive and small @xmath152 phenomena .
the diffractive scattering of the fast outgoing quarks on spectators in the target in turn causes shadowing in the dis cross section .
thus the depletion of the nuclear structure functions is not intrinsic to the wave function of the nucleus , but is a coherent effect arising from the destructive interference of diffractive channels induced by final - state interactions .
this is consistent with the glauber - gribov interpretation of shadowing as a rescattering effect .
similar effects can be present in the photon structure function ; _ i.e. _ , the photon structure function will be modified by rescattering of the struck quark with the photon s spectator system .
measurements from the hermes and smc collaborations show a remarkably large azimuthal single - spin asymmetries @xmath153 and @xmath154 of the proton in semi - inclusive pion leptoproduction @xmath155 dae sung hwang and ivan schmidt and i @xcite have shown that final - state interactions from gluon exchange between the outgoing quark and the target spectator system lead to single - spin asymmetries in deep inelastic lepton - proton scattering at leading twist in perturbative qcd ; _ i.e. _ , the rescattering corrections are not power - law suppressed at large photon virtuality @xmath151 at fixed @xmath156 .
the existence of such single - spin asymmetries requires a phase difference between two amplitudes coupling the proton target with @xmath157 to the same final - state , the same amplitudes which are necessary to produce a nonzero proton anomalous magnetic moment .
we have shown that the exchange of gauge particles between the outgoing quark and the proton spectators produces a coulomb - like complex phase which depends on the angular momentum @xmath158 of the proton s constituents and is thus distinct for different proton spin amplitudes .
the single - spin asymmetry which arises from such final - state interactions does not factorize into a product of distribution function and fragmentation function , and it is not related to the transversity distribution @xmath159 which correlates transversely polarized quarks with the spin of the transversely polarized target nucleon .
these effects highlight the unexpected importance of final and initial state interactions in qcd observables .
final state interactions will also lead to new types of single spin asymmetries in photon - photon collisions .
for example , in @xmath160 and @xmath161 we expect @xmath162-odd correlations of the type @xmath163 where @xmath164 is the polarization of the real photon , @xmath165 is the beam direction of an incident virtual photon , and @xmath166 is the direction of a produced quark or hadron .
the resulting asymmetry of the photon polarized normal to the production plane will be leading twist .
as in the proton target case , the single - spin asymmetry will be sensitive to orbital angular momentum in the photon wavefunction and details of the photon structure at the amplitude level .
the high energies of a photon - photon collider will make the study of double diffractive @xmath167 and semi - inclusive single diffractive processes @xmath168 in the regge regime @xmath169 interesting . here
@xmath170 if @xmath171 is taken larger than the qcd confinement scale , then one has the potential for a detailed study of fundamental pomeron processes and its gluonic composition . as in the case of large angle
exclusive @xmath15 processes , the scattering amplitude is computed by convoluting the hard scattering pqcd amplitude for @xmath172 with the vector meson distribution amplitudes .
the two gluon exchange contribution dominates in the regge regime @xcite , giving a characteristic exclusive process scaling law of order @xmath173 ginzburg , ivanov and serbo @xcite have emphasized that the corresponding @xmath174 pseudoscalar and tensor meson channels can be used to isolate the odderon exchange contribution , contributions related at a fundamental level to three gluon exchange .
in addition , the photon can diffractively dissociate into quark pairs @xmath175 by coulomb scattering on the incoming electron .
this measures the transverse derivative of the photon wavefunction @xmath176 this is the analog of the e791 experiment at fermilab @xcite which resolved the pion light - front wavefunction by diffractive dissociation @xmath177 on a nuclear target .
the results of the diffractive pion experiment are consistent with color transparency , and the momentum partition of the jets conforms closely with the shape of the asymptotic distribution amplitude , @xmath178 , corresponding to the leading anomalous dimension solution @xcite to the perturbative qcd evolution equation .
two - photon annihilation @xmath179 hadrons for real and virtual photons can thus provide some of the most detailed and incisive tests of qcd . among the processes of special interest are : 1 .
the production of four jets such as @xmath180 can test fermi - color statistics for charm quarks by checking for the interference effects of like sign quarks @xcite .
the total two - photon annihilation hadronic cross section @xmath181 which is related to the light - by - light hadronic contribution to the muon anomalous moment ; 3 . the formation of @xmath182 hadronic resonances , which can reveal exotic states such as @xmath183 hybrids and discriminate gluonium formation @xcite
. the production of the @xmath184 and @xmath185 states are essentially unexplored in qcd @xcite .
single - hadron processes such as @xmath186 , which test the transition from the anomaly - dominated pion decay constant to the short - distance structure of currents dictated by the operator - product expansion and perturbative qcd factorization theorems ; 5 .
hadron pair production processes such as @xmath187 , which at fixed invariant pair mass measures the @xmath188 crossing of the virtual compton amplitude @xcite .
when one photon is highly virtual , these exclusive hadron production channels are dual to the photon structure function @xmath189 in the endpoint @xmath190 region at fixed invariant pair mass .
the leading twist - amplitude for @xmath191 is sensitive to the @xmath192 moment of the @xmath193 distribution amplitude @xmath194 of the two - pion system @xcite , the time - like extension of skewed parton distributions .
in addition one can measure the pion charge asymmetry in @xmath195 arising from the interference of the @xmath196 compton amplitude with the time - like pion form factor @xcite . at the unphysical point @xmath197 ,
the amplitude is fixed by the low energy theorem to the hadron charge squared .
the ratio of the measured @xmath198 and @xmath199 cross sections is anomalous at threshold , a fact which may be associated with the soliton structure of baryons in qcd @xcite ; 6
. exclusive or semi - inclusive channels can also be studied by coalescence of the produced quarks .
an interesting example is higher generation final state such as @xmath200 which can have very complex angular structure @xcite .
7 . as pointed out by hwang @xcite
, one can study deeply virtual compton scattering on a photon target in @xmath42 collisions to determine the light - cone wavefunctions and other features of the photon @xcite .
at large momentum transfer , the angular distribution of hadron pairs produced by photon - photon annihilation are among the best determinants of the shape of the meson and baryon distribution amplitudes @xmath201 and @xmath202 which control almost all exclusive processes involving a hard scale @xmath203 .
the determination of the shape and normalization of the distribution amplitudes , which are gauge - invariant and process - independent measures of the valence wavefunctions of the hadrons , has become particularly important in view of their importance in the analysis of exclusive semi - leptonic and two - body hadronic @xmath12-decays @xcite .
there has also been considerable progress both in calculating hadron wavefunctions from first principles in qcd and in measuring them using diffractive di - jet dissociation .
hadron pair production from two - photon annihilation plays a crucial role in unravelling the perturbative and non - perturbative structure of qcd , first by testing the validity and empirical applicability of leading - twist factorization theorems , second by verifying the structure of the underlying perturbative qcd subprocesses , and third , through measurements of angular distributions and ratios which are sensitive to the shape of the distribution amplitudes . in effect , photon - photon collisions provide a microscope for testing fundamental scaling laws of pqcd and for measuring distribution amplitudes .
two - photon reactions , @xmath204 at large @xmath205 and fixed @xmath206 , provide a particularly important laboratory for testing qcd since these cross - channel `` compton '' processes are the simplest calculable large - angle exclusive hadronic scattering reactions .
the helicity structure , and often even the absolute normalization can be rigorously computed for each two - photon channel @xcite . in the case of meson pairs , dimensional counting predicts that for large @xmath85 , @xmath207 scales at fixed @xmath208 or @xmath206 up to factors of @xmath209 .
the angular dependence of the @xmath204 amplitudes can be used to determine the shape of the process - independent distribution amplitudes , @xmath210 .
an important feature of the @xmath211 amplitude for meson pairs is that the contributions of landshoff pitch singularities are power - law suppressed at the born level even before taking into account sudakov form factor suppression .
there are also no anomalous contributions from the @xmath190 endpoint integration region .
thus , as in the calculation of the meson form factors , each fixed - angle helicity amplitude can be written to leading order in @xmath212 in the factorized form @xmath213 ; @xmath214 $ ] : @xmath215 where @xmath216 is the hard - scattering amplitude @xmath217 for the production of the valence quarks collinear with each meson , and @xmath218 is the amplitude for finding the valence @xmath219 and @xmath220 with light - cone fractions of the meson s momentum , integrated over transverse momenta @xmath221 .
the contribution of non - valence fock states are power - law suppressed .
furthermore , the helicity - selection rules @xcite of perturbative qcd predict that vector mesons are produced with opposite helicities to leading order in @xmath212 and all orders in @xmath222 .
the dependence in @xmath149 and @xmath223 of several terms in @xmath224 is quite similar to that appearing in the meson s electromagnetic form factor .
thus much of the dependence on @xmath201 can be eliminated by expressing it in terms of the meson form factor .
in fact , the ratio of the @xmath225 and @xmath226 amplitudes at large @xmath85 and fixed @xmath227 is nearly insensitive to the running coupling and the shape of the pion distribution amplitude : @xmath228 the comparison of the pqcd prediction for the sum of @xmath229 plus @xmath230 channels with cleo data @xcite is shown in fig . [
fig : cleo ] .
the cleo data for charged pion and kaon pairs show a clear transition to the scaling and angular distribution predicted by pqcd @xcite for @xmath231 gev .
it is clearly important to measure the magnitude and angular dependence of the two - photon production of neutral pions and @xmath232 in view of the strong sensitivity of these channels to the shape of meson distribution amplitudes ( see figs .
[ fig : piangle ] and [ fig : rhoangle ] ) .
qcd also predicts that the production cross section for charged @xmath233-pairs ( with any helicity ) is much larger than for that of neutral @xmath233 pairs , particularly at large @xmath206 angles .
similar predictions are possible for other helicity - zero mesons . for an alternative model based on the qcd
handbag " diagram , see the references @xcite .
the leading - twist qcd predictions for exclusive two - photon processes such as the photon - to - pion transition form factor and @xmath38 hadron pairs are based on rigorous factorization theorems .
the data from the cleo collaboration on @xmath234 and the sum of @xmath235 and @xmath236 channels are in excellent agreement with the qcd predictions .
it is particularly compelling to see a transition in angular dependence between the low energy chiral and pqcd regimes .
the problem of setting the renormalization scale of the coupling for exclusive amplitudes is discussed in the references @xcite .
the success of leading - twist perturbative qcd scaling for exclusive processes at presently experimentally accessible momentum transfer can be understood if the effective coupling @xmath237 is approximately constant at the relatively small scales @xmath238 relevant to the hard scattering amplitudes @xcite .
evidence that the qcd coupling has an ir fixed point at small scales has been presented in the references @xcite ; the evolution of the quark distribution amplitudes in the low-@xmath238 domain also needs to be minimal .
sudakov suppression of the endpoint contributions is also strengthened if the coupling is frozen because of the exponentiation of a double logarithmic series .
conformal symmetry can be used as a guide to organize the hard scattering calculation and to determine the leading contributions to the hadron distribution amplitudes @xcite .
as noted above , the analysis of exclusive @xmath12 decays has much in common with the analysis of exclusive two - photon reactions @xcite .
for example , consider the three representative contributions to the decay of a @xmath12 meson to meson pairs illustrated in fig .
[ fig : b ] . in fig .
[ fig : b](a ) the weak interaction effective operator @xmath239 produces a @xmath240 in a color octet state .
a gluon with virtuality @xmath241 is needed to equilibrate the large momentum fraction carried by the @xmath23 quark in the @xmath242 wavefunction .
the amplitude then factors into a hard qcd / electroweak subprocess amplitude for quarks which are collinear with their respective hadrons : @xmath243 \to [ q(y ) \bar u(1-y)]_1 [ q(z ) \bar q(1-z)]_2)$ ] convoluted with the distribution amplitudes @xmath244 @xcite of the incident and final hadrons : @xmath245 here @xmath246 are the light - cone momentum fractions carried by the valence quarks .
baryon pair production in two - photon annihilation is also an important testing ground for qcd @xcite .
the calculation of @xmath216 for compton scattering requires the evaluation of 368 helicity - conserving tree diagrams which contribute to @xmath247 at the born level and a careful integration over singular intermediate energy denominators @xcite .
brooks and dixon @xcite have completed a recalculation of the proton compton process at leading order in pqcd , extending and correcting earlier work .
it is useful to consider the ratio @xmath248 since the power - law fall - off , the normalization of the valence wavefunctions , and much of the uncertainty from the scale of the qcd coupling cancel .
the scaling and angular dependence of this ratio is sensitive to the shape of the proton distribution amplitudes .
the perturbative qcd predictions for the phase of the compton amplitude phase can be tested in virtual compton scattering by interference with bethe - heitler processes @xcite .
the cross section for @xmath199 has been measured at lep by the opal and l3 collaborations @xcite .
the results are in essential agreement with perturbative qcd scaling and the predictions for the angular dependence .
however , the normalization of the pqcd three - quark predictions @xcite appear to underestimate the measured exclusive cross section by an order of magnitude
. this could be evidence for an effective quark - diquark structure of the proton wavefunction @xcite .
however , the normalization discrepancy with the leading - twist qcd predictions could be due to several reasons .
( 1 ) note that the rate is proportional to @xmath249 the qcd coupling is assumed to have a nominal value @xmath250 @xcite which could be an underestimate of the effective value of the running qcd coupling at the relevant scales . ( 2 )
the leading order perturbative qcd predictions lead to real amplitudes .
the imaginary parts which appear from the analytic continuation of the qcd coupling and final state qcd interactions can also give a significant correction , as is known to be the case in @xmath251 @xcite .
( 3 ) the relatively slow fall - off of the pauli form factor as observed in polarization transfer experiments at jefferson lab , suggests significant higher twist contributions .
all of these arguments suggest that the most reliable test of qcd are obtained from the ratio : @xmath252 .
in addition measurements of the proton polarization single - spin asymmetry provides important information on the phase of the timelike pair production amplitudes @xcite .
it is also interesting to measure baryon and isobar pair production in two photon reactions near threshold .
ratios such as @xmath253 can be as large as @xmath254 in the quark model since the three - quark wavefunction of the @xmath255 is expected to be symmetric .
such large ratios would not be expected in soliton models @xcite in which intermediate multi - pion channels play a major role .
pobylitsa _ et al .
_ @xcite have shown how the predictions of perturbative qcd can be extended to processes such as @xmath256 where the pion is produced at low velocities relative to that of the @xmath257 or @xmath258 by utilizing soft pion theorems in analogy to soft photon theorems in qed .
the distribution amplitude of the @xmath259 composite is obtained from the proton distribution amplitude from a chiral rotation .
a test of this procedure in inelastic electron scattering at large momentum transfer @xmath260 and small invariant @xmath261 mass has been remarkably successful .
many tests of the soft meson procedure are possible in multiparticle @xmath16 and @xmath15 final states .
one of the formidable challenges in qcd is the calculation of non - perturbative wavefunctions of hadrons from first principles .
the calculations of the pion distribution amplitude by dalley @xcite and by burkardt and seal @xcite using light - cone and transverse lattice methods is particularly encouraging .
the predicted form of @xmath262 is somewhat broader than but not inconsistent with the asymptotic form favored by the measured normalization of @xmath263 and the pion wavefunction inferred from diffractive di - jet production .
clearly much more experimental input on hadron wavefunctions is needed , particularly from measurements of two - photon exclusive reactions into meson and baryon pairs at the high luminosity @xmath12 factories .
for example , as shown in fig .
[ fig : piangle ] , the ratio @xmath264 is particularly sensitive to the shape of pion distribution amplitude . at fixed pair mass , and high photon virtuality , one can study the distribution amplitude of multi - hadron states @xcite .
two - photon annihilation will provide much information on fundamental qcd processes such as deeply virtual compton scattering and large angle compton scattering in the crossed channel .
i have also emphasized the interrelation between the wavefunctions measured in two - photon collisions and the wavefunctions needed to study exclusive @xmath12 and @xmath265 decays .
much of the most interesting two - photon annihilation physics is accessible at low energy , high luminosity @xmath16 colliders , including measurements of channels important in the light - by - light contribution to the muon @xmath2662 and the study of the transition between threshold production controlled by low - energy effective chiral theories and the domain where leading - twist perturbative qcd becomes applicable .
the threshold regime of hadron production in photon - photon and @xmath16 annihilation , where hadrons are formed at small relative velocity , is particularly interesting as a test of low energy theorems , soliton models , and new types of resonance production .
such studies will be particularly valuable in double - tagged reactions where polarization correlations , as well as the photon virtuality dependence , can be studied .
the light - front quantization of gauge theories can be carried out in an elegant way using the dirac method to impose the light - cone gauge constraint and eliminate dependent degrees of freedom @xcite .
unlike the case in equal - time quantization , the vacuum remains trivial .
since only physical degrees of freedom appear , unitarity is maintained .
one can verify the qcd ward identities for the physical light - cone gauge and compute the qcd @xmath267 function .
srivastava and i @xcite have extended the light - front quantization procedure to the standard model .
the spontaneous symmetry breaking of the gauge symmetry is due to a zero mode of the scalar field rather than vacuum breaking .
the goldstone component of the scalar field provides mass to the @xmath79 and @xmath268 gauge bosons as well as completing its longitudinal polarization .
the resulting theory is free of faddeev - popov ghosts and is unitary and renormalizable .
the resulting rules give an elegant new way to compute standard model processes using light - front hamiltonian theory .
the light - front method suggests the possibility of developing an event amplitude generator " for high energy processes such as photon - photon collisions by calculating amplitudes for specific parton spins using light - front time - ordered perturbation theory @xcite .
the positivity of the @xmath269 light - front momenta greatly constrains the number of contributing light - front time orderings .
since particle states are labelled by the spin projection @xmath270 , total angular momentum constraints are readily implemented .
the renormalized amplitude can be obtained diagram by diagram by using the alternating denominator " method which automatically subtracts the relevant counterterm .
the dlcq method also provides a simple way to discretized the light - front momentum variables , while maintaining frame - independence .
the resulting renormalized amplitude can be convoluted with the light - front wavefunctions to simulate hadronization and hadron matrix elements .
i thank professor clem heusch for organizing this workshop . work supported by the department of energy under contract number de - ac03 - 76sf00515 .
s. j. brodsky and p. m. zerwas , _ nucl .
instrum .
meth . _ * a355 * , 19 ( 1995 ) [ arxiv : hep - ph/9407362 ] .
this report is an update of s. j. brodsky , _ int .
j. mod .
phys . _ * a18 * , 2871 ( 2003 ) [ arxiv : hep - ph/0204197 ] . s. j. brodsky , t. kinoshita and h. terazawa , phys . rev .
d * 4 * , 1532 ( 1971 ) .
v. m. budnev , i. f. ginzburg , g. v. meledin and v. g. serbo , phys .
rept . * 15 * , 181 ( 1974 ) .
i. f. ginzburg , g. l. kotkin , v. g. serbo and v. i. telnov , _ jetp lett . _ * 34 * , 491 ( 1981 ) [ _ pisma zh . eksp.teor
. fiz . _ * 34 * , 514 ( 1981 ) ] .
v. telnov , _
17th international conference on high - energy accelerators ( heacc 98 ) , dubna , russia , 7 - 12 sep 1998 _ [ arxiv : hep - ex/9810019 ] .
d. asner , s. boege , j. early , j. gronberg , k. skulina , k. van bibber and t. markiewicz , pac-2001-fpah055 _ presented at ieee particle accelerator conference ( pac2001 ) , chicago , illinois , 18 - 22 jun 2001_. d. asner _
et al . _ , arxiv : hep - ex/0111056 . s. j. brodsky , slac - pub-6314 _ presented at the 2nd international workshop on physics and experiments with linear e+ e- colliders , waikoloa , hi , 26 - 30 apr 1993_. m. s. chanowitz , _ nucl
. instrum .
_ * a355 * , 42 ( 1995 ) [ arxiv : hep - ph/9407231 ] .
j. f. gunion and h. e. haber , _ phys .
_ * d48 * , 5109 ( 1993 ) . s. d. drell and a. c. hearn , _ phys . rev .
lett . _ * 16 * , 908 ( 1966 ) .
s. b. gerasimov , _ sov .
j. nucl .
phys . _ * 2 * , 430 ( 1966 ) [ yad .
598 ( 1966 ) ] .
s. j. brodsky and j. r. hiller , _ phys .
_ * d46 * , 2141 ( 1992 ) .
w. k. tung , _ phys .
_ * 176 * , 2127 ( 1968 ) .
s. j. brodsky and s. d. drell , _ phys .
rev . _ * d22 * , 2236 ( 1980 ) .
e. yehudai , _ phys .
rev . _ * d44 * , 3434 ( 1991 ) . j. layssac and f. m. renard ,
rev . _ * d64 * , 053018 ( 2001 ) . g. jikia and a. tkabladze ,
_ * b332 * , 441 ( 1994 ) [ arxiv : hep - ph/9312274 ] . g. j. gounaris , p. i. porfyriadis and f. m. renard , _ eur .
j. _ * c20 * , 659 ( 2001 ) [ arxiv : hep - ph/0103135 ] . g. v. jikia , _ nucl
_ * b374 * , 83 ( 1992 ) .
e. boos , m. dubinin , a. pukhov , m. sachwitz and h. j. schreiber , _ eur . phys .
j. _ * c21 * , 81 ( 2001 ) [ arxiv : hep - ph/0104279 ] .
s. j. brodsky , t. g. rizzo and i. schmidt , _ phys .
rev . _ * d52 * , 4929 ( 1995 )
. t. g. rizzo , [ arxiv : hep - ph/9907395 ] .
i. f. ginzburg , g. l. kotkin , s. l. panfil and v. g. serbo , _ nucl .
_ * b228 * , 285 ( 1983 ) .
see also s. y. choi and f. schrempp , _ phys .
* b272 * , 149 ( 1991 ) ; m. raidal , _ nucl .
phys . _ * b441 * , 49 ( 1995 ) , [ arxiv : hep - ph/9411243 ] .
s. j. brodsky , t. a. degrand , j. f. gunion and j. h. weis , _ phys .
rev . _ * d19 * , 1418 ( 1979 ) .
m. drees and t. han , _ phys .
lett . _ * 76 * , 3076 ( 1996 ) [ arxiv : hep - ph/9512361 ] .
s. j. brodsky , j. rathsman and c. merino , _ phys .
lett . _
* b461 * , 114 ( 1999 ) . s. j. brodsky , f. hautmann and d. e. soper , _ phys .
rev . _ * d56 * , 6957 ( 1997 ) .
j. bartels , a. de roeck and h. lotter , _ phys .
_ * b389 * , 742 ( 1996 ) .
a. v. lipatov and n. p. zotov , _ liverpool 2000 , deep inelastic scattering , _ pp .
157 - 158 v. s. fadin and l. n. lipatov , _ phys .
lett . _ * b429 * , 127 ( 1998 ) [ arxiv : hep - ph/9802290 ] .
s. j. brodsky , v. s. fadin , v. t. kim , l. n. lipatov and g. b. pivovarov , _ jetp lett .
_ * 70 * , 155 ( 1999 ) [ arxiv : hep - ph/9901229 ] .
s. j. brodsky , p. hoyer , n. marchal , s. peigne and f. sannino , [ arxiv : hep - ph/0104291 ] .
s. j. brodsky , t. kinoshita and h. terazawa , _ phys .
lett . _ * 27 * , 280 ( 1971 )
. t. f. walsh , _ phys .
lett . _ * b36 * , 121 ( 1971 ) .
m. krawczyk , _ ambleside 2000 , structure and interactions of the photon _ [ arxiv : hep - ph/0012179 ] .
e. witten , _ nucl .
phys . _ * b120 * , 189 ( 1977 ) . c. peterson , p. m. zerwas and t. f. walsh , _ nucl
* b229 * , 301 ( 1983 ) . s. j. brodsky , d. s. hwang and i. schmidt , _ phys .
_ * b530 * , 99 ( 2002 ) . v.
l. chernyak and i. r. zhitnitsky , _ nucl .
_ * b222 * , 382 ( 1983 ) . i. f. ginzburg , d. y. ivanov and v. g. serbo , _ phys . atom .
* 56 * , 1474 ( 1993 ) [ yad .
, 45 ( 1993 ) ] .
d. ashery , [ e791 collaboration ] , [ hep - ex/9910024 ] . s. j. brodsky and g. p. lepage , _ phys
* 53 * , 545 ( 1979 ) ; _ phys .
lett . _ * 87b * , 359 ( 1979 ) ; g. p. lepage and s. j. brodsky , _ phys .
_ * d22 * , 2157 ( 1980 ) .
m. r. pennington , _ ambleside 2000 , structure and interactions of the photon _ 388 - 393 , [ arxiv : hep - ph/0009267 ] .
m. acciarri _ et al .
_ [ l3 collaboration ] , _ phys . lett . _ * b501 * , 173 ( 2001 ) [ hep - ex/0011037 ] .
a. heister _ et al .
_ [ aleph collaboration ] , _ phys .
lett . _ * b530 * , 56 ( 2002 ) . s. j. brodsky and g. p. lepage ,
_ * d24 * , 1808 ( 1981 ) .
d. muller , d. robaschik , b. geyer , f. m. dittes , and j. horejsi , _ fortsch .
phys . _ * 42 * , 101 ( 1994 ) , [ hep - ph/9812448 ] .
m. diehl , t. gousset and b. pire , _ phys .
_ * d62 * , 073014 ( 2000 ) [ hep - ph/0003233 ]
. m. diehl , t. gousset , b. pire and o. teryaev , _ phys .
lett . _ * 81 * , 1782 ( 1998 ) [ arxiv : hep - ph/9805380 ] . s. j. brodsky and g. p. lepage , _ phys
. rev . _ * d24 * , 2848 ( 1981 ) . j. dominick _ et al . _
[ cleo collaboration ] , _ phys .
_ d * d50 * , 3027 ( 1994 ) ; j. boyer _ et al .
_ , _ phys .
lett . _ * 56 * , 207 ( 1980 ) ; tpc / two gamma collaboration ( h. aihara _
_ ) , _ phys . rev .
lett . _ * 57 * , 404 ( 1986 ) .
s. j. brodsky , s. menke , c. merino and j. rathsman , _ phys .
* d67 * , 055008 ( 2003 ) [ arxiv : hep - ph/0212078 ] .
s. j. brodsky , y. frishman , g. p. lepage , and c. sachrajda , _ phys .
* 91b * , 239 ( 1980 ) .
g. r. farrar and h. zhang , _ phys .
lett . _ * 65 * , 1721 ( 1990 ) . a. s. kronfeld and b. nizic , _ phys .
* d44 * , 3445 ( 1991 ) .
p. a. guichon and m. vanderhaeghen , _ prog .
_ * 41 * , 125 ( 1998 ) .
g. abbiendi _ et al . _
[ opal collaboration ] , _ eur .
j. _ * c28 * , 45 ( 2003 ) [ arxiv : hep - ex/0209052 ] .
p. achard _ et al .
_ [ l3 collaboration ] , phys .
b * 571 * , 11 ( 2003 ) [ arxiv : hep - ex/0306017 ] .
g. r. farrar , e. maina and f. neri , nucl .
b * 259 * , 702 ( 1985 ) [ erratum - ibid .
b * 263 * , 746 ( 1986 ) ] .
|
high intensity back - scattered laser beams will allow the efficient conversion of a substantial fraction of the incident lepton energy into high energy photons , thus significantly extending the physics capabilities of an @xmath0 linear collider .
the annihilation of two photons produces @xmath1 final states in virtually all angular momentum states .
an important physics measurement is the measurement of the higgs coupling to two photons .
the annihilation of polarized photons into the higgs boson determines its fundamental @xmath2 coupling as well as determining its parity .
other novel two - photon processes include the two - photon production of charged pairs @xmath3 @xmath4 , @xmath5 and supersymmetric squark and slepton pairs .
the one - loop box diagram leads to the production of pairs of neutral particles such as @xmath6 and @xmath7 at the next order one can study higgstrahlung processes , such as @xmath8 since each photon can be resolved into a @xmath4 pair , high energy photon - photon collisions can also provide a remarkably background - free laboratory for studying possibly anomalous @xmath9 collisions and annihilation . in the case of qcd , each photon can materialize as a quark anti - quark pair which interact via multiple gluon exchange .
the diffractive channels in photon - photon collisions allow a novel look at the qcd pomeron and odderon
. the @xmath10 odderon exchange contribution can be identified by looking at the heavy quark asymmetry . in the case of @xmath11 collisions
, one can measure the photon structure functions and its various components .
exclusive hadron production processes in photon - photon collisions provide important tests of qcd at the amplitude level , particularly as measures of hadron distribution amplitudes which are also important for the analysis of exclusive semi - leptonic and two - body hadronic @xmath12-decays .
slac - pub-10410 + april 2004 * high energy photon
photon + collisions at a linear collider * stanley j. brodsky + stanford linear accelerator center , stanford university + stanford , california 94309 + e - mail : [email protected] invited talk , presented at the + 5th international workshop on electron - electron interactions at tev energies + santa cruz , california + 1214 december 2003
|
it was suggested by wmap that the amplitude of cosmic microwave background ( cmb ) anisotropies has a dipolar directional dependence @xcite .
this anomalous signature has been recently confirmed by planck @xcite .
although it is possible that this anisotropy is a foreground effect , it is an interesting question whether or not we can explain this feature consistently within the paradigm of the inflationary universe .
the basic issue is that a dipolar modulation of the amplitude of the curvature perturbation requires a seed dipolar perturbation .
this seed dipolar perturbation itself is known to produce quadrupole and octupole moments of cmb anisotropy , known as the grishchuk - zeldovich ( gz ) effect @xcite .
various scenarios that explain the dipolar statistical anisotropy have been proposed @xcite and various features have been investigated @xcite .
the proposal in ref .
@xcite advocates the use of the super - curvature perturbation @xcite produced in the open inflation scenario as the seed for the dipolar modulation .
the amplitudes of monopole , dipole , and quadrupole fluctuations @xmath0 @xmath1 caused by the super - curvature mode of a free scalar field in the small mass limit are , respectively , evaluated by taking the small radius limit of the corresponding normalized mode functions as @xcite @xmath2 where @xmath3 and @xmath4 are the radii of the last scattering surface and the curvature radius in the comoving conformal coordinate , and @xmath5 and @xmath6 are the mass and the hubble expansion rate in the false vacuum before bubble nucleation .
thus , one can basically avoid the constraint due to the gz effect by choosing @xmath7 to be sufficiently small .
( although in ref .
@xcite a constraint on the amplitude of the super - curvature perturbation was imposed , such a condition on the magnitude of fluctuations beyond the current horizon scale is usually thought to be unnecessary in the context of the inflationary universe .
the quasi - open inflation scenario @xcite is a typical example . )
however , we suspect that one important constraint has been overlooked . it is a genuine non - linear effect
. if the amplitude of a dipole moment is large , a non - negligible quadrupole moment is naturally induced by non - linearity .
taking into account the constraint from the amplitude of this induced quadrupole moment , we find that almost all models proposed so far are almost excluded by observations . in this letter , we propose a simple alternative scenario that can naturally evade this constraint .
in refs .
@xcite , in order to explain the observed large statistical anisotropy , the simplest curvaton scenario was studied .
the only difference from the ordinary curvaton scenario is that the curvaton has a large - magnitude super - horizon fluctuation . in ref .
@xcite , this super - horizon fluctuation was identified with a super - curvature mode , whose amplitude can naturally be large compared with sub - curvature modes if the creation of the one - bubble open universe occurs in the sea of false vacuum whose energy scale is much higher than that inside the bubble . as this generation mechanism of a large super - horizon fluctuation seems most natural , we focus on possible scenarios in this context .
first let us discuss the model based on the simplest curvaton scenario .
we denote the super - curvature fluctuation of the curvaton as @xmath8 .
more precisely , its distribution measured at the current horizon radius is @xmath9 , where the axis of the angular coordinates @xmath10 is aligned to the direction of the dipolar statistical anisotropy .
a typical amplitude of @xmath8 is given by @xmath11 in eq .
( [ scamp ] ) , whereas @xmath12 represents the ordinary mode fluctuation whose amplitude is @xmath13 , where @xmath14 is the hubble rate at the time when the relevant length scale crosses the horizon scale during the inflation .
we assume @xmath15 .
the inflation inside the nucleated bubble is driven by an inflaton field @xmath16 , but the details of the model concerning the @xmath16-field are not important at all . for simplicity , we assume that the super - curvature perturbation is completely dominated by the curvaton field @xmath17 and that the inflaton @xmath16 is in the slow - roll regime when the relevant length scale crosses the horizon
. the observed curvature perturbation can be evaluated by using the @xmath18-formula ( see , for example , ref .
@xcite ) , @xmath19 we assume that the leading - order power spectrum of the curvature perturbation is given by the terms with the first derivatives of @xmath20 , except for extremely low @xmath21 modes like dipole or quadrupole .
then , we have @xmath22 the dipolar statistical anisotropy is caused by the term including @xmath8 in eq .
( [ deltan ] ) .
the leading term is given by @xmath23 the ratio of these two quantities is constrained by the observation as @xcite .
we thank m. bucher for making this point clear to us .
see ref .
@xcite for a more detailed discussion . ]
@xmath24 here we neglected the non - linear terms in the sachs - wolfe relation between @xmath25 and the observed temperature fluctuation @xmath26 since the effect due to this non - linearity is suppressed by @xmath27 and it can be absorbed by the redefinition of @xmath28 .
on the other hand , this model predicts the local - type non - gaussianity , @xmath29 where we have used eq .
( [ statng ] ) in the second and third equalities .
furthermore , a quadrupolar anisotropy is induced by the second - order effect of @xmath8 which originates from the term @xmath30 in eq .
( [ deltan ] ) .
its magnitude is estimated as @xmath31 we note that this is not the so - called grishchuk - zeldovich ( gz ) effect @xcite .
we also mention that although the functional dependence of the quadrupole is similar to that obtained in ref .
@xcite , our estimate gives a more robust bound while that in ref .
@xcite is based on a rather crude argument concerning the convergence of perturbative expansion which can not be directly compared with the observational data . as mentioned earlier ,
if @xmath32 is composed of the super - curvature mode in the open inflation scenario , the constraint from the gz effect can be evaded .
combining eqs .
( [ fnl ] ) and ( [ ampa20 ] ) , we find @xmath33 thus one can not make both the quadrupole and local - type non - gaussianity very small simultaneously .
this gives a bound slightly tighter and more robust than that obtained in eq .
( 59 ) of ref .
@xcite . here
we note that the contribution of @xmath34 does not relax the above constraint . from the observed quadrupole anisotropy , we may bound the induced quadrupole as @xmath35 where the observed magnitude of the squared quadrupole temperature anisotropy is @xmath36 and we inserted a threshold number @xmath37 of order unity .
this implies the constraint @xmath38 which should be satisfied in any kind of model . inserting eq .
( [ ampbeta ] ) into eq .
( [ quadfnl ] ) , we obtain @xmath39 the required value of @xmath40 is already difficult to reconcile with the planck observation ( @xmath41 ) @xcite .
we emphasize that this constraint differs from the one discussed in the literature @xcite .
if we require the above constraint on the non - gaussianity to be satisfied , we should take @xmath42 .
this implies , from eq .
( [ ampbeta ] ) , that @xmath43 must be accidentally cancelled by the completely independent cmb quadrupole contribution to give the observed small value of @xmath44 .
secondly , the value of @xmath8 must be fine - tuned so that @xmath43 is almost exactly @xmath45 .
thus it is manifest that any models of this kind require extreme fine - tuning .
0 furthermore , when the temperature perturbation is dominated by the curvaton , there should be the dipole temperature perturbation @xmath46 caused by @xmath8 , which means that the peculiar velocity field relative to the cmb rest frame is shifted by the order of @xmath47km / sec .
this will be in conflict with the current observation of the magnitude and the direction of the cmb dipole , which is consistent with the clustering dipoles determined by the distribution of galaxies @xcite .
we found it difficult to construct a viable model to explain the dipolar statistical anisotropy in the context of simple curvaton models using the non - linear coupling @xmath28 .
one might wonder if using the cross term @xmath48 instead of @xmath28 could improve the situation .
it turns out that this possibility for generating the dipolar statistical anisotropy is also almost ruled out .
this is because one can show that the amplitude of the dipole anisotropy @xmath49 inevitably becomes too large if the dipolar statistical anisotropy is generated by the cross term @xmath48 . in order to produce the dipolar statistical anisotropy , we must have @xmath50 however , within one expansion time during the inflation , @xmath51 , the average value of the inflaton @xmath16 changes by @xmath52 , where a dot denotes the cosmic time derivative . therefore
, as long as the cross term @xmath48 exists , this shift of the background value of @xmath16 generates @xmath53 , @xmath54 hence the dipole temperature perturbation caused by @xmath8 becomes @xmath55 this constraint is one order of magnitude larger than @xmath56 , which is obtained from the analysis of the doppler boosting by planck @xcite .
we note that the above argument also applies to a single - field model , in which the role of the @xmath17-field is also played by the inflaton @xmath16 . for this case , @xmath48 and @xmath8 in the above argument
are replaced with @xmath57 and @xmath58 , respectively .
consequently , single - field models are also in conflict with observation .
if we introduce the third field , it may be possible to construct models that generate the dipolar statistical anisotropy using the non - linear terms in the @xmath18 formula . however
, as it seems difficult to avoid fine - tuning in such models , we do not pursue this direction further .
the above consideration teaches us that it is not easy to generate the dipolar statistical anisotropy by using cross terms in the expansion of the @xmath18 formula , and simultaneously satisfy the observational constraint that the local - type non - gaussianity is small .
therefore we pursue a different approach .
we consider the possibility of modulating the amplitude of the quantum fluctuations at horizon crossing without affecting the inflationary dynamics .
first we examine a model in which the hubble expansion rate is modulated , but the number of @xmath59-folds @xmath20 is not affected by this modulation .
a model lagrangian is given by @xmath60 here the @xmath61-field is supposed to have a large amplitude of the super - curvature perturbation , but not to affect the inflationary dynamics .
namely , it is assumed that the curvature perturbation is generated by the inflaton @xmath16 .
this means @xmath62 . on the other hand ,
the hubble expansion rate @xmath14 is a function of not only @xmath16 but also @xmath61 , @xmath63 . for @xmath62 ,
this implies @xmath64 . in other words , if we draw contours of @xmath65 on the @xmath66 plane , @xmath14 on the contour of a given value of @xmath20 may depend on @xmath61 .
then the inflaton fluctuation at around @xmath67 , @xmath68 also depends on @xmath61 , where @xmath69 is the value of the @xmath59-folding number at which the comoving scale of the current hubble horizon size left the horizon during inflation . at first sight , for a potential of the form given by eq .
( [ prodpot ] ) , this model seems viable because the @xmath59-folding number @xmath20 becomes @xmath61-independent , @xmath70 under the slow - roll approximation for @xmath16 , where the subscript @xmath16 denotes the @xmath16-derivative , @xmath71 , and @xmath72 is the value of @xmath16 at the end of inflation . unfortunately , however , the above formula for @xmath20 has non - negligible higher - order corrections .
in fact , there always exists a correction of the form @xmath73^\phi_{\phi_{\rm f}}\,,\end{aligned}\ ] ] which depends on @xmath74 .
therefore the curvature perturbation induced by the super - curvature perturbation @xmath75 becomes as large as @xmath76 .
hence this model produces too large an amplitude of the dipole anisotropy @xmath77 , which is already ruled out observationally @xcite .
finally we propose a viable model . as before
, we assume the existence of a field @xmath61 that carries a super - curvature perturbation @xmath75 .
in addition , we introduce a curvaton @xmath17 .
the idea is to modulate the dynamics of the curvaton @xmath17 by using the @xmath61-field without affecting the inflationary dynamics .
this can be realized by assuming that the kinetic term of @xmath17 depends on @xmath61 .
a model lagrangian takes the form @xmath78 owing to the function @xmath79 , the amplitude of the fluctuations of @xmath17 is modulated as @xmath80 . with an appropriate choice of @xmath79 , this model can easily explain the dipolar statistical anisotropy .
assuming the hierarchy @xmath81 inside the bubble , one can make @xmath61 decay quickly during inflation so that it will not affect the dynamics during or after inflation except for the modulation of the fluctuation amplitude @xmath12 near or beyond the current hubble horizon scale .
we may also assume that the final curvature perturbation is dominated by the curvaton @xmath17 and ignore the fluctuations of the inflaton @xmath16 .
then , we can easily suppress the local - type non - gaussianity .
in this model the origin of the modulation is assumed to die out during the inflation .
therefore , the amplitude of the dipolar statistical anisotropy becomes smaller on smaller length scales .
this seems to be favored by observational data @xcite .
we would like to thank m. bucher , j. garriga , r. gobbetti and j. soda for insightful discussions and comments .
this work was supported in part by the grant - in - aid for scientific research ( nos .
21244033 , 21111006 , 24103006 and 24103001 ) , and in part by phy-0855447 from the national science foundation .
sk is grateful for the hospitality extended to her by the yukawa institute for theoretical physics ( yitp ) , kyoto university , during her stay when this work was initiated .
this work was completed during the molecule - type yitp workshop
the cmb and theories of the primordial universe " .
j. hoftuft , h. k. eriksen , a. j. banday , k. m. gorski , f. k. hansen and p. b. lilje , astrophys .
j. * 699 * ( 2009 ) 985 [ arxiv:0903.1229 [ astro-ph.co ] ] ; c. l. bennett , r. s. hill , g. hinshaw , d. larson , k. m. smith , j. dunkley , b. gold and m. halpern _ et al .
_ , astrophys . j. suppl .
* 192 * ( 2011 ) 17 [ arxiv:1001.4758 [ astro-ph.co ] ] .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1303.5083 [ astro-ph.co ] .
l. p. grishchuk and i. b. zeldovich , sov
* 22 * ( 1978 ) 125 .
a. l. erickcek , m. kamionkowski and s. m. carroll , phys .
d * 78 * ( 2008 ) 123520 [ arxiv:0806.0377 [ astro - ph ] ] .
a. l. erickcek , s. m. carroll and m. kamionkowski , phys .
d * 78 * ( 2008 ) 083012 [ arxiv:0808.1570 [ astro - ph ] ] .
a. r. liddle and m. cortes , arxiv:1306.5698 [ astro-ph.co ] . a. l. erickcek , c. m. hirata and m. kamionkowski , phys . rev .
d * 80 * , 083507 ( 2009 ) [ arxiv:0907.0705 [ astro-ph.co ] ] .
l. wang and a. mazumdar , phys .
d 88 , * 023512 * ( 2013 ) [ arxiv:1304.6399 [ astro-ph.co ] ] .
z. -g .
liu , z. -k .
guo and y. -s .
piao , arxiv:1304.6527 [ astro-ph.co ] . j. mcdonald , arxiv:1305.0525 [ astro-ph.co ] . m. h. namjoo , s. baghram and h. firouzjahi , arxiv:1305.0813 [ astro-ph.co ] . a. mazumdar and l. wang , arxiv:1306.5736 [ astro-ph.co ] . g. damico , r. gobbetti , m. kleban and m. schillo , arxiv:1306.6872 [ astro-ph.co ] . y. -f . cai , w. zhao and y. zhang , arxiv:1307.4090 [ astro-ph.co ] . d. h. lyth , arxiv:1304.1270 [ astro-ph.co ] .
a. a. abolhasani , s. baghram , h. firouzjahi and m. h. namjoo , arxiv:1306.6932 [ astro-ph.co ] .
z. chang , x. li and s. wang , arxiv:1307.4542 [ astro-ph.co ] .
m. sasaki , t. tanaka and k. yamamoto , phys .
2979 ( 1995 ) [ gr - qc/9412025 ] . j. garcia - bellido , j. garriga and x. montes , phys .
d * 57 * , 4669 ( 1998 ) [ hep - ph/9711214 ]
. t. tanaka , t. suyama and s. yokoyama , class .
* 27 * , 124003 ( 2010 ) [ arxiv:1003.5057 [ astro-ph.co ] ] .
j. m. maldacena , jhep * 0305 * , 013 ( 2003 ) . c. m. hirata , `` constraints on cosmic hemispherical power anomalies from quasars , '' jcap * 0909 * ( 2009 ) 011 [ arxiv:0907.0703 [ astro-ph.co ] ] .
s. flender and s. hotchkiss , arxiv:1307.6069 [ astro-ph.co ] .
|
the presence of a dipolar statistical anisotropy in the spectrum of cosmic microwave background ( cmb ) fluctuations was reported by the wilkinson microwave anisotropy probe ( wmap ) , and has recently been confirmed in the planck 2013 analysis of the temperature anisotropies . at the same time , the planck 2013 results report a stringent bound on the amplitude of the local - type non - gaussianity . we show that the non - linear effect of the dipolar anisotropy generates not only a quadrupole moment in the cmb but also a local - type non - gaussianity .
consequently , it is not easy to build models having a large dipolar modulation and at the same time a sufficiently small quadrupole and level of local bispectral anisotropy to agree with the present data .
in particular , most models proposed so far are almost excluded , or are at best marginally consistent with observational data .
we present a simple alternative scenario that may explain the dipolar statistical anisotropy while satisfying the observational bounds on both the quadrupole moment and local - type non - gaussianity .
|
underground muons are the remnants of the air showers initiated by the collisions of primary cosmic rays with air nuclei
. these secondary cosmic ray muons can easily cross large amounts of matter and penetrate into underground laboratories .
the energy spectra of underground muons depend on the energy spectra and composition of primary cosmic rays , on their interactions with air nuclei and on the muon energy loss in the rock . in this paper the measurement of the energy spectra of underground single and double muons , performed with a @xmath1 transition radiation detector ( trd ) , that was installed in the empty upper part of the macro detector @xcite ,
is presented .
an analysis using part of the single muon data sample was already published @xcite .
the final sample is approximately 10 times larger and a correct evaluation of the systematics has been performed , thus allowing us to make more reliable conclusions about the spectra and the average residual energy of single and double muon events . the trd sub - detector , described in sec .
[ sec : trd ] , uses for triggering purpose and for the measurement of the event multiplicity the larger area ( @xmath19 ) of the streamer tube and scintillation counters systems . the data selection is described in sec .
[ sec : data ] . to evaluate the muon energy spectra from the trd data , we used two complementary approaches described in sec .
[ sec : spectra ] . in the first case ,
an unfolding procedure was applied , while in the second case a parameterization for the underground muon energy spectrum was assumed and a best fit of the spectral parameters was carried out . to study how the energy spectra of underground muons are related to the primary cosmic ray spectrum and composition , a dedicated monte carlo simulation , described in sec .
[ sec : mc ] , was performed .
the results of the comparisons between data and monte carlo are discussed in sec .
[ sec : discussion ] .
the `` all - particle '' flux of the primary cosmic radiation can be described by an inverse power law energy spectrum @xcite , with differential flux given by : @xmath20 where @xmath21 for @xmath22 , @xmath23 for @xmath24 and @xmath25 for @xmath26 . from the primary spectrum ,
it is possible to evaluate the energy spectrum at the earth surface of secondary uncorrelated muons , which is given , with good approximation , by @xcite : @xmath27 where @xmath28 is the muon energy at the surface , @xmath29 and @xmath30 .
the first and the second term in parenthesis of equation [ eq : musurf ] represent the contributions of muons from @xmath31 and @xmath32 decays , respectively . in the limit of high energies , an approximate expression of the muon surface energy spectrum
has the simple form : @xmath33 where @xmath34 .
the surface muon spectral index is therefore related to the primary cosmic ray spectral index .
the energy spectrum of underground muons can be derived taking into account the process of energy loss in the rock , which is assumed to have the form : @xmath35 where @xmath36 is a thin rock slab ( usually in @xmath37 ) , @xmath38 is the contribution from the ionization energy loss and @xmath39 is the contribution from radiative processes ( bremsstrahlung , pair production and muon hadroproduction ) . the parameters @xmath38 and @xmath40 are functions of the muon energy , but for practical purposes can be assumed as constants @xcite .
the quantity @xmath41 is called the critical energy and is defined as the energy value above which the radiative processes become dominant . with the above assumptions ,
the general solution of equation [ eq : dedx ] is : @xmath42 where @xmath43 is the muon energy after crossing the rock slant depth @xmath44
. the underground muon energy spectrum can be thus obtained from equations [ eq : musurf1 ] and [ eq : muund ] using the following relation : @xmath45_{\mathcal{e } = \mathcal{e } ( e_{\mu } ) } \cdot \frac{d\mathcal{e } } { de_{\mu}}\ ] ] and is given by : @xmath46 from eq .
[ eq : undspec ] , the average underground muon energy at depth @xmath11 is : @xmath47 and its asymptotic value is @xmath48 . at great depths @xmath11 , the underground muon energy spectrum given by eq .
[ eq : undspec ] is almost flat for @xmath49 , and then decreases with energy .
transition radiation ( tr ) is the process of the emission of x - ray photons occurring when an ultrarelativistic charged particle crosses the boundary between two materials with different dielectric constants .
the most important features of tr are that the tr yield is roughly proportional to the lorentz factor @xmath50 of the radiating particle over a wide range of @xmath50 , and the emission probability of a tr x - ray is of the order of @xmath51 .
if the rest mass @xmath52 of the radiating particle is known , a measurement of its lorentz factor @xmath50 also allows one to evaluate the energy as @xmath53 .
trds can provide energy measurements over ranges typically spanning one order of magnitude . due to the characteristic dependence of the tr yield on the lorentz factor @xmath50 , trds were proposed @xcite for the measurement of energies of underground cosmic ray muons in the @xmath54 region .
the trd operated inside the macro @xcite detector in the gran sasso underground laboratory ( lngs ) collected data from april 1995 to june 2000 .
it was designed to be sensitive to the energy region between @xmath55 and @xmath56 .
the detector consisted of three modules covering a total horizontal area of @xmath57 . each module
was composed by eleven @xmath58 thick radiator layers , interleaved by ten planes of @xmath59 proportional tubes @xmath60 long and with a square cross section of @xmath61 .
a detailed description of the detector is given in @xcite .
the radiator material ( ethafoam 220 ) contains cells of @xmath62 wall thickness and @xmath63 spacing , ensuring a threshold lorentz factor @xmath64 and a saturation lorentz factor @xmath65 , that correspond to the muon energy range between @xmath66 and @xmath67 .
the proportional tubes were filled with an @xmath68 gas mixture and were operated at a gain of @xmath69 .
they were equipped with a cluster counting read - out electronics .
wire signals were sharply differentiated and compared to a threshold corresponding to an energy deposit of @xmath70 . in this way it was possible to discriminate @xmath71-ray background from x - ray photoelectrons producing pulses exceeding the threshold amplitude . for each event
the pulses with amplitude greater than the threshold ( `` hits '' ) were counted in all the proportional tubes .
the third trd module was partially equipped with ten @xmath72 thick aluminum foils of @xmath73 area , that were inserted between each radiator and the tube plane below , in the terminal part of the module .
the aluminum foils absorbed the tr emitted by muons in the upstream radiator layers . using this technique a sample of muons
were collected with only the ionization loss measured .
a reduced scale prototype of the trd was exposed to a pion / electron beam @xcite to evaluate the detector response function .
the physical observable related to the energy of the particles crossing the detector is the total number of hits produced in the proportional tubes . for a sample of particles crossing the detector with a fixed energy and at a fixed angle ,
the number of hits are roughly poisson distributed , with an average value of few units , which depends on the beam energy and crossing angle . in fig . [
fig : trdcalib ] the average number of measured hits in the proportional tubes is plotted as a function of the lorentz factor @xmath50 of the incident particles , for several beam crossing angles . below the threshold value ( @xmath74 ) there is only the contribution of the ionization energy loss ; for @xmath75 , the tr contribution is also present and the average number of hits increases logarithmically with @xmath50 .
we considered the data collected by all the three macro trd modules during the acquisition period from april 1995 to december 2000 .
two classes of events were analyzed : 1 .
`` single muons '' , i.e. events with one muon in macro crossing one of the trd modules ; 2 .
`` double muons '' , i.e. events with two muons in macro and only one muon crossing one of the trd modules , like the one shown in figure [ fig : evdisp ] . to evaluate the muon energies
, we associated to each muon track the hits produced in the trd proportional tubes .
the muon was tracked with the standard macro procedure @xcite , which uses the information of the streamer tube system .
the distribution of the distances of the trd hits from the expected position , calculated with the muon reconstructed track , has a gaussian shape with a standard deviation @xmath76 .
we associated to each muon all the proportional tubes hits within @xmath77 from the track . to avoid badly reconstructed tracks ,
only muons crossing at least three layers of streamer tubes in the lower part of macro were used .
muons accompanied with electromagnetic showers initiated in the rock surrounding the detector were also discarded . since the trd was calibrated ( see fig .
[ fig : trdcalib ] ) with particles crossing all the ten layers of proportional tubes at angles below @xmath78 , in the present analysis only muons crossing the whole detector with zenith angle smaller than @xmath78 were included .
runs in which the trd modules were affected by stability problems or malfunctioning were discarded . also runs with reconstructed muon rates differing more than three standard deviations from the average values were disregarded , as well as those runs whose duration was less than @xmath79 .
the final data samples consist of @xmath80 single muons and @xmath81 double muons , for a total life time of @xmath82 days ( see table [ tab : summary ] ) .
[ fig : hitdist ] shows the distributions of the number of hits produced in the trd proportional tubes along the muon tracks for the final data samples .
since the second and third trd modules were equipped with a different read - out electronics from that of the first module , two different response functions ( described in the next section ) were necessary to analyze the data samples from the first module and those from the second and third modules .
the results of the two analysis were then combined .
we define @xmath83 as the distributions of the trd hits for a sample of muons with zenith angle @xmath84 that crossed a rock thickness @xmath11 .
the rock thickness @xmath11 was calculated from the direction @xmath85 using the gran sasso map @xcite and was converted into standard rock according the prescriptions of @xcite .
these distributions can be related to the residual muon energy spectra @xmath86 by : @xmath87 where @xmath88 is the tr detector response function , i.e. the probability to observe @xmath89 hits along the track of a muon with underground energy @xmath90 , crossing the detector at a zenith angle @xmath84 . to reconstruct the muon energy spectrum @xmath86 starting from the measured hit distributions ( fig .
[ fig : hitdist ] ) , once the trd response function @xmath88 is known , eq .
[ eq : hit ] must be inverted .
we derived two matrices @xmath88 ( one for the first tr module and another for the second and third modules ) on the basis of the calibration data taken by a reduced scale prototype exposed to a pion - electron test beam at cern @xcite .
we simulated , with a full geant - based @xcite monte carlo , a muon sample distributed according a flat energy and solid angle spectrum ( i.e. @xmath91 ) .
on the basis of the calibration data shown in fig .
[ fig : trdcalib ] , seven energy bins and four angular bins were defined .
the first energy bin covers the range from @xmath92 to @xmath93 ( i.e. @xmath94 ) , and the last one covers the range above the trd saturation ( @xmath95 , i.e. @xmath96 ) . the number of hits produced in the trd by a simulated muon of energy @xmath97 crossing the detector at zenith angle @xmath84 was extracted by the corresponding calibration data set . the trd response function ( a @xmath98 matrix )
was calculated as : @xmath99 where @xmath100 is the number of simulated muons with residual energy @xmath97 and zenith angle @xmath84 , producing @xmath89 hits , with @xmath101 and @xmath102 is the total number of simulated muons with energy @xmath97 crossing the trd at a zenith angle @xmath84 .
the simulated data had the same format of experimental data and were processed by the same analysis tools used for real data .
the accuracy and the time stability of the trd response function also needed to be taken into account .
although the nominal gas gain of the trd proportional tubes was the same of the tubes used in the prototype , during the long time - scale data acquisition period their operating conditions were affected by fluctuations .
this was due to drifts of some important parameters , like the gas mixture composition , the atmospheric pressure and the temperature .
a first check of the trd response function for energies below the threshold for the emission of tr was made using stopping muons , i.e. muons that are absorbed in the lower part of macro , that can be easily tagged by imposing simple geometrical cuts .
since the average residual energy of these muons is below a few @xmath103 , we simulated a sample of muons with a flat energy spectrum up to @xmath104 .
the same algorithm for the selection of stopping muons was applied to both the real and the simulated data samples .
the measured hit distributions of stopping muons are well reproduced by the monte carlo simulation ( see table [ tab : mustop ] ) , thus confirming the reliability of the trd calibration in the low energy region .
a further check of the detector response function was performed using the data from the part of the third module equipped with aluminum foils .
since the aluminum foils absorbed the tr produced in the upstream radiator , this data sample allows to check the detector response for muons releasing energy in the proportional tubes only by ionization . to take into account the weak dependence of the trd response on ionization ( see fig . [
fig : trdcalib ] ) , we simulated a sample of muons with the characteristic energy spectrum at macro depth ( eq . [ eq : undspec ] ) and with the same angular distribution as real data . in fig .
[ fig : trdal ] the values of the average number of hits along the muon track versus the muon zenith angle from the monte carlo and from the real data are shown .
the @xmath105 value ( evaluated using only statistical errors ) is @xmath106 , showing that there is a good agreement between the data and the simulation . from this analysis we estimated the systematic uncertainty arising from the fluctuations of the detector operating conditions .
for each of the four angular bins used for evaluating the trd response function , we compared the measured hit distributions @xmath107 with the simulated ones .
we remarked that the differences between the experimental average values and the monte carlo predictions are within @xmath108 .
for this reason , to take into account the fluctuations of the detector operating conditions , we associated to the average values of each hit distribution of the calibration data set , an uncertainty of @xmath108 .
these uncertainties will propagate to the systematic errors affecting the measurement of the residual energy of muons .
the systematic errors quoted in the present analysis ( see sec . [
sec : spectra ] ) are greater than the ones quoted in our previous work @xcite .
when that analysis was performed , the trd module equipped with aluminum foils was still not taking data . for this reason , we could estimate the systematic uncertainties only on the basis of the beam test data and we decided to associate to the calibration data set an uncertainty of @xmath109 , that is lower than the value we adopted for this analysis .
as shown in section [ sec : trdresp ] , the residual energy spectra of underground muons are related to the measured hit distributions in the trd by eq .
[ eq : hit ] .
two different methods were applied to reconstruct the underground muon energy spectra by inversion of eq .
[ eq : hit ] .
the first approach uses the same unfolding procedure described in @xcite . in the second approach a parameterization for the muon energy spectrum
was assumed and the parameters were derived using a best fit procedure . unfolding techniques are widely applied in problems where matrix inversion is required @xcite . to reconstruct the energy spectra of single and double muons starting from the hit distributions of fig .
[ fig : hitdist ] , we applied an iterative unfolding procedure @xcite .
as an initial energy spectrum ( used as a starting point for the unfolding procedure ) , we assumed eq .
[ eq : undspec ] for both the single and the double muon events , with parameters : @xmath110 , @xmath111 and @xmath112 .
the final reconstructed spectrum does not depend on the choice of the initial spectrum .
it only affects the time needed for the procedure to converge @xcite .
the energy spectrum reconstructed after each iteration is used as input for the next iteration . at the end of each iteration a @xmath113 test is performed between the reconstructed energy spectra and the input energy spectra .
the @xmath113 is defined as : @xmath114^{2 } } { \sigma_{n}^{2}(e_{i } \mid h_{j } , \theta_{k } ) + \sigma_{n-1}^{2}(e_{i } \mid h_{j } , \theta_{k})}\ ] ] where @xmath115 and @xmath116 are the energy spectra reconstructed after the @xmath117-th and the @xmath118-th iteration and @xmath119 , while @xmath120 are the associated errors .
the iterative procedure stops when the energy distribution reconstructed after the @xmath121 iteration has a @xmath113 probability greater than @xmath122 to be equivalent to the one reconstructed after the @xmath123 iteration .
the unfolding procedure was separately applied to the data samples from the first and from the second and third trd modules and the results were combined .
[ fig : diffspectra ] shows the reconstructed differential energy spectra of single and double muons at the gran sasso underground laboratory depth .
the error bars in the figure were calculated by adding in quadrature statistical and systematic errors .
the systematic errors are originated by fluctuations of the trd response function , and were estimated by a @xmath124 and a @xmath125 variation of the calibration data .
the unfolding procedure was replicated using the `` perturbed '' detector response functions ; the systematic uncertainty was evaluated as the difference between the two results .
the unfolding procedure reconstructs the shapes of the spectra up to @xmath56 . for energies greater than @xmath56 ,
where the trd response is saturated , the spectral shapes can not be reconstructed , and only the number of events can be evaluated .
the average value of the energy of underground muons with energy above threshold , @xmath126 , was calculated from eq .
[ eq : undspec ] as : @xmath127 the average muon energy was evaluated as : @xmath128 where @xmath129 is the average energy of muons with energies up to @xmath130 , @xmath131 and @xmath126 are the fraction and the average energy of muons with energies greater than @xmath56 , respectively . for a @xmath132 variation of the parameters @xmath133 , @xmath40 and @xmath134 , as is typically quoted by various authors ( e.g. @xcite ) , the uncertainties on the average muon energies are less than @xmath135 and are significantly smaller than our quoted errors .
the results are shown in table [ tab : energy ] .
the average energy of single muons and double muons are respectively of @xmath136 and @xmath137 .
the single muon result is not in contradiction with the result of our previous analysis @xcite , where the systematic effects were underestimated .
an alternative way for evaluating the muon energy spectra from the measured hit distributions is that of assuming eq .
[ eq : undspec ] as an analytic description for the spectra , and to derive the parameters using a best - fit . substituting the trial energy spectrum into eq .
[ eq : hit ] and summing over the zenith angles we get : @xmath138^{-\alpha } \label{eq : hit2}\ ] ] where the normalization constants @xmath139 represent the number of muons detected in each bin of depth and zenith angle .
we fixed the values of the @xmath40 and @xmath134 parameters to @xmath140 and @xmath141 @xcite ; @xmath133 was left as the free parameter for the single and double muon data .
for each value of @xmath133 in the range from @xmath142 to @xmath143 and step @xmath144 we built a set of hit distributions @xmath145 according eq .
[ eq : hit2 ] , and for each set of distributions we evaluated the @xmath113 as : @xmath146^{2 } } { \sigma^{2}(k \mid h ) + \hat{\sigma}^{2}(k \mid h)}\ ] ] where @xmath147 is the measured hit distribution for muons ( single or double ) crossing a rock slant depth @xmath11 ; @xmath148 and
@xmath149 are the poissonian errors on @xmath147 and @xmath145 , respectively . the curves representing the @xmath113 as a function of @xmath133 are continuous both for single and double muons and show a well defined minimum .
the value of the best fit spectral index for single muons is @xmath150 , with @xmath151 , while for the double muons is @xmath152 , with @xmath153 .
the statistical error on @xmath133 was evaluated using the values @xmath154 and @xmath155 corresponding to @xmath156 .
the systematic error was estimated from the positions of the new minima of @xmath133 obtained using the two trd response functions evaluated from the sets of `` perturbed '' calibration data .
the result obtained for the single muon spectral index is consistent with that obtained from the macro measurement of the underground muon intensity as a function of the rock depth @xcite .
it is also consistent with the results of the nusex experiment @xcite , that found a spectral index of @xmath157 for a sample of events mainly composed by single muons .
[ fig : confronto ] shows a comparison between the differential muon energy spectra reconstructed with the two methods described in sections [ sec : unfo ] and [ sec : fit ] both for single and double muons . in the case of single muons
we find a value of @xmath158 , while in the case of double muons we find @xmath159 .
in the case of single muon data , there is a good agreement between the result obtained using the unfolding procedure and that obtained using the best fit method . in the case of double muon data , there are some discrepancies , especially for low residual energies . to understand these difference , it should be noted that the result of the unfolding procedure does not depend on the trial spectrum .
the spectrum used for the fit ( eq . [ eq : undspec ] ) is derived assuming that the surface muon energy spectra obey a power law with spectral index @xmath133 , and in the muon propagation formula , @xmath134 and @xmath40 are constant , i.e. the cross sections for the radiative processes do not depend on the muon energy . if the correct expression ( eq . [ eq : musurf ] ) is assumed for the surface energy spectrum and the dependence of @xmath134 and @xmath40 on the muon energy is taken into account , it is impossible to obtain an analytical expression for the underground muon spectrum .
hence , eq . [ eq : undspec ] represents only a useful parameterization for the muon underground spectra , whose accuracy is limited by the hypotheses from which it has been derived .
the difference between our results obtained using the two techniques mentioned in sections [ sec : unfo ] and [ sec : fit ] are ascribed to these approximations .
we can conclude that the parameterization is not completely suitable for describing the underground muon energy spectra , especially for events with large underground multiplicity .
as shown in section [ sec : undmuon ] , the underground muon energy spectrum is related to the primary cosmic ray spectrum . in order to investigate the relationship between the underground muon energy and the primary cosmic ray spectrum and composition
, we performed a monte carlo simulation using different composition models for primary cosmic rays .
the interactions of cosmic rays in the atmosphere were simulated with the hemas code @xcite , while the process of muon propagation in the rock was simulated using the propmu code @xcite .
two extreme composition models were assumed for primary cosmic rays : * the `` light model '' @xcite , i.e. a proton - rich composition model ; * the `` heavy model '' @xcite , i.e. a fe - rich composition model .
these models assume that cosmic rays are composed by five main mass groups ( @xmath15,@xmath160,@xmath161,@xmath162 and @xmath163 ) .
the energy spectrum of each component is described by means of power laws given by : @xmath164 in eq .
[ eq : compo ] there are 5 parameters for each primary component ( i.e. the normalization constants @xmath165 and @xmath166 , the spectral indices @xmath167 and @xmath168 and the cutoff energy @xmath169 ) .
these parameters are not independent : usually the constant @xmath166 is expressed as a function of the others by imposing the continuity of the function @xmath170 at @xmath171 .
the all - particle spectrum is evaluated by adding the contributions from all the mass groups . in table
[ tab : models ] the values of all the parameters are summarized for each component of the primary cosmic rays in both the composition models we adopted . the light model is different from the heavy model because there is an extra - component of protons .
[ fig : evsnmu ] shows the predictions obtained using the light and heavy composition models for the average energy of the underground muons as a function of the muon multiplicity .
the gap between the predictions from the two extreme models increases with the muon multiplicity .
the two dark points represent our measured values , averaged on the whole rock depth range from @xmath2 to @xmath3 .
to explain the average energy behavior , in fig .
[ fig : mcsummary ] we plotted , for the two composition models and as a function of the underground muon multiplicity , for the @xmath15 , @xmath160 and @xmath163 components : @xmath172 the average energies of underground muons ; @xmath173 the relative contribution of these three mass groups and @xmath174 the average energies of the parent cosmic rays to events for a fixed underground muon multiplicity . in @xmath172 and @xmath174 the values for the all - particle spectrum are also shown . from this figure
the following conclusions can be drawn : * single muons are originated mainly by primary protons ; * the contributions of primary heavy elements become relevant when the muon multiplicities increase , while the contribution of protons decreases ; * in the `` heavy '' composition model the contribution of @xmath163 nuclei is relevant from low multiplicities ( it is @xmath175 for @xmath176 and tends to @xmath177 for high multiplicities ) ; * in the light model , the main contribution is always that of protons ; * for a fixed muon multiplicity , muons originated by light primaries are more energetic than muons originated by heavy primaries . from the simulation we get that the average energies of underground muons produced by each component are almost independent on the primary composition model .
this is due to the fact that the differences between the spectral indices of the primary components in the models we are examining are small for energies below the `` knee '' , as can be seen from table [ tab : models ] . as a consequence ,
the muons produced e.g. by he nuclei have the same average energy , independently on the choice of the composition model .
hence , the differences between the predictions of the cosmic ray composition models are mainly due to differences in the weights of the various components in each model and not to differences in their spectra . from fig .
[ fig : mcsummary]@xmath7 , the primary energy needed to produce one underground muon at the macro depth is smaller than @xmath178 , while the primary energy needed to produce two underground muons is about @xmath179 .
this means that an analysis of events with only one or two underground muons is sensitive to the energy region below the knee of the primary spectrum .
our measurements were compared with the predictions from the simulation .
[ fig : evsnmu ] shows a comparison between the results obtained applying the the analysis technique described in sec .
[ sec : unfo ] and the monte carlo predictions concerning the average energies of single and double muons , while in fig .
[ fig : datamc ] the average energies of single and double muons are plotted as a function of the rock thickness crossed by muons .
single muon data are not in contradiction with the predictions from both the composition models . in fact , since the energy gaps between the predictions from the two extreme composition models are of the same order of magnitude as the error bars , single muons do not allow one to discriminate between different cosmic ray composition models . the sensitivity of this measurement to the primary cosmic ray composition
is strongly limited by systematic errors associated to the trd response function . also in the case of double muons , our data do not allow to perform a discrimination between the cosmic ray composition models . in this case ,
although the errors associated to the average muon energies are smaller than the gap between the predictions from the two extreme composition models , experimental data do not provide a clean signature in favour of a given composition model .
another comparison between the trd data and the monte carlo predictions can be done by looking at the spectral indices . for each composition model
we fitted the underground energy spectra of single and double muons with the formula [ eq : undspec ] , assuming the same values of @xmath134 and @xmath40 as in sec .
[ sec : fit ] and we compared the fit results with the trd data .
table [ tab : hemasfit ] shows the comparison of the data with the monte carlo predictions . as in the previous case , the single and double muon
spectral indices do not allow to perform a study of the cosmic ray composition .
the sensitivity of our measurement to the primary cosmic ray composition is mainly limited by its precision and by the poor statistics of the high multiplicity muon events , that does not allow to reconstruct their spectra .
a detector with a larger area than the macro trd and with a reduced systematics could enhance the precision of the measurement of the single and double muon energies and also allow to measure the energies of high multiplicity muons , that are more sensitive to the cosmic ray composition in the energy region of the knee .
the macro trd allowed the measurement of the energies of muons penetrating in the gran sasso underground laboratories , in the standard rock depth range from @xmath2 to @xmath3 . for reconstructing the muon energy spectra we used an unfolding method and a best fit procedure .
the average energies of single and double muons , evaluated with the unfolding technique described in sec . [ sec : unfo ] , are @xmath180 and @xmath181 .
the spectral indices for the approximate parameterization of the muon energy spectra ( eq . [ eq : undspec ] ) , evaluated with the best fit procedure described in sec .
[ sec : fit ] are @xmath150 and @xmath152 .
we also performed a monte carlo simulation to study how the muon energy spectra depend on the primary cosmic ray energy spectra and composition , on their interactions with the atmosphere and on the muon propagation in the rock .
our data show that double muons are more energetic than single ones in the standard rock depth range from @xmath2 to @xmath182 , as predicted by our monte carlo simulation .
the monte carlo simulation also showed that a measurement of the energies of underground high multiplicity muons could provide useful information about the primary cosmic ray composition .
in fact , we remarked that the differences between the predictions from the various composition models become relevant at high muon multiplicities , where the contribution from heavy primaries is more significant .
the energy spectra of single and double muons reconstructed from our data are in agreement with the monte carlo predictions obtained assuming two extreme cosmic ray composition models .
this measurement do not allow to perform a cosmic ray composition study because the errors are compatible with the gaps between the predictions from the two models .
however , a measurement of the energy spectra of high multiplicity underground muons could provide a useful tool for investigating the primary cosmic ray composition in the energy region above the knee of the spectrum .
|
the macro detector was located in the hall b of the gran sasso underground laboratories under an average rock overburden of @xmath0 .
a transition radiation detector composed of three identical modules , covering a total horizontal area of @xmath1 , was installed inside the empty upper part of the detector in order to measure the residual energy of muons .
this paper presents the measurement of the residual energy of single and double muons crossing the apparatus .
our data show that double muons are more energetic than single ones .
this measurement is performed over a standard rock depth range from @xmath2 to @xmath3 . the macro collaboration + = 10000 , , , , , , , , , , , % authorc .
bloise@xmath4 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , and .
+ 1
. dipartimento di fisica delluniversit di bari and infn , 70126 bari , italy + 2 .
dipartimento di fisica delluniversit di bologna and infn , 40126 bologna , italy + 3 .
physics department , boston university , boston , ma 02215 , usa + 4 .
california institute of technology , pasadena , ca 91125 , usa + 5 .
department of physics , drexel university , philadelphia , pa 19104 , usa + 6 .
laboratori nazionali di frascati dellinfn , 00044 frascati ( roma ) , italy + 7 .
laboratori nazionali del gran sasso dellinfn , 67010 assergi ( laquila ) , italy + 8 . depts . of physics and of astronomy , indiana university , bloomington , in 47405 , usa + 9 .
dipartimento di fisica delluniversit dellaquila and infn , 67100 laquila , italy + 10 .
dipartimento di fisica delluniversit di lecce and infn , 73100 lecce , italy + 11 .
department of physics , university of michigan , ann arbor , mi 48109 , usa + 12 .
dipartimento di fisica delluniversit di napoli and infn , 80125 napoli , italy + 13 .
dipartimento di fisica delluniversit di pisa and infn , 56010 pisa , italy + 14 .
dipartimento di fisica delluniversit di roma `` la sapienza '' and infn , 00185 roma , italy + 15 .
physics department , texas a&m university , college station , tx 77843 , usa + 16 .
dipartimento di fisica sperimentale delluniversit di torino and infn , 10125 torino , italy + 17 .
l.p.t.p , faculty of sciences , university mohamed i , b.p .
524 oujda , morocco + 18 .
dipartimento di fisica delluniversit di roma tre and infn sezione roma tre , 00146 roma , italy + @xmath5 also universit della basilicata , 85100 potenza , italy + @xmath6 also infn milano , 20133 milano , italy + @xmath7 also istituto iasf / cnr , 40129 bologna , italy + @xmath8 also universit di trieste and infn , 34100 trieste , italy + @xmath9 also dipartimento di energetica , universit di roma , 00185 roma , italy + @xmath10 also institute for space sciences , 76900 bucharest , romania + @xmath11 macalester college , dept . of physics and astr . , st .
paul , mn 55105 + @xmath12 also department of physics , pennsylvania state university , university park , pa 16801 , usa + @xmath13also dipartimento di fisica delluniversit della calabria , rende ( cosenza ) , italy + @xmath14 also dipartimento di scienze e tecnologie avanzate , universit del piemonte orientale , alessandria , italy + @xmath15 also u. minn .
duluth physics dept . ,
duluth , mn 55812 + @xmath16 also dept . of physics , mit , cambridge ,
ma 02139 + @xmath17 also intervideo inc . , torrance ca 90505 usa + @xmath18 also resonance photonics , markham , ontario , canada +
|
since the invention of turbo codes @xcite and the rediscovery of low - density parity - check ( ldpc ) codes @xcite , constructing practical good codes has been being an active research topic in our field .
recent developments include the invention of polar codes @xcite and flourishment of spatially coupled ldpc ( sc - ldpc ) codes ( first introduced as ldpc convolutional codes @xcite and later recast as sc - ldpc codes @xcite ) , both of which are provable capacity - achieving @xcite over memoryless binary - input symmetric - output channels . despite this success in theory ,
more flexible constructions are still desired in practice .
especially , it is often desirable in practice to design codes that support a variety of code rates but maintain essentially the same encoding / decoding hardware structure .
one way to achieve this is the use of rate - compatible codes , which can be constructed from a mother code by using the puncturing and/or extending techniques .
the former starts with a low - rate mother code and punctures some coded bits to achieve higher rates @xcite , while the latter starts with a high - rate code and extends its parity - check matrix to achieve lower rates @xcite . both puncturing and extending require optimizations .
for example , the puncturing patterns for rate - compatible punctured convolutional ( rcpc ) codes in @xcite were selected by maximizing the average free distance , while the puncturing distributions for rate - compatible ldpc codes in @xcite were optimized by density evolution . in @xcite ,
the incremental protomatrices for protograph - based raptor - like ( pbrl ) ldpc codes were chosen by maximizing the density evolution threshold . to reduce the construction complexity caused by the optimizations , one can use random puncturing , as proposed in @xcite .
however , similar to the conventional punctured ldpc codes @xcite , the performance of the randomly punctured ldpc codes degrades significantly when the puncturing fraction increases beyond a threshold . to the best of our knowledge ,
no methods were reported along with simulations in the literature that can construct good rate - compatible codes over all rates of interest in the interval ( 0,1 ) .
recently , a coding scheme called block markov superposition transmission ( bmst ) of short codes ( referred to as _ basic codes _ ) was proposed @xcite , which has a good performance over the binary - input additive white gaussian noise ( awgn ) channel .
it has been pointed out in @xcite that any short code ( linear or nonlinear ) with fast encoding algorithm and efficient soft - in soft - out ( siso ) decoding algorithm can be chosen as the basic code .
a bmst code is indeed a convolutional code with extremely large constraint length , which has a simple encoding algorithm and a low complexity sliding window decoding algorithm .
more importantly , bmst codes have near - capacity performance ( observed by simulation and confirmed by extrinsic information transfer ( exit ) chart analysis @xcite ) in the waterfall region of the bit - error - rate ( ber ) curve and an error floor ( predicted by analysis ) that can be controlled by the encoding memory . in @xcite ,
short hadamard transform ( ht ) codes are taken as the basic codes , resulting in a class of multiple - rate codes with fixed code length , referred to as bmst - ht codes .
an even simpler construction for multiple - rate bmst codes was proposed in @xcite , where the involved basic codes consist of repetition ( r ) codes and single - parity - check ( spc ) codes , resulting in bmst - rspc codes .
different from bmst - ht codes which adjust their code rates by setting properly the number of frozen bits in the short ht codes , bmst - rspc codes adjust the code rates by time - sharing between the r code and the spc code .
the construction of bmst codes is flexible , in the sense that it applies to all code rates of interest in the interval ( 0,1 ) .
however , original bmst codes @xcite are neither rate - compatible nor systematic .
note that systematic codes may be more attractive in practical applications since the information bits can be extracted directly from the estimated codeword .
even worse , original bmst codes do not perform well over block fading channels due to errors propagating to successive decoding windows . in this paper , we propose systematic bmst of repetition codes , referred to as systematic bmst - r codes . for encoding ,
the information sequence is partitioned equally into blocks and transmitted directly , while their replicas are interleaved and transmitted in a block markov superposition manner . for decoding ,
a sliding window decoding algorithm with a tunable decoding delay can be implemented , as with sc - ldpc codes @xcite .
systematic bmst - r codes not only preserve the advantages of the original non - systematic bmst codes , namely , low encoding complexity , effective sliding window decoding algorithm and predictable error floors , but also have improved decoding performance especially in short - to - moderate decoding latency . the main contributions of this paper include : 1 .
we propose systematic rate - compatible bmst - r codes by using both extending and puncturing .
the construction requires no optimization but applies universally to all code rates varying continuously " from zero to one .
we propose an upper bound on the ber of a systematic bmst - r code ensemble under maximum _ a posteriori _ ( map ) decoding , which can be evaluated by calculating partial input - redundancy weight enumerating function ( irwef ) with truncated information weight .
we propose a lower bound on the ber of a systematic bmst - r code ensemble under map decoding , which depends on the encoding memory and code rate .
the derived lower bound reveals connections among ber , encoding memory and code rate , which provides a way to design good systematic bmst - r codes and also allows us to make trade - offs among efficiency , performance and complexity .
we investigate the impact of various parameters on the performance of systematic bmst - r codes , and then present a performance comparison of systematic bmst - r codes and sc - ldpc codes on the basis of equal decoding latency .
simulation results show that : 1 ) the upper and lower bounds are tight in the high signal - to - noise ratio ( snr ) region ; 2 ) with a moderate decoding delay , the ber curves can match the respective lower bounds in the low ber region , implying that the iterative sliding window decoding algorithm is near optimal ; 3 ) systematic bmst - r codes perform well ( within one db away from the corresponding shannon limits ) in a wide range of code rates , confirming the effectiveness of the construction procedure ; and 4 ) over both awgn channels and block fading channels , systematic bmst - r codes , overcoming the weakness of non - systematic bmst codes , can have better performance than sc - ldpc codes in the waterfall region under the equal decoding latency constraint .
the rest of the paper is structured as follows . in section [ secii ] , we present the encoding and decoding algorithms of systematic bmst - r codes . in section [ seciii ] , we analyze the performance and complexity of systematic bmst - r codes . numerical analysis and performance comparison
are presented in section [ seciv ] .
finally , some concluding remarks are given in section [ sec : conclusion ] .
let @xmath0 be the binary field .
let @xmath1 , @xmath2 , @xmath3 be the information sequence to be transmitted , where @xmath4 is the information subsequence of length @xmath5 .
the encoding algorithm of a systematic bmst - r code of rate @xmath6 with encoding memory @xmath7 is described as follows ( see fig .
[ bmst_encoder ] for reference ) , where @xmath8 @xmath9 are interleavers of size @xmath5 . and encoding memory @xmath10 , where the information subsequence @xmath11 at time @xmath12
is encoded into the subcodeword @xmath13 for transmission . ]
encoding of systematic bmst - r codes[alg : encoding ] 1 .
* initialization : * [ step : encoding_initialize ] for @xmath14 and @xmath15 , set @xmath16 .
* loop : * [ step : encoding_iteration ] for @xmath17 , * repeat @xmath11 @xmath18 times such that @xmath19 and @xmath20 for @xmath15 ; * for @xmath15 , 1 . for @xmath21 , interleave @xmath22 into @xmath23 using the @xmath24-th interleaver @xmath8 ; 2 . compute @xmath25 .
* take @xmath26 as the @xmath12-th block of transmission .
the above encoding structure can implement all code rates of the form @xmath6 , @xmath27 .
if @xmath28 of @xmath5 bits in @xmath29 are randomly punctured resulting in @xmath30 , we can implement a code rate @xmath31 , where @xmath32 is the puncturing fraction . in practice ,
the code need to be terminated .
this can be done easily by driving the encoder to the zero state with a zero - tail of length @xmath33 after @xmath34 blocks of data .
that is , for @xmath35 , @xmath36 , @xmath37 , @xmath38 , we set @xmath39 , compute @xmath40 following * loop * in algorithm [ alg : encoding ] , and then take the redundant check part of @xmath40 as the @xmath12-th block of transmission . the rate of the resulting terminated systematic bmst - r code is @xmath41 which is less than that of the unterminated code .
however , the rate loss is negligible for large @xmath34 . in summary ,
all code rates of interest in the interval ( 0,1 ) can be implemented by adjusting the _ repetition degree _ @xmath18 and the _ puncturing fraction _
@xmath42 , all with the encoding structure as shown in fig .
[ bmst_encoder ] , where stands for the optional puncturing .
assume that the subcodeword @xmath40 is modulated using binary phase - shift keying ( bpsk ) with 0 and 1 mapped to @xmath43 and @xmath44 , respectively , and transmitted over an awgn channel , resulting in a received vector @xmath45 expressed as @xmath46 for @xmath47 , where @xmath48 is the @xmath49-th component of @xmath45 and @xmath50 is a sample from an independent gaussian random variable with distribution @xmath51 .
the decoding algorithm for systematic bmst - r codes can be described as an iterative message processing / passing algorithm over the associated forney - style factor graph , which is also known as a normal graph @xcite . in the normal graph ,
edges represent variables and vertices ( nodes ) represent constraints . all edges connected to a node must satisfy the specific constraint of the node .
a full - edge connects to two nodes , while a half - edge connects to only one node . a half - edge is also connected to a special symbol , called a dongle " , that denotes coupling to other parts of the transmission system ( say , the channel or the information source ) @xcite .
[ bmst_decoder ] shows the normal graph of a systematic bmst - r code with @xmath52 , @xmath53 and @xmath54 .
it is indeed a high - level normal graph , where each edge represents a sequence of random variables .
there are four types of nodes in the normal graph of the systematic bmst - r code .
* * node * @xmath55 : all edges ( variables ) connected to node @xmath55 must sum to the all - zero vector . the message updating rule at node @xmath55 is similar to that of a check node in the factor graph of a binary ldpc code .
the only difference is that the messages on the half - edges are obtained from the channel observations . * * node * @xmath56 : all edges ( variables ) connected to node @xmath56 must take the same ( binary ) values .
the messages on the half - edges are obtained from both the channel observations and the information source . to avoid confusion and messy plots , are assumed to be independent and uniformly distributed over @xmath57 . ] the message updating rule at node @xmath56 is the same as that of a variable node in the factor graph of a binary ldpc code . * * node * : the node represents the @xmath24-th interleaver , which interleaves or de - interleaves the input messages . *
* node * : two edges ( variables ) connected to node must satisfy the constraint specified by the puncturing rules . , @xmath53 and @xmath54 . ] the normal graph of a systematic bmst - r code can be divided into _
layers _ , where each layer typically consists of a node of type , @xmath58 nodes of type , @xmath59 nodes of type , and a node of type ( see fig . [ bmst_decoder ] ) .
similar to sc - ldpc codes , an iterative sliding window decoding algorithm with decoding delay @xmath60 performing over a subgraph consisting of @xmath61 consecutive layers can be implemented for systematic bmst - r codes . for each window position , the sliding window decoding algorithm can be implemented using the parallel ( flooding ) updating schedule within the decoding window . the first layer in any window is called the _
target layer_. decoding proceeds until a fixed number of iterations has been performed or certain given stopping criterion is satisfied , in which case the window shifts to the right by one layer and the symbols corresponding to the target layer shifted out of the window are decoded . from fig .
[ bmst_encoder ] , we can see that systematic bmst - r codes resemble the classical rcpc codes @xcite .
evidently , we can start from a rate @xmath6 systematic bmst - r code ( the mother code ) , where @xmath18 is as large as required . by puncturing can be removed . ]
, one can obtain all code rates of interest from @xmath6 to 1 , all of which can be implemented with essentially the same pair of encoder and decoder .
the difference between systematic bmst - r codes and rcpc codes is also obvious .
the encoding of systematic bmst - r codes is block - oriented and the decoding is typically not implementable by the viterbi algorithm @xcite due to the huge constraint length induced by the block - oriented encoding process . alternatively
, systematic bmst - r codes are decodable with a sliding window decoding algorithm , which is similar to sc - ldpc codes .
more generally , systematic bmst - r codes can be viewed as a special class of spatially coupled codes , since spatial coupling can be interpreted as introducing memory among successive independent transmissions , where extra edges are allowed to be added during the coupling process @xcite .
in contrast to sc - ldpc codes , which are usually defined by the null space of a sparse parity - check matrix , systematic bmst - r codes are easily described using generator matrices .
further , since the encoder for a systematic bmst - r code is non - recursive , an all - zero tail can be added to drive the encoders to the zero state at the end of the encoding process .
this is different from sc - ldpc codes , where the tail is usually non - zero and depends on the encoded information bits ( see section iv of @xcite ) . as a result , the encoding procedure for systematic bmst - r codes is simpler than for sc - ldpc codes . when described in terms of generator matrices , systematic bmst codes
can also be viewed as a special class of spatially coupled low - density generator - matrix ( sc - ldgm ) codes @xcite . however , as an ensemble , systematic bmst - r codes are different from sc - ldgm codes .
sc - ldgm code ensembles are usually defined in terms of their node distributions , while systematic bmst - r code ensembles are defined in terms of their interleavers ( see fig .
[ bmst_encoder ] ) . as another evidence that systematic bmst - r codes are different from existing codes
, we would like to emphasize that systematic bmst - r codes have a simple lower bound on the ber performance , as described in the next section .
a reasonable criterion for a construction to be good is its ability to make trade - offs between complexity and performance .
specifically , if the error performance required by the user is relaxed or , if the gap between the code rate and the capacity is more tolerant , the encoding / decoding complexity should be reduced . in this section
, we will find a relation of the performance to the complexity for _ terminated _ systematic bmst - r codes .
we start with a general systematic linear block code .
a binary linear block code @xmath62 $ ] is a @xmath63-dimensional subspace of @xmath64 .
an encoding algorithm can be described simply by @xmath65 where @xmath66 is the information vector , @xmath67 is the associated codeword , and @xmath68 is a generator matrix of size @xmath69 with rank of @xmath63 .
define @xmath70 let @xmath71 be the minimum hamming weight of @xmath72 , i.e. , @xmath73 where @xmath74 represents the hamming weight .
obviously , the minimum hamming weight @xmath75 of the linear block code @xmath76 can be given by @xmath77 assume that the codeword @xmath78 is modulated using bpsk and transmitted over an awgn channel , resulting in a received vector @xmath79 .
a decoding algorithm is defined as a mapping @xmath80 where @xmath81 .
given the signal mapping @xmath82 and @xmath83 , the snr is given by @xmath84 in db , where @xmath85 is the variance of the noise .
suppose that @xmath86 is distributed uniformly at random over @xmath87 .
let @xmath88 be the error event that the decoder output @xmath89 is not equal to the encoder input vector @xmath86 , and let @xmath90 be the error event that the @xmath91-th estimated bit @xmath92 at the decoder is not equal to the @xmath91-th input bit @xmath93 .
obviously , @xmath94 .
then , under the given decoding algorithm @xmath95 , we can define frame error probability @xmath96 and bit - error probability @xmath97 from the definitions of ber and fer , we have @xmath98 we also have @xmath99 thus , we have @xmath100 the maximum - likelihood ( ml ) decoding algorithm selects a codeword @xmath101 such that @xmath102 for all codewords @xmath103 . if ties happen , the ml decoding algorithm can randomly select one candidate as the decoder output .
since @xmath86 is distributed uniformly at random over @xmath104 , the ml decoding algorithm is optimal in the sense that it minimizes the fer . to minimize the ber , the map decoding algorithm computes @xmath105 for all @xmath91 . for each @xmath106 ,
the map decoding algorithm outputs @xmath107
if @xmath108 and @xmath109 otherwise . the irwef of a systematic block code can be given as @xcite @xmath110 where @xmath111 , @xmath112 are two dummy variables and @xmath113 denotes the number of codewords having input ( information bits ) weight @xmath91 and redundancy ( parity check bits ) weight @xmath49 .
the irwef can also be written in a more compact form as @xmath114 where @xmath115 is the conditional redundancy weight enumerating function ( crwef ) , which enumerates redundancy weight for a given input weight @xmath91 .
since map decoding is optimal in the sense that it minimizes the ber , an upper bound on ber performance under any decoding algorithm is applicable to the map decoding algorithm . in the following ,
we consider a suboptimal list decoding algorithm .
a list decoding algorithm for the purpose of performance analysis[alg : listdecoding ] 1 .
make hard decisions on the information part of the received vector @xmath79 , resulting in a vector @xmath116 of length @xmath63 .
then the channel becomes a memoryless binary symmetric channel ( bsc ) with cross probability @xmath117 2 .
list all sequences of length @xmath63 within the hamming sphere with center at @xmath116 of radius @xmath118 .
the resulting list is denoted as @xmath119 .
3 . encode each sequence in @xmath119 by the encoding algorithm of the systematic code , resulting in a list of codewords , denoted as @xmath120 .
4 . find the codeword @xmath121 that is closest to @xmath79 .
output the information part @xmath122 of @xmath123 as the decoding result .
the above list decoding algorithm is similar to but different from the algorithm presented in @xcite . the _ list region _ in @xcite is an @xmath124-dimensional hamming sphere with center at the hard decision of the whole received sequence , while the list region here is a @xmath125-dimensional hamming sphere with center at the hard decision of the information part of the received sequence . by analyzing the ber performance of the proposed list decoding algorithm
, we have the following theorem .
[ theoremupperbound ] for any integer @xmath126 , the bit - error probability of systematic codes under map decoding is upper - bounded by @xmath127 @xmath128 consider the list decoding algorithm ( algorithm [ alg : listdecoding ] ) .
the decoding error occurs in two cases under the assumption that the all - zero codeword is transmitted . 1 .
the all - zero sequence of length @xmath63 is not in the list @xmath119 , i.e. , the hard - decisions have @xmath129 errors . in this case
, the decoder output has at most @xmath130 erroneous bits .
hence , the bit - error probability , denoted as @xmath131 , is upper - bounded by @xmath132 2 .
the all - zero sequence of length @xmath63 is in the list @xmath119 , but the all - zero codeword @xmath133 is not the closest one to @xmath79 . in this case , the bit - error probability , denoted as @xmath134 , is upper - bounded by @xmath135 in summary , for any given radius @xmath136 , we have @xmath137 @xmath138 combining ( [ list ] ) and the fact that @xmath139 , we complete the proof . from theorem [ theoremupperbound
] , we have the following three corollaries . [ corollary1 ] @xmath140 it
can be proved by simply setting @xmath141 in ( [ upperbound ] ) .
[ corollary2 ] @xmath142 by simply setting @xmath143 in ( [ upperbound ] ) , we have @xmath144 [ corollary3 ] assuming that we know only the truncated irwef @xmath145 , @xmath146 of systematic codes , we have @xmath147 @xmath148 it is obvious and omitted here . * remarks .
* corollary [ corollary1 ] is the well - known union bound , while corollary [ corollary2 ] is almost trivial , which can be easily understood by noting that setting @xmath143 in algorithm [ alg : listdecoding ] is equivalent to taking directly the hard decisions @xmath116 as the decoding result @xmath122 ( one of the simplest sub - optimal decoding algorithms ) . given that only the truncated irwef is available , corollary [ corollary3 ] is the tightest upper bound of this type .
there exist several lower bounds on fer under ml decoding @xcite .
however , lower bounds on ber are rarely mentioned in the literature .
any lower bound on @xmath149 can be adapted to a lower bound on ber by noticing that @xmath150 from ( [ eq1 ] ) .
the simplest lower bound on fer under ml decoding over awgn channels is given by @xmath151 which leads to @xmath152 logically , it is not safe to conclude from the above derivation that the lower bound ( [ ml_propo ] ) applies to map decoding .
this is subtle due to the fact that ml decoding is not optimal for minimizing the bit - error probability . in the following
, we will show that the lower bound on @xmath153 ( [ ml_propo ] ) is indeed a lower but usually loose bound on @xmath154 by proving an improved lower bound .
was found with proof in the literature . ] to see the looseness of the lower bound , we consider the following toy example .
let @xmath155 with @xmath156 and @xmath157 with @xmath158 be two codes .
define @xmath159 , whose codewords are in a cartesian product form @xmath160 , where @xmath161 and @xmath162 for @xmath163 . obviously , the code @xmath76 has minimum hamming weight @xmath156 .
however , for bpsk modulation over an awgn channel , the ber for the code @xmath76 is dominated by the code @xmath164 rather than the code @xmath165 . to be precise ,
@xmath166 which implies that the lower bound @xmath167 can be very loose in the low snr region . in the following
we present an improved lower bound under map decoding .
[ maplowerbound ] the bit - error probability for the linear block code @xmath76 under map decoding can be lower - bounded by @xmath168 it suffices to prove that @xmath169 for each given @xmath91 @xmath170 .
let @xmath171 be a codeword such that @xmath172 .
there must exist an invertible matrix @xmath173 of size @xmath174 such that @xmath175 with the first row of @xmath176 being @xmath177 .
assume @xmath178 be the information vector and @xmath179 be the codeword to be transmitted .
define @xmath180 .
the map decoder for a binary linear block code computes @xmath181 . we know
that if @xmath182 , the decoding output is correct for this considered bit . in the meanwhile , we assume a _ genie - aided decoder _ , which computes @xmath183 with @xmath184 available . likewise ,
if @xmath185 , the decoding output is correct for this considered bit . for a specific @xmath186 and @xmath79 , it is possible that @xmath187
however , the expectation @xmath188 = i \left(u_i ; \boldsymbol{v } ' | \boldsymbol{y}\right ) \geq 0,\end{aligned}\ ] ] where @xmath189 is the conditional mutual information .
this implies that the genie - aided decoder performs statistically no worse than the map decoder of the binary linear block code . under the condition that @xmath190 is available ,
there exist only two codewords whose hamming distance is @xmath71 .
thus , the bit - error probability with the genie - aided decoder for the binary - input awgn channels is @xmath191 .
it follows that @xmath192 * remarks .
* theorem [ maplowerbound ] also applies to _ non - systematic _ linear block codes .
however , it does not apply to non - linear codes , indicating that the proof is not that simple as considering only the two closest codewords . from theorem [ maplowerbound ]
, we have the following three corollaries . [ ml_lowerbound2map ]
@xmath193 combining ( [ dmin ] ) and theorem [ maplowerbound ] , we have @xmath194 [ cyclic_lowerbound ] if a code has the property that @xmath195 for all @xmath91 , we have @xmath196 it is obvious and omitted here .
[ ldgm_lowerbound ] if the row weights of the generator matrix @xmath197 for a linear block code @xmath76 are @xmath198 , we have @xmath199 this can be proved by noting that @xmath200 and that @xmath201 is a decreasing function .
* remarks .
* corollary [ ml_lowerbound2map ] shows that the lower bound ( [ ml_propo ] ) on @xmath153 is also a lower bound on the @xmath154 , while corollary [ cyclic_lowerbound ] indicates that the lower bound ( [ ml_propo ] ) can be very loose .
corollary [ ldgm_lowerbound ] indicates that an ldgm code may have a higher error floor compared to an ldpc code , since the generator matrix for an ldpc code is typically high - density . to apply the derived bounds to systematic bmst - r codes ,
we need calculate the irwef . for systematic bmst - r codes
, we have @xmath202 @xmath203 where the summation is over all possible data sequences @xmath186 with @xmath204 for @xmath205 .
since it is a sum of products , @xmath206 can be computed in principle by a trellis - based algorithm over the polynomial ring . for specific interleavers ,
the trellis has a state space of size @xmath207 , which makes the computation intractable for large @xmath33 . to circumvent this issue
, we turn to an ensemble of systematic bmst - r codes by assuming that all the interleavers ( see fig . [ bmst_encoder ] for reference ) are chosen at each time independently and uniformly at random , and that @xmath30 is obtained by randomly puncturing @xmath28 of @xmath5 bits in @xmath29 . with the assumption that all interleavers are uniform interleavers @xcite
, we can see that @xmath208 , for @xmath15 , is a random variable which depends _ only _ on the hamming weights @xmath209 .
this admits a reduced - complexity trellis representation of the average irwef of the defined systematic bmst - r code ensemble .
the trellis is time - invariant . at stage @xmath12 ,
the trellis has @xmath210 states , each of which corresponds to a vector of hamming weights @xmath211 .
a state @xmath212 at stage @xmath12 and a state @xmath213 at stage @xmath214 are connected ( with a branch denoted by @xmath215 ) if and only if @xmath216 for @xmath217 , where @xmath218 and @xmath219 are the @xmath49-th components of @xmath212 and @xmath213 , respectively .
evidently , emitting from ( or entering into ) each state , there are @xmath220 branches . associated with a branch @xmath215
are a deterministic input weight @xmath221 but a random redundancy weight due to the existence of random interleavers .
the weight distribution of the parity check vector @xmath222 is given by @xmath223 where @xmath224 is interpreted as the probability of current outputs @xmath222 having weight @xmath225 given that the weight vector of previous @xmath7 input blocks @xmath226 and the current input weight @xmath227 . by symmetry , it is easy to see that the weight distribution of @xmath228 for @xmath229 is the same as @xmath230 .
since the parity check vector @xmath30 is obtained by randomly puncturing @xmath28 of @xmath5 bits in @xmath29 , the weight distribution of @xmath30 is given by is equal to zero for @xmath231 or @xmath232 .
] @xmath233 to calculate the probability @xmath224 , we define @xmath234 as the probability that a vector of length @xmath5 has weight @xmath225 , given that the vector is obtained by superimposing two randomly interleaved vectors of ( respective ) weights @xmath235 and @xmath236 .
define @xmath237 .
following the same lines as the method in section iv - b of @xcite , the probability @xmath234 is given by @xmath238 then , @xmath224 can be calculated as described in algorithm [ alg : f_pqr ] .
computing the probability @xmath224[alg : f_pqr ] 1 .
initialize a vector @xmath239 of length @xmath220 such that its components are all zero except that the @xmath221-th component is 1 . 2 .
for @xmath240 , @xmath241 , @xmath37 , @xmath242 , compute @xmath243 for @xmath244 , where @xmath245 is the @xmath246-th component of @xmath247 and @xmath248 is the @xmath49-th component of @xmath212 .
we have @xmath249 for @xmath250 , @xmath241 , @xmath37 , @xmath5 .
finally , @xmath251 can be calculated recursively by performing a forward trellis - based algorithm @xcite over the polynomial ring in algorithm [ alg : iowef ] .
computing irwef of systematic bmst - r codes[alg : iowef ] 1 .
initialize @xmath252 if @xmath212 is the all - zero state ; otherwise , initialize @xmath253 .
2 . for @xmath254 , @xmath241 , @xmath37 , @xmath38 , for each state @xmath213 , @xmath255 where @xmath256 is the first component of @xmath213 .
3 . at time @xmath257
, we have @xmath258 .
* remarks .
* the summation for step 2 ) in algorithm [ alg : iowef ] is over @xmath220 possible states @xmath212 for a given state @xmath213 .
the computation of algorithm [ alg : iowef ] becomes more complicated and even intractable for large @xmath7 and/or @xmath5 due to the huge number of trellis states @xmath210 .
fortunately , as shown in section [ subsec : upperbound ] , we can calculate bounds by the use of a truncated irwef , which can be obtained by removing certain states and branches from the trellis . specifically , for a given truncating parameter @xmath259 which corresponds to the maximum input weight , we remove all the branches @xmath215 with @xmath260 and keep only those terms @xmath261 with @xmath262 of the polynomial @xmath263 for @xmath264 . from corollary [ corollary3 ] ,
the upper bounds may be improved by increasing the truncating parameter @xmath259 , which usually needs more computational and memory loads . however , the lower bound ( as well as the upper bound in the high snr region ) is dominated by the crwefs with input weights 1 and 2 , which can be given explicitly as below .
we first show the crwefs for a systematic bmst - r code ensemble without puncturing .
we have @xmath265 for the crwef @xmath266 , we consider the following three cases . 1
. the two non - zero information bits are in the same layer . in this case , the corresponding crwef @xmath267 is given by @xmath268 2 .
the two non - zero information bits are in two different layers with a gap @xmath246 ( spaced away from @xmath269 layers ) satisfying that @xmath270 . in this case , the corresponding crwef @xmath271 is given by @xmath272 @xmath273 3 .
the two non - zero information bits are in two different layers with a gap @xmath246 satisfying that @xmath274 . in this case , the corresponding crwef @xmath275 is given by @xmath276 in summary , the crwef @xmath266 for a systematic bmst - r code ensemble without puncturing is given by @xmath277 then , we consider the crwefs for a systematic bmst - r code ensemble with @xmath28 bits in each layer punctured .
taking into account the puncturing effect , when @xmath28 bits of a sequence with length @xmath5 and weight 1 are randomly punctured , the resulting weight enumerator @xmath278 is given by @xmath279 when @xmath28 bits of a sequence with length @xmath5 and weight 2 are punctured , the resulting weight enumerator @xmath280 is given by @xmath281 @xmath282 then we have @xmath283 @xmath284 for the crwef @xmath266 , we consider the following three cases . 1
. the two non - zero information bits are in the same layer . in this case , the corresponding crwef @xmath267 is given by @xmath285 2 .
the two non - zero information bits are in two different layers with a gap @xmath246 satisfying that @xmath270 . in this case , the corresponding crwef @xmath271 is given by @xmath286 3 .
the two non - zero information bits are in two different layers with a gap @xmath246 satisfying that @xmath274 . in this case , the corresponding crwef @xmath275 is given by @xmath287 @xmath288 in summary , the crwef @xmath266 for a systematic bmst - r code ensemble with @xmath28 bits in each layer punctured is given by @xmath289 from theorem [ maplowerbound ] , we know that the bit - error probability for systematic codes under map decoding can be lower - bounded in terms of the minimum hamming weights @xmath71 of @xmath72
. however , these minimum weights are usually not available for a general code .
if this is the case , we can use instead the row weights of the generator matrix to calculate a looser lower bound as shown in corollary [ ldgm_lowerbound ] , where the @xmath91-th row weight can be determined by setting the binary information data @xmath186 to a nonzero sequence with only the @xmath91-th component @xmath290 .
then we have the following two corollaries . [ corollaryensemblelowerbound ]
the bit - error probability of a systematic bmst - r code ensemble under map decoding can be lower - bounded by @xmath291 @xmath292 where @xmath42 is the puncturing fraction .
[ corollary8 ] the bit - error probability of a systematic bmst - r code ensemble without puncturing ( i.e. , with puncturing fraction @xmath300 ) under map decoding can be lower - bounded by @xmath301 for a specific @xmath91 @xmath302 , we can see from ( [ eq : a1 ] ) that the @xmath91-th row of the generator matrix has a deterministic weight @xmath303 thus , the error probability of the @xmath91-th estimated bit of the systematic bmst - r code ensemble under map decoding can be lower - bounded by @xmath304 it follows that the bit - error probability of the systematic bmst - r code ensemble without puncturing under map decoding can be lower - bounded by @xmath305 * remarks .
* corollaries [ corollaryensemblelowerbound ] and [ corollary8 ] also hold for systematic bmst - r codes with specific interleavers for the reason that the interleavers have no effect on the row weights of the generator matrix .
since the lower bound ( [ codelowerbound ] ) without puncturing is equivalent to the lower bound ( [ ensemblelowerbound ] ) with puncturing fraction @xmath300 , in the rest of the paper , we consider for generality the lower bound ( [ ensemblelowerbound ] ) . , @xmath306 and @xmath307 . ] , @xmath306 and @xmath307 . ]
the implementation complexity of systematic bmst - r codes can be analyzed as with their non - systematic counterpart . for encoding ,
the information sequence is partitioned equally into blocks and transmitted directly , while their replicas are interleaved and transmitted in a block markov superposition manner .
this shows that the encoding complexity grows linearly with the encoding memory @xmath10 . for decoding ,
a sliding window decoding algorithm with a tunable decoding delay can be implemented over the normal graph ( see fig .
[ bmst_decoder ] ) . the decoding complexity for node and node of systematic bmst - r codes grows linearly with the encoding memory @xmath10 .
furthermore , similar to non - systematic bmst codes , a decoding delay @xmath308 is required to achieve the lower bound on the performance . as a result
, the decoding complexity for systematic bmst - r codes can be roughly given as @xmath309 , or equivalently , @xmath310 .
the above analysis shows that the decoding complexity is closely related to the repetition degree @xmath18 and the encoding memory @xmath10 , both of which in turn determine the lower bound ( [ ensemblelowerbound ] ) .
this allows us to make trade - offs among efficiency , performance and complexity . to be precise , we consider the following two cases . 1 . fixed snr .
we can observe from the lower bound ( [ ensemblelowerbound ] ) that , for a given snr and ber , the required encoding memory @xmath10 decreases as the repetition degree @xmath18 increases ( accordingly , the code rate decreases ) , resulting in a lower complexity .
[ fig_complexity_diffrate ] shows the decoding complexity for systematic bmst - r codes as a function of code rate that requires an snr of 2 db to achieve bers of @xmath311 , @xmath306 and @xmath307 .
as expected , for fixed ber , the greater the code rate is , the higher the decoding complexity is .
we also see that for fixed code rate , the higher the performance requirement ( equivalently , the more stringent the ber ) is , the higher the decoding complexity is
fixed code rate .
we can observe from the lower bound ( [ ensemblelowerbound ] ) that , for a given rate and ber , the required encoding memory @xmath10 decreases as the snr increases , resulting in a lower decoding complexity .
this is reasonable since more excessive snr is available compared to the corresponding shannon limit .
[ fig_complexity_rate0.5 ] shows the decoding complexity for rate 1/2 systematic bmst - r codes as a function of snr with bers of @xmath311 , @xmath306 and @xmath307 .
as expected , for fixed ber , the greater the snr is , the lower the decoding complexity is .
we also observe that for fixed snr , the more stringent the ber is , the higher the decoding complexity is .
in this section , all simulations are performed assuming bpsk modulation and transmitted over an awgn channel , unless otherwise specified .
the @xmath59 random interleavers ( randomly generated but fixed ) of size @xmath5 are used for encoding . the iterative sliding window decoding algorithm for systematic bmst - r codes
is performed using the parallel ( flooding ) updating schedule within the decoding window with a maximum iteration number of 18 , and the entropy stopping criterion @xcite with a preselected threshold of @xmath312 is employed . in this subsection
, we present an example to study the performance bounds on ber of systematic bmst - r codes .
we consider systematic bmst - r codes with repetition degree @xmath313 and puncturing fraction @xmath300 .
the decoding delay @xmath60 for the sliding window decoding is specified as @xmath314 .
[ example1 ] of systematic bmst - r code ensembles with encoding encoding memory @xmath315 , @xmath53 and @xmath316 in example 1 .
assume @xmath317 blocks of information data , where the information subsequence has length @xmath318 .
the systematic bmst - r code with @xmath315 is equivalent to the independent transmission of rate 0.5 repetition code .
the truncating parameter is set to @xmath319 .
the code rates of systematic bmst - r code ensembles with @xmath315 , @xmath53 and @xmath316 are 0.5 , 0.4878 and 0.4762 , respectively . ] , @xmath53 and @xmath316 in example 1 .
the systematic bmst - r code with @xmath315 is equivalent to the independent transmission of rate 0.5 repetition code .
assume @xmath317 blocks of information data , where the information subsequence has length @xmath318 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the decoding delay @xmath60 is specified as @xmath314 .
the truncating parameter is set to @xmath319 .
the code rates of systematic bmst - r code ensembles with @xmath315 , @xmath53 and @xmath316 are 0.5 , 0.4878 and 0.4762 , respectively . ]
assume that there are @xmath317 blocks of information data to be transmitted , where the information subsequence have length @xmath318 .
we consider systematic bmst - r code ensembles with encoding memory @xmath315 , @xmath53 and @xmath316 , whose code rates are 0.5 , 0.4878 and 0.4762 , respectively . here , the systematic bmst - r code with @xmath315 is equivalent to the independent transmission of rate 0.5 repetition code .
assume that we only calculate the truncated irwef @xmath145 , @xmath320 .
that is , the truncating parameter is set to @xmath319 .
[ fig_spectrum ] shows the spectrum @xmath321 of systematic bmst - r code ensembles , where @xmath322 from fig .
[ fig_spectrum ] , we see that the spectrum of the systematic bmst - r code ensembles with @xmath53 and @xmath316 have less number of codewords with small hamming weights .
we also see that the minimum hamming distances of systematic bmst - r code ensembles with encoding memory @xmath315 , @xmath53 and @xmath316 are 2 , 3 and 4 , respectively .
these indicate that the systematic bmst - r codes have potentially better performance than the independent transmission system .
the simulation results are shown in fig .
[ fig_performancebounds ] , where we observe that 1 .
the lower and upper bounds on the ber performance of systematic bmst - r codes are tight in the high snr region .
2 . the simulated ber performance curves match well with the bounds in the high snr region , indicating that the sliding window decoding algorithm is near optimal in the high snr region .
the systematic bmst - r codes outperform the independent transmission system ( i.e. , the systematic bmst - r code with @xmath315 ) .
furthermore , the systematic bmst - r code with encoding memory @xmath316 outperforms the systematic bmst - r code with @xmath53 . taking into account the rate loss
, the systematic bmst - r code with @xmath316 obtains a net gain of 1.175 db in terms of @xmath323 at a ber of @xmath307 , compared to the systematic bmst - r code with @xmath53 .
given the tightness of the lower bound ( [ ensemblelowerbound ] ) as demonstrated in example [ example1 ] , we have the following simple procedure to construct good codes .
let @xmath324 be the target code rate and @xmath325 be the target ber .
the object is to construct a code with rate @xmath326 , which can approach the shannon limit at the target ber .
a systematic bmst - r code has the following five parameters : repetition degree @xmath18 , information subsequence length @xmath5 , puncturing length @xmath28 , data block length @xmath34 , and encoding memory @xmath7 .
these parameters can be determined as follows . 1 .
determine the repetition degree @xmath18 and puncturing fraction @xmath42 such that @xmath327 .
choose sufficiently large information subsequence length suffices to approach the shannon limit within around half db . ]
@xmath5 and puncturing length @xmath28 such that @xmath328 ; 2 .
find the shannon limit for the given code rate @xmath329 and target ber @xmath325 ; 3 .
determine the minimum encoding memory @xmath7 such that the lower bound of @xmath154 in ( [ ensemblelowerbound ] ) at the shannon limit is not greater than the preselected target ber @xmath325 ; 4 .
choose a data block length @xmath34 such that the rate loss ( i.e. , @xmath330 ) due to the termination is small ; 5 .
generate @xmath59 interleavers randomly .
evidently , the above procedure requires no optimization and hence can be easily implemented .
the encoding memories for some systematic bmst - r codes required to approach the corresponding shannon limits at given target bers are shown in table [ table1 ] . as expected , the lower the target ber is , the greater the required encoding memory @xmath10 is . [ cols="^,^,^,^,^ " , ] in this subsection , we study the impact of various parameters ( e.g. , encoding memory @xmath10 , information subsequence length @xmath5 and decoding delay @xmath60 ) on the ber performance of systematic bmst - r codes with fixed code rate .
note that , as pointed out in section [ subsec : encoding ] , varying repetition degree @xmath18 and puncturing fraction @xmath42 results in systematic codes with different rates .
for simplicity , we focus on @xmath331 systematic bmst - r codes with repetition degree @xmath313 and puncturing fraction @xmath300 .
three regimes are considered : ( 1 ) fixed @xmath10 and @xmath5 , increasing @xmath60 , ( 2 ) fixed @xmath10 and @xmath60 , increasing @xmath5 , and ( 3 ) fixed @xmath5 , increasing @xmath10 ( and hence @xmath60 ) .
systematic bmst - r codes decoded with different decoding delays @xmath60 in example 2 .
information subsequence length @xmath332 , encoding memory @xmath333 and data block length @xmath334 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the corresponding window decoding thresholds and the lower bound are also plotted . ]
consider a systematic bmst - r code with information subsequence length @xmath332 , encoding memory @xmath333 and data block length @xmath334 .
the ber performance of the systematic bmst - r code decoded with different decoding delays @xmath60 is shown in fig .
[ fig_impactparameters_diffdelay ] . the asymptotic threshold performance obtained by using the exit
chart analysis method in @xcite is also included . from fig .
[ fig_impactparameters_diffdelay ] , we observe that 1 .
the ber performance of the systematic bmst - r code decoded with different delays @xmath60 matches well with the corresponding window decoding thresholds in the high snr region .
the ber performance in the waterfall region improves as the decoding delay @xmath60 increases , but it does not improve much further beyond a certain decoding delay ( roughly @xmath335 ) .
the ber performance in the error floor region improves as the decoding delay @xmath60 increases , and it matches well with the lower bound for the systematic bmst - r code with @xmath333 when @xmath60 increases up to a certain point ( roughly @xmath336 ) .
systematic bmst - r codes with different information subsequence lengths @xmath5 in example 3 .
the codes are constructed with encoding memory @xmath333 and data block length @xmath334 , and decoded with decoding delay @xmath336 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the corresponding lower bound is also plotted . ]
consider systematic bmst - r codes constructed with encoding memory @xmath333 , data block length @xmath334 and decoded with decoding delay @xmath336 .
the ber performance of systematic bmst - r codes constructed with different information subsequence lengths @xmath5 is shown in fig .
[ fig_impactparameters_diffcartesianproductorder ] , where we observe that 1 . increasing the information subsequence length @xmath5 can improve waterfall region performance . as expected , this improvement saturates for sufficiently large @xmath5 .
for example , the improvement at a ber of @xmath307 from @xmath337 to @xmath338 , both decoded with @xmath339 , is about 0.25 db , while the improvement decreases to about 0.05 db from @xmath340 to @xmath341 .
2 . the error floors , which are determined by the encoding memory and code rate ( see corollary [ corollaryensemblelowerbound ] ) ,
can not be lowered by increasing @xmath5 .
systematic bmst - r codes constructed with different encoding memories @xmath10 and decoded with decoding delay @xmath342 in example 4 .
the information subsequence length of the involved systematic bmst - r codes is @xmath343 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the corresponding window decoding thresholds and lower bounds for systematic bmst - r codes are also plotted . ]
consider systematic bmst - r codes constructed with information subsequence length @xmath343 and encoding memories @xmath344 , @xmath345 and @xmath346 . the performance with sufficiently large decoding delay @xmath342 of the systematic bmst - r codes is shown in fig .
[ fig_impactparameters_diffmemory ] , where we observe that 1 .
the ber performance of systematic bmst - r codes matches well with the corresponding window decoding thresholds in the high snr region .
2 . for a high target ber ( roughly above @xmath311 )
, the ber performance improves as the encoding memory @xmath10 increases , which is consistent with the threshold analysis performance .
note that this phenomenon does not exist for non - systematic bmst codes whose performance degrades slightly as @xmath10 increases ( see section v - c of @xcite ) .
the error floor can be lowered by increasing the encoding memory @xmath10 ( and hence the decoding delay @xmath60 ) .
in addition to decoding performance , the latency introduced by employing channel coding is a crucial factor in the design of a practical communication system , such as personal wireless communication and real - time audio and video . in this subsection
, we study the ber performance of systematic bmst - r codes based on their decoding latency . for finite - length systematic bmst - r codes ,
non - systematic bmst - r codes in @xcite , @xmath347-regular sc - ldpc codes , and @xmath348-regular sc - ldpc codes as a function of decoding latency .
all the codes have rate 0.49 .
the decoding delays for @xmath347-regular sc - ldpc codes and @xmath348-regular sc - ldpc codes are @xmath349 and @xmath350 , respectively .
the encoding memories for non - systematic bmst - r codes and systematic bmst - r codes are 8 and 16 , respectively .
the values of the information subsequence length and decoding delay for the non - systematic bmst - r codes are chosen such that the combination gives the best decoding performance .
the decoding delays for the systematic bmst - r codes are @xmath339 , @xmath351 , @xmath37 , @xmath352 .
the codeword is modulated using bpsk and transmitted over an awgn channel . ]
we consider rate @xmath331 systematic bmst - r codes with encoding memory @xmath333 , repetition degree @xmath313 and puncturing fraction @xmath300 .
the decoding latency of the sliding window decoder , in terms of bits , is given by @xmath353 the snr required to achieve a ber of @xmath307 as a function of decoding latency is shown in fig .
[ fig_equaldecodinglatency ] .
we observe that the performance of systematic bmst - r codes ( with fixed information subsequence length @xmath5 ) improves as the decoding delay @xmath60 ( and hence the latency ) increases , but it does not improve much further beyond a certain decoding delay .
moreover , beyond a certain latency , using a greater information subsequence length @xmath5 with a smaller decoding delay @xmath60 gives better performance .
for example , the systematic bmst - r code constructed with a greater information subsequence length @xmath332 and decoded with a smaller decoding delay @xmath354 outperforms the systematic bmst - r code constructed with a small information subsequence length @xmath355 and decoded with a greater decoding delay @xmath356 ( both have the same decoding latency of 12000 bits ) .
we also compare the performance of systematic bmst - r codes , non - systematic bmst - r codes in @xcite , and sc - ldpc codes when the decoding latencies are equal , as shown in fig .
[ fig_equaldecodinglatency ] .
all the codes have rate 0.49 .
we restrict consideration to @xmath347-regular sc - ldpc codes with two component submatrices @xmath357 $ ] and @xmath358 $ ] , and @xmath348-regular sc - ldpc codes with two component submatrices @xmath359 $ ] and @xmath360 $ ] .
the decoding delays for @xmath347-regular sc - ldpc codes and @xmath348-regular sc - ldpc codes are @xmath349 and @xmath350 , respectively , which are good choices to achieve optimum performance when the decoding latencies are fixed .
the encoding memories for non - systematic bmst - r codes and systematic bmst - r codes are 8 and 16 , respectively .
the values of the information subsequence length and decoding delay for the non - systematic bmst - r codes are chosen such that the combination gives the best decoding performance ( see section vi - a of @xcite ) .
the decoding delays for the systematic bmst - r codes are @xmath339 , @xmath351 , @xmath37 , @xmath352 .
we observe that the systematic bmst - r codes perform better than both the non - systematic bmst - r codes and the sc - ldpc codes .
for example , in the decoding latency of 12000 bits , the systematic bmst - r code with information subsequence length @xmath332 and decoding delay @xmath354 gains 0.12 db , 0.21 db and 0.24 db , respectively , compared to the non - systematic bmst - r code , @xmath347-regular sc - ldpc code , and @xmath348-regular sc - ldpc code . in this subsection , we show the performance of systematic bmst - r codes with different rates by varying repetition degree @xmath18 and puncturing fraction @xmath42 . and data block length @xmath361 . the repetition degree @xmath18 ,
encoding memories @xmath7 and puncturing fraction @xmath42 are specified in the legends .
the decoding delay is specified as @xmath362 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the corresponding lower bounds ( dotted magenta ) for systematic bmst - r codes are also plotted .
the rates of the systematic bmst - r codes corresponding to the ber curves from left to right in the figure are @xmath363 and @xmath364 . ] for systematic bmst - r codes with information subsequence length @xmath365 .
the codeword is modulated using bpsk and transmitted over an awgn channel .
the performances of three ar4ja ldpc codes with code rates @xmath366 , @xmath367 and @xmath368 in the ccsds standard @xcite , and five pbrl ldpc codes @xcite with code rates @xmath369 , @xmath370 , @xmath366 , @xmath367 , and @xmath368 , all of which have information length @xmath371 , are also included . ] consider systematic bmst - r codes with information subsequence length @xmath365 and data block length @xmath361 .
the encoding memories @xmath7 for systematic bmst - r codes required to approach the shannon limits at a target ber of @xmath307 are determined following the procedure described in section [ seciv - a ] .
the decoding delay is specified as @xmath362 .
simulation results for systematic bmst - r codes with different rates are shown in fig .
[ fig_bermultirate ] .
we observe that the performances for all code rates are almost the same as that for uncoded code in the relatively low snr region .
this is different from non - systematic bmst codes whose performance in the relatively low snr region is very bad due to error propagation .
we also observe that , as the snr increases , the performance curves of the systematic bmst - r codes drop down to the respective lower bounds for all considered code rates .
to evaluate the bandwidth efficiency , we plot the required snr to achieve a ber of @xmath307 of the systematic bmst - r codes with information subsequence length @xmath365 against the code rate in fig .
[ fig_requiredsnrmultirate ] , where we observe that the systematic bmst - r codes achieve the ber of @xmath307 within one db from the shannon limits for all considered code rates . in fig .
[ fig_requiredsnrmultirate ] , we also include the simulation results of three ar4ja ldpc codes with code rates @xmath366 , @xmath367 and @xmath368 in the ccsds standard @xcite , and five pbrl ldpc codes @xcite with code rates @xmath369 , @xmath370 , @xmath366 , @xmath367 , and @xmath368 , all of which have information length @xmath371 .
we observe that the systematic bmst - r codes have a similar performance as both ar4ja ldpc codes and pbrl ldpc codes over such code rates .
note that no simulation results were reported for ar4ja ldpc codes and pbrl ldpc codes with rates less than @xmath369 , while codes of all rates of interest in the interval ( 0,1 ) can be constructed using the systematic bmst - r construction .
actually , to the best of our knowledge , no other methods were reported along with simulations in the literature that can construct good rate - compatible codes over such a wide range of code rates .
all the examples above assume that the subcodewords are modulated using bpsk and transmitted over an awgn channel . in this subsection ,
we study the performance of systematic bmst - r codes transmitted over a block fading channel .
the word - error - rate ( wer ) is defined as the ratio between the number of erroneous subcodewords and the total number of subcodewords transmitted .
assume that the subcodeword @xmath40 is modulated using bpsk with 0 and 1 mapped to @xmath43 and @xmath44 , respectively , and transmitted over a block fading channel , resulting in a received vector @xmath45 expressed as @xmath372 for @xmath373 , where @xmath48 is the @xmath49-th component of @xmath45 , @xmath50 is a sample from an independent gaussian random variable with distribution @xmath51 , and @xmath374 is a fading coefficient . in this paper
, we consider a rayleigh fading channel , where @xmath374 is a sample from a rayleigh distribution @xmath375 with @xmath376=1 $ ] . for block fading channels with a coherence period of @xmath377 symbols , we assume that @xmath374 ( perfectly known at the receiver ) remains constant over @xmath377 symbols within the same period and is independent identically distributed across different coherence periods . systematic bmst - r codes with information subsequence length @xmath378 , repetition degree @xmath313 and puncturing fraction @xmath300 over a block fading channel .
the encoding memories are @xmath379 and @xmath380 .
the decoding delay is specified as @xmath342 . ] , encoding memory @xmath344 , repetition degree @xmath313 , and puncturing fraction @xmath300 , and decoded with decoding delay @xmath381 .
the @xmath347-regular sc - ldpc codes is constructed with the protograph lifting factor 100 and three component submatrices @xmath382 $ ] , and decoded with decoding delay @xmath381 .
the decoding latencies of two codes are the same . ] consider @xmath331 systematic bmst - r codes with information subsequence length @xmath378 , repetition degree @xmath313 and puncturing fraction @xmath300 .
the subcodewords are modulated using bpsk and transmitted over a block fading channel with a coherence period of @xmath383 symbols .
that is , a subcodeword @xmath40 has @xmath384 independent fading values .
the wer curves of systematic bmst - r codes constructed with encoding memory @xmath379 and @xmath380 , and decoded with decoding delay @xmath342 are shown in fig . [ fig_pbmst_diff_m_fadingchannel ] , where we observe that the performance of systematic bmst - r code improves with increasing encoding memory @xmath10 until @xmath344 and then it degrades slightly as @xmath10 increases further .
this implies that @xmath344 is a good choice for optimum performance .
the performance comparison of the systematic bmst - r code and the sc - ldpc code over a block fading channel is shown in fig .
[ fig_pbmst_scldpc_fadingchannel ] .
the systematic bmst - r code is constructed with information subsequence length @xmath378 , encoding memory @xmath344 , repetition degree @xmath313 , and puncturing fraction @xmath300 , and decoded with decoding delay @xmath381 .
the @xmath347-regular sc - ldpc code is constructed with the protograph lifting factor 100 and three component submatrices @xmath382 $ ] , and decoded with decoding delay @xmath381 presented in @xcite .
thus , the decoding latencies of two codes are the same .
we see from fig .
[ fig_pbmst_scldpc_fadingchannel ] that , in the low wer region , the systematic bmst - r code performs better than the @xmath347-regular sc - ldpc code under the equal decoding latency constraint .
for example , at a wer of @xmath306 , the systematic bmst - r code gains about one db compared to the equal latency @xmath347-regular sc - ldpc code .
these results confirm that systematic bmst - r codes without any further optimization can perform well over block fading channels .
in this paper , we have proposed systematic block markov superposition transmission ( bmst ) of repetition codes , referred to as systematic bmst - r codes . using both extending and puncturing , systematic bmst - r codes support a wide range of code rates but maintain essentially the same encoding / decoding hardware structure .
the systematic bmst - r codes not only preserve the advantages of the original non - systematic bmst codes , namely , low encoding complexity , effective sliding window decoding algorithm and predictable error floors , but also have improved decoding performance especially in short - to - moderate decoding latency . a simple lower bound and an upper bound
were derived to analyze the performance of systematic bmst - r codes under map decoding .
simulation results show that , over an awgn channel , 1 ) the performance of systematic bmst - r codes around or below the ber of @xmath307 can be predicted by the lower bound ; 2 ) systematic bmst - r codes can approach the shannon limit at a ber of @xmath307 within one db for a wide range of code rates ; and 3 ) systematic bmst - r codes can outperform both non - systematic bmst codes and sc - ldpc codes in the waterfall region under the equal decoding latency constraint .
simulation results also show that , systematic bmst - r codes without any further optimization can outperform @xmath347-regular sc - ldpc codes over a block fading channel .
a final note is that the construction of systematic bmst - r codes can be extended to high - order abelian groups since only addition is required during the encoding process .
the authors would like to thank prof .
costello from university of notre dame for his helpful comments on the performance lower bounds . the second author , ever visiting university of notre dame for one year as an exchange student , is also grateful to prof .
costello for his insightful supervision on the related research .
the authors would also like to thank dr . chulong liang and dr .
jia liu for helpful discussions .
s. kudekar , t. j. richardson , and r. l. urbanke , `` threshold saturation via spatial coupling : why convolutional ldpc ensembles perform so well over the bec , '' _ ieee trans .
theory _ , vol .
57 , no . 4 , pp . 803834 ,
m. lentmaier , a. sridharan , d. j. costello , jr . , and k. s. zigangirov , `` iterative decoding threshold analysis for ldpc convolutional codes , '' _ ieee trans .
inf . theory _
56 , no .
52745289 , oct .
s. kudekar , t. j. richardson , and r. l. urbanke , `` spatially coupled ensembles universally achieve capacity under belief propagation , '' _ ieee trans .
theory _ , vol .
59 , no . 12 , pp .
77617813 , dec . 2013 .
x. ma , c. liang , k. huang , and q. zhuang , `` block markov superposition transmission : construction of big convolutional codes from short codes , '' _ ieee trans .
inf . theory _ ,
61 , no . 6 , pp .
31503163 , june 2015 .
a. r. iyengar , m. papaleo , p. h. siegel , j. k. wolf , a. vanelli - coralli , and g. e. corazza , `` windowed decoding of protograph - based ldpc convolutional codes over erasure channels , '' _ ieee trans .
inf . theory _ ,
58 , no . 4 ,
pp . 23032320 , apr .
2012 .
a. e. pusane , a. j. felstrm , a. sridharan , m. lentmaier , k. s. zigangirov , and d. j. costello , jr .
, `` implementation aspects of ldpc convolutional codes , '' _ ieee trans . commun . _ , vol . 56 , no . 7 , pp .
10601069 , july 2008 .
s. kumar , a. j. young , n. macris , and h. d. pfister , `` threshold saturation for spatially coupled ldpc and ldgm codes on bms channels , '' _ ieee trans .
theory _ , vol .
60 , no . 12 , pp .
73897415 , dec . 2014 .
i. sason and s. shamai , `` performance analysis of linear codes under maximum - likelihood decoding : a tutorial , '' in _ foundations and trends in communications and information theory _ , vol . 3 , no . 1 - 2.1em plus
0.5em minus 0.4emdelft , the netherlands : now , 2006 , pp . 1225 .
k. huang , d. g. m. mitchell , l. wei , x. ma , and d. j. costello , jr .
, `` performance comparison of ldpc block and spatially coupled codes over gf@xmath385 , '' _ ieee trans .
_ , vol .
63 , no . 3 , pp . 592604 , mar .
2015 . _ tm synchronization and channel coding
summary of concept and rationale _ , consultative committee for space data systems ( ccsds ) std .
, nov . 2012 .
[ online ] .
available : http://public.ccsds.org/publications/archive/130x1g2.pdf n. ul hassan , m. lentmaier , i. andriyanova , and g. p. fettweis , `` improving code diversity on block - fading channels by spatial coupling , '' in _ proc .
symp . on inf .
theory _ , honolulu , hi , june 2014 , pp . 23112315 .
|
in this paper , we propose systematic block markov superposition transmission of repetition ( bmst - r ) codes , which can support a wide range of code rates but maintain essentially the same encoding / decoding hardware structure .
the systematic bmst - r codes resemble the classical rate - compatible punctured convolutional ( rcpc ) codes , except that they are typically non - decodable by the viterbi algorithm due to the huge constraint length induced by the block - oriented encoding process .
the information sequence is partitioned equally into blocks and transmitted directly , while their replicas are interleaved and transmitted in a block markov superposition manner . by taking into account that the codes are systematic
, we derive both upper and lower bounds on the bit - error - rate ( ber ) under maximum _ a posteriori _ ( map ) decoding .
the derived lower bound reveals connections among ber , encoding memory and code rate , which provides a way to design good systematic bmst - r codes and also allows us to make trade - offs among efficiency , performance and complexity .
numerical results show that : 1 ) the proposed bounds are tight in the high signal - to - noise ratio ( snr ) region ; 2 ) systematic bmst - r codes perform well in a wide range of code rates ; and 3 ) systematic bmst - r codes outperform spatially coupled low - density parity - check ( sc - ldpc ) codes under an equal decoding latency constraint
. block markov superposition transmission ( bmst ) , lower bounds , maximum _ a posteriori _ ( map ) decoding , rate - compatible codes , upper bounds , sliding window decoding , systematic codes .
|
the difficulty of directly observing molecular hydrogen ( @xmath0 ) , the major constituent of the interstellar medium in galaxies , and ways of doing so indirectly are reviewed elsewhere in this volume ( combes 2000 ) .
usually , @xmath0 cloud properties are derived by extrapolation from more easily conducted co observations .
for instance , observed co cloud sizes and velocity widths yield total molecular gas masses under the assumption of virial equilibrium .
however , in extragalactic systems especially , this method is beset by pitfalls ( see israel , 1997 , hereafter is97 ) and requires high linear resolutions ( i.e. use of interferometer arrays ) .
more seriously , _ the fundamental assumption of virialization appears to be false_. as individual components ( ` clumps ' ) have velocities of only a few km s@xmath2 and co complex sizes are 50100 pc , crossing times are comparable to co complex _ lifetimes _ of only a few times 10@xmath3 years or less ( leisawitz et al .
1989 ; fukui et al .
2000 ; see also elmegreen 2000 ) . as equilibrium can not be reached in a single crossing time or less , the virial theorem is not applicable to such complexes .
indeed , the elongated and interconnected filamentary appearance of many large co cloud complexes do not suggest virialized systems ( see also maloney 1990 ) .
the observed co intensity is the weighted product of co brightness temperature and emitting surface area ; actual co column densities are completely hidden by high optical depths .
however , in large beams co cloud ensembles may be assumed to be statistically identical so that co intensities scale with co mass within the beam , i.e. beam - averaged co column density .
if we can determine the proportionality , the @xmath0-to - co conversion factor @xmath4 , subsequent co measurements can be used to find the appropriate @xmath0 column density and mass . in the milky way ,
the calibration of @xmath4 is controversial by a factor of about two ( cf .
combes 2000 ) , and frequently based on application of the virial theorem ( but see preceding paragraph ... ) . in other extragalactic environments ,
the assumption of statistical co cloud ensemble similarity becomes questionable .
very clumpy , even fractal molecular clouds are very sensitive to e.g. variations in radiation field intensity and metallicity .
as @xmath0 and co , supposedly tracing @xmath0 , react differently to such variations , @xmath4 is also very sensitive to them ( maloney @xmath5 black 1988 ) .
the determination of the dependence of @xmath4 on metallicity and radiation field intensity , needed to correctly estimate amounts of @xmath0 in environments ( dwarf galaxies , galaxy centers ) different from the solar neighbourhood thus requires @xmath0 mass determinations independent of co.
fortunately , @xmath0 and hi column densities are traced by optically thin continuum emission from associated dust particles . unfortunately
, dust emissivities depend strongly on temperature , dust particle properties are not accurately known and dust - to - gas ratios are frequently uncertain .
the effect of these uncertainties are minimized if we can avoid the need for determining absolute values of the dust column density and the dust - to - gas ratio .
far - infrared / submillimeter continuum fluxes and hi intensities from spatially nearby positions , preferably in dwarf galaxies that lack strong temperature or metallicity gradients , can be used to obtain reasonably accurate @xmath0 column densities ( is97 ) .
the ratio of dust continuum emission to hi column density at locations lacking substantial molecular gas provides _ a measure _ for the dust - to - gas column density ratio . without requiring its absolute value
, we can apply this measure to a nearby location rich in molecular gas to find the total hydrogen column density and , after subtraction of hi , the @xmath0 column density .
division by the co intensity yields the local value of @xmath4 in absolute units with an accuracy better than a factor of two ( is97 ) .
individual molecular cloud complexes in the nearby magellanic clouds were used by is97 to determine the effects of radiation field intensity ( as sampled by far - infrared surface brightness ) on @xmath4 . over a large range of intensities , @xmath4 is linearly proportional to the radiative energy available per nucleon ( @xmath6 ) .
quiescent regions in the lmc yield @xmath4 values close to those of the solar neighbourhood , whereas a value 40 times higher is obtained for the radiation - saturated 30 doradus region .
the more metal - poor smc exhibits higher @xmath4 values , but again linearly proportional to @xmath6 .
to further study the relation between @xmath4 and metallicity , we have added several recent results to the database given by is97 . these include data for ngc 7331 ( 3 points ; israel @xmath5 baas 1999 ) , the milky way center and centers of ngc 253 ( both from dahmen et al .
1998 ) , ngc 891 ( gulin et al . 1993 ; israel et al . 1999 ) , ngc 3079 ( braine et al . 1997
; israel et al .
1998a ) as well as ic 10 ( madden et al . 1997 ) and d 478 in m 31 ( israel et al .
although they were obtained somewhat differently from those in is97 , they are quite comparable ( figs . 1 and 2 ) . ( 6.5,9 ) ( 6.5,9 ) in fig .
1 , radiation - corrected values @xmath7 = @xmath8 are plotted against metallicity [ o]/[h ] . in fig .
2 , values of @xmath4 are plotted in in its usual form .
1 and 2 yield the relations : @xmath9/[h])\,+\,33.9\ ] ] and @xmath10/[h])\,+\,12.2 \label{xplot}\ ] ] with a larger sample size , these results differ only slightly from those published by is97 .
the points representing high - metallicity regions in ngc 7331 extend rather well along the relation defined by the low - metallicity dwarfs , as do those representing the galaxy centers with a larger scatter .
both correlations are highly significant .
thus , _ eqn .
( 3.2 ) should in general be used to convert co intensities observed in large beams to obtain @xmath0 column densities _ within a factor of about two .
note that the result may greatly differ from that obtained by applying ` standard ' milky way conversion factors ( i.e. lower by a factor of 410 for high - metallicity galaxy centers and higher by a factor of 10100 for low - metallicity irregular dwarf galaxies ) . in fig .
2 , we have also included @xmath4 values derived by virial theorem application to co clouds mapped with interferometer arrays taken from wilson ( 1995 replacing her m 31 and m 33 metallicities by those from garnett et al .
1999 ) , taylor @xmath5 wilson ( 1998 ) and taylor et al .
these points define a different dependence of @xmath4 on metallicity , with a much shallower slope of only -1.0 .
generally , these @xmath4 values are much lower than those in is97 .
a _ steep _ dependence of @xmath4 on metallicity can be understood within the context of photon - dominated regions ( pdrs ) . in weak radiation fields and at high metallicities , neither @xmath0 or co suffers much from photo - dissociation , and the co volume will fill practically the whole @xmath0 volume . however , when radiation fields become intense , co photo - dissociates more rapidly than @xmath0 because it is less strongly selfshielding .
thus , the projected co emitting projected area will shrink and no longer fill that of @xmath0 .
the observed co intensity , proportional to the shrinking emitting area , then requires use of a _ higher _ @xmath4 factor to obtain the correct , essentially unchanged @xmath0 mass . we have found that at constant metallicity , @xmath4 must be increased linearly with radiation field intensity
. we may somewhat quantitatively estimate the effects of metallicity on co ( self)shielding and thereby on @xmath4 . from garnett et al .
( 1999 ) we find that over the range covered by figs . 1 and 2 , log [ c]/[h ] @xmath11 1.7 log [ o]/[h ] .
thus , co abundances drop significantly more rapidly than metallicity [ o]/[h ] , as do dust abundances given by @xmath12 @xmath11 2 log [ o]/[h ] ( lisenfeld @xmath5 ferrara 1998 ) .
thus , a ten times lower metallicity ( cf .
figs 1 and 2 ) implies a [ co]/[@xmath0 ] ratio lower by a factor of 50 , and less co shielding by a factor of 5000 !
the precise effect on @xmath4 depends on the nature of the cloud ensemble , but at lower metallicities pdr effects very quickly increase in magnitude . in a standard @xmath0 column , there is less co to begin with , and this smaller amount is even less capable of resisting further erosion by photodissociation . with decreasing metallicity , co is losing both its selfshielding and its dustshielding , so that even modestly strong radiation fields completely dissociate extended but relatively low - density diffuse co gas , leaving only embedded smaller higher - density co clumps intact .
as co intensities primarily sample emitting surface area , the loss of extended diffuse co strongly reduces them , even when actual co mass loss is still modest .
further metallicity decreases cause further erosion and molecular clumps of ever higher column density lose their co gas .
co is thus occupying an ever - smaller fraction of the @xmath0 cloud which still fills most of the pdr .
its destruction releases a large amount of atomic carbon which is ionized and forms a large and bright cloud of cii filling the entire pdr .
this and the expected anticorrelation between co and cii intensities is indeed observed in the magellanic clouds and in ic 10 ( is97 ; israel et al .
1996 ; madden et al . 1997 ; bolatto et al .
2000 ) . as the strongly selfshielding @xmath0 is still filling most of the pdr ( cf .
maloney @xmath5 black 1988 ) , the appropriate value of @xmath4 becomes ever higher . in the extreme case of total co dissociation
, any amount of @xmath0 left defines an infinitely large value of @xmath4 ! in contrast , use of e.g. interferometer maps to find resolved co clouds for the determination of @xmath4 introduces a strong bias in low - metallicity environments . in the pdr ,
only those subregions are selected which have most succesfully resisted co erosion , with co still filling a relatively large fraction of the local @xmath0 volume .
the relatively low @xmath4 values thus derived , although appropriate for the selected pdr subregions , are not at all valid for the remaining pdr volume where co has been much weakened or has disappeared ; the pdr has a much higher overall @xmath4 value than the selected subregion .
it is because of this bias that the array - derived points in fig .
2 are much lower than the large - beam points and exhibit a much weaker dependence on metallicity .
incidentally , it also explains the suggested dependence of @xmath4 on observing beamsize ( rubio et al .
1993 ) .
2000 _ ap . j. _ in prep .
1997 _ a&a _ * 326 * , 963975
. 2000 _ these proceedings _ , * * * * * * 1998 _ a&a _ * 351 * , 959976 . 2000 _ ap . j. _ , in press 2000 _
p. a. .s .
j. _ in press 1999 _ ap .
j. _ * 513 * , 168179 .
1993 a&a * 279 * , l37l40 .
1997 _ a&a _ * 328 * , 471482 , ( is97 ) .
1996 _ ap . j. _ * 465 * , 738747 .
1998a _ a&a _ * 336 * , 433444 . 1998b _
a&a _ * 339 * , 398404 . 1999
_ a&a _ * 344 * , l83l86 . 1999 _ a&a _ * 351 * , 1020 .
1989 _ ap . j. suppl . _ * 70 * , 731812 1998 _ ap .
j. _ * 496 * , 145154 .
1997 _ ap .
j. _ * 483 * , 200209 .
1990 _ ap . j. l. _ * 348 * , l9l12 1988 _ ap . j. _ * 325 * , 389401 . 1993 _ a&a _ * 271 * , 917 .
1998 _ ap .
j. _ * 494 * , 581586 .
a&a _ * 349 * , 424435 . 1995 _ ap . j. _ * 448 * , l97l100 .
|
we derive and discuss the strong dependence on metallicity of the co to @xmath0 conversion factor @xmath1/[h]$ ] appropriate to extragalactic objects , as well as the weaker dependence found for such objects from interferometer measurements . 2 2 # 1#1 0123456789 abcdef
@mtlplain@loaded = cmssi10 = cmssi10 = cmssi10 = cmmib10 = mtssi10 = mtssbx10 = mtssbi10 = mtmib10 @mtlplain@loaded
|
the atmospheric neutrino anomaly observed by super - kamiokande(sk ) and other recent underground experiments strongly suggests @xmath3 neutrino oscillation .
the allowed region of the oscillation parameters obtained by sk are in the range of @xmath4 ev@xmath5 and @xmath6 0.88 with 90% confidence level@xcite , where @xmath7 is the mass difference squared between two neutrino mass eigenstates and @xmath8 is the mixing angle between two neutrinos .
the principal goal of the k2k experiment is to confirm neutrino oscillation with a man - made neutrino beam and to measure the oscillation parameters .
[ fig1 ] is a schematic of the setup .
we use the 12 gev kek - ps as a neutrino source and sk as the far detector .
the distance between kek and sk is 250@xmath9 , and the neutrino beam has an average energy of 1.4 gev .
the neutrinos are produced by charged pion decays and expected to be 99% pure @xmath10 with an angular deviation @xmath11 3 @xmath12 . in order to measure the effects of oscillation we compare the @xmath10 spectrum observed by sk to the one measured in the front detectors at the point of production .
[ fig2](a ) shows the monte carlo simulation of reconstructed neutrino spectra at sk for @xmath13 protons on target(@xmath14 ) .
this corresponds to approximately 5 years of measurement .
oscillated spectra for four specific oscillation parameter sets are shown .
the effects of oscillation are seen in the figures as a significant divergence from the spectra of the null oscillation case , which is given by open histograms .
k2k s 90% confidence level sensitivity covers almost all of the 90% allowed regions obtained by kamiokande and sk , as shown in fig .
[ fig2](b ) .
[ fig3 ] shows the record of protons on target . after the completion of the front detector construction in feb .
1999 , the fast extraction of protons for the experiment started on feb .
3 . on mar .
4 , the neutrino beam daq started with a horn current of 175 ka and with proton intensity of 3@xmath010@xmath15 .
after engineering runs to study neutrino beam operations in april through may , stable data taking began in june with an aluminum target with 2@xmath16 , a horn current of 200 ka , and a proton intensity of 4.5@xmath17 . after the summer shutdown , continuous data taking began again in november , this time with an aluminum target of @xmath18 and a horn current of 250 ka with 5@xmath17 . in 1999
we accumulated 7.2 @xmath19 in total .
the near detector system consists of a 1kt water cherenkov detector(1kt ) and a fine - grained detector(fgd ) .
the latter consists of a scintillating fiber tracker(sft)@xcite , plastic scintillator veto counters , an electromagnetic calorimeter of 600 lead glass blocks , and a muon ranger of 12 iron plates ( 10@xmath204 + 20@xmath208 ) with drift chambers(muc ) .
the sft is composed of 19 layers of 6@xmath21-thick water containers sandwiched with 20@xmath0(@xmath22-@xmath23 ) layers of 700@xmath24 scintillating fibers .
muc contained events are used to check the stability of the neutrino beam .
this is because the large fiducial volume mass of the muc results in a very large event rate .
the beam center is observed to be stable to within 1@xmath25 .
in addition , spill by spill beam centering , monitored by the profile of the muons produced by pion - decay that penetrate the beam dump , was also found to be within 0.5@xmath12 .
[ fig4](a ) is an example of quasi - elastic(qe ) neutrino event candidates with vertex in the sft .
the primary goal of the fgd is to reconstruct the neutrino spectrum by using qe samples .
( b ) shows the reconstructed muon energy distribution for single track events , and ( c ) is @xmath26 distribution for 2 track samples , where @xmath27 is angular difference between reconstructed proton track and that calculated from muon momentum .
the single track samples and 2 track samples with @xmath28 contain a large fraction of qe events , @xmath2960% and @xmath2980% respectively .
we also study the ratio of inelastic events to qe events , in order to reduce the uncertainty of the calculated neutrino interaction cross section to less than 10% .
this information is important not only for k2k , but also for the sk atmospheric neutrino analyses .
table [ table1 ] gives summary of the front detector results .
event numbers normalized by mc predictions agree very well to each other .
|
the first results of the k2k(kek to kamioka ) long - baseline neutrino oscillation experiment are presented in this talk . in 1999 7.2@xmath010@xmath1 protons on target were delivered to the experiment . during this period of running
there were 3 events fully contained in the super - kamiokande inner detector fiducial area which occurred during the beam spill timing window . in the case of no oscillations
the expected number of events during this period was 12.3@xmath2 .
the near detectors located at kek also have begun detailed measurements of neutrino interactions in water at around 1 gev .
|
car parking , first considered in a mathematical way by rnyi @xcite in 1958 , gives rise to interesting models that in several variations have been applied in many fields of science . in the original car parking problem ,
unit length cars are appearing with constant rate in time and with constant density in space on the line where they try to park .
a new car is allowed to park only in case there is no intersection with previously parked cars . otherwise the attempt is rejected .
rnyi proved that the density of cars has the limit @xmath1 , the so - called parking constant . in the simplest discrete version of the car parking problem ,
cars of length @xmath2 try to park at their midpoints randomly on @xmath0 .
this model has been solved analytically as well @xcite .
this model belongs to a wider class of more complicated models of deposition with exclusion interaction .
usually such models are not analytically solvable . in physical chemistry `` cars ''
become particles which are deposited in layers on a substrate , a process called random sequential adsorption ( rsa ) . a variety of related models are studied . for a review of recent developments
see @xcite .
moreover , models with more complicated graphs e.g. ( random ) trees have been investigated @xcite,@xcite,@xcite,@xcite .
multilayer variations of the model are used to describe multilayer adsorption of particles on a substrate @xcite and the sequential frequency assignment process @xcite which appears in telecommunication . in these papers
it is also observed that the density in higher layers increases up from the first layer , which at first seems rather counterintuitive .
heuristic arguments for monotonicity of densities were found in specific models @xcite , but no rigorous proofs could be given yet .
moreover privman finds numerically a scaling behavior of the density in a similar rsa model @xcite with slightly different adhesion rules which is notoriously difficult to explain mathematically . in the present paper we aim for a rigorous investigation and treat two versions of the discrete two - line car - parking problem with cars of length 2 .
first we describe the dynamics of the car parking process without screening and also with screening .
then we provide the _ solutions _ of these models by reducing them to closed finite dimensional systems of odes for densities of local patterns , see theorems [ theo1 ] and [ theo2 ] . that it is possible to find a finite - dimensional dynamical description is quite remarkable .
it is not obvious , and in fact our method ceases to work for a three - line extension of the model without screening where an infinite system appears .
a second remarkable fact is that , even without screening , the second line density is higher than the first .
cars do not communicate or plan a common strategy and their arrival is random , but they seem to use the resources in the second line more efficiently , once they have been rejected in the first line .
we will define a markov jump process on the ( suitably coded ) occupation numbers @xmath3 . here
the _ spin _
@xmath4 denotes the joint occupation numbers at vertex @xmath5 at height @xmath6 and @xmath2 .
it is useful for short notation to interpret the occupation numbers at various heights as binary digits and write ordinary natural numbers .
that is we write @xmath7 so that @xmath8 .
the dynamics of the process is defined in terms of the generator which is given by the right hand side of the differential equation @xmath9 c(s;m_{k-1},m_{k},m_{k+1 } ) \label{generator } \end{split}\ ] ] with @xmath10 denoting the configuration which has been obtained by @xmath11 by changing the configuration in @xmath5 to @xmath12 . here
@xmath13 denotes the expected value with respect to the process , started at the initial configuration @xmath11 .
the rates are either equal to zero or one .
they are @xmath6 precisely in the following cases .
1 . @xmath14 adding a car in the first line at site @xmath5 . for the model without screening we have @xmath15 indeed , this occurs when the site itself is empty on the first and second line and the nearest neighbors are empty in the first line , see figure [ fig:01 ] for an example .
+ ( 150,30)(0,-7 ) ( 20,5)(30,0)4(0,0)(5,0)5(0,1)10 ( 20,5)(30,0)4(1,0)20 ( 20,10)(30,0)4(1,0)20 ( 55,10 ) ( 50,5 ) ( 95,10 ) ( 100,5 ) ( 115,10 ) ( 110,5 ) ( 125,10 ) ( 130,5 ) ( 51,11)(.1,0)80(0,1)3 ( 50,6)(.1,0)40(0,1)3 ( 91,11)(.1,0)80(0,1)3 ( 96,6)(.1,0)40(0,1)3 ( 111,11)(.1,0)80(0,1)3 ( 110,6)(.1,0)40(0,1)3 ( 121,11)(.1,0)80(0,1)3 ( 126,6)(.1,0)40(0,1)3 ( 22.5,0)(30,0)4(0,0 ) ( 29.5,0)(30,0)4(0,0 ) ( 34.5,0)(30,0)4(0,0 ) + in the screening model however , cars in the second line will obstruct cars from reaching the first line .
therefore , in the screening model we have as the only nonvanishing rate @xmath16 2 .
@xmath17 adding a car in the second line at @xmath5 while the first line was empty at the site @xmath18 indeed , this occurs when there was a supporting site @xmath19 or @xmath20 or both with one car in the first line .
this is true for both models .
@xmath21 adding a car in the second line while the first line was full at the site @xmath22 indeed , this occurs when there are no obstructing cars right and left at height @xmath2 .
there can be no obstructing cars right and left at height @xmath6 because there could not be a car in the first line at @xmath5 otherwise .
this is true for both models .
all other transitions are impossible .
this generator defines a markov jump process on the infinite graph by standard theory @xcite , such that ( [ generator ] ) holds for any local function @xmath23 .
we provide a closed system of differential equations for the densities of occupied sites , involving densities of finitely many local patterns , in both models .
first we need some definitions . here and in the following we use for the densities at single sites , and triples of sites the notation @xmath24
further we need the following `` one - sided densities '' @xmath25 where @xmath26 denotes the poisson counting process of events of car arrivals at site @xmath27 . as our main result
we show that the time - evolution of these densities gives rise to a closed ode .
[ theo1 ] two - line parking without screening .
+ the time evolution of the probability vector @xmath28 obeys the following system of differential equations .
@xmath29 with initial conditions @xmath30 , where the vector @xmath31 obeys the linear ode @xmath32 with initial conditions @xmath33 , + and finally , @xmath34 is obeying the equation @xmath35 with @xmath36 . the system above can be solved numerically and the results are depicted in figure [ fig : d ] . as it can be seen in the right figure , surprisingly the value of @xmath37 has a slightly higher limit than @xmath38 .
this clearly means that the second line has a higher limit density of cars than the first line .
this result is independently confirmed by simulations of the parking process measuring the empirical densities . and @xmath37 to show they tend to different values . , title="fig:",height=207 ] and @xmath37 to show they tend to different values .
, title="fig:",height=207 ] @xmath39 a similar system of equations holds for the model with screening . recall that in this model cars are not allowed to pass cars on the second line to reach a void on the first line .
this results in less possibilities of filling voids of the first line than in the model treated above .
in fact we can derive the odes of this model by simply deleting those terms in ( [ osqode ] ) that represent the possibility of `` jumping '' over a car in the second line to reach a void on the first line .
so , we get [ theo2 ] two - line parking with screening + the time evolution of the probability vector @xmath40 obeys the following system of differential equations @xmath41 with initial conditions @xmath42 , where the vector @xmath43 obeys the linear ode @xmath44 with initial conditions @xmath45 , + and finally , @xmath46 is obeying the equation @xmath47 with @xmath48 .
the following lemmas are used to prove our theorems .
[ d ] the probability vector @xmath49 obeys @xmath50 * remark : * summing over the four right hand sides we get zero , due to the fact that we have summed a probability vector .
it is also interesting to check that @xmath51 recovers the ode for the density in the first line . + * proof : * fix an arbitrary vertex .
let us call this vertex @xmath52 .
starting from the dynamics ( [ generator ] ) and using symmetries we have @xmath53 indeed , the first three terms correspond to adding a car in the first line , the next two terms correspond to adding a car in the second line , see figure [ fig : d0 ] .
( 150,30)(0,-7 ) ( 10,5)(20,0)7(0,1)10 ( 10,10)(20,0)7(1,0)12 ( 16,5)(20,0)7(0,1)10 ( 22,5)(20,0)7(0,1)10 ( 10,5)(20,0)7(1,0)12 ( 30,10 ) ( 62,10 ) ( 70,10 ) ( 82,10 ) ( 90,5 ) ( 122,5 ) ( 130,5 ) ( 142,5 ) ( 8.5,0 ) ( 28.5,0 ) ( 48.5,0 ) ( 68.5,0 ) ( 88.5,0 ) ( 108.5,0 ) ( 128.5,0 ) ( 30,11)(.1,0)50(0,1)3 ( 57,11)(.1,0)50(0,1)3 ( 70,11)(.1,0)50(0,1)3 ( 77,11)(.1,0)50(0,1)3 ( 90,6)(.1,0)50(0,1)3 ( 117,6)(.1,0)50(0,1)3 ( 130,6)(.1,0)50(0,1)3 ( 137,6)(.1,0)50(0,1)3 the other three differential equations are derived in a similar way .
the triple site densities @xmath54 and the one - sided densities as defined in [ def : dt ] and [ osq ] respectively are related in the following way @xmath55 for @xmath56 . [ d&osq ] * proof : * we note that for the mentioned choices of @xmath57 conditioning on non - arrival at zero does not change the probability , that is @xmath58 in the next step we note that , conditional on the event that no car has arrived at the site @xmath52 , the dynamics for the two sides that are emerging from @xmath52 is independent .
consequently we have @xmath59 this concludes the proof of the lemma .
next we look at the time - evolution of the `` one - sided densities '' .
[ lem : osq ] the vector @xmath31 obeys the ode @xmath61 with initial conditions @xmath62 and @xmath63 . *
remark 1 : * note that combining the equations of @xmath64 and @xmath65 readily gives @xmath66 which is a known result for the first line in a semi - infinite chain @xcite . * remark 2 : * note also that because latexmath:[$r_t : = p_t(m_{1}=1 , m_{2}=0 * proof : * to derive odes for these densities we employ the generator of the process , while putting the term at the site @xmath52 to sleep , and correspondingly the spin at zero to be the constant @xmath68 . for the first quantity we get @xmath69 next we have @xmath70 and @xmath71 finally , we get @xmath72 using conditioning on non - arrival again we get @xmath73 clearly we have @xmath74 because there is precisely one car at @xmath6 if and only if precisely one car arrived conditioning on no cars at @xmath52 and @xmath2 .
this shows that the last ode is correct and concludes the proof of the lemma.@xmath60 the only remaining term whose time - evolution we need to consider is @xmath34 .
@xmath34 is a solution of the differential equation @xmath35 [ d010 ] * proof : * we note that @xmath75 the first term is for adding a car at the central site from the vacuum , the second for adding a car at the central site at height one .
the last two terms are for adding a car to the right of the central site . as we already know
we have @xmath76 using conditioning on non - arrival at @xmath2 we get , by reflection invariance @xmath77 for the last term we get in the same way @xmath78 @xmath60 + * proof of theorem [ theo1 ] : * combining the results of lemma [ d ] , [ d&osq ] , [ lem : osq ] and [ d010 ] proves theorem 1 .
@xmath60 + * proof of theorem [ theo2 ] : * the proof follows from theorem [ theo1 ] by deleting every term that represents the possibility of skipping a second line car to reach a void in the first line .
this results in deleting @xmath79 from the first two equations of [ lem : osq ] , and @xmath80 and @xmath81 from [ d0 ] and [ d1 ] .
@xmath60 @xmath39
we introduced two extensions of the classic parking problem to a two - line model i.e. a model with screening and a model without screening . for both models we derived closed systems of finite - dimensional odes
from which the time - evolution of the densities in the first and second line can be obtained .
interestingly , the numerical solution of the ode shows that the final densities in the second line are higher than those in the first line , for both models .
the increase factor in the model without screening is approximately @xmath82 it is known by analytical computations @xcite that @xmath83 approaches @xmath84 for large @xmath85 , which provides a checkup for the numerics . in the screening model
we find @xmath86 in other words , in both models the cars seem to exploit the resources in the second line in a ( slightly ) more efficient way than in the first line .
the authors thank aernout van enter and herold dehling for interesting discussions .
|
we consider two variations of the discrete car parking problem where at every vertex of @xmath0 a car arrives with rate one , now allowing for parking in two lines .
+ a ) the car parks in the first line whenever the vertex and all of its nearest neighbors are not occupied yet .
it can reach the first line if it is not obstructed by cars already parked in the second line ( `` screening '' ) .
+ b ) the car parks according to the same rules , but parking in the first line can not be obstructed by parked cars in the second line ( `` no screening '' ) . in both models ,
a car that can not park in the first line will attempt to park in the second line . if it is obstructed in the second line as well , the attempt is discarded .
+ we show that both models are solvable in terms of finite - dimensional odes .
we compare numerically the limits of first and second line densities , with time going to infinity .
while it is not surprising that model a ) exhibits an increase of the density in the second line from the first line , more remarkably this is also true for model b ) , albeit in a less pronounced way .
* ams 2000 subject classification : * 82c22 , 82c23 . car parking problem , random sequential adsorption , sequential frequency assignment process , particle systems .
|
transport networks occur in a vast variety of natural systems . from river basins to leaf venation , we observe a spectrum of solutions to the problem of optimal distribution through a landscape , given a set of constraints on the construction of links .
previous work on treelike networks originates in the study of river basins , which can be modeled as convergent random walks , or equivalently as flow through a linear slope field with additive noise .
this approach , the scheidegger model @xcite , does not incorporate the effects of convergent and divergent topologies , which occur both in geomorphology ( e.g. endorheic basins ) and in biologically relevant systems .
the vasculature of the retina , for instance , may be viewed as divergent for the arteries , which originate at the center , and convergent for the veins , which terminate there .
another example of this morphology occurs in the functional units of organs , such as the liver , have distinct zones where the arteries originate and the veins terminate , respectively . convergence and divergence in river networks originates in the presence of curvature in the underlying topography .
thus , to begin to understand the morphologies of transport networks found in these biologically - relevant examples , we investigate the effects of curving the manifold in which we embed the scheidegger model , from a plane into a cone @xcite .
the study of river networks depends upon the ability to impose a hierarchy on the streams , dividing the main channel from its tributaries . ordering by magnitude , where each node is labeled with the number of drains , becomes rapidly intractable , due to the exponential growth of binary trees .
strahler stream ordering assigns all tips a value of one . at every junction ,
the resulting stream takes the value of the larger of its two inputs , unless they are the same , in which case it is augmented by one @xcite .
binary trees grow exponentially ( the number of nodes within a distance @xmath2 of the center goes as @xmath3 ) and can not be embedded linearly into euclidean space , which grows only algerbraically ( the volume of space goes as @xmath4 , where @xmath5 is the number of spatial dimensions ) .
strahler stream ordering , which requires pairs of order @xmath6 streams to create one of order @xmath7 , is inherently logarithmic and able to fit " a binary tree into euclidean space @xcite .
horton first examined the scaling of the length of , number of and area drained by a link in a river network with its strahler order .
his work and subsequent examinations of examples around the world showed that these three quantities nearly always scale in exponential fashion , defining an empirical range of length , bifurcation , and area ratios respectively @xcite .
these scaling relationships originate in the logarithmic nature of strahler stream ordering .
the addition of curvature , which changes the scaling of area with radius from @xmath8 , affects the ability to accommodate a binary tree .
spaces with negative intrinsic curvature grow exponentially and have no need for logarithmic embedding .
there is thus an underlying connection between horton s scaling laws and curvature of the embedding topology .
numerous other scaling relationships ( usually power law ) have been observed in river networks(see table in @xcite ) .
notably , it has also been found that @xmath0 , with @xmath9 the mainstream length and @xmath10 the basin area , a relationship known as hack law @xcite . by relating basin morphology to a correlate of flow volume ( the area drained ) , hack s law provides a quantification of basin shape , useful for vascular and other distribution trees@xcite .
scheidegger ( 1967 ) proposed a simplified model of river networks , in order to better understand the origin of these scaling laws .
consider a hexagonal grid of points , tilted out of the plane .
each point drains downhill and moves with equally probability to either the right or the left .
basins are collections of convergent random walks .
the hack exponent is 2/3 : a basin is the area between two random walks ( the neighboring streams ) .
this distance is itself a random walk ; the area between two such random walks should scale as @xmath11 @xcite . the scheidegger model
may also be created by adding a slope field plus random noise to a lattice and requiring downhill flow from every point @xcite .
this form is both a simplification of the physics of streamflow and amenable to embedding in convergent and divergent topologies . adding a radial instead of a unidirectional slope embeds the model in a cone .
a negative radial slope models the ( temporary ) divergent flow off of a mountaintop , while a positive value represents the flow into an endorheic basin or a deep valley .
the effects of curvature have been investigated previously for their effects on channelization and stream head formation , but not for their effects on network morphology @xcite .
our procedure was to generate a grid of @xmath12 random points in an annular region of outer radius @xmath13 and inner radius @xmath14 with connectivities defined through the delaunay triangulation . to eliminate edge effects ,
points were placed in the square region of width @xmath15 centered on the origin and then removed if they fell outside the desired annulus .
each point was assigned a height equal to its euclidean distance from the origin times the radial slope of the simulation , @xmath16 , plus frozen white noise , @xmath17 , giving : @xmath18 .
each point was then connected to its neighbor with the lowest value .
if all neighbors had a larger z value , then the point was assumed to be the mouth of a river network .
these points almost always only occurred on the edge of the region , except in the flattest cases , see below . from the resulting connectivities matrix
, we could recursively calculate the magnitude and strahler order numbers at each point in the grid .
additionally , all points that flowed into a single mouth were collected as a basin allowing us to collect volume , length , and order statistics .
the following results are robust to lattice geometry . unless otherwise noted , all simulations were conducted with the following parameters : a radius of 67 units , a radial slope of -1 or 1 , and a noise strength of 0.1 .
at positive and negative values of the radial slope , m , two distinct phases emerge . in the case of a convergent network , @xmath19 , there are relatively few , but large basins , whereas for a divergent network,@xmath20 , there are many more , considerably smaller and narrower basins
. see figure [ networks ] , where each basin is represented by a single color .
the basins in the divergent case all appear to share the same characteristic , leaflike shape : narrow at the top , widening in the center , and tapering towards the mouth , reminiscent of tilings of the poincare disk @xcite . ) and divergent ( @xmath21 ) embedding topologies on the top and bottom , respectively .
basins , defined as collections of points with the same outlet , are colored the same .
note the differences in shape and size of basins . in the convergent case
, most basins take the form of large sectors of the annular region , whereas in the divergent case there are basins at all length scales , with a characteristic leaflike shape ( e.g. the heavy - outlined basins).[networks ] ] iterating simulations from @xmath22 to @xmath23 , to investigate the crossover region , we found a sharp increase in the average number of basins as we approached m=0 with the power - law scaling indicative of a singularity , @xmath24 , ( see figure [ phasetrans ] ) .
additional properties such as the average basin length exhibit a zero , @xmath25 , when m vanishes , again with power law scaling .
these results lead us to conclude that there is a phase transition at zero radial slope for our modified scheidegger model . from both sides ,
the number of basins approaches a singularity and the average of basin length falls to zero , both in power law fashion , with exponents respectively @xmath26 and @xmath27 .
results here come from averaging 40 simulations at @xmath28 .
[ phasetrans ] ] we used finite - size scaling analysis to determine the critical exponent(s ) . conducting simulations at exponentially distributed network radii from 4 to 100 , we determined a value of 1 for the finite - size scaling exponent , as predicted by the conventional theory of finite - size effects @xcite . from these results
, we found a value of @xmath26 for @xmath29 , the critical exponent for @xmath30 . extrapolating to the thermodynamic limit
, we can estimate @xmath30 in the divergent case as @xmath31 and in the convergent case , @xmath32 .
hence , the ratio of basins to lattice points vanishes . beyond the average number and size of a basin , we examine the distribution of basin sizes and shapes . we took the cases @xmath33 at a simulation radius of 67 and aggregated 40 simulation runs .
given the wide dispersal in basin sizes , we binned the data logarithmically and examined the logarithm of the number of counts . in the convergent case , there was a sharp enhancement of basins at the largest size , but little other structure . in the divergent case
, the data fell onto a line of slope @xmath34 , indicating a scale - free distribution with exponent @xmath35 ( see fig [ hack ] , upper panel ) , quantifying the self similarity of basin shapes seen in figure [ networks ] .
examining @xmath36 vs. @xmath37 , in the convergent case , we observe a knee in the distribution , as finite size effects greatly limited mainstream length , see figure [ hack ] , lower panel .
the largest basins were three decades greater in area and two in length than the smallest , matching the scheidegger prediction of @xmath38 . in the divergent case , finite - size effects were generally much less important .
we found a ( nearly ) linear relationship between @xmath10 and @xmath39 .
since basins on a divergent cone tend to be narrow and elongated with little branching , this value is not unexpected .
the hack exponent provides a useful , and easily calculated , metric for the differentiation of convergent and divergent basins .
vs @xmath40 ; data points are x s and a linear regression is overlaid .
note the differing slopes for the convergent case it is approximately 2/3 , the scheidegger value .
for the divergent case it is nearly unity.[hack ] ] the maximum strahler order observed in networks we could reliably simulate , was found to be 6 and to only occur in convergent topologies .
with such a low maximum order , it is impossible to properly define and calculate the bifurcation and other hortonian ratios @xcite .
we examined instead the distribution of strahler numbers and found no difference in the pdf of strahler numbers for 5th and 6th order basins between divergent and convergent networks .
while the size and shape of basins may vary wildly due to the embedding manifold , the hierarchy of the links within each basin does not change .
when one embeds the scheidegger model in a curved manifold , it is not the stream ordering properties that change , but rather the way that the manifold is tiled by the river basins .
we have observed severe changes in the number and length of basins and in the hack exponent , but not the distribution of strahler orders . the same within - basin ordering principles
yield grossly different bulk morphologies at the level of basins .
in essence , the curving of the embedding manifold affects the shapes of the basins themselves , while the area within each basin may be treated as locally flat .
we observe two distinct phases that can only be transformed into one another after passage through a singularity , where the average number of basins blows up to order @xmath41 from order @xmath42 in convergent and divergent embedding topologies . given the minimal specifications of our implementation of the scheidegger model , a noisy set of heights , z , and flow between neighbors chosen by @xmath43 , we may expect that the general features are reflected in natural systems .
specifically , we predict that convergent and divergent networks will exhibit different morphologies , quantified by the hack exponent .
a potential system for testing our model is vascular networks .
as mentioned previously , divergent topography does not lead to permanent channelization , preventing us from comparing the flow off of mountaintops to that into endorheic basins .
two - dimensional vascular networks are typified by leaf venation ; that system , however , does not exhibit a difference between convergence and divergence : flow both to and from the stem is carried in the same veins .
nearly two - dimensional systems may be found in certain tissues , such as the retina . at the smallest level , however , vascular networks become loopy meshes . at larger scales , these networks are treelike .
we hypothesize a morphological difference between arterial and venous trees . assuming a uniform drainage density , here a constant level of blood demand throughout thetissue in question
, we may compare the length of these trees with the cross - sectional area of the terminal arteries and veins , as it correlates with flow volume @xcite .
given a large enough sample of arteries and veins , the distribution of basin " sizes , which in vasculature would correspond with vessel cross - sectional area , could be investigated .
we predict a scale - free distribution of terminal artery sizes and a distribution peaked at the system size for veins . given the simplicity and generality of the assumptions of our model , the absence of these properties would warrant careful study of the mechanisms shaping the growth and form of vasculature . while this system does not pose a perfect test of our model , as it exists in three dimensions , and at its lowest level ( the capillaries )
, it is full of loops , naive models of angiogenesis , in which vascularature growth is dependent on the concentration of vegf , or other chemical signals , map directly onto our version of the scheidegger model @xcite .
the concentration of a signaling molecule originates in a certain point ( or area ) , providing a radial slope due to diffusion , and poisson noise provides the the fluctuations to impart randomness into the gradient .
while nearly all vascular networks exist in three dimensions , we do not expect the dramatic differences between convergent and divergent networks to disappear with the addition of another euclidean dimension .
as binary trees grow exponentially , the difference between two and three dimensions that grow algerbraically should be immaterial .
the differences between convergence , where streams are forced together , and divergence , where the additional space would amplify the scale - free distribution of basin sizes should be amplified . at vanishing @xmath16 , @xmath30
will still be of order @xmath44 , preserving the phase transition .
a proper understanding of distribution networks , then , must account for the severe differences between convergence and divergence , or explain their absence , given that they arise from the most basic theoretical assumptions .
jo and mm would like to thank eleni katifori , alex petroff , and carl modes for helpful discussions throughout
. supported in part by the national science foundation under grant nsf phy-1058899
simulations were conducted in matlab with general procedure as follows : a set of @xmath12 random points were generated in the region @xmath45$]x@xmath45 $ ] , with @xmath13 equal to 67 unless otherwise noted ( hence 89780 points were initially generated ) .
all points outside the circle @xmath46 were then discarded .
the remaining points had nearest neighbors defined by the delaunay triangulation .
each point was then assigned a height equal to @xmath47 , with @xmath17 being frozen white noise , whose strength could be adjusted in each simulation .
each point was then connected to the lowest of its nearest neighbors ( minimum @xmath48 value ) .
those points that did not have a neighbor with a lower @xmath48 value were presumed to be basin mouths ( or lakes , see below ) . after generating the connectivity matrix
, we could calculate all of the necessary network quantities .
magnitude and strahler number were calculated recursively : we found the tips of all streams by searching the connectivity matrix for points with no inflow .
these were assigned a value of one .
we then took their mouths and ran the same code , augmenting the magnitude number by one , and keeping the strahler number the same . at all merges ,
the magnitude numbers were added and the strahler numbers were augmented according to the rule : @xmath49 or @xmath50 if @xmath51 .
basins were identified by a recursive algorithm acting in the opposite direction : starting with the mouths , defined as points with no outflows .
all points flowing into each mouth were found form the connectivity matrix ; the find inflow algorithm was then run on these points .
lists of the point addresses within each basin could be converted into mainstream lengths , by calculating the euclidean distance between mouth and furthest stream tip , and basin lengths , by summing the number of points .
the basin area came directly from the magnitude number at the mouth , as a uniform drainage density was assumed .
we performed these simulations over hundreds of trials at values of @xmath16 from -1.5 to 1.5 , usually sampling every 0.05 . data on the hack s law relationship and the distribution of basin sizes came from simulations at @xmath28 .
for all other simulations we sampled at logarithmically - distributed basin sizes from @xmath52 to @xmath53 @xmath54 .
preliminary simulations on square and hexagonal grids were also conducted . however , the degree of noise necessary to observe a difference in behavior with changing curvature made simulating nearly flat topographies extremely difficult , as the number of , lakes " or basins that terminated not at the edge of the annular region was improperly large . to remove a lake
, one must simulate filling by raising the z - value at a point to epsilon greater than the smallest of its neighbors and then recalculating flows .
this procedure leads to the creation of cycles of 3 or 4 points .
for instance , consider a and b flowing into c , where c is a local minimum .
increasing the height of c causes it to flow into a. if a now is a local minimum , it will flow into b after filling , creating a cycle . for small enough values of m , however , our model appropriately returns a swamp " of many short basins terminating in lakes .
the delaunay triangulated random grid , by adding noise in the locations of points and thus the difference in height between neighbors , effectively added randomness to the system without creating undue lakes .
the algorithm to find magnitude and strahler numbers for each lattice point and link in the network is recursive by necessity .
additionally , the matrix of connectivities is @xmath55x@xmath55 .
we were thus limited in the size of the networks we could simulate to approximately @xmath28 . for the purposes of finite size scaling analysis ( see below ) , we simulated networks at @xmath53 , however these simulations took several hours each , due to paging .
we conducted simulations for @xmath52 to @xmath53 ( the different colored lines ) at logarithmically - spaced intervals ( a factor of 1.5 , rounded down ) in the range @xmath56 to @xmath57 stepping by 0.05 .
we then logarithmically regressed versus @xmath2 for each point with @xmath58 .
this gave us a finite - size scaling exponent @xmath59 .
the results of @xmath60 vs. @xmath16 for each value of @xmath2 are plotted in figure s1 .
we then logarithmically regressed the rescaled @xmath30 with @xmath2 to find the critical exponent , @xmath29 .
this procedure was repeated for @xmath61 , to find @xmath62 .
the original , planar , scheidegger model is not recovered at @xmath63 .
the model we simulate lacks a uniform slope in either the x or y direction , which is implicit in the original scheidegger model .
this uniform slope could be included by altering our slope field to read : @xmath64 .
geometrically , this modification would tilt the conical manifold we simulate towards either the positive or negative y - direction .
one may envision a tilted coffee filter that is flattening as m approaches zero .
in essence , it would add an additional field " which would split and shift the location of the critical point to @xmath65 .
for@xmath66 , the model would function as before . at the first critical point
crossed , half the manifold would be nearly flat and would exhibit a singularity in number of basins , whereas the other half would be tilted at the sum of the two slopes . as m approached zero , we would recover the original scheidegger model . at the second critical point , the opposite half of the manifold would be nearly flat and would exhibit the same singularity . the total effect would be akin to the addition of an external field to the ferromagnetic phase transition .
22ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.53.1510 [ * * , ( ) ] link:\doibase 10.1103/physreva.37.3110 [ * * , ( ) ] link:\doibase 10.1103/physreva.46.r4471 [ * * , ( ) ] @noop _ _ , ed .
( , , ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , ed .
( , ) @noop `` , '' ( ) @noop _ _ ( , ) @noop _ _ , ed .
( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * ( )
|
transport networks are found at the heart of myriad natural systems , yet are poorly understood , except for the case of river networks . the scheidegger model , in which rivers are convergent random walks , has been studied only in the case of flat topography , ignoring the variety of curved geometries found in nature .
embedding this model on a cone , we find a convergent and a divergent phase , corresponding to few , long basins and many , short basins , respectively , separated by a singularity , indicating a phase transition .
quantifying basin shape using hacks law @xmath0 gives distinct values for @xmath1 , providing a method of testing our hypotheses .
the generality of our model suggests implications for vascular morphology , in particular differing number and shapes of arterial and venous trees .
|
recently , it has been given much attention to the study of topological defects in the context of brane - world models , due to its property of allowing the localization of several different types of fields .
as extended defects in field theory the domain walls have been used in high - energy physics to represent brane scenarios with extra dimensions @xcite .
the bloch walls , which could be seen as chiral interfaces @xcite , are used in the context of extra dimensions to construct a @xmath0 model of two scalar fields coupled with gravity , the so called bloch brane @xcite .
the bloch brane model is generated dynamically and has internal structure .
the asymptotic bulk metric is a slice of a five - dimensional anti - de sitter ( ads ) spacetime , denoted by @xmath1 .
such scenario may be used to mimic a brane - world containing internal structure @xcite , which have implications on the density of matter - energy along the extra dimension @xcite .
the appearance of the internal structure could be also observed by a splitting effect on the curvature invariant . on the other hand , dutra et al .
have showed that the bloch brane scenario addressed in @xcite , holds more general soliton solutions @xcite .
the brane configurations obtained from these new solutions were named by the authors as _ degenerate bloch branes _ due to the existence of a degeneracy parameter that is not present in the lagrangian density .
such defects present more details as the appearing of two - kink solutions .
these additional features were interpreted as the formation of a double wall structure .
when the degeneracy parameter approaches to a critical value , the brane splits in two and its separation become larger .
this effect will contribute to the emergence of massive graviton resonances .
the phenomenon of two separate interfaces on the defect is known in condensed matter physics as complete wetting @xcite . moreover , the same behavior was considered as a critical phenomena of phase transition on thick branes in warped geometries @xcite . in the above scenarios
some authors have investigated the localization of several types of bulk fields .
namely , fermion fields @xcite , gauge fields @xcite and graviton zero mode @xcite .
however , the study of the graviton massive spectrum has not been properly studied .
it is worthwhile to mention that , the search for resonances in warped spacetimes @xcite has received attention because they give us important information about the interaction of kaluza - klein ( kk ) massive spectrum with the four - dimensional brane . specifically in the study of gravity localization
, the presence of a resonance at zero energy is related to the existence of a large - distance region on which the 4d laws of gravity are valid @xcite .
therefore , if the resonance width becomes very large it results in nonphysical effects . to the best of our knowledge
the first work about localization of graviton zero mode on the bloch brane was the work of bazeia and gomes @xcite .
their results were also confirmed on the degenerate bloch branes by dutra et al .
however , the important issues concerning the analysis of the massive spectrum and the search for resonances in these scenarios have not been addressed .
therefore , in the present work we propose to study the graviton massive spectrum and search for resonant states on the two field thick brane scenarios described above .
additionally , we intend to verify the relation between the brane thickness and the internal structure over possible detected resonances .
we organized this work as follows . in sec .
[ sec.bloch ] , we review the bloch brane scenario and search for new resonant structures in this model in sec .
[ sec.massive ] .
we addressed the degenerate bloch brane solutions in the beginning of the sec .
[ sec.other ] and we search for new graviton resonances in this more general setup .
finally , we present our results and conclusions in sec .
[ sec.conc ] .
the two - field thick brane scenario that we consider is composed by two fields @xmath2 and @xmath3 coupled to gravity which depend only on the extra dimension @xmath4 .
such model was previously studied in references @xcite .
their action is given as follows @xmath5,\ ] ] where @xmath6 is the scalar curvature and the spacetime is an ads @xmath7 with metric @xmath8 the minkowski spacetime metric is @xmath9 with signature @xmath10 and the indices @xmath11 , @xmath12 run over @xmath13 to @xmath14 .
the corresponding equations of motion are @xmath15 where prime stands for derivative with respect to @xmath4 . a method for solving the coupled differential equations system has been developed in the context of thick branes @xcite .
it consists of a appropriate redefinition of the potential @xmath16 as @xmath17-\frac{1}{3}w^{2}\ ] ] in terms of a superpotential @xmath18 this implies that the resulting first - order equations can be written as @xmath19 , @xmath20 and @xmath21 , from which we find the solutions that describe our brane model , namely @xmath22 @xmath23 and @xmath24.\label{warp}\ ] ] from eq .
, we can see that for the limit @xmath25 the one - field scenario is recovered . for certain values of the coupling parameter @xmath26 that controls the brane thickness , we have a splitting of the defect and the appearance of an internal structure .
this characteristic is also evident on the curvature invariant . for this geometry , for instance
, we obtain @xmath27.\ ] ] the ricci scalar is finite , which we can observe through the fig .
( [ fig : gravbloch0 ] ) . for @xmath28 , with @xmath29
there is a maximum at @xmath30 .
when @xmath26 arrives to the interval @xmath31 , the maximum splits into two separate maxima , which indicates the presence of the internal structure .
these effects will influence the behavior of the massive modes , which will be investigated in the following section . with @xmath32 .
we note the appearance two separate minima at z=0 , indicating the presence of internal structure.,title="fig : " ] + with @xmath32 .
we note the appearance two separate minima at z=0 , indicating the presence of internal structure.,title="fig : " ] +
the seminal work of bazeia and gomes @xcite considering the bloch brane model performed an analysis in the zero modes of the graviton but have not confirmed the existence of resonant modes . from now
, we will take this scenario again focusing our attention in the massive spectrum and seeking resonant modes for the graviton . in order to search for resonances in the massive spectrum
we must obtain a shredinger - like equation to the graviton on the fifth dimension .
initially , we perform a metric perturbation using @xmath33 , where @xmath34 represents the graviton with the axial gauge @xmath35 . when we set the metric fluctuation as transverse and traceless ( tt ) , namely @xmath36 , its equations of motion take the simplified form @xcite : @xmath37 where @xmath38 is the four - dimensional wave operator . using the transformation @xmath39 and choosing an ansatz containing a bulk wave function times a space plane wave , @xmath40
, we can rewrite the eq . as a schrdinger - like equation given by @xmath41 with the potential @xmath42 .
( [ eq : schro ] ) leads to no tachyonic states and owns a normalizable zero mode solution as was showed in ref .
@xcite .
when @xmath43 , the potential represents only a small perturbation and the solutions will acquire plane wave structure . in fig .
[ fig : gravblochpot ] we plot the potential @xmath44 varying @xmath26 . as in the curvature scalar ( fig .
[ fig : gravbloch0 ] ) we note the appearance of a splitting effect due to the thickness of the defect .
the minimum of the potential separates in two as we reduce @xmath26 , acquiring the shape of a double - well - type potential .
similar feature was also found in the potential of the shredinger equation for the tt sector of the metric perturbations in a thick brane scenario generated by one scalar field @xcite . for some specific energies , the solutions of eq .
could exhibit large amplitudes inside the brane in comparison with its values far from the defect .
these solutions are interpreted as resonant modes and reveal states with large probabilities to be found on the brane @xcite . from the eq .
( [ eq : schro ] ) we can consider @xmath45 as the probability for finding the mode on the brane , where @xmath46 is a normalization constant . in this way , we must know the variation of this quantity in terms of @xmath47 in order to find possible resonant modes .
for this purpose , we first consider two different methods to detect resonant modes .
one of them uses directly the probability of finding the mode on the brane from the square of the normalized wave function @xcite , namely @xmath48 a second method uses the relative probability @xcite : @xmath49 in both approaches , the massive modes are considered in a box with borders @xmath50 far from the turning points .
the difference between the two methods is that in the case of relative probability @xmath51 we have a narrow integration range around the brane @xmath52 conveniently chosen such that @xmath53 @xcite .
the main motivation for using this approach is that , since thick branes are objects spread around the origin of the extra dimension , the probability of finding the modes must be integrated in a small region around the location of the defect in order to detect resonances .
furthermore , according to the formal quantum theory of resonance , we can study the probability of finding the massive modes around the vicinity of the brane position within a relatively large region @xcite .
we must choose one of the probability methods based on their efficiency to detect in a better way the resonant modes on the kk spectrum ( if they exist ) .
such structures are characterized by large amplitudes inside the brane in comparison with the amplitude outside .
thus we first compare the two functions @xmath54 and @xmath51 to choose the more suitable .
due to the coordinate transformations and the complexity of the @xmath55 function , we can not find @xmath56 analytically .
therefore , we solve the eq . by numerical integration and compare the functions @xmath51 and @xmath54 .
the result is showed in fig .
[ fig : gravbloch1 ] .
as expected , the two methods indicate a resonance at @xmath57 . with @xmath58 and @xmath59 with @xmath60 .
we have used @xmath61.,title="fig : " ] + with @xmath58 and @xmath59 with @xmath60 .
we have used @xmath61.,title="fig : " ] + it is worth to mentioning that the two functions @xmath51 and @xmath54 acquire different values when @xmath62 @xcite . in this region , where the solutions present plane wave structures , we have @xmath63 @xcite and @xmath64 @xcite . we also observed in fig .
[ fig : gravbloch1 ] that the resonance peak is more pronounced and broader on the @xmath51 function .
this characteristic shows that using @xmath51 we can detect more easily possible narrow resonance structures .
then , we will adopt @xmath51 in our analysis henceforth . since we consider the kk modes @xmath65 in a box with borders @xmath50 , beyond which @xmath65 are turned into plane waves ,
the @xmath66 value should not interfere with the mass of the resonant modes to maintain the consistency of the method .
this is a fundamental characteristic because the physical information is contained in the value of the resonance peak , which should not depends on @xmath66 , since it is chosen sufficiently large . in fig .
[ fig : gravbloch2 ] , we plot the function @xmath51 with @xmath58 and @xmath59 with @xmath60 .
as expected , the position of the resonant mode in @xmath51 is not altered when we vary @xmath66 .
we must now analyze the behavior of the function @xmath51 as well as its dependence on the thickness of the membrane .
the real parameter @xmath26 controls the thickness of the defect and may give rise to an internal structure on the membrane .
this characteristic can be verified by analyzing the structure of the potential as well as the matter energy density in terms of @xmath26 , as shown in ref .
@xcite . in order to verify the effects of the brane thickness on the graviton massive modes ,
we analyze the behavior of the function @xmath51 in terms of @xmath26 .
the result is shown on the left side of the fig .
[ fig : gravbloch3 ] .
we also plot the correspondent resonant modes on right side of the same figure .
( left ) and the zero modes solutions ( right ) with @xmath67 and @xmath68.,title="fig : " ] + by varying the @xmath51 function in terms of @xmath26 we find resonances at @xmath57 . we can estimate the lifetime @xmath69 of a resonance by @xmath70 , where @xmath71 is the width at half maximum in mass of the peaks in @xmath51 @xcite . for small @xmath26 ,
the resonance becomes very broad characterizing states with lifetimes too short to result in any physical effect .
this characteristic can be confirmed observing the @xmath72 function correspondent to the resonant modes at @xmath57 , as can be seen in fig .
[ fig : gravbloch3 ] ( right side ) . in this case , the zero modes are scattered along the extra dimension leading to non - localized zero modes .
in this section we search for graviton resonances in more general thick brane scenarios generated by two scalar fields . as a matter of fact , after the work of bazeia and gomes in bloch branes @xcite , the existence of new field solutions coming from the bloch brane scenario was shown by dutra and collaborators @xcite . in these works ,
the authors directed their analysis to the stability and zero mode solutions of the graviton , but now we generalize the analysis to the massive spectrum in a quantum mechanics scenario . on these new solutions , which were called degenerate bloch walls ( dbw ) ,
the brane thickness is controlled by the variation of a parameter associated to the domain wall degeneracy .
another important characteristic of the dbw is the appearance of two - kink solutions indicating the formation of two branes . particularly , we choose the solutions presented in ref .
@xcite , where the action and the metric are the same in eq .
and eq . respectively
therefore , the schrdinger - like eq . is also maintained . with the dbw field solutions and the corresponding warp factors , we reevaluated the @xmath51 function defined in , in order to search for resonances on the massive spectrum over these new conditions .
first , we consider the potential @xmath73 which is written in terms of a superpotential as @xmath74 -\frac{4}{3}w^{2},\ ] ] which now takes the form @xmath75,\ ] ] from where we can recover the eq . by choosing @xmath76 , @xmath77 and @xmath78 . the motivation to consider the superpotential of this form is that now we can control the warp factor behavior with more parameters than the @xmath26 parameter as in the previous section .
this characteristic will bring us richer solutions that will give rise to new resonance structures . for more details on how to obtain the warp factors
, the reader should consult the refs .
@xcite .
( _ right _ ) for the dbw(1 ) solution .
we have used @xmath79.,title="fig : " ] + the first case that we will consider , named dbw(1 ) , is obtained for @xmath80 and @xmath81 and it results on the following warp factor @xcite @xmath82^{4a^2/9 } \nonumber \\ & & \times \exp \left\ { \frac{2a^2\left [ c_{0}^{2 } - 4a^2-c_{0}\left ( \sqrt{c_{0}^{2}-4a^2}\right ) \cosh ( 2a\mu y)\right ] } { 9\left [ \left ( \sqrt { c_{0}^{2}-4a^2}\right ) \cosh ( 2a\mu y)-c_{0}\right ] ^{2}}\right\ } , \end{aligned}\ ] ] where @xmath83 is chosen such that @xmath84 . the warp factor and the resulting potential @xmath44 of the schrdinger - like eq .
are plotted in fig .
[ fig : resd1 ] on the left side and in the center , respectively .
as we can observe , the @xmath85 constant plays a role similar to the parameter @xmath26 that controls the thickness of the standard bloch brane solution .
however , the integration constant @xmath85 , unlike the parameter @xmath26 , is not present in lagrangian density .
when @xmath85 approximates to its critical value @xmath86 the minimum of the potential splits in two and we observe the appearance of a flat region on the warp factor .
another interesting feature is the formation of a two - kink solution in @xmath87 when @xmath85 approximates to its critical value .
such feature was understood on the work @xcite as the formation of a double wall structure along the extra dimension . from the @xmath55 solution in eq .
, we compute the relative probability again and the results are showed on the right side of the fig .
[ fig : resd1 ] . sweeping the region where @xmath88 , the function @xmath51 show us resonance peaks at @xmath57 .
such resonances correspond to the graviton zero modes . varying @xmath85 and approaching the critical value , the resonances are not significantly changed .
new results are obtained when we consider another degenerate bloch wall , named dbw(2 ) .
such solution is obtained when @xmath89 and @xmath90 .
the corresponding warp factor is @xmath91 ^{\frac{16a^{2}}{9 } } \nonumber \\ & & \times \exp{\left\{-\frac{4a^{2}\left[1 + 8a^{2}c_{0}+\left ( \sqrt{1 - 16c_{0}a^2}\right ) \cosh ( 4\mu ay)\right]}{9 \left [ 1+\left ( \sqrt { 1 - 16c_{0}a^2}\right ) \cosh ( 4\mu a y)\right ] ^{2 } } \right\}}. \label{2.93}\end{aligned}\ ] ] in this case , we also find a double kink profile to @xmath87 when the integration constant approaches to the critical value . and the solution @xmath65 for @xmath92 ( _ left _ ) and @xmath93 ( _ right _ ) .
we have used @xmath94 and @xmath95 in the two graphics.,title="fig : " ] + the potential and some solutions to @xmath65 are plotted in fig .
[ fig : pot ] for @xmath94 and @xmath96 . when @xmath97 the solutions oscillate rapidly along the extra dimension , what is observed on the right side of the fig .
[ fig : pot ] .
however , as noted in the plot on the left , for @xmath98 the amplitude of the wave solutions are suppressed in the region between the maxima of the potential and the modes could potentially exhibit a resonance structure @xcite .
for @xmath96 we detect two resonances , as showed on the fig .
[ fig : resd2 ] . beyond the peak corresponding to the zero mode ( @xmath57 ) we find another resonance at @xmath99
such result can be understood as the existence of graviton massive modes highly coupled to the brane .
when the constant @xmath85 approximates its critical value , the peak at @xmath57 is kept and the number of resonances with @xmath100 increases .
the presence of resonances at @xmath57 for the cases considered shows the consistency of the relative probability method applied .
indeed , the bloch branes support graviton zero modes localized and therefore there should be resonances corresponding to the zero modes .
the number of resonances increases when @xmath85 tends to @xmath101 and their lifetimes are increased as well . with @xmath102 , for example , we have two peaks of probability beyond the peak at @xmath57 .
such massive modes have high amplitudes at @xmath103 and should interact with the brane . for dbw(2 ) solution .
we have used @xmath104 and @xmath105.,title="fig : " ] +
we have considered a thick brane generated by two scalar fields in which we search for graviton resonances .
we have reviewed the brane setup known as bloch brane in the sec .
2 . we know from the literature @xcite that the real parameter @xmath26 controls the bloch brane thickness , which in the interval @xmath106 produces a splitting on the defect .
we have analyzed such property in terms of the curvature invariant .
the results show that when we increase the brane thickness we observe a splitting effect on the maximum of the curvature invariant and the raising of a flat region at y=0 . from a metric perturbation
, we find a schrdinger - like equation of motion . since the previous works @xcite have studied the zero mode cases , we started the analysis to identify resonances comparing two probability methods from @xmath100 schredinger - like equations .
the outcomes indicated that the direct probability method @xmath54 has shown a resonance correctly .
however , in the @xmath51 function of relative probability the resonance peak was more pronounced indicating this process as more effective for detecting possible narrow resonances .
we tested the efficacy of the relative probability method to identify resonances by changing the integration range @xmath107 . in this case , the resonance position was not changed . the relative probability function @xmath51 in eq .
provide us a resonance at @xmath57 .
this result is in agreement with previous results which demonstrated the existence of a zero mode located @xcite . on the other hand ,
looking for the effects of the internal structure on the resonance we have noted that when we reduce @xmath26 , which makes the brane thicker , the width of the resonance is increased .
thus reducing @xmath26 will decreases the resonance lifetime .
the same effect was observed at the zero mode which corresponds to the resonances , as shown in fig .
[ fig : gravbloch3 ] .
furthermore , reducing @xmath26 and making the brane thicker causes a splitting of the defect that delocalizes the zero modes .
similar feature was also observed in the ref .
@xcite as a phase transition in thick branes .
the existence of a resonance at zero energy is related to the existence of scales on which 4d gravity is recovered . on this way , when the resonances acquire large lifetimes we can restore the rs scenario @xcite . otherwise , when the width of the resonance increases too much
, we will no longer have a region of 4d gravity @xcite .
the results obtained show that when we increase the thickness of the bloch brane ( reducing @xmath26 ) , we reduce the lifetime of the graviton resonance . in this case and we can not guarantee the existence of a long - distance scale where the newtonian potential is valid . in the same way ,
if we made the brane thicker , the localization of the zero modes could be jeopardized .
we conclude that in the limit when @xmath108 the bloch brane supports a region where gravity is effectively 4d
. however , when @xmath109 the brane becomes very thick and this result is no more valid .
we have extended the investigation to the degenerate bloch brane solutions . in the first degenerate case
examined we have a critical constant @xmath85 that plays a role similar to the parameter @xmath26 in sec .
when @xmath85 approximates to its critical value @xmath110 the brane becomes thick and the minimum of the potential splits in two .
the analysis shows resonances at @xmath57 corresponding to the graviton zero modes .
they should be understood the same manner as those found in the basic bloch brane setup .
we find interesting and new issues in the second class of degenerate solutions .
we observe massive graviton resonances in addition to those at @xmath57 .
the appearance of resonant modes happens when the degeneracy parameter approaches its critical value .
the number of resonances and their respective lifetimes increases when @xmath85 tends to @xmath101 .
another striking point is that when the brane is thicker ( @xmath111 ) the resonance lifetimes increases .
such behaviour is opposite to that find in the first bloch brane scenario examined .
the presence of those structures in the massive spectrum is related to the existence of massive ( quasilocalized ) modes highly coupled to the brane .
we find massive resonances to the graviton field in another context of thick branes in ref .
@xcite . _
note added_. -during the process of revision of this work we find another study @xcite of resonances in two - field thick brane scenarios .
the authors also noticed the presence of graviton resonances .
the authors thank the fundao cearense de apoio ao desenvolvimento cientfico e tecnolgico ( funcap ) , the coordenao de aperfeioamento de pessoal de n ' ivel superior ( capes ) , and the conselho nacional de desenvolvimento cient ' ifico e tecnol ' ogico ( cnpq ) for financial support .
|
this work presents new results about the graviton massive spectrum in two - field thick branes . analyzing the massive spectra with a relative probability method we have firstly showed the presence of resonance structures and
obtained a connection between the thickness of the defect and the lifetimes of such resonances .
we obtain another interesting results considering the degenerate bloch brane solutions . in these thick brane models
, we have the emergence of a splitting effect controlled by a degeneracy parameter .
when the degeneracy constant tends to a critical value , we have found massive resonances to the gravitational field indicating the existence of modes highly coupled to the brane .
we also discussed the influence of the brane splitting effect over the resonance lifetimes .
|
circumstellar disks surrounding t tauri and herbig ae / be stars are thought to the birth sites of planetary systems and they are called protoplanetary disks . the physical properties of the protoplanetary disks have been inferred from analyses of their spectral energy distributions ( seds ) .
analyses of observations at millimeter / sub - millimeter wavelengths suggest that the disks have masses of 0.001 - 0.1 @xmath8 and their life time is @xmath9yr ( e.g. , strom et al .
1989 ; beckwith et al . 1990 ; wyatt et al .
2003 ) .
recent observations with high - resolution at various wavelengths gave us much information on disk spatial structure ( e.g. , grady et al .
1999 ; fukagawa et al .
2004 ; mundy et al .
1996 ; kitamura et al . 2002 ; eisner et al . 2003 ) .
for example , kitamura et al . (
2002 ) derived disk outer radii and surface density distributions from results of their image survey of 13 t tauri disks and sed data .
their results show that the radial power - law index of the surface density @xmath10 ranges from 0 to -1 , which is less steep than -1.5 in the minimum mass solar nebula ( mmsn ) ( hayashi 1981 ) . because the obtained surface densities at 100 au is almost consistent with the mmsn , the small radial power - law index of @xmath1 means too small surface density at @xmath11au for planet formation in those disks . in the most analyses of disk observations including kitamura et al .
( 2002 ) , dust growth have not been taken into account .
however , the grain growth is indicated by the variety of silicate feature at 10@xmath12 m observed in t tauri disks ( przygodda et al .
2003 ; honda et al .
2003 ; meeus et al .
. dust growth usually decreases the opacity and the optical thickness of disks .
if grains grow considerably in the observed disks , the estimated disk mass mentioned above would be underestimated ( dalessio et al .
2001 ; natta et al .
2004 ) . furthermore ,
dust growth is generally fast in the inner disk region compared with the outer region ( weidenschilling 1980 ; nakagawa et al .
rapid decrease in the optical thickness in the inner region would make the radial distribution of the optical thickness less steep than that of the surface density .
the rapid dust growth in inner regions may also be responsible to inner holes identified by calvet et al .
( 2002 ) in tw hydra and by bouwman et al .
( 2003 ) in hd1000546 .
furthermore , the spitzer space telescope started to provide lots of new data on disk inner holes ( forrest at al .
2004 ; meyer et al . 2004
; uchida et al . 2004 ) .
hence , dust growth should be taken into account in the analyses of disk observations .
recent models of protoplanetary disks succeeded in accounting for most properties in the observed seds ( e.g. , chiang & goldreich 1997 ; chiang et al . 2001 ; dalessio et al .
2001 ; dullemond et al .
2001 ; dullemond & dominik 2004a ) .
although they also examined the effect of dust growth , they a priori adopted a single power - law size distribution for a whole disk and assumed that the grain number at each size is proportional to the gas density .
however , the size distribution would be considerably different at each position because of a large variety of the growth time in a whole disk , as mentioned above . furthermore
, dust settling would change the vertical distribution of dust grain in disks .
dullemond & dominik ( 2004b ) showed the importance of dust settling at the self - shadowing process in passive disks . to derive information correctly from observations , the more realistic model of dust growth and settling
should be used . dust growth and settling in protoplanetary disks
have been investigated by many authors for both laminar disks ( e.g. , safronov 1969 ; weidenschilling 1980 , 1997 ; nakagawa et al . 1981 , 1986 ) and turbulent disks ( e.g. , weidenschilling 1984 ; mizuno et al . 1988 ; schmitt et al .
1997 ) . in laminar disks or those with sufficiently weak turbulence ,
dust grains grow to mm - size when they settle to the mid - plane of the gaseous disk . as a result of dust settling
, a dust - rich layer is formed near the mid - plane . in the dust layer , a large number of planetesimals are thought to form as the first step of planetary formation ( e.g. , goldreich & ward 1973 ; weidenschilling 1997 ; youdin & shu 2002 ) . because most previous studies focused on the dust layer and the planetesimal formation , they did not examine in detail a relatively small amount of micron - sized grains that remain floating after the dust layer formation . however , as suggested by nakagawa et al ( 1981 ) and some authors , the floating small grains would play an important role in the radiative transfer in protoplanetary disks even though their total mass is small . the change in seds due to settling of such floating grains
was examined by miyake & nakagawa ( 1995 ) and by dullemond & dominik ( 2004b ) .
however , they did not consider settling coupled with growth .
the variation in the disk optical thickness due to dust growth was also evaluated in some studies on dust growth and settling ( weidenschilling 1984 , 1997 ; weidenschilling & cuzzi 1993 ; schmitt et al .
however , these evaluations were done only in short - term calculations ( @xmath13yr ) or only at limited radial positions ( i.e. , 1 and 30 au ) . in order to take into account dust growth in analyses of disk observations ,
it is necessary to clarify the general property in the long - term evolution of floating grains and disk optical depth in various disks . in this series of papers , we investigate dust growth and settling in protoplanetary disks , focusing on floating small grains rather than large grains in dust layers .
we perform numerical simulations of dust growth and settling for various disk models to obtain the general property of the spatial and size distribution of dust grains . using the obtained spatial and size distribution of grains ,
we further examine time - evolution of the disk optical thickness and seds due to dust growth and settling .
our results show that , because of relatively rapid dust growth in the inner disk region , the disk optical thickness decreases more rapidly in the inner region than in the outer .
this makes the radial distribution of the optical thickness less steep , as expected above . the dust emission at millimeter / sub - millimeter wavelengths declines on the time scale @xmath14 yr even without any disk depletion .
therefore , the observed decay of dust emission does not necessarily indicate disk depletion . in the present paper
, we consider laminar disks as the first step of this series of papers .
as shown by dullemond & dominik ( 2004b ) , vertical stirring by turbulence is effective on dust settling even in weakly turbulent disks with the parameter @xmath15 for the disk viscosity ( shakura & sunyaev 1973 ) .
furthermore , since turbulence would play a crucial role in the issue of planetesimal formation in a dust layer by self - gravitational instability , it can not be neglected on that issue . however , on the disk optical thickness , essential ingredients would be included in laminar disk cases .
thus the calculations with laminar disks would give deep insights into time - evolution of the optical thickness and disk seds , although the time scale can be affected by turbulence .
hence , the case in laminar disks should be firstly examined as a simple case .
the case in turbulent disks will be investigated in the subsequent paper . in the next section ,
we describe basic equations and our numerical model of growth and settling .
we show numerical results for various disk models in section 3 .
we also derive an analytical solution of the grain size distribution , which agrees well with numerical results . in section 4 , we examine time - evolution of the disk optical thickness and seds due to dust growth and settling . in the calculation of seds , we used the two - layer model proposed by chiang & goldreich ( 1997 ) and refined by chiang et al .
( 2001 ) . in the last section ,
we summarize our results and discuss the effect of turbulence on dust growth and settling .
we consider a passive protoplanetary disk with no turbulence around a central stars with one solar mass ( @xmath16 ) .
it is assumed that the disk has an inner edge at 0.1 au and an outer edge at 150 au . to describe the disk
, we use a cylindrical coordinate system @xmath17 of which the origin is located at the central star . the @xmath18-axis
coincide with the rotation axis of the disk . in numerical simulations of dust growth and settling , we assume that the gaseous disk is isothermal in the vertical direction ( i.e. , in the @xmath18-direction ) for simplicity . as seen in the two - layer model of passive disks , however , the surface layer directly irradiated by the central star is much hotter than the disk interior .
nevertheless , our modeling of a vertical isothermal disk would be valid because dust growth and settling in laminar disks are little affected by the disk temperature , as seen in the next section . in the evaluation of seds in section 4 , of course , the vertical temperature distribution will be taken into account with the two - layer model . under this assumption , the vertical density distribution @xmath19 of the gas component
is given in the hydrostatic equilibrium by @xmath20 where @xmath21 is the gas surface density and @xmath22 is the pressure scale height of the disk given by @xmath23 .
furthermore , @xmath24 @xmath25 is the keplerian angular velocity and @xmath26 is the isothermal sound speed , where @xmath27 , @xmath28 , and @xmath29 are the gravitational constant , boltzmann s constant , and the atomic mass unit , respectively ; and @xmath30 and @xmath12 are the disk temperature and the mean molecular weight , respectively .
we set @xmath31 . dust growth and settling change the disk temperature .
thus , exactly speaking , we need to solve coupled equations of dust growth , settling and the disk temperature .
however , since a change in the disk temperature has a little effect on dust growth and settling , we use a tentative disk temperature for simulations of dust growth and settling .
the tentative disk temperature is calculated in the same manner as the calculation of the temperature of the disk interior of the two - layer model ( i.e. , using equations [ [ eq : ti ] ] and [ [ eq : angle ] ] in the appendix ) but , in this calculation , we assumed that the disk is optically thick ( @xmath32 ) and that the height of the surface layer , @xmath33 , is equal to @xmath34 .
note that we will correctly evaluate the temperatures of the disk interior and the surface layer according to the two - layer model in the calculation of the disk sed ( section 4 ) .
the deviation of the tentative temperature from that of the disk interior reevaluated in section 4 is within 30% .
as for the surface density distribution of gaseous disks , we adopt a power - law distribution of @xmath35 the gas surface density at 1au , @xmath36 , and the power - law index @xmath37 are parameters . in the standard case of the present paper ,
we adopt the values of the mmsn , that is , @xmath38g@xmath39 and @xmath40 .
the dust - to - gas ratio @xmath41 and the solid material density @xmath42 are set to be 0.014 and 1.4g@xmath39 , respectively in the standard case .
these values are consistent with the solar abundance when h@xmath43o ice is included in grains ( pollack et al.1994 ) .
we also examine cases with a different value of @xmath41 or @xmath42 . at @xmath44 , h@xmath43o evaporates and we should decrease the dust - gas ratio and increase the solid material density .
however , we neglected these changes in the simulations of dust growth and settling for simplicity . to describe dust growth and settling
, we adopted the basic equations used by nakagawa et al .
let @xmath45 denote the number density of grains with masses between @xmath46 and @xmath47 at a height @xmath18 .
thus @xmath48 describes the distribution of grains with respect to @xmath46 ( or the size ) and @xmath18 .
the statistical coagulation equation is used to describe evolution of the mass distribution due to collisional growth . to describe dust settling as well as growth
, a vertical advection term is added to the coagulation equation .
that is , the following equation is used : @xmath49 \nonumber \\ & & \quad = \frac{1}{2 } \int_0^m k_{m',m - m ' } n(m',z ) n(m - m',z ) dm ' \nonumber \\ & & \quad \quad - n(m , z ) \int_0^{\infty } k_{m , m ' } n(m',z ) dm ' , \label{eq : coagsedi1}\end{aligned}\ ] ] where the kernel , @xmath50 , is related to the coalescence rate between grains with masses @xmath46 and @xmath51 , and @xmath52 is the settling velocity of dust grains . in the above ,
the first term on the right - hand side indicates the formation of grains of the mass @xmath46 by collisions between the smaller grains and the second term represents the consumption of grains with the mass @xmath46 due to collisions with other grains .
the second term on the left - hand side is the advection one , which describes the vertical mass transport . in equation ( [ eq : coagsedi1 ] ) , we neglected the advection term due to the radial mass transport because it is effective only for relatively large grains that settle down to the dust layer .
we also neglected the diffusion term due to turbulence since laminar disks are considered . the kernel @xmath53 is given by @xmath54 , where @xmath55 is the collisional cross section between grains with masses @xmath46 and @xmath56 , @xmath57 is the sticking probability , and @xmath58 is the relative velocity between grains with @xmath46 and @xmath56 . for dust grains , @xmath55 is equal to the geometrical cross section .
we assume spherical grains in the present study .
then we obtain the cross section @xmath55 as @xmath59 ^ 2 $ ] , where @xmath60 is the radius of grains with @xmath46 .
the relative velocity @xmath58 will be explained in the next subsection .
the sticking probability @xmath57 is difficult to be determined because it depends on its size , shape , material properties ( e.g. , weidenschilling & cuzzi 1993 ) . in the present study
, we regard @xmath57 as a parameter and assume that it is independent of the grain size . in a laminar disk , grains settle toward its mid - plane .
the settling velocity of each grain is given by the terminal velocity , which is determined by the balance between the gas drag force and the vertical component of the central star s gravity .
as well as such a settling velocity , nakagawa et al .
( 1986 ) also calculate the horizontal components of the grain velocity , assuming an system of equal - sized grains . here
, we describe motion of grains with an arbitrary size distribution we first explain the gas drag forces on grains . for a spherical grain with a mass @xmath46 , the gas drag force is given by @xmath61 and the coefficient @xmath62 is given by @xmath63 \displaystyle \sqrt{\frac{8}{\pi } } \frac{3 l c{_{\text{s}}}}{2 \rho{_{\text{s } } } a(m)^2 } & \quad ( a > \frac{3}{2 } l ) , \end{array } \right .
\label{eq : gas_drag4}\ ] ] where @xmath64 and @xmath65 are the velocities of gas and the grain , respectively and @xmath66 is the mean free path . in equation ( [ eq : gas_drag4 ] )
, the first case corresponds to epstein s law and the second is stokes law .
the mean free path is given by @xmath67 , where @xmath68 is the molecular collision cross section given by @xmath69@xmath70 ( chapman & cowling 1970 ) . at 1au and 10au on the mid - plane in the standard disk , the mean free paths are evaluated to be about 1 cm and 500 cm , respectively .
next , we describe motion of gas and grains .
the velocities of gas and grains , @xmath64 and @xmath65 , are obtained in a similar way to nakagawa et al .
( 1986 ) , though they considered equal - sized grains and focused on the dust layer around the mid - plane . when grains have a size distribution , the equations of motion for gas and grains are given by @xmath71 the gas pressure @xmath72 is given by @xmath73 .
at the surface layer directly irradiated by the central star , grains suffers a strong radiative pressure ( takeuchi & lin 2003 ) .
however , we neglected the radiation pressure in equation ( [ eq : motion_dust ] ) for simplicity .
we introduce a differentially rotating coordinate system with the angular velocity @xmath74@xmath75 $ ] to eliminate the gravity of the central star in the equations of motion . instead of @xmath64 and @xmath65 , we use the velocities , @xmath76 and @xmath77 , in the rotational frame .
since @xmath78 and @xmath79 are much smaller than the keplerian velocity , we can neglect second - order terms of them in the equations of motion .
then , setting @xmath80 , we can rewrite equations ( [ eq : motion_gas ] ) and ( [ eq : motion_dust ] ) in the rotating cylindrical coordinates as ( cf .
nakagawa et al . 1986 ) @xmath81 and @xmath82 respectively . from equations
( [ eq : v_r1 ] ) and ( [ eq : v_o1 ] ) , the @xmath83- and @xmath84-components of the grain velocity are obtained as @xmath85 where @xmath86 .
substituting equations ( [ eq : v_r2 ] ) and ( [ eq : v_o2 ] ) into ( [ eq : v_r1 ] ) and ( [ eq : v_o1 ] ) , we also have for the @xmath83- and @xmath84-components of the gas velocity @xmath87 where @xmath88 note that @xmath89 is dependent on @xmath18 as well as @xmath83 .
it should be also noted that our definition of @xmath89 is different from that in takeuchi & lin ( 2002 ) by the factor 2 but the same as nakagawa et al .
the vertical velocity of the gas , @xmath90 , is always negligibly small compared with @xmath91 as shown by nakagawa et al .
hence , from equation ( [ eq : v_z1 ] ) we have @xmath92 the obtained velocities agree with those in nakagawa et al .
( 1986 ) in the case of the dust layer ( i.e. , @xmath93 ) with equal - sized grains .
we use equations ( [ eq : v_r2])-([eq : v_z2 ] ) at the calculation of the grain velocity in the simulations of dust growth and settling .
we finally explain the relative velocity between grains appearing in the kernel @xmath53 .
to evaluate the relative velocity between @xmath12m - sized grains , we have to take account of brownian motion . the mean relative velocity due to brownian
motion is given by ( suttner & yorke 2001 ) @xmath94 then the total relative velocity between grains is defined by @xmath95 we calculated evolution of the grain distribution due to growth and settling , by solving equation ( [ eq : coagsedi1 ] ) with the grain velocities in the subsection 2.3 . because we neglected the radial drift of grains , equation ( [ eq : coagsedi1 ] ) can be integrated at each radial position , independently . in our simulation ,
the disk extending from 0.1 au to 150 au is logarithmically divided into 21 ring regions .
the median radius of the @xmath96-th ring region is given by @xmath97 au and its width is defined by @xmath98 .
equation ( [ eq : coagsedi1 ] ) is integrated with respect to time with the first order accuracy .
the change in the number density @xmath99 is given by the sum of two changes .
one is the change due to coagulation , which comes from two terms on the right - hand side of equation ( [ eq : coagsedi1 ] ) , and the other is the change due to settling , which comes from the second term on the left - hand side .
these two changes are calculated separately at each time step .
the former change is calculated with the fixed bin scheme ( nakagawa et al .
1981 ; ohtsuki et al . 1990 ) .
this scheme is valid for calculations of dust growth because runaway growth does not occur ( tanaka & nakazawa 1994 ) . the latter change is due to the advection term . to calculate it accurately and stably ,
we used the spatially third - order muscl - type ( monotonic upwind scheme for conservation law - type ) scheme developed by van leer ( 1977 ) .
as an initial condition for numerical calculations , all grains have the radius of @xmath100 m initially .
we assume the dust - to - gas ratio is uniform initially .
the density and the temperature of the gaseous disk are not varied with time during numerical simulations for simplicity . in the numerical integration of equation ( [ eq : coagsedi1 ] ) , the @xmath18-axis from @xmath101 to @xmath102
is divided into 251 equally spaced grids at each radial position . as for the coordinate of the grain mass , we divide it into 500 discrete fixed mass bins .
the minimum mass bin is that of grains with the radius of @xmath100 m and its representative mass , @xmath103 , is @xmath104 .
the representative masses , @xmath105 , of the smallest six bins are set to be @xmath106 ( @xmath107 . the larger mass bins are spaced logarithmically with a multiplication factor 1.15 ( i.e. , @xmath108 ) .
time steps @xmath109 are determined as follows . as the courant condition
, we made @xmath109 smaller than or equal to @xmath110 , where @xmath111 is the space between vertical grids and @xmath112 is the settling velocity of grains in the @xmath113-th mass bin of the @xmath114-th vertical grid .
we further restricted @xmath109 to solve the grain growth correctly .
let @xmath115 denote the change in the number density due to coagulation at the mass bin of @xmath105 in the @xmath96-th vertical grid at a time step .
the change @xmath115 is proportional to @xmath109 in our scheme .
we regulated the time interval @xmath109 so that the ratio @xmath116 is kept smaller than 0.025 at all mass bins if @xmath115 is negative .
because of this condition , however , @xmath109 tends to be very small by mass bins that contain negligibly small amount of grains . to avoid it ,
if the mass contained in a mass bin is smaller than @xmath117 times the total mass of the simulation , we removed all grains from the mass bin at each time step to prevent it from regulating @xmath109 .
such removal of grains hardly breaks the mass conservation . in our simulation with double precision ,
the relative error of the mass conservation is the order of @xmath118 .
furthermore , we checked the coagulation part of our algorithm , by comparing numerical results with analytic solutions to the coagulation equation for the kernel @xmath119 , as done by ohtsuki et al .
initially , all particles have an unit mass and they have an unit number density in this test . figure [ fig:1 ] shows the result of the comparison between numerical results and analytic solutions .
numerical results agrees with analytic solutions much better than those of ohtsuki et al .
( 1990 ) because we adopted a small mass multiplication factor of 1.15 .
before showing the numerical results , we mention the growth time and the settling time , which have already been examined in the previous studies ( e.g. , safronov 1969 ; weidenschilling 1980 ; nakagawa et al .
1981 , 1986 ) .
according to nakagawa et al .
( 1981 ) , the growth time of grains , @xmath120 , is given by @xmath121 at the last equation , the grain number density @xmath122 is estimated to be @xmath123 ; the collisional cross section @xmath124 is given by @xmath125 under the assumption of similar - size collisions ; and the settling velocity ( eq . [ [ eq : v_z2 ] ] ) at @xmath126 is substituted to the relative velocity @xmath127 .
the growth time is independent of the grain size , the disk surface density , and the temperature .
the settling time , @xmath128 , is given by @xmath129 where we used equation ( [ eq : v_z2 ] ) at the last equation , assuming @xmath126 .
the settling time is also independent of the disk temperature and decreases with an increase in grain radius .
the initial grain size is so small ( i.e. , sub - micron size ) that the growth time is much shorter than the settling time in an early stage .
hence such grains grow considerably before settling down by a vertical distance comparable to @xmath18 .
( of course , since the grain growth is driven primarily by the settling velocity , the settling is also essential to the growth of such small grains . ) as grains grow , the settling time eventually becomes comparable to the growth time .
then the grains do not grow much more before settling down to the dust layer ( nakagawa et al .
we can estimate such a critical grain radius , @xmath130 , by equating the settling time with the growth time . from equations ( [ eq :
t_g ] ) and ( [ eq : t_settle ] ) , we have @xmath131 , where @xmath132 is the dust surface density
. the radius @xmath130 corresponds to safronov ( 1969 ) s maximum radius to which grains can grow before settling to the dust layer .
actually , @xmath130 deviates only by a factor @xmath133 from the maximum radius obtained by safronov ( 1969 ) with more exact treatment .
now we show our numerical results on grain growth and settling . instead of the mass distribution @xmath48
, we consider the distribution of the grain radius , @xmath134 , which is defined by @xmath135 .
furthermore , by integrating @xmath134 with respect to the height @xmath18 , we define the surface number density of grains with each size , @xmath136 .
that is , @xmath137 the surface number density @xmath136 represents the vertically averaged size distribution .
the dimension of @xmath136 is length@xmath138 because @xmath139 has the unit of number@xmath140length@xmath141 .
figure [ fig:2 ] shows evolution of the surface number density @xmath136 at 8au in the standard disk ( 2.1 ) , with @xmath142 .
panel ( a ) of the figure shows the size distributions in the early stage .
grains grow from 0.1@xmath12m - size to micron size in the first thousand years . at 3000 yr
, the high - mass end of the distribution reaches several hundreds @xmath12 m . from equation ( [ eq : t_g ] ) , the growth time is estimated to be 400yr at 8au .
this value is consistent with the growth of the high - mass end from 1000yr to 3000yr in figure [ fig:2]a .
we found that the coalescence between grains with comparable masses is dominant during this time interval in the simulation .
thus the assumption of equal - sized collisions in equation ( [ eq : t_g ] ) would be valid . at 8000yr
, the size distribution has a bump at the high mass - end .
grains in the bump exist in the dust layer , which formed near the mid - plane at that time .
the formation time of the dust layer is roughly given by @xmath143 , regardless of the disk model and the radial position in disks .
this result is consistent with results of the previous studies ( e.g. , safronov 1969 ; weidenschilling 1980 ; nakagawa et al .
1986 ) .
evolution after the formation of the dust layer is shown in figure [ fig:2]b . in this stage ,
the size distribution consists of two parts : large grains in the dust layer and small grains floating above the layer .
the floating grains gradually settle down onto the dust layer without considerable growth .
relatively large floating grains settle onto the layer earlier and the upper cut - off of the size distribution of floating grains shifts to the smaller size . at @xmath144yr ,
only micron - sized grains remain floating . in this way
, the time - evolution can be divided into two stages : the growth stage ( fig . [ fig:2]a ) and the settling stage ( fig .
[ fig:2]b ) .
the dust layer is formed at the transition between the two stages in the settling stage , the size distribution of floating grains is well - described by a power - law distribution of @xmath145 .
dashed lines in figure [ fig:2 ] are the size distribution given by equation ( [ eq : size_dis ] ) , which will be derived in the subsection 3.3 . at @xmath146yr ,
the size distribution has a bump at @xmath147 m .
grains in this bump exist at @xmath148 .
these grains are removed by the coalescence with larger floating grains settling faster in the early part of the settling stage .
the size distribution of floating grains is also characterized by its upper cut - off . the grain radius at the upper cut - off , @xmath149 ,
can be analytically estimated .
since @xmath149 decreases due to grain settling , the radius @xmath150 is estimated from @xmath151 . introducing a numerical factor in this relation for fitting with the numerical results , we obtain @xmath152 in figure [ fig:2]b , @xmath149 of equation ( [ eq : a_max ] )
is plotted with vertical dotted and dashed lines at each time .
the upper cut - off radii in the numerical results are well - described by equation ( [ eq : a_max ] ) .
at the low - mass end , the surface number density decreases with time even in the settling stage .
this is due to grain growth by brownian motion .
as grains approach the mid - plane , their settling velocities decrease while brownian motion do not vary .
thus , since the collision rate between grains is proportional to their relative velocity , brownian motion is effective on grain growth in the settling stage especially for small grains . since smaller grains remain floating at a higher altitude , growth due to brownian motion affects the opacity at the disk surface .
furthermore , growth due to brownian motion is remarkable in an inner part of the disk rather than an outer part because of high temperature at the inner part .
brownian motion and settling deplete floating grains at the low - mass and the high - mass ends , respectively .
these reduce the disk optical thickness in the settling stage . in the dust layer , grains grow to more than 100-meter - size at @xmath153 yr ( see fig .
[ fig:2]b ) .
however , grain growth in the layer would not be calculated accurately because of the following two problems .
the first problem is insufficient vertical resolution .
the vertical resolution @xmath154 , which is @xmath155 in our calculation , gives the minimum thickness of the dust layer .
because of this limit , the dust - gas ratio has an upper limit of the order of unity in our simulations .
nakagawa et al . ( 1986 ) investigated dust growth in the dust layer getting thinner infinitely due to dust settling and showed that dust growth terminates when the gas - dust ratio becomes much larger than unity .
such termination of dust growth does not occur in our simulations since the coarse resolution keep the dust - gas ratio @xmath156 . however , the infinitely shrinking dust layer assumed by nakagawa et al .
( 1986 ) would be unrealistic .
many authors reported that turbulence due to the shear instability prevents the further dust settling when the gas - dust ratio exceeds unity ( e.g. , weidenschilling 1980 ; cuzzi et al .
hence termination of dust growth proposed by nakagawa et al .
( 1986 ) would not occur in the realistic disk .
the second and more serious problem is neglect of the radial drift of grains .
if the radial drift is taken into account , particles with meter - sized would rapidly fall to the central star and be lost ( e.g. , adachi et al . 1976 ) . in spite of above problems in our calculation ,
the inaccuracy in the dust layer would not significantly influence our results on the disk optical thickness because the dust layer has only a minor contribution to the disk optical thickness in most cases . in figure [ fig:3 ] ,
vertical density distributions of dust grains are shown .
the vertical axis in the figure is @xmath18 and the horizontal axis is the density of grains , @xmath157 , given by @xmath158 .
panel ( a ) shows evolution in the growth stage and panel ( b ) corresponds to the settling stage . at 8000 yr
, the density of grains is significantly enhanced near the mid - plane and the dust layer forms there .
grains are depleted with time especially at a high altitude of @xmath159 .
the dust growth and settling are expected to be dependent on the following physical parameters : the disk surface densities of gas and dust grains , the temperature , the material density of grains , the sticking probability , and the radial position in the disk ( i.e. , @xmath83 ) .
here we examine the dependence of the size distributions of floating grains on these physical parameters . in figure [ fig:4 ] , we show size distributions with different values of the physical parameters .
first we examine the dependence on the surface density of dust ( @xmath132 ) , the disk temperature , and the radial position . to do this ,
another simulation similar to the above was done at 32au in the standard disk .
furthermore , two more runs were done at 8 au in disks with the dust surface density ( or equivalently the dust - to - gas ratio @xmath41 ) ten times as large as that of the standard disk and with the temperature three times higher . in figure
[ fig:4]a , these three cases were compared with the standard case in figure [ fig:2 ] at @xmath160yr .
these all cases have the same size distribution for floating grains in the settling stage .
the upper cut - off radii of the size distributions also have the same value .
this is because @xmath41 and @xmath30 are not included in equation ( [ eq : a_max ] ) and @xmath161 is independent of @xmath83 in the standard disk .
we also mention distributions in the growth stage . in the growth stage ,
the case at 32au and the case with a high dust - to - gas ratio have different size distributions from others because of their different growth time . on the other hand ,
the difference in the disk temperature hardly affects the grain size distribution in the growth stage , too . in figure [ fig:4]b
, we showed the dependence on the gas surface density and the material density of grains .
we did runs with a low material density of @xmath162g@xmath39 and with the surface densities of both gas and grains ten times as large as those of the standard disk .
these runs also result in the same power - law size distribution .
the upper cut - off radius in these cases are larger than that of the standard case but they are also well - described with equation ( [ eq : a_max ] ) .
finally we examine the dependence on the sticking probability . performing two more runs with @xmath163 and 0.3 , we compared these results with the standard case in figure [ fig:4]c .
the power - law distributions form even for different sticking probabilities but the absolute values of the distribution are inversely proportional to the sticking probability . in all simulations , the mass of the central star was set to be one solar mass . the central star s mass is included only in @xmath24 in the equations governing the simulations .
if the time and velocities are scaled by @xmath164 and @xmath165,respectively , @xmath24 appears only in the ratio @xmath166 . in the above , we found that the numerical results does not depend on @xmath166 ( or the disk temperature ) .
hence , the above numerical results can be used for stars with different masses if we use the scaled time . in the above we showed that , in the settling stage , the size distribution of floating grains is universally given by a power - law distribution with the index of @xmath167 , that is , @xmath168 .
the coefficient @xmath169 of the distribution is inversely proportional to the sticking probability and independent of the disk properties and the material density of grains .
the upper cut - off of the size distribution is well - described by equation ( [ eq : a_max ] ) .
here we analytically derive the grain size distribution in the settling stage .
first we examine the settling process . here
we neglect grain growth since it hardly occurs for floating grains in the settling stage .
the settling velocity is given by equation ( [ eq : v_z2 ] ) .
we consider epstein s regime for gas drag in equation ( [ eq : gas_drag4 ] ) because most floating grains are smaller than cm - size in the settling stage .
the vertical distribution of the gas density is generally given by @xmath170 this expression is valid even for the cases of vertically non - isothermal disks . in non - isothermal cases , we set @xmath22 to be the value at the mid - plane to make it independent of @xmath18 . in the isothermal case , the function @xmath171 is given by @xmath172 /\sqrt{2\pi}$ ] .
the grain settling is described by the equation @xmath173 in epstein s regime , the drag coefficient @xmath174 is inversely proportional to @xmath175 and the ratio @xmath176 is independent of @xmath18 even for non - isothermal disks .
we introduce the normalized vertical axis @xmath177 and the non - dimensional time @xmath178 .
then equation ( [ eq : settle1 ] ) is reduced to @xmath179 and the solution is expressed as @xmath180 .
that is , @xmath181 the above solution also depends on the initial height @xmath182 .
figure [ fig:5 ] shows the solution @xmath183 for various @xmath184 in the isothermal case .
if the initial height is high enough ( @xmath185 ) , the solution is almost independent of @xmath186 except at an early stage .
this is because the gas density is very low at a high altitude and grains settle down rapidly in epstein s regime .
the small floating grains in figure [ fig:2]b come from the high altitude . on the other hand ,
most grains initially located at a low altitude ( @xmath187 ) are removed by the coagulation with larger ones settling from a high altitude .
thus we only consider grains initially located at a high altitude and neglect the dependence of the solution on @xmath182 .
equation ( [ eq : zsol ] ) represents well - defined trajectory of each grain .
this indicates that there exist grains with a definite size at each height at certain time .
that is , the settling causes size segregation in the disk .
actually , in our numerical simulations , the altitude of each floating grain is well - described by equation ( [ eq : zsol ] ) and the size segregation occurs .
this size segregation makes the relative velocity between colliding grains small and prevents grain growth considerably in the settling stage .
secondly , we derive a simple relation between the spatial and size distribution @xmath134 and the vertically integrated size distribution @xmath136 .
because of the size segregation , the spatial and size distribution @xmath188 is proportional to the delta function @xmath189 , where the function @xmath190 is given by solving equation ( [ eq : zsol ] ) with respect to @xmath175 . in a vertical section of ( @xmath18 , @xmath191 ) , there exist grains with sizes ( @xmath190 , @xmath192 ) , where @xmath193 is given by @xmath194 ( see eq .
[ [ eq : zsol ] ] ) . noting that grains with a same size exist at two altitudes ( i.e. , above and below the mid - plane ) , we obtain @xmath195 .
this equation determines @xmath188 as @xmath196 using equation ( [ eq : n - ns ] ) , we can derive the expression of the power - law size distribution of @xmath136 in the settling stage . in the growth stage ,
the growth time @xmath120 is shorter than @xmath128 .
dust settling , however , decreases the grain number density at each altitude and increases the growth time . at the transition to the settling stage ,
the two time scales equal to each other and , in the settling stage , the growth time further increases because of the size segregation .
hence the size distribution is fixed at the transition ( see also fig .
[ fig:2 ] ) and the fixed size distribution is obtained from the equation of @xmath197 .
we reevaluate @xmath120 , by setting the grain number density @xmath198 in equation ( [ eq : t_g ] ) and using equation ( [ eq : n - ns ] ) for @xmath188 .
the time @xmath199 can be replaced by @xmath128 since equation ( [ eq : n - ns ] ) is derived from the solution of settling .
then we obtain @xmath200 and the equation , @xmath197 , yields @xmath201 the obtained size distribution is dependent only on the sticking probability . in figures
[ fig:2 ] and [ fig:4 ] , equation ( [ eq : size_dis ] ) is plotted by dashed lines . in all cases ,
equation ( [ eq : size_dis ] ) agrees well with the numerically obtained size distributions of floating grains in the settling stage . in the above derivation
, we only consider grains initially located at a high altitude ( @xmath202 ) to make the solution almost independent of @xmath184 .
thus grains initially located at @xmath203 are not included in the size distribution of equation ( [ eq : size_dis ] ) . in figure
[ fig:2]b , such grains are located in the bump at several @xmath12 m at 10@xmath204yr . since most of them are removed by the coagulation with larger grains in an early part of the settling stage ,
equation ( [ eq : size_dis ] ) is thus valid in most of the settling stage . )
is dependent of @xmath184 even for large @xmath184 .
thus the above derivation would be invalid in stokes regime .
] furthermore , since we did not assume vertically isothermal disks , equation ( [ eq : size_dis ] ) would be valid for vertically non - isothermal disks like the two - layer model , too .
we also mention the temperature dependence of the spatial and size distribution @xmath134 . as seen from equations ( [ eq : n - ns ] ) and
( [ eq : settle1 ] ) , @xmath134 depends on the temperature . however , if the spatial and size distribution is considered as a function of @xmath175 and the normalized vertical axis @xmath205 ( i.e. , @xmath206 ) , the product @xmath207 is independent of the temperature .
this invariance of @xmath207 is also seen in numerical simulations . in figure
[ fig:3 ] , we also plotted the result of the high temperature case to compare with the standard case . in both of the growth stage and the settling stage ,
the disk temperature hardly affect the values of the product @xmath208 @xmath209 .
this result comes from the invariance of @xmath207 . in our simulation of dust growth and settling
, we used a tentative disk temperature .
owing to the invariance of @xmath207 , however , our numerical results can be used for disks with different temperature distributions .
in order to evaluate the disk optical thickness , we adopted the following grain model .
grain compositions and abundance were set to be the same as pollack et al .
that is , the compositions are h@xmath43o - ice , organics , olivine , pyroxene , metallic iron , and troilite .
we assumed that grains consist of a mixture of these compositions instead of the core - mantle type grains adopted in chiang et al .
the model of well - mixed grains would be suitable for grown grains ( miyake & nakagawa 1993 ) . as the optical constants of olivine and pyroxene , we used data by dorschner et al .
( 1995 ) . for h@xmath43o - ice
, we used the same data as miyake & nakagawa ( 1993 ) . for troilite ,
data by begemann et al .
( 1994 ) was used at wavelengths from 10@xmath12 m to 500@xmath12 m .
other optical constants are set to be the same as those in pollack et al .
( 1994 ) . for such grains , we calculated the grain emissivity ( or the absorption efficiency ) @xmath210 for each grain radius @xmath175 , using the mie theory ( bohren & huffman 1983 ) . at high temperature of @xmath211k
, we used the grain emissivity of grains without ice or organics while we used that including ice or organics at @xmath212k . by the use of the emissivity @xmath210 and the spatial and size distribution @xmath213 ,
the opacity @xmath214 is expressed by @xmath215 since @xmath216 ( see eq .
[ [ eq : dens ] ] or [ [ eq : rho_gas2 ] ] ) , we can see that the opacity is expressed with the product @xmath207 , which is independent of the disk temperature .
the disk optical thickness @xmath217 is given by @xmath218 . at each heliocentric distance @xmath83 in the standard disk
, we performed simulations of dust growth and settling and calculated the disk optical thickness . figure [ fig:6 ] shows the radial distributions of the optical thickness in the standard disk for wavelengths @xmath219 m and 1 mm . from @xmath220 to @xmath0yr , the radial distribution of the optical thickness is almost flat for both wavelengths .
this is because the vertically integrated size distribution @xmath136 of floating grains and its upper cut - off @xmath149 are independent of the radial position @xmath83 .
large grains in the dust layer have only a minor contribution to the optical thickness though they are effective in the inner part ( @xmath22110au ) for @xmath222 mm at @xmath223yr . for @xmath224
m , the optical thickness is much smaller in the inner disk than the outer disk .
this is because grain growth due to brownian motion depletes the micron - sized grains in the inner disk ( see fig . [ fig:2]b ) .
this extremely optically thin region may corresponds to inner holes recently observed by the spitzer space telescope ( forrest at al .
2004 ; meyer et al . 2004
; uchida et al . 2004 ) . in most previous analysis of observations ,
the optical thickness is assumed to be proportional to the disk surface density . under this assumption ,
the radial power - law index of the surface density @xmath225 is estimated to be from 0 to -1 by recent observations with high - resolution at millimeter wavelengths ( kitamura et al . 2002 ) . however , this assumption is invalid because of grain growth .
our model shows that the flat radial distribution of the disk optical thickness forms even in disks in which the surface density is proportional to @xmath226 like the mmsn .
we also performed simulations for another disk , in which the surface density is proportional to @xmath227 .
the gas surface density at 1au , @xmath36 , is set to be the same as the standard disk . in figure
[ fig:7 ] , we display the radial distributions of the optical thickness of this disk for @xmath228 mm .
the optical thickness is almost proportional to @xmath229 at @xmath230yr in this disk .
this is because the upper cut - off radius @xmath149 of the size distribution is also proportional to @xmath229 at a given time in this disk ( see eq .
[ [ eq : a_max ] ] ) . from these results
, it is found that the difference in the power - law index between @xmath217 at millimeter wavelength and @xmath21 is about 1.5 at @xmath231yr .
using our results on grains and the two - layer model proposed by chiang & goldreich ( 1997 ) , we calculate disk temperatures and spectral energy distributions of disks .
in the two layer model , the disk consists of two surface layers ( i.e. , the upper and the lower ) and the disk interior .
then the disk structure is characterized by the heights of the surface layers , @xmath232 , and temperatures of the surface layers and the disk interior , @xmath233 , @xmath234 .
we neglected the effect of the inner rim where silicate grains evaporate because it has a only minor contribution to the sed for t tauri disks ( dullemond et al .
we also neglected heating due to the disk accretion as chiang & goldreich ( 1997 ) did .
equations governing the two - layer model have already been described in detail by chiang et al .
( 2001 ) and , thus , we briefly describe them in appendix a. we numerically solved the equations and obtained the structure of the standard disk .
the effective temperature , @xmath235 , and the radius , @xmath236 of the central star are set to be 4000k and 0.0097au , respectively .
figure [ fig:8 ] shows time - evolution of the height of the upper surface layer @xmath33 .
the surface layer descends because of grain settling and growth .
the descent of the surface layer is remarkable in an inner part of the disk .
this is because of the significant reduction of the optical thickness in the inner disk mentioned above . actually , the disk is optically thin to the starlight in the region of @xmath237au ( @xmath238au ) at @xmath0yr ( @xmath239yr ) and the disk interior disappears there .
the grazing angle between the starlight and the disk surface given by equation ( [ eq : angle ] ) decreases because of the descent of the disk surface .
however , the grazing angle remains positive and the self - shadowing does not occur .
the temperatures in the surface layer and the disk interior are shown in figure [ fig:9 ] . because the temperature of the surface layer depends on the size of growing grains ( see eq .
[ [ eq : ts ] ] ) , it varies with time .
the temperature of the disk interior is much lower than that of the surface layer , as shown by chiang & goldreich ( 1997 ) .
the ratio of @xmath240 is about 4 in this case .
the temperature of the disk interior does not vary much with time in spite of the substantial descent of the surface layer .
in other disk models examined below , we also found similar disk structure to the standard disk . using the obtained disk structure , we calculate the disk sed with eqs .
( [ eq : fs ] ) and ( [ eq : fi ] ) .
we assume that the disk is face - on ( the inclination angle @xmath96 is zero ) .
figure [ fig:10 ] shows evolution of the seds from the standard disk and the central star from @xmath220yr to @xmath239yr .
the sed does not change much until @xmath241yr .
after that , the energy flux from the disk decreases especially at @xmath6 m .
the decreases in the disk optical thickness mainly reduces the flux from the disk interior , which has a major contribution to the sed at @xmath6 .
the feature at 40@xmath12 m , which comes from icy grains in the surface layer , does not decline much .
one the other hand , the silicate feature at 10@xmath12 m depletes at @xmath223yr .
this is due to the `` inner hole " ( @xmath242 ) of the disk optical thickness mentioned above .
this result may explain the depletion of silicate feature in coku tau/4 reported by forrest et al .
the disk life time estimated from the observed energy fluxes at millimeter / sub - millimeter wavelengths is about @xmath7yr ( e.g. , wyatt at al .
2003 ) . in our result on the standard disk , on the other hand ,
the depletion time at the wavelengths is several@xmath243yr .
we examine the possibilities to reconcile the discrepancy below .
we performed three more similar simulations of grain growth and settling .
the first one is for a disk ten times as massive as the standard one , the second one is for the standard disk but with the sticking probability @xmath163 , and the third one is for a disk with the disk outer radius @xmath244au . in figure [ fig:11 ] , we compare the seds in those cases with the standard disk at @xmath223yr . in the case of the massive disk , fluxes at sub - millimeter
/ millimeter wavelengths are larger than those in the standard disk .
this is because of slow settling in the massive disk ( see eq . [ [ eq : t_settle ] ] ) .
the cases with a small sticking probability and with a large outer radius also result in larger fluxes at the long wavelengths .
this is due to the slow grain growth at the outer disk radius ( see eq . [ [ eq : t_g ] ] ) . in figure
[ fig:12 ] , the fluxes in these cases are displayed as a function of time .
the energy flux from the massive disk is larger than that from the standard disk initially and starts to decrease at a few @xmath241yr similar to the standard disk . in the cases with a small @xmath57 and with a large @xmath245 ,
on the other hand , the flux starts to decrease more lately .
this is because of slow grain growth at the outer disk radii .
the depletion time in these cases is almost consistent with that inferred from the sed observations .
our results indicate that the decrease in the observed energy fluxes at millimeter / sub - millimeter wavelengths can be explained by grain growth and settling without any disk depletion .
hence the gaseous disks themselves can survive more than @xmath239yr . in order to determine the disk life time , which is crucial for formation of gas giant planets , it is necessary to observe the decay of gas components .
we performed numerical simulations of dust growth and settling in passive disks and obtained the spatial and size distributions of grains .
as the first step of series of papers , we consider laminar disks in the present paper . using the obtained grain distributions and the two - layer model , we also examined evolution of disk structure and seds due to dust growth and settling .
our results are summarized as follows : 1 .
evolution of grains in laminar disks is divided into the growth stage and the subsequent settling stage . at the end of the growth stage ,
large grains settle to the mid - plane and form a dust layer , while small grains remain floating above the layer .
the size distribution , @xmath136 , of the floating grains is well - described by the power - law distribution analytically obtained as @xmath246 ( eq . [ [ eq : size_dis ] ] ) with the upper cut - off @xmath149 ( eq . [ [ eq : a_max ] ] ) .
this size distribution is kept in the settling stage . in the settling stage ,
relatively large floating grains settle earlier and the upper cut - off radius of the size distribution decreases as equation ( [ eq : a_max ] ) .
the validity of equations ( [ eq : size_dis ] ) and ( [ eq : a_max ] ) was confirmed with numerical simulations for wide parameter ranges .
the approximate solution of the spatial and size distribution , @xmath134 , were obtained analytically , too ( eq . [
[ eq : n - ns ] ] ) .
the optical thickness of disks considerably decreases because of dust growth .
the optical thickness is governed by floating small grains rather than large grains in the dust layer .
rapid grain growth in the inner part of disks makes the radial distribution of the disk optical thickness less steep than that of the disk surface density @xmath1 . for disks with @xmath2 ,
the radial distribution of the optical thickness is almost flat for all wavelengths at @xmath231yr .
the difference in the power - law index between the optical thickness at @xmath222 mm and the surface density is generally about @xmath247 .
this difference between the optical thickness and the surface density should be taken into account in analyses of observations with high - resolution , which will be done by spitzer space telescope or atacama large millimeter array ( alma ) . at @xmath248yr , the optical thickness for infrared radiation
is extremely small in the inner region ( @xmath5 3au ) .
this region may corresponds to inner holes observed by the spitzer space telescope ( forrest at al .
2004 ; meyer et al . 2004
; uchida et al . 2004 ) .
the surface layer directly irradiated from the central star descends because of grain settling and growth .
the descent of the surface layer is remarkable in an inner part of the disk because grain growth due to brownian motion depletes the micron - sized grains there .
the temperature of the disk interior , @xmath249 , is much lower than that of the surface layer , @xmath249 , as shown by chiang & goldreich ( 1997 ) .
the ratio of @xmath240 is about 4 in our simulations .
4 . the energy flux from the disk decreases due to grain growth and settling especially at @xmath6 m . in the standard disk with the sticking probability @xmath142 and the outer radius @xmath250150au ,
the depletion time of the energy fluxes at sub - millimeter/ millimeter wavelengths is obtained as several@xmath243yr . in the case with a small @xmath57 or with a larger @xmath245 ,
the depletion time is longer .
our results indicate that the decrease in the observed energy fluxes at millimeter / sub - millimeter wavelengths can be explained by grain growth and settling without disk depletion .
the gaseous disks themselves can survive more than @xmath239yr . the silicate feature at 10@xmath12 m depletes on the time scale of @xmath239yr because of the formation of the extremely optical thin inner region . in the present paper
, we consider only laminar disks . here
, we comment on the effect of turbulence on dust growth and disk structure .
vertical stirring by turbulence would prevent dust settling . as shown by dullemond & dominik ( 2004b ) , vertical stirring by turbulence against dust settling is effective even in weakly turbulent disks with the parameter @xmath15 for the disk viscosity .
in turbulent disks , vertical stirring is expected to keep the surface layer high compared with laminar disks . on the other hand
, turbulence also increases the relative velocity between grains .
since the collisional probability is proportional to the relative velocity , turbulence accelerates grain growth , too ( e.g. , weidenschilling 1984 ; mizuno et al . 1988 ) .
acceleration of grain growth would result in a rapid decrease in the opacity and the disk optical thickness .
this would make the surface layer lower .
it is not clear whether the turbulence makes the surface layer higher with stirring or lower with acceleration of grain growth .
furthermore , in sufficiently strong turbulence , collisional disruption due to high relative velocity creates small grains and would increases the optical thickness . to clarify which effect is dominant , it is necessary to perform our simulations of grain growth and settling , by taking into account above effects of turbulence . in the next paper ,
we investigate cases of turbulent disks and examine more realistic evolution of the optical thickness and seds in protoplanetary disks .
we thank y. kitamura , m. honda , and m. momose for valuable comments .
we also thank the anonymous reviewer for many valuable comments and suggestions .
this work is supported by grant - in - aid of the japanese ministry of education , science , and culture ( 15740116 and 16340054 )
here we briefly explain the equations governing the structure of a passive disk in the two - layer model .
the optical thicknesses of the surface layer and the disk interior , @xmath251 , @xmath252 are given by @xmath253 respectively , where the height @xmath254 is the boundary between the surface layer and the disk interior in the normalized vertical axis .
the opacity @xmath214 is given by equation ( [ eq : kappa ] ) by the use of our numerical results . by the introduction of the normalized axis , the optical thicknesses of them are independent of the disk temperature .
the planck mean optical thicknesses of them are @xmath255 the boundary @xmath254 is indirectly determined by the equation @xmath256 where @xmath235 is the effective temperature of the central star and @xmath257 is the grazing angle ( i.e. the angle between the starlight and the disk surface ) . by the use of @xmath258 , the grazing angle is given by @xmath259 where @xmath236 is the radius of the central star .
when the disk optical thickness is small enough , equation ( [ eq : zs ] ) does not have a positive solution of @xmath254 .
this corresponds to the case with no disk interior . in this case
we set @xmath260 and , thus , the optical thicknesses of the disk interior , @xmath252 and @xmath261 , are zero .
the grains in the surface layer have various temperatures depending on their radii . instead of the various grain temperatures , we use the mean temperature of the surface layer @xmath249 defined by @xmath262 the temperature of the disk interior @xmath263
is given by @xmath264 } { 1-\exp[-2\tau{_{\text{p , i}}}(t{_{\text{i}}})]}. \label{eq : ti}\ ] ] in equations ( [ eq : ts ] ) and ( [ eq : ti ] ) , we doubled the left - hand sides , assuming that the entire star is visible from the surface layer ( dullemond et al .
furthermore , we inserted the factor 2 into the exponential functions in equation ( [ eq : ti ] ) in order to take into account radiative transfer in the oblique direction in the disk interior .
this is because , in the direction with the angle @xmath265 to the vertical axis , the optical thickness of the disk interior is double of that in the vertical direction .
equations ( [ eq : zs])-([eq : ti ] ) are coupled with one another .
we numerically solved these equations for a whole disk .
the derivative @xmath266 in equation ( [ eq : angle ] ) is evaluated with the central difference using the both sides of grids though the one - side difference is used at the innermost and the outermost grids . by the use of the solutions to the above equations , the emissions ( or the energy fluxes ) at a frequency @xmath267 from the surface layers and the disk interior , @xmath268 , @xmath269 ,
are given by @xmath270 2\pi r dr \label{eq : fs}\ ] ] and @xmath271 2\pi r dr , \label{eq : fi}\ ] ] respectively , where @xmath272 is the distance from the observer to the source and @xmath96 is the inclination angle of the disk ( @xmath273 means face - on ) .
in the above we assumed that @xmath274 .
the energy flux from the whole disk is given by their sum .
99 adachi , i. , hayashi , c. , and nakazawa , k. 1976 , prog .
phys . , 56 , 1756 beckwith , s. v. w. , sargent , a. i. , chini , r. s. , & guesten , r. 1990 , , 99 , 924 begemann , b. , dorschner , t. , henning , t. , mutschke , h. , & thamm , e. 1994 , , 423 , l71 bohren , c. f. , & huffman , d. r. 1983 , absorption and scattering of light by small particles ( new york : wiley ) bouwman , j. , de koter , a. , dominik , c. , and waters , l. b. f. m. 2003 , a&a , 401 , 577 calvet , n. , dalessio , p. , hartmann , l. , wilner , d. , walsh , a. , and sitko , m. 2002 , , 568 , 1008 chapman , s. , & cowling , t. g. 1970 , the mathematical theory of non - uniform gases ( london : cambridge univ . press ) chiang , e. i. , & goldreich , p. 1997
, , 490 , 368 .
chiang , e. i. , joung , m. k. , creech - eakman , m. j. , qi , c. , kessler , j. e. , blake , g. a. , & van dishoeck , e. f. 2001 , , 547 , 1077 cuzzi , j. n. , dobrovolskis , a. r. , and champney , j. m. 1993 , icarus , 106 , 102 dalessio , p. , calvet , n. , & hartmann , l. 2001 , , 553 , 321 dorschner , t. , begemann , b. , henning , t. , jger , c. , & mutschke , h. 1995 , a&a , 300 , 503 dullemond , c. p. , dominik , c. , & natta , a. 2001 , , 560 , 957 dullemond , c. p. , & dominik , c. 2004a , a&a , 417 , 159 dullemond , c. p. , & dominik , c. 2004b , a&a , 421 , 1075 eisner , j. a. , lane , b. f. , akeson , r. l. , hillenbrand , l. a. , & sargent , a. i. 2003 , , 588 , 360 forrest , w. j. , et al .
2004 , , 154 , 443 fukagawa , m. , et al .
2004 , , 605 , l53 goldreich , p. , & ward , w. r. 1973 , , 183 , 1051 grady , c.a .
, woodgate , b. , bruhweiler , f.c . ,
boggess , a. , plait , p. , lindler , d.j . , clampin , m. , & kalas , p. 1999 , 523
, l151 hayashi , c. 1981 , prog .
suppl . , 70 , 35 honda , m. , kataza , h. , okamoto , y. k. , miyata , t. , yamashita , t. , sako , s. , takubo , s. , & onaka , t. 2003 , , 585 , l59 kitamura , y. , momose , m. , yokogawa , s. , kawabe , r. , tamura , m. , & ida , s. 2002 , , 581 , 357 meeus , g. , sterzik , m. , bouwman , j. , and natta , a. 2003 , a&a , 409 , 25 meyer , m. r. , et al .
2004 , , 154 , 422 miyake , k. , & nakagawa , y. 1993 , icarus , 106 , 20 miyake , k. , & nakagawa , y. 1995 , , 441 , 361 mizuno , h. , markiewicz , w. j. , & voelk , h. j. 1988 , a&a , 195 , 183 mundy , l. g. , et al .
1996 , , 464 , l169 nakagawa , y. , nakazawa , k. , & hayashi , c. 1981 , icarus , 45 , 517 nakagawa , y. , sekiya , m. & hayashi , c. 1986 , icarus , 67 , 375 natta , a. , testi , l. , neri , r. , shepherd , d. s. , & wilner , d. j. 2004 , a&a , 416 , 179 ohtsuki , k. , nakagawa , y. , & nakazawa , k. 1990 , icarus , 83 , 205 pollack , j. b. , hollenbach , d. , beckwith , s. , simonelli , d. p. , roush , t. , & fong , w. 1994 , , 421 , 615 przygodda , f. , van boekel , r. , brahm , p. , melnikov , s. y. , waters , l. b. f. m. , & leinert , ch .
2003 , a&a , 412 , l43 safronov , v. s. 1969 , evolution of the protoplanetary cloud and formation of the earth and the planets ( nasa tech .
f-677 ) ( nauka : moscow ) schmitt , w. , henning , t. , & mucha , r. 1997 , a&a , 325 , 569 shakura , n. i. , & sunyaev , r. a. 1973 , a&a , 24 , 337 strom , k. m. , strom , s. e. , edwards , s. , cabrit , s. , & skrutskie , m. f. 1989 , , 97 , 1451 suttner , g. , & yorke , h. w. 2001 , , 551 , 461 takeuchi , t. , & lin , d. n. c. 2002 , , 581 , 1344 takeuchi , t. , & lin , d. n. c. 2003 , , 593 , 524 tanaka , h. , & nakazawa , k. 1994 , icarus , 107 , 404 uchida , k. i. , et al .
2004 , 154 , 439 van leer , b. 1977 , j. comput .
phys . , 23 , 276 weidenschilling , s. j. 1980 , icarus , 44 , 172 weidenschilling , s. j. 1984 , icarus , 60 , 553 weidenschilling , s. j. 1997 , icarus , 127 , 290 weidenschilling , s. j. , & cuzzi , j. n. 1993 , protostar and planets iii , ed .
e. h. levy & j. l. lunine ( tucson : univ .
arizona press ) , 1031 wyatt , m. c. , dent , w. r. f. , & greaves , j. s. 2003 , mnras , 342 , 876 .
youdin , a. n. , & shu , f. h. 2002 , , 580 , 494
|
dust growth and settling considerably affect the spectral energy distributions ( seds ) of protoplanetary disks .
we investigated dust growth and settling in protoplanetary disks through numerical simulations to examine time - evolution of the disk optical thickness and seds . in the present paper
, we considered laminar disks as a first step of a series of papers . as a result of dust growth and settling , a dust layer forms around the mid - plane of a gaseous disk . after the formation of the dust layer , small dust grains remain floating above the layer .
although the surface density of the floating small grains is much less than that of the dust layer , they govern the disk optical thickness and the emission .
size distributions of the floating grains obtained from numerical simulations are well - described by an universal power - law distribution , which is independent of the disk temperature , the disk surface density , the radial position in the disk , etc .
the floating small grains settle onto the dust layer in a long time scale compared with the formation of the dust layer .
typically , it takes @xmath0yr for micron - sized grains . rapid grain growth in the inner part of disks makes the radial distribution of the disk optical thickness less steep than that of the disk surface density , @xmath1 .
for disks with @xmath2 , the radial distribution of the optical thickness is almost flat for all wavelengths at @xmath3yr . at @xmath4yr , the optical thickness of the inner disk ( @xmath5 a few au ) almost vanishes , which may correspond to disk inner holes observed by spitzer space telescope .
furthermore , we examined time - evolution of disk seds , using our numerical results and the two - layer model .
the grain growth and settling decrease the magnitude of the seds especially at @xmath6 m .
our results indicate that grain growth and settling can explain the decrease in observed energy fluxes at millimeter / sub - millimeter wavelengths with time scales of @xmath7yr without depletion of the disks .
|
many remaining regions in the parameter space of the minimal supersymmetric standard model ( mssm ) , which yield the observed thermal relic density for neutralino dark matter , rely on very specific mechanisms , such as higgs - resonant annihilation in the so - called funnel region , or sfermion co - annihilation . in @xcite
we identified new regions , where the dark matter particle is a mixed as opposed to pure wino , has mass in the tev region , and yields the observed relic density .
these new regions are driven to the correct relic abundance by the proximity of the resonance of the sommerfeld effect due to electroweak gauge boson exchange .
in such situations , the annihilation cross section is strongly velocity dependent , and the present - day annihilation cross section is expected to be relatively large , potentially leading to observable signals in indirect searches for dark matter ( dm ) . on the other hand
, a substantial higgsino fraction of a mixed dark matter particle leads to a large , potentially observable dark matter - nucleon scattering cross section . in this paper
we address the question of which part of this region survives the combination of direct and indirect detection constraints .
for the latter we consider diffuse gamma - rays from the dwarf spheroidal galaxies ( dsphs ) , galactic cosmic rays ( crs ) and cosmic microwave background ( cmb ) anisotropies .
these have been found to be the most promising channels for detecting or excluding the pure - wino dm model @xcite .
stronger limits can be obtained only from the non - observation of the gamma - line feature and to a lesser extent from diffuse gamma - rays both originating in the galactic centre ( gc ) .
indeed , it has been shown @xcite that the pure - wino model is ruled out by the absence of an excess in these search channels , unless the galactic dark matter profile develops a core , which remains a possibility .
since the viability of wino - like dm is a question of fundamental importance , we generally adopt the weaker constraint in case of uncertainty , and hence we take the point of view that wino - like dm is presently not excluded by gamma - line and galactic diffuse gamma - ray searches .
future results from the erenkov telescope array ( cta ) are expected to be sensitive enough to resolve this issue ( see e.g. @xcite ) , and will either observe an excess in gamma - rays or exclude the dominantly wino dm mssm parameter region discussed in the present paper . imposing the observed relic density as a constraint ,
the pure - wino dm model has no free parameters and corresponds to the limit of the mssm when all other superpartner particles and non - standard higgs bosons are decoupled .
departing from the pure wino in the mssm introduces many additional dimensions in the mssm parameter space and changes the present - day annihilation cross section , branching ratios ( brs ) for particular primary final states , and the final gamma and cr spectra leading to a modification of the limits .
the tools for the precise computation of neutralino dark matter ( co- ) annihilation in the generic mssm when the sommerfeld enhancement is operative have been developed in @xcite and applied to relic density computations in @xcite .
the present analysis is based on an extension of the code to calculate the annihilation cross sections for all exclusive two - body final states separately , rather than the inclusive cross section .
further motivation for the present study is provided by the spectrum of the cosmic antiproton - to - proton ratio reported by the ams-02 collaboration @xcite , which appears to be somewhat harder than expected from the commonly adopted cosmic - ray propagation models . in @xcite
it has been shown that pure - wino dm can improve the description of this data .
although our understanding of the background is insufficient to claim the existence of a dark matter signal in antiprotons , it is nevertheless interesting to check whether the surviving mixed - wino dm regions are compatible with antiproton data .
the outline of this paper is as follows . in section [ sec : fluxes ] we summarize the theoretical input , beginning with a description of the dominantly wino mssm parameter region satisfying the relic - density constraint , then providing some details on the computation of the dm annihilation rates to primary two - body final states .
the following section [ sec : methodology ] supplies information about the implementation of the constraints from diffuse gamma - rays from the dsphs , galactic crs , direct detection and the cmb , and the data employed for the analysis .
the results of the indirect detection analysis are presented in section [ sec : results ] as constraints in the plane of the two most relevant parameters of the mssm , the wino mass parameter @xmath11 and @xmath12 , where @xmath13 is the higgsino mass parameter . in section
[ sec : maximalplot ] the indirect detection constraints are combined with that from the non - observation of dark matter - nucleon scattering . for the case of @xmath1
we demonstrate the existence of a mixed wino - higgsino region satisfying all constraints , while for @xmath0 we show that there is essentially no remaining parameter space left . section [ sec : conclusions ] concludes .
in @xcite the sommerfeld corrections to the relic abundance computation for tev - scale neutralino dark matter in the full mssm have been studied . the ability to perform the computations for mixed dark matter at a general mssm parameter space point
@xcite revealed a large neutralino mass range with the correct thermal relic density , which opens mainly due to the proximity of the resonance of the sommerfeld effect and its dependence on mssm parameters . in this subsection
we briefly review the dominantly - wino parameter region identified in @xcite , which will be studied in this paper .
`` dominantly - wino '' or `` wino - like '' here refers to a general mssm with non - decoupled higgs bosons , sfermions , bino and higgsinos as long as the mixed neutralino dark matter state is mainly wino .
we also require that its mass is significantly larger than the electroweak scale .
the well - investigated pure - wino model refers to the limit in this parameter space , when all particles other than the triplet wino are decoupled . despite the large number of parameters needed to specify a particular mssm completely , in the dominantly - wino region
, the annihilation rates depend strongly only on a subset of parameters .
these are the wino , bino and higgsino mass parameters @xmath11 , @xmath14 and @xmath13 , respectively , which control the neutralino composition and the chargino - neutralino mass difference , and the common sfermion mass parameter @xmath15 . in this work
we assume that the bino is much heavier that the wino , that is , the lightest neutralino is a mixed wino - higgsino .
effectively a value of @xmath16 larger than @xmath11 by a few 100 gev is enough to decouple the bino in the tev region .
the wino mass parameter determines the lightest neutralino ( lsp ) mass , and the difference @xmath17 the wino - higgsino admixture . in the range @xmath18 tev considered here ,
the relation @xmath19 remains accurate to a few gev , when some higgsino fraction is added to the lsp state , and values of @xmath20 gev imply practically decoupled higgsinos .
increasing the higgsino component of the wino - like lsp lowers its coupling to charged gauge bosons , to which wino - like neutralinos annihilate predominantly , and therefore increases the relic density .
larger mixings also imply that the mass difference between the lightest chargino and neutralino increases , which generically reduces the size of the sommerfeld enhancement of the annihilation cross section .
these features are apparent in the contours of constant relic density in the @xmath12 vs. @xmath11 plane for the wino - higgsino case shown in @xcite , which are almost straight for large @xmath12 , but bend to lower values of @xmath21 as @xmath12 is reduced .
a representative case is reproduced in fig .
[ fig : oldlux ]
. the contours also bend towards lower @xmath11 when sfermions become lighter , as they mediate the t- and u - channel annihilation into sm fermions , which interferes destructively with the s - channel annihilation , effectively lowering the co - annihilation cross section . by choosing small values of @xmath15 ( but larger than @xmath22 to prevent sfermion co - annihilation , not treated by the present version of the code ) , lsp masses as low as @xmath23 tev are seen to give the correct thermal density , to be compared with the pure - wino result , @xmath24 tev . for @xmath25 tev
a resonance in the sommerfeld - enhanced rates is present , which extends to larger @xmath11 values as the higgsino fraction is increased .
the enhancement of the cross section in the vicinity of the resonance makes the contours of constant relic density cluster around it and develop a peak that shifts @xmath21 to larger values .
in particular , the largest value of @xmath11 , which gives the correct thermal relic density , is close to 3.3 tev , approximately 20% higher than for the pure - wino scenario .
the influence of the less relevant mssm higgs mass parameter @xmath26 is also noticeable when the lsp contains some higgsino admixture , which enhances the couplings to the higgs ( and z ) bosons in s - channel annihilation .
this is more pronounced if @xmath26 is light enough such that final states containing heavy higgs bosons are kinematically accessible .
the corresponding increase in the annihilation cross section results in positive shifts of around 100 to 250 gev in the value of @xmath11 giving the correct relic density on decreasing @xmath26 from 10 tev to 800 gev . in summary ,
a large range of lightest neutralino masses , @xmath27 tev , provides the correct relic density for the mixed wino - higgsino state as a consequence of the sommerfeld corrections .
the mssm parameter points considered in this paper have passed standard collider , flavour and theoretical constraints as discussed in @xcite . in the dominantly - wino parameter space ,
most of the collider and flavour constraints are either satisfied automatically or receive mssm corrections that are suppressed or lie within the experimental and theoretical uncertainties . ref .
@xcite further required compatibility with direct dark matter detection constraints by imposing that the dm - nucleon spin - independent cross section was less than twice the lux limits reported at the time of publication @xcite .
this did not affect the results significantly , see fig .
[ fig : oldlux ] , as in most of the parameter space of interest the scattering cross section was predicted to be much above those limits .
recently the lux collaboration has presented a new limit , stronger than the previous one by approximately a factor of four @xcite , potentially imposing more severe constraints on the dominantly - wino neutralino region of the mssm parameter space .
the details of the implementation of the limits from indirect detection searches for the mixed wino , which were not included in our previous analysis , and from the new lux results are given in section [ sec : methodology ] . vs. @xmath28 plane for @xmath0 , as computed in @xcite .
the ( green ) band indicates the region within @xmath29 of the observed dark matter abundance .
parameters are as given in the header , and the trilinear couplings are set to @xmath30 tev for all sfermions except for that of the stop , which is fixed by the higgs mass value .
the black solid line corresponds to the old lux limit @xcite on the spin - independent dm - nucleon cross section , which excludes the shaded area below this line . relaxing the old lux limit by a factor of two to account for theoretical uncertainties eliminates the direct detection constraint on the shown parameter space region .
[ fig : oldlux ] ] the annihilation of wino - like dm produces highly energetic particles , which subsequently decay , fragment and hadronize into stable sm particles , producing the cr fluxes .
the primary particles can be any of the sm particles , and the heavy mssm higgs bosons , @xmath31 and @xmath32 , when they are kinematically accessible .
we consider neutralino dark matter annihilation into two primary particles .
the number of such exclusive two - body channels is 31 , and the corresponding neutralino annihilation cross sections are computed including sommerfeld loop corrections to the annihilation amplitude as described in @xcite . as input for this calculation we need to provide the tree - level exclusive annihilation rates of _ all _ neutral neutralino and chargino pairs , since through sommerfeld corrections the initial lsp - lsp state can make transitions to other virtual states with heavier neutralinos or a pair of conjugated charginos , which subsequently annihilate into the primaries .
the neutralino and chargino tree - level annihilation rates in the mssm have been derived analytically in @xcite , and including @xmath33-corrections in @xcite , in the form of matrices , where the off - diagonal entries refer to the interference of the short - distance annihilation amplitudes of different neutralino / chargino two - particle states into the same final state . for the present analysis
the annihilation matrices have been generalized to vectors of matrices , such that the components of the vector refer to the 31 exclusive final states .
the large number of different exclusive final states can be implemented without an increase in the cpu time for the computation relative to the inclusive case .
since the information about the exclusive annihilation rates only enters through the ( short - distance ) annihilation matrices , the two - particle wave - functions that account for the ( long - distance ) sommerfeld corrections only need to be computed once . on the contrary , since the @xmath33-corrections to the annihilation of dm in the present universe are very small , they can be neglected , which results in a significant reduction in the time needed to compute the annihilation matrices .-
corrections , we did not make use of this simplification in the present analysis . ]
it further suffices to compute the present - day annihilation cross section for a single dark matter velocity , and we choose @xmath34 .
the reason for this choice is that the sommerfeld effect saturates for very small velocities , and the velocity dependence is negligible for velocities smaller than @xmath35 .
the energy spectrum @xmath36 of a stable particle @xmath37 at production per dm annihilation can be written as @xmath38 where @xmath39 , and @xmath40 represents the contribution from each two - body primary final state @xmath41 with branching fraction @xmath42 to the spectrum of @xmath37 after the decay , fragmentation and hadronization processes have taken place .
we compute @xmath42 from our mssm sommerfeld code as described above and use the tables for @xmath40 provided with the pppc4dmid code @xcite , which include the leading logarithmic electroweak corrections through the electroweak fragmentation functions @xcite .
two comments regarding the use of the spectra provided by the pppc4dmid code are in order .
the code only considers primary pairs @xmath41 of a particle together with its antiparticle , both assumed to have the same energy spectrum . for wino - like dm there
exist primary final states with different species , _
i.e. _ @xmath43 with @xmath44 , such as @xmath45 and @xmath46 . in this case , we compute the final number of particles @xmath37 produced from that channel as one half of the sum of those produced by channels @xmath47 and @xmath48 .
this is justified , since the fragmentation of particles @xmath49 and @xmath50 is independent .
a second caveat concerns the heavy mssm higgs bosons that can be produced for sufficiently heavy neutralinos .
these are not considered to be primary channels in the pppc4dmid code , which only deals with sm particles .
a proper treatment of these primaries would first account for the decay modes of the heavy higgs bosons , and then consider the fragmentation and hadronization of the sm multi - particle final state in an event generator . instead of a full treatment ,
we replace the charged higgs @xmath32 by a longitudinal - polarized @xmath51-boson , and the neutral heavy higgses @xmath52 by the light higgs @xmath53 when computing the spectra in @xmath54 .
this approximation is not very well justified .
however , the branching ratios of the dominantly - wino neutralino to final states with heavy higgses are strongly suppressed , and we could equally have set them to zero without a noticeable effect on our results .
the branching fractions of primary final states obtained from our code are shown in the left panel of fig .
[ fig : fluxes ] as a function of the higgsino fraction for a wino - like lsp with 2 tev mass .
the pure wino annihilates mostly to @xmath55 and to a lesser extent to other pairs of gauge bosons , including the loop - induced photon final state , which is generated by the sommerfeld correction .
the annihilation to fermions is helicity or @xmath56-wave suppressed .
the suppression is lifted only for the @xmath57 final state as the higgsino admixture increases , in which case this final state becomes the second most important .
except for this channel , the dominant branching fractions are largely independent of the higgsino fraction .
the annihilation to @xmath55 is always dominant and above 75% .
the final spectra of photons , positrons and antiprotons per annihilation at production for small ( solid lines ) and large ( dashed lines ) higgsino mixing are plotted in the right panel of fig .
[ fig : fluxes ] .
the spectra in these two extreme cases are very similar , because @xmath58 is the dominant primary final state largely independent of the wino - higgsino composition , and also the number of final stable particles produced by the sub - dominant primary channels do not differ significantly from each other .
the inset in the right - hand plot shows that the relative change between the mixed and pure wino case varies from about @xmath59 to about @xmath60 over the considered energy range .
concerning the variation with respect to the dm mass , the most important change is in the total annihilation cross section , not in the spectra @xmath36 .
the branching ratios @xmath42 to primaries depend on the lsp mass in the tev regime only through the sommerfeld corrections , which can change the relative size of the different channels .
however , since for wino - like neutralinos annihilation into @xmath55 dominates the sum over @xmath41 in ( [ eq : spectra ] ) , the dependence of the final spectra on @xmath21 is very mild . and sfermions .
@xmath61 refers to the higgsino fraction of the lightest neutralino in the convention of @xcite .
right : comparison of @xmath62 , @xmath63 and gamma - ray spectra per annihilation at production of a 50% mixed wino - higgsino ( dashed ) to the pure - wino ( solid ) model .
the gamma - line component is not shown . in the inset at the bottom of the plot
the relative differences between the two spectra are shown.[fig : fluxes],title="fig : " ] and sfermions .
@xmath61 refers to the higgsino fraction of the lightest neutralino in the convention of @xcite .
right : comparison of @xmath62 , @xmath63 and gamma - ray spectra per annihilation at production of a 50% mixed wino - higgsino ( dashed ) to the pure - wino ( solid ) model . the gamma - line component is not shown .
in the inset at the bottom of the plot the relative differences between the two spectra are shown.[fig : fluxes],title="fig : " ]
in this section we discuss our strategy for determining the constraints on mixed - wino dark matter from various indirect searches .
while the analysis follows that for the pure wino @xcite , here we focus on the most relevant search channels : the diffuse gamma - ray emission from dsphs , antiprotons and positron crs , and the cmb .
moreover , since we consider wino - like dm with a possibly significant higgsino admixture , we implement the direct detection constraints as well .
the propagation of charged crs in the galaxy is best described within the diffusion model with possible inclusion of convection . in this framework the general propagation equation takes the form @xcite @xmath64 where @xmath65 is the number density of the @xmath49-th particle species with momentum @xmath56 and corresponding velocity @xmath66 , written in cylindrical coordinates @xmath67 , @xmath68 the inelastic scattering cross section , @xmath69 the production cross section of species @xmath49 by the fragmentation of species @xmath50 , and @xmath70 , @xmath71 are the lifetimes related to decays of @xmath49 and production from heavier species @xmath50 , respectively .
we solve with the help of the dragon code @xcite , assuming cylindrical symmetry and no convection , @xmath72 . with the galacto - centric radius @xmath73 , the height from the galactic disk @xmath74 and rigidity @xmath75
, we adopt the following form of the spatial diffusion coefficient : @xmath76 the momentum - space diffusion coefficient , also referred to as reaccelaration , is related to it via @xmath77 , where the alfvn velocity @xmath78 represents the characteristic velocity of a magnetohydrodynamic wave .
the free parameters are the normalization @xmath79 , the spectral indices @xmath80 and @xmath81 , the parameters setting the radial scale @xmath82 and thickness @xmath83 of the diffusion zone , and finally @xmath78 .
we fix the normalization at @xmath84 gv .
the diffusion coefficient is assumed to grow with @xmath73 , as the large scale galactic magnetic field gets weaker far away from the galactic center .
the source term is assumed to have the form @xmath85 where @xmath86 parametrizes the spatial distribution of supernova remnants normalized at @xmath87 , and @xmath88 is the injection spectral index for species @xmath49 . for protons and helium we modify the source term to accommodate for two breaks in the power - law , as strongly indicated by observations .
leptons lose energy very efficiently , thus those which are very energetic need to be very local , while we do not observe nor expect many local sources of tev scale leptons .
this motivates multiplying by an additional exponential cut - off in energy , @xmath89 , with @xmath90 set to 50 tev for electron and positron injection spectra .
we employ the gas distribution @xmath91 derived in @xcite and adopt the standard force - field approximation @xcite to describe the effect of solar modulation .
the modulation potential is assumed to be a free parameter of the fit and is allowed to be different for different cr species .
= 0.16 cm in @xcite 11 benchmark propagation models with varying diffusion zone thickness , from @xmath92 kpc to @xmath93 kpc , were identified by fitting to the b / c , proton , helium , electron and @xmath94 data . since
then the ams-02 experiment provided cr spectra with unprecedented precision , which necessitates modifications of the above benchmark models .
following the same procedure as in @xcite we choose three representative models , which give a reasonable fit to the ams-02 data , denoted thin , med and thick , corresponding to the previous @xmath92 kpc , @xmath95 kpc and @xmath96 kpc models .
the relevant parameters are given in table [ tab : propagation ] . in fig .
[ fig : crs ] we show the fit to the b / c and the ams-02 proton data @xcite and superimpose the older data from pamela @xcite . in all these cases , as well as for the lepton data @xcite ,
the measurements used in the fits were from ams-02 results only .
, title="fig : " ] , title="fig:",height=249 ] in the fit we additionally assumed that the normalization of the secondary cr antiprotons can freely vary by 10% with respect to the result given by the dragon code .
this is motivated by the uncertainty in the antiproton production cross sections .
the impact of this and other uncertainties has been studied in detail in e.g. @xcite . as we will show below
, the dm contribution to the lepton spectra is of much less importance for constraining the parameter space of our interest , therefore , we do not discuss the lepton backgrounds explicitly .
all the details of the implementation of the lepton limits closely follow @xcite , updated to the published ams-02 data @xcite . recently the _
fermi_-lat and magic collaborations released limits from the combination of their stacked analyses of 15 dwarf spheroidal galaxies @xcite . here
we use the results of this analysis to constrain the parameter space of the mixed wino - higgsino neutralino . to this end
we compute all exclusive annihilation cross sections for present - day dm annihilation in the halo and take a weighted average of the limits provided by the experimental collaborations . as discussed in section [ sec : model ] ,
the tev scale wino - like neutralino annihilates predominantly into @xmath55 , @xmath97 and @xmath98 , with much smaller rates into leptons and the lighter quarks . in the results from @xcite
only the @xmath55 , @xmath99 , @xmath100 and @xmath101 final states are given .
however , as the predicted spectrum and number of photons from a single annihilation is not significantly different for the hadronic or leptonic final states , we adopt the approximation that the limits from annihilation into @xmath97 are the same as from @xmath55 , while those from @xmath98 and @xmath102 are the same as @xmath99 .
the differences in the number of photons produced between these annihilation channels in the relevant energy range are maximally of order @xmath103 for @xmath55 vs. @xmath97 and @xmath98 vs. @xmath99 . comparing @xmath99 to light quarks these can rise up to factor 2 , however due to helicity suppression these channels have negligible branching fractions .
hence , the adopted approximation is expected to be very good and , the corresponding uncertainty is significantly smaller than that related to the astrophysical properties of the dsphs ( parametrised by the @xmath104-factors ) . the annihilation of dark matter at times around recombination can affect the recombination history of the universe by injecting energy into the pre - recombination photon - baryon plasma and into the post - recombination gas and background radiation , which has consequences for the power and polarization spectra of the cmb @xcite .
in particular , it can result in the attenuation of the temperature and polarization power spectra , more so on smaller scales , and in a shift of the te and ee peaks .
these effects can be traced back to the increased ionization fraction and baryon temperature , resulting in a broadening of the surface of last scattering , which suppresses perturbations on scales less than the width of this surface . therefore the cmb temperature and polarization angular power spectra can be used to infer upper bounds on the annihilation cross section of dark matter into a certain final state for a given mass .
when majorana dark matter particles annihilate , the rate at which energy @xmath105 is released per unit volume @xmath106 can be written as @xmath107 where @xmath108 is the critical density of the universe today , and experiment provides constraints on @xmath109 , which describes the effects of the dm .
these effects are found to be well enough accounted for when the @xmath74 dependence of @xmath110 is neglected , such that a limit is obtained for the constant @xmath111 .
the latest 95% c.l .
upper limit on @xmath111 was obtained by planck @xcite , and we adopt their most significant limit @xmath112 from the combination of tt , te , ee + lowp + lensing data .
the constant @xmath111 can further be expressed via @xmath113 where @xmath114 , parametrizing the fraction of the rest mass energy that is injected into the plasma or gas , must then be calculated in order to extract bounds on the dm annihilation cross section in the recombination era . in our analysis , for @xmath114
we use the quantities @xmath115 from @xcite for a given primary annihilation channel @xmath41 .
we then extract the upper limit on the annihilation cross section at the time of recombination by performing a weighted average over the contributing annihilation channels , as done for the indirect detection limits discussed in section [ sec : dsphs ] . as the sommerfeld effect saturates before this time , @xmath116 at recombination is the same as the present - day cross section . in the future
the cross section bound can be improved by almost an order of magnitude , until @xmath111 is ultimately limited by cosmic variance . direct detection experiments probe the interaction of the dark matter particle with nucleons . for the parameter space of interest here ,
the bounds on spin - independent interactions , sensitive to the t - channel exchange of the higgs bosons and to s - channel sfermion exchange are more constraining than those on spin - dependent interactions .
the coupling of the lightest neutralino to a higgs boson requires both a higgsino and gaugino component , and is therefore dependent on the mixing .
note that the relative size of the higgs yukawa couplings means that the contribution due to the higgs coupling to strange quarks dominates the result . in the pure - wino limit , when the sfermions are decoupled and the coupling to the higgs bosons vanishes , the direct detection constraints are very weak as the elastic scattering takes place only at the loop level @xcite .
allowing for a higgsino admixture and/or non - decoupled sfermions introduces tree - level scattering processes mediated by higgs or sfermion exchange .
direct detection experiments have recently reached the sensitivity needed to measure such low scattering cross sections and with the new data released by the lux @xcite and pandax @xcite collaborations , a portion of the discussed parameter space is now being probed . in the analysis below
we adopt the lux limits @xcite , being the strongest in the neutralino mass range we consider . in order to be conservative , in addition to the limit presented by the collaboration we consider a weaker limit by multiplying by a factor of two .
this factor two takes into account the two dominant uncertainties affecting the spin - independent cross section , i.e. the local relic density of dark matter and the strange quark content of the nucleon .
the former , @xmath117 , results in an uncertainty of @xmath118 @xcite and the latter result contributes an uncertainty on the cross section of about @xmath119 @xcite , which on combination result in weakening the bounds by a factor of two ( denoted as @xmath120 on the plots ) . for the computation of the spin - independent scattering cross section for every model point we use micromegas @xcite .
note that the sommerfeld effect does not influence this computation and the tree - level result is expected to be accurate enough . since only mixed higgsino - gaugino neutralinos couple to higgs bosons ,
the limits are sensitive to the parameters affecting the mixing . to be precise , for the case that the bino is decoupled ( @xmath121 ) and @xmath122 , the couplings of the higgs bosons @xmath123 to the lightest neutralino are proportional to @xmath124 where @xmath125 , and it is further assumed that @xmath26 is heavy such that @xmath126 can be computed in the decoupling limit @xmath127 .
when @xmath128 increases , the light higgs coupling @xmath129 decreases for @xmath0 and increases for @xmath1 . on the other hand the coupling @xmath130 increases in magnitude with @xmath128 for both @xmath0 and @xmath1 , but is positive when @xmath0 and negative for @xmath1 .
in addition , in the decoupling limit the coupling of the light higgs to down - type quarks is sm - like , and the heavy higgses couple to down - type quarks proportionally to @xmath128 .
the sfermion contribution is dominated by the gauge coupling of the wino - like component neutralino to the sfermion and the quarks .
we remark that for the parameter range under consideration there is destructive interference between the amplitude for the higgs and sfermion - exchange diagrams for @xmath0 , and for @xmath1 when @xcite @xmath131 provided @xmath132 and @xmath133 . for these cases
lower values of the sfermion masses reduce the scattering cross section . in fig .
[ fig : ddlimits ] we show the resulting limits from lux data in the @xmath12 vs. @xmath11 plane for different choices of @xmath134 , @xmath26 , @xmath15 , and the sign of @xmath13 .
the above discussion allows us to understand the following trends observed : * on decreasing @xmath134 and @xmath26 the direct detection bound becomes stronger for positive @xmath13 and weaker for negative @xmath13 .
note that for @xmath1 the cross section decreases , and the bound weakens , due to the destructive interference between the @xmath135 and @xmath136 contributions as the relative sign between the couplings @xmath129 and @xmath137 changes . *
the direct detection bound weakens for less decoupled sfermions when there is destructive interference between the t - channel higgs - exchange and s - channel sfermion - exchange diagrams .
this always occurs for @xmath0 , while for @xmath1 one requires small heavy higgs masses .
for instance , for @xmath138 the maximum value of @xmath26 giving destructive interference is slightly above 500 gev , while for @xmath139 one needs @xmath140 gev .
since we consider a point in the @xmath12 vs. @xmath11 plane to be excluded only if it is excluded for _ any _ ( allowed ) value of the other mssm parameters , this means that the bounds from direct detection experiments are weakest for @xmath1 in combination with low values of @xmath15 , @xmath26 and @xmath128 , and for @xmath0 in combination with high values of @xmath26 and @xmath128 but low values of @xmath15 . , while the right of @xmath1.[fig : ddlimits],title="fig : " ] , while the right of @xmath1.[fig : ddlimits],title="fig : " ]
in this section we first determine the region of the @xmath12 vs. @xmath11 plane which satisfies the relic density constraint and is allowed by the gamma - ray limits from dwarf spheroidals , the positron limits from ams-02 , and the cmb limits .
flux several earlier observations provide data extending to higher energies than the ams-02 experiment , though with much larger uncertainties .
we do not include these data in our analysis , because for the dm models under consideration , the strongest lepton limits arise from energies below about 100 gev , in particular the from observed positron fraction ( see fig . 7 of @xcite ) .
] we also determine the regions preferred by fits to ams-02 antiproton results . over a large part of the considered @xmath12 vs. @xmath11 plane
, the observed relic density can be obtained for some value of the sfermion masses and other mssm parameters . for the remaining region of the plane , where the relic density constraint is not fulfilled for thermally produced neutralino dark matter
, we consider both , the case where the dark matter density is that observed throughout the plane , in which case it can not be produced thermally , and the case where it is always thermally produced , for which the neutralino relic density does not always agree with that observed , and the limits must be rescaled for each point in the plane by the relic density calculated accordingly . that the neutralino dark matter is not thermally produced , or that it only constitutes a part of the total dark matter density are both viable possibilities .
we then consider various slices through this plane for fixed values of @xmath12 , and show the calculated present - day annihilation cross section as a function of @xmath141 together with the same limits and preferred regions as above , both for the case that the limits are and are not rescaled according to the thermal relic density .
in this section we present our results on the limits from indirect searches for wino - like dm in the mssm , assuming the relic density is as observed .
that is , for most parameter points the dm must be produced non - thermally or an additional mechanism for late entropy production is at play .
we show each of the considered indirect search channels separately in the @xmath12 vs. @xmath11 plane ( including both @xmath0 and @xmath1 ) , superimposing on this the contours of the correct relic density for three choices of the sfermion mass .
note that while the indirect detection limits are calculated for @xmath142 tev , the effect of the choice of sfermion mass on them is minimal , and therefore we display only the relic density contours for additional values of @xmath15 . vs. @xmath12 plane .
left : limits from dsphs ( upper ) and the cmb ( lower ) .
the shaded regions are excluded , different shadings correspond to the dm profile uncertainty .
right : the region excluded by ams-02 leptons ( upper ) , and the best fit contours for antiprotons ( lower ) , where the green solid lines show the thin and thick propagation models , while the dotted lines around them denote the 1@xmath143 confidence intervals .
contours where the observed relic density is obtained for the indicated value of the sfermion mass are overlaid .
[ fig : higgsino],title="fig : " ] vs. @xmath12 plane . left : limits from dsphs ( upper ) and the cmb ( lower ) .
the shaded regions are excluded , different shadings correspond to the dm profile uncertainty .
right : the region excluded by ams-02 leptons ( upper ) , and the best fit contours for antiprotons ( lower ) , where the green solid lines show the thin and thick propagation models , while the dotted lines around them denote the 1@xmath143 confidence intervals .
contours where the observed relic density is obtained for the indicated value of the sfermion mass are overlaid .
[ fig : higgsino],title="fig : " ] + vs. @xmath12 plane . left : limits from dsphs ( upper ) and the cmb ( lower ) .
the shaded regions are excluded , different shadings correspond to the dm profile uncertainty .
right : the region excluded by ams-02 leptons ( upper ) , and the best fit contours for antiprotons ( lower ) , where the green solid lines show the thin and thick propagation models , while the dotted lines around them denote the 1@xmath143 confidence intervals .
contours where the observed relic density is obtained for the indicated value of the sfermion mass are overlaid .
[ fig : higgsino],title="fig : " ] vs. @xmath12 plane . left : limits from dsphs ( upper ) and the cmb ( lower ) .
the shaded regions are excluded , different shadings correspond to the dm profile uncertainty .
right : the region excluded by ams-02 leptons ( upper ) , and the best fit contours for antiprotons ( lower ) , where the green solid lines show the thin and thick propagation models , while the dotted lines around them denote the 1@xmath143 confidence intervals .
contours where the observed relic density is obtained for the indicated value of the sfermion mass are overlaid .
[ fig : higgsino],title="fig : " ] in fig .
[ fig : higgsino ] we show the exclusions from dsphs , @xmath63 , and the cmb separately in the @xmath12 vs. @xmath11 plane . for the positrons we show two limits , obtained on assuming the thin and thick propagation models described in section [ sec : bckgr ] .
we see that the most relevant exclusions come from the diffuse gamma - ray searches from dsphs . here
we show three lines corresponding to the limit on the cross section assuming the navarro - frenk - white profile in dsphs , and rescaling this limit up and down by a factor 2 .
this is done in order to estimate the effect of the uncertainty in the @xmath104-factors .
for instance , the recent reassessment @xcite of the @xmath104-factor for ursa minor inferred from observational data suggests 2 to 4 times smaller limits than those commonly quoted . in order to provide conservative bounds
, we adopt the weakest of the three as the reference limit .
we then compare ( lower right plot ) this weakest limit from dsphs to the preferred region obtained on fitting to the ams-02 antiproton results , showing the results for both thin and thick propagation models .
we find that there are parts of the mixed wino - higgsino and dominantly wino neutralino parameter space both below and above the sommerfeld resonance region , where the relic density is as observed and which are compatible with the non - observation of dark matter signals in indirect searches . in the lower right plot of fig .
[ fig : higgsino ] we see that these further overlap with the regions preferred by fits to the antiproton results . in the smaller region above the resonance , this overlap occurs when the sfermions are decoupled , and hence corresponds to an almost pure - wino case , whereas below the resonance the overlap region is spanned by varying the sfermion masses from @xmath144 to being decoupled .
the latter region requires substantial higgsino - mixing of the wino , and extends from @xmath145 tev to about 2.5 tev , thus allowing dominantly - wino dark matter in a significant mass range .
let us comment on the improvement of the fit to the antiproton measurements found for some choices of the parameters . in fig .
[ fig : antip ] we show examples of antiproton - to - proton ratio fits to the data from the background models ( left ) and including the dm component ( right ) . although the propagation and antiproton production uncertainties can easily resolve the apparent discrepancy of the background models vs. the observed data @xcite , it is nevertheless interesting to observe that the spectral shape of the dm component matches the observed data for viable mixed - wino dark matter particles .
, title="fig : " ] , title="fig : " ] in this section we present our results for the limits from indirect searches on wino - like dm , assuming the relic density is always thermally produced . in other words , for the standard cosmological model , these constitute the limits on the parameter space of the mssm , since even if the neutralino does not account for all of the dark matter , its thermal population can give large enough signals to be seen in indirect searches . in this case
a parameter - space point is excluded , if @xmath146 where @xmath147 is the theoretically predicted present - day cross section and @xmath148 the limit quoted by the experiment .
this is because the results presented by the experiments assume the dm particle to account for the entire observed relic density .
therefore if one wishes to calculate the limits for dark matter candidates which only account for a fraction of the relic density , one needs to rescale the bounds by the square of the ratio of observed relic density @xmath149 to the thermal relic density @xmath150 .
viewed from another perspective , the results below constitute astrophysical limits on a part of the mssm parameter space , which is currently inaccessible to collider experiments , with the only assumption that there was no significant entropy production in the early universe after the dm freeze - out .
vs @xmath12 plane for the case where the limits are rescaled according to the thermal relic density for a given point in the plane .
details are as in fig .
[ fig : higgsino ] .
[ fig : higgsino_rescaled],title="fig : " ] vs @xmath12 plane for the case where the limits are rescaled according to the thermal relic density for a given point in the plane .
details are as in fig .
[ fig : higgsino ] .
[ fig : higgsino_rescaled],title="fig : " ] + vs @xmath12 plane for the case where the limits are rescaled according to the thermal relic density for a given point in the plane .
details are as in fig . [
fig : higgsino ] .
[ fig : higgsino_rescaled],title="fig : " ] vs @xmath12 plane for the case where the limits are rescaled according to the thermal relic density for a given point in the plane .
details are as in fig .
[ fig : higgsino ] .
[ fig : higgsino_rescaled],title="fig : " ] in fig .
[ fig : higgsino_rescaled ] , as in the previous subsection , we show the exclusions from dsphs , @xmath63 , and the cmb individually in the @xmath12 vs. @xmath11 plane .
the limits are calculated as for fig .
[ fig : higgsino ] .
we then compare the weakest limit from dsphs to the preferred region obtained on fitting to the ams-02 antiproton results , where we show the results for both thin and thick propagation models .
again we find that parameter regions exist where the relic density is correct and which are not excluded by indirect searches .
the marked difference between the previous and present results is that in fig .
[ fig : higgsino_rescaled ] the region of the plots for lower @xmath11 is not constrained by the indirect searches , because in this region the thermal relic density is well below the measured value and therefore the searches for relic neutralinos are much less sensitive . in the bottom lower plot of fig .
[ fig : higgsino_rescaled ] we see that the unconstrained regions overlap with the regions preferred by fits to the antiproton results .
while the limits themselves do not depend on the sfermion mass , the thermal relic density does , and therefore the rescaling of the limits via induces a dependence on the sfermion mass .
therefore the intersection of the lines of correct relic density for @xmath151 tev with the preferred region from antiproton searches is not meaningful , and we do not show them in the plots .
( black ) is shown as a function of @xmath141 for the higgsino admixture @xmath12 as indicated .
this is compared with exclusion limits from dsphs ( brown ) , positrons ( blue dashed ) and the cmb ( magenta dot - dashed ) , along with the preferred regions from antiproton searches ( pale green ) adopting the thin and thick models .
we also show the dsphs exclusion limits multiplied and divided by 2 ( brown ) , the weaker of which is the thicker line
. the observed relic density is assumed .
the blue shaded region indicates where the relic density can correspond to the observed value by changing @xmath15 .
[ fig : sectionplots],title="fig : " ] + ( black ) is shown as a function of @xmath141 for the higgsino admixture @xmath12 as indicated .
this is compared with exclusion limits from dsphs ( brown ) , positrons ( blue dashed ) and the cmb ( magenta dot - dashed ) , along with the preferred regions from antiproton searches ( pale green ) adopting the thin and thick models .
we also show the dsphs exclusion limits multiplied and divided by 2 ( brown ) , the weaker of which is the thicker line .
the observed relic density is assumed .
the blue shaded region indicates where the relic density can correspond to the observed value by changing @xmath15 .
[ fig : sectionplots],title="fig : " ] , but the thermal relic density is assumed and the limits are rescaled according to . note the different value of @xmath12 in the lower plot compared to the previous figure .
the black - dashed vertical line indicates where the relic density is equal to that observed for the sfermion mass value @xmath142 tev .
[ fig : sectionplots2],title="fig : " ] + , but the thermal relic density is assumed and the limits are rescaled according to . note the different value of @xmath12 in the lower plot compared to the previous figure .
the black - dashed vertical line indicates where the relic density is equal to that observed for the sfermion mass value @xmath142 tev .
[ fig : sectionplots2],title="fig : " ] in order to understand how the limits and the present - day annihilation cross section depend on the mass of the dm candidate , we take slices of the @xmath12 vs. @xmath11 plane for fixed values of @xmath12 , and plot @xmath152 ( black ) as a function of @xmath11 , which is approximately equal to the lsp mass @xmath153 in the range shown in figs . [ fig : sectionplots ] and [ fig : sectionplots2 ] . as in figs .
[ fig : higgsino ] and [ fig : higgsino_rescaled ] we show the limits from dsphs ( brown ) , positrons ( blue dashed ) and the cmb ( magenta dot - dashed ) , along with the preferred regions from antiproton searches ( pale green ) adopting the thin and thick propagation models .
we consider three choices of @xmath154 : a very mixed neutralino lsp , @xmath155 gev where @xmath13 is negative , a mixed case @xmath156 gev where @xmath13 is positive , and an almost pure - wino scenario , @xmath157 gev .
the blue shaded region indicates where the relic density can correspond to the observed value by changing @xmath15 . for fig .
[ fig : sectionplots ] we adopt the unrescaled limit , that is , two sections of fig .
[ fig : higgsino ] . in the case of the very mixed wino - higgsino shown in the upper panel
there is a wide range of neutralino masses for which the black curve lies below the conservative dsphs limit and simultaneously within the range of correct relic density spanned by the variation of the sfermion mass .
this is different for the almost pure - wino scenario shown in the lower panel , where only a small mass region survives the requirement that the conservative dsphs limit is respected and the observed relic density is predicted . moreover , in this mass region the sfermions must be almost decoupled .
[ fig : sectionplots2 ] shows two cases of mixed wino - higgsino dark matter , which exhibit similar features , but now for the case of assumed thermal relic density , such that the limits are rescaled .
it is evident from both figures that for lower values of @xmath12 , larger regions in @xmath11 can provide both the correct relic density and present - day cross section below the dsphs bounds .
we also see that while the correct relic density can be attained at the sommerfeld resonance , the mass regions compatible with indirect search constraints typically lie below the sommerfeld resonance , as was evident from figs .
[ fig : higgsino ] and [ fig : higgsino_rescaled ] .
we have seen in the previous section that there is a sizeable mixed wino - higgsino mssm parameter space where the lightest neutralino has the correct relic abundance and evades indirect detection constraints . a significant higgsino fraction might , however , be in conflict with the absence of a direct detection signal . in this section
we therefore combine the exclusion limits from indirect searches studied in the previous section with those coming from the latest lux results for direct detection , in order to determine the allowed mixed wino - higgsino or dominantly - wino dark matter parameter space . to this end
we first determine the maximal region in this space that passes relic density and indirect detection limits in the following way . for a given @xmath12
we identify two points in @xmath26 , @xmath15 and @xmath128 within the considered parameter ranges , i.e. @xmath158 , @xmath159 and @xmath160 , gev to 800 gev would result in a barely noticeable change to the boundaries marked by p2 . ] corresponding to maximal and minimal values of @xmath11 , for which the relic density matches the observed value .
two distinct areas of parameter space arise : the first is larger and corresponds to a mixed wino - higgsino whereas the second is narrower and corresponds approximately to the pure wino .
the relic density criterion therefore defines one ( almost pure wino ) or two ( mixed wino - higgsino ) sides of the two shaded regions , shown in figs .
[ fig : maxregion_positivemu ] and [ fig : maxregion_negativemu ] , corresponding to the pure and mixed wino .
the dsphs limit defines the other side in the almost pure - wino region , while the remaining sides of the mixed wino - higgsino area are determined by the dsphs limit ( upper ) , the condition @xmath161 , and the antiproton search ( the arc on the lower side of the mixed region beginning at @xmath162tev ) .
we recall that we consider the central dsphs limit and those obtained by rescaling up and down by a factor of two ; the shading in grey within each region is used to differentiate between these three choices .
next we consider the exclusion limits in the @xmath11 vs. @xmath12 plane from the 2016 lux results , which have been obtained as outlined in section [ sec : dd ] .
as discussed there , the sign of @xmath13 can strongly influence the strength of the direct detection limits and consequently the allowed parameter space for mixed wino - higgsino dm .
we therefore consider the two cases separately .
vs @xmath12 plane for @xmath0 where the relic density is as observed and the limit from dsphs diffuse gamma searches is respected within parameter ranges considered . the darker the grey region , the more stringent is the choice of the bound as described in the text .
the grey lines mark the _ weakest possible _ limit of the region excluded by the 2016 lux results and the same limit weakened by a factor of two as indicated .
the limit from the previous lux result is the dotted line .
the different bounds are calculated at different parameter sets p1 , p2 and p3 , as indicated . [ fig : maxregion_positivemu ] ] out of the two distinct regions described above , the close - to - pure wino and the mixed wino - higgsino , only the former survives after imposing the direct detection constraints , see fig . [
fig : maxregion_positivemu ] .
if conservative assumptions are adopted for direct detection _ and _ dsphs limits a small triangle at the top of the mixed region is still allowed .
the fact that the direct detection constraints mainly impact the mixed rather than the pure wino region was discussed in section [ sec : dd ] , and is understood in that the higgs bosons only couple to mixed gaugino - higgsino neutralinos .
note that the direct detection limits presented on the plot are for the choice of mssm parameters giving the weakest possible constraints .
this is possible because the boundaries of the maximal region allowed by indirect searches do not depend as strongly on the parameters governing the wino - higgsino mixing as the spin - independent scattering cross section does .
the only exceptions are the boundaries of the mixed - wino region , arising from the relic density constraint , which indeed depend strongly on @xmath15 . however , as varying these boundaries does not significantly change the allowed region , since it is mostly in the part excluded by the lux data , we choose to display the lux bound for a value of @xmath15 different from that defining these boundaries .
therefore , all in all , the case of the mixed wino - higgsino with @xmath0 is verging on being excluded by a combination of direct and indirect searches , when imposing that the lightest neutralino accounts for the entire thermally produced dark matter density of the universe .
note , however , that the small close - to - pure wino region is not affected by direct detection constraints . when @xmath1 the spin - independent cross section decreases , particularly for smaller values of @xmath128 .
this allows for parameter choices with small @xmath12 giving viable neutralino dm , in agreement with the direct detection constraint .
indeed , for appropriate parameter choices the direct detection limits are too weak to constrain any of the relevant regions of the studied parameter space . in particular , the weakest possible limits correspond to @xmath163 , @xmath164 tev and @xmath165 .
note that for @xmath166 tev a significantly lower value of @xmath128 would be in conflict with constraints from heavy higgs searches at the lhc .
vs @xmath154 plane for @xmath1 , obtained as in fig .
[ fig : maxregion_positivemu ] .
the limit from the 2016 lux result weakened by a factor of two is not visible within the ranges considered in the plot .
the different bounds are calculated at different parameter sets p1 , p2 and p3 , as indicated .
[ fig : maxregion_negativemu ] ] the result of varying @xmath26 , @xmath15 and @xmath128 is a sizeable mass region for viable mixed - wino dark matter in the mssm , ranging from @xmath167 to 3 tev , as shown in fig .
[ fig : maxregion_negativemu ] .
the parameter @xmath12 for the higgsino admixture varies from close to 0 gev to 210 gev below the sommerfeld resonance , and from 200 gev upwards above , when the most conservative dsphs limit ( shown in light grey ) is adopted .
we note that in determining the viable mixed - wino parameter region we did not include the diffuse gamma - ray and gamma line data from observations of the galactic center , since the more conservative assumption of a cored dark matter profile would not provide a further constraint .
however , future gamma data , in particular cta observations of the galactic center , are expected to increase the sensitivity to the parameter region in question to the extent ( _ cf . _
@xcite ) that either a dominantly - wino neutralino dark matter would be seen , or the entire plane shown in fig .
[ fig : maxregion_negativemu ] would be excluded even for a cored profile .
this study was motivated by the wish to delineate the allowed parameter ( in particular mass ) range for a wino - like dark matter particle in the mssm , only allowing some mixing with the higgsino .
more generically , this corresponds to the case where the dark matter particle is the lightest state of a heavy electroweak triplet with potentially significant doublet admixture and the presence of a scalar mediator .
the sommerfeld effect is always important in the tev mass range , where the observed relic density can be attained , and has been included in this study extending previous work @xcite .
our main results are summarized in figs .
[ fig : maxregion_positivemu ] and [ fig : maxregion_negativemu ] , which show the viable parameter region for the dominantly - wino neutralino for the cases @xmath0 and @xmath1 , respectively . after imposing the collider and flavour constraints ( both very weak ) , we considered the limits from diffuse gamma - rays from the dwarf spheroidal galaxies ( dsphs ) , galactic cosmic rays and cosmic microwave background anisotropies
. we also calculated the antiproton flux in order to compare with the ams-02 results .
the choice of indirect search constraints is influenced by the attitude that the fundamental question of the viability of wino - like dark matter should be answered by adopting conservative assumptions on astrophysical uncertainties .
the non - observation of an excess of diffuse gamma - rays from dsphs then provides the strongest limit .
it turns out that in addition to these indirect detection bounds , the direct detection results have a significant impact on the parameter space , particularly for the @xmath0 case where the mixed higgsino - wino region is almost ruled out as shown in fig .
[ fig : maxregion_positivemu ] . in the @xmath1 case
the limits are weaker as seen in fig .
[ fig : maxregion_negativemu ] , and a sizeable viable region remains .
note that the region of the @xmath12 vs. @xmath11 plane constrained by direct detection is complementary to that constrained by indirect detection .
therefore while for @xmath0 , ( almost ) the entire mixed region is ruled out , for @xmath1 there is a part of parameter space where @xmath168 tev which is in complete agreement with all current experimental constraints .
let us conclude by commenting on the limits from line and diffuse photon spectra from the galactic center .
if a cusped or mildly cored dm profile was assumed , the h.e.s.s
. observations of diffuse gamma emission @xcite would exclude nearly the entire parameter space considered in this paper , leaving only a very narrow region with close to maximal wino - higgsino mixing .
the limits from searches for a line - like feature @xcite would be even stronger , leaving no space for mixed - wino neutralino dm . however , a cored dm profile remains a possibility , and hence we did not include the h.e.s.s . results . in other words , adopting a less conservative approach , one would conclude that not only the pure - wino limit of the mssm , but also the entire parameter region of the dominantly - wino neutralino , even with very large higgsino or bino admixture , was in strong tension with the indirect searches .
therefore , the forthcoming observations by cta should either discover a signal of or definitively exclude the dominantly - wino neutralino .
we thank a. ibarra for comments on the manuscript , and a. goudelis and v. rentala for helpful discussions .
this work is supported in part by the gottfried wilhelm leibniz programme of the deutsche forschungsgemeinschaft ( dfg ) and the excellence cluster `` origin and structure of the universe '' at technische universitt mnchen .
ah is supported by the university of oslo through the strategic dark matter initiative ( sdi ) .
we further gratefully acknowledge that part of this work was performed using the facilities of the computational center for particle and astrophysics ( c2pap ) of the excellence cluster .
m. beneke , a. bharucha , f. dighera , c. hellmann , a. hryczuk , s. recksiegel and p. ruiz - femenia , _ relic density of wino - like dark matter in the mssm _ , _ jhep _ * 03 * ( 2016 ) 119 [ http://arxiv.org/abs/1601.04718 [ 1601.04718 ] ] .
a. hryczuk , i. cholis , r. iengo , m. tavakoli and p. ullio , _ indirect detection analysis : wino dark matter case study _ , _ jcap _ * 1407 * ( 2014 ) 031 [ http://arxiv.org/abs/1401.6212[1401.6212 ] ] .
t. cohen , m. lisanti , a. pierce and t. r. slatyer , _ wino dark matter under siege _ , _ jcap _ * 1310 * ( 2013 ) 061 [ http://arxiv.org/abs/1307.4082[1307.4082 ] ] . j. fan and m. reece , _ in wino veritas ?
indirect searches shed light on neutralino dark matter _ , _ jhep _ * 1310 * ( 2013 ) 124 [ http://arxiv.org/abs/1307.4400[1307.4400 ] ] .
l. roszkowski , e. m. sessolo and a. j. williams , _
prospects for dark matter searches in the pmssm _ , _ jhep _ * 02 * ( 2015 ) 014 [ http://arxiv.org/abs/1411.5214[1411.5214 ] ]
. v. lefranc , e. moulin , p. panci , f. sala and j. silk , _ dark matter in @xmath169 lines : galactic center vs dwarf galaxies _ , http://arxiv.org/abs/1608.00786[1608.00786 ] .
m. beneke , c. hellmann and p. ruiz - femenia , _ non - relativistic pair annihilation of nearly mass degenerate neutralinos and charginos i. general framework and s - wave annihilation _ , _ jhep _ * 1303 * ( 2013 ) 148 [ http://arxiv.org/abs/1210.7928[1210.7928 ] ] . c. hellmann and p. ruiz - femenia , _ non - relativistic pair annihilation of nearly mass degenerate neutralinos and charginos ii .
p - wave and next - to - next - to - leading order s - wave coefficients _ , _ jhep _ * 1308 * ( 2013 ) 084 [ http://arxiv.org/abs/1303.0200[1303.0200 ] ] .
m. beneke , c. hellmann and p. ruiz - femenia , _ non - relativistic pair annihilation of nearly mass degenerate neutralinos and charginos iii .
computation of the sommerfeld enhancements _ , _ jhep _ * 05 * ( 2015 ) 115 [ http://arxiv.org/abs/1411.6924[1411.6924 ] ] .
m. beneke , c. hellmann and p. ruiz - femenia , _ heavy neutralino relic abundance with sommerfeld enhancements - a study of pmssm scenarios _ , _ jhep _ * 1503 * ( 2015 ) 162 [ http://arxiv.org/abs/1411.6930 [ 1411.6930 ] ] .
collaboration , m. aguilar _ et .
_ , _ antiproton flux , antiproton - to - proton flux ratio , and properties of elementary particle fluxes in primary cosmic rays measured with the alpha magnetic spectrometer on the international space station _ , _ phys .
. lett . _ * 117 * ( 2016 ) , no . 9 091103 .
m. ibe , s. matsumoto , s. shirai and t. t. yanagida , _ wino dark matter in light of the ams-02 2015 data _ , _ phys .
* d91 * ( 2015 ) , no . 11 111701 [ http://arxiv.org/abs/1504.05554[1504.05554 ] ] .
collaboration , d. s. akerib _ et .
_ , _ first results from the lux dark matter experiment at the sanford underground research facility _ , _ phys .
* 112 * ( 2014 ) 091303 [ http://arxiv.org/abs/1310.8214[1310.8214 ] ] .
d. s. akerib _ et .
_ , _ results from a search for dark matter in lux with 332 live days of exposure _ , http://arxiv.org/abs/1608.07648[1608.07648 ] .
m. cirelli , g. corcella , a. hektor , g. hutsi , m. kadastik , p. panci , m. raidal , f. sala and a. strumia , _ pppc 4 dm i d : a poor particle physicist cookbook for dark matter indirect detection _ , _ jcap _ * 1103 * ( 2011 ) 051 [ http://arxiv.org/abs/1012.4515[1012.4515 ] ] .
[ erratum : _ jcap _ * 1210 * , e01 ( 2012 ) ] .
p. ciafaloni , d. comelli , a. riotto , f. sala , a. strumia and a. urbano , _ weak corrections are relevant for dark matter indirect detection _ , _ jcap _ * 1103 * ( 2011 ) 019 [ http://arxiv.org/abs/1009.0224 [ 1009.0224 ] ]
. v. s. berezinsky , s. v. bulanov , v. a. dogiel and v. s. ptuskin , _ astrophysics of cosmic rays _ , north - holland , amsterdam 1990 , edited by v. l. ginzburg . c. evoli , d. gaggero , d. grasso and l. maccione , _ cosmic - ray nuclei , antiprotons and gamma - rays in the galaxy : a new diffusion model _ , _ jcap _ * 0810 * ( 2008 ) 018 [ http://arxiv.org/abs/0807.4730 [ 0807.4730 ] ] .
m. tavakoli , _ three dimensional distribution of atomic hydrogen in the milky way _ , http://arxiv.org/abs/1207.6150[1207.6150 ] .
m. pohl , p. englmaier and n. bissantz , _
3d distribution of molecular gas in the barred milky way _ , _ astrophys .
j. _ * 677 * ( 2008 ) 283291 [ http://arxiv.org/abs/0712.4264[0712.4264 ] ] .
l. j. gleeson and w. i. axford , _ solar modulation of galactic cosmic rays _
, _ astrophys .
j. _ * 154 * ( 1968 ) 1011 .
f. donato , n. fornengo , d. maurin and p. salati , _ antiprotons in cosmic rays from neutralino annihilation _ , _ phys .
* d69 * ( 2004 ) 063501 [ http://arxiv.org/abs/astro-ph/0306207 [ astro - ph/0306207 ] ] .
a. oliva , _ ams results on light nuclei : measurement of the cosmic rays boron - to - carbon ration with ams-02 _ , presentation at _ `` ams days at cern the future of cosmic ray physics and latest results '' _ , cern , geneva , switzerland , 15 - 17 april 2015 .
collaboration , m. aguilar _ et .
al . _ , _ precision measurement of the proton flux in primary cosmic rays from rigidity 1 gv to 1.8 tv with the alpha magnetic spectrometer on the international space station _ , _ phys .
* 114 * ( 2015 ) 171103 .
collaboration , o. adriani _ et .
_ , _ pamela measurements of cosmic - ray proton and helium spectra _ , _ science _ * 332 * ( 2011 ) 6972 [ http://arxiv.org/abs/1103.4055[1103.4055 ] ]
. o. adriani _ et .
_ , _ measurement of boron and carbon fluxes in cosmic rays with the pamela experiment _ , _ astrophys .
j. _ * 791 * ( 2014 ) , no . 2 93 [ http://arxiv.org/abs/1407.1657[1407.1657 ] ] .
collaboration , l. accardo _ et .
_ , _ high statistics measurement of the positron fraction in primary cosmic rays of 0.5500 gev with the alpha magnetic spectrometer on the international space station _ , _ phys .
* 113 * ( 2014 ) 121101 .
collaboration , m. aguilar _ et .
_ , _ precision measurement of the ( @xmath170 ) flux in primary cosmic rays from 0.5 gev to 1 tev with the alpha magnetic spectrometer on the international space station _ , _ phys
* 113 * ( 2014 ) 221102 .
j. j. engelmann , p. ferrando , a. soutoul , p. goret and e. juliusson , _ charge composition and energy spectra of cosmic - ray for elements from be to ni - results from heao-3-c2 _ , _ astron .
astrophys . _
* 233 * ( 1990 ) 96111 .
h. s. ahn _ et .
_ , _ measurements of cosmic - ray secondary nuclei at high energies with the first flight of the cream balloon - borne experiment _ , _ astropart .
_ * 30 * ( 2008 ) 133141 [ http://arxiv.org/abs/0808.1718[0808.1718 ] ] .
s. p. swordy , d. mueller , p. meyer , j. lheureux and j. m. grunsfeld , _ relative abundances of secondary and primary cosmic rays at high energies _ , _ the astrophysical journal _ * 349 * ( 1990 ) 625633 .
r. kappl , a. reinert and m. w. winkler , _
ams-02 antiprotons reloaded _ , _ jcap _ * 1510 * ( 2015 ) , no . 10 034 [ http://arxiv.org/abs/1506.04145[1506.04145 ] ] . c. evoli , d. gaggero and d. grasso , _ secondary antiprotons as a galactic dark matter probe _ , _ jcap _ * 1512 * ( 2015 ) , no . 12 039 [ http://arxiv.org/abs/1504.05175[1504.05175 ] ] .
g. giesen , m. boudaud , y. gnolini , v. poulin , m. cirelli , p. salati and p. d. serpico , _
ams-02 antiprotons , at last !
secondary astrophysical component and immediate implications for dark matter _ , _ jcap _ * 1509 * ( 2015 ) , no . 09 023 [ http://arxiv.org/abs/1504.04276 [ 1504.04276 ] ] .
collaboration , m. l. ahnen _ et .
_ , _ limits to dark matter annihilation cross - section from a combined analysis of magic and fermi - lat observations of dwarf satellite galaxies _ , _ jcap _ * 1602 * ( 2016 ) , no . 02 039 [ http://arxiv.org/abs/1601.06590 [ 1601.06590 ] ]
. n. padmanabhan and d. p. finkbeiner , _ detecting dark matter annihilation with cmb polarization : signatures and experimental prospects _ , _ phys
* d72 * ( 2005 ) 023508 [ http://arxiv.org/abs/astro-ph/0503486[astro-ph/0503486 ] ] .
s. galli , f. iocco , g. bertone and a. melchiorri , _ cmb constraints on dark matter models with large annihilation cross - section _ , _ phys .
_ * d80 * ( 2009 ) 023505 [ http://arxiv.org/abs/0905.0003[0905.0003 ] ] .
t. r. slatyer , n. padmanabhan and d. p. finkbeiner , _ cmb constraints on wimp annihilation : energy absorption during the recombination epoch _ , _ phys .
_ * d80 * ( 2009 ) 043526 [ http://arxiv.org/abs/0906.1197[0906.1197 ] ] .
collaboration , p. a. r. ade _ et .
_ , _ planck 2015 results .
cosmological parameters _ , _ astron .
astrophys . _
* 594 * ( 2016 ) a13 [ http://arxiv.org/abs/1502.01589[1502.01589 ] ] .
m. s. madhavacheril , n. sehgal and t. r. slatyer , _ current dark matter annihilation constraints from cmb and low - redshift data _ , _ phys .
_ * d89 * ( 2014 ) 103508 [ http://arxiv.org/abs/1310.3815 [ 1310.3815 ] ] .
j. hisano , k. ishiwata , n. nagata and t. takesako , _ direct detection of electroweak - interacting dark matter _ , _ jhep _ * 07 * ( 2011 ) 005 [ http://arxiv.org/abs/1104.0228[1104.0228 ] ] .
collaboration , a. tan _ et .
al . _ , _ dark matter results from first 98.7-day data of pandax - ii experiment _ , _ phys .
* 117 * ( 2016 ) , no . 12 121303 [ http://arxiv.org/abs/1607.07400 [ 1607.07400 ] ] .
j. bovy and s. tremaine , _ on the local dark matter density _ , _ astrophys
. j. _ * 756 * ( 2012 ) 89 [ http://arxiv.org/abs/1205.4033[1205.4033 ] ] .
s. drr _ et .
_ , _ lattice computation of the nucleon scalar quark contents at the physical point _ , _ phys .
* 116 * ( 2016 ) , no . 17 172001 [ http://arxiv.org/abs/1510.08013[1510.08013 ] ] .
g. belanger , f. boudjema , p. brun , a. pukhov , s. rosier - lees _ et .
_ , _ indirect search for dark matter with micromegas2.4 _ , _ comput.phys.commun .
_ * 182 * ( 2011 ) 842856 [ http://arxiv.org/abs/1004.1092[1004.1092 ] ] .
g. belanger , f. boudjema , a. pukhov and a. semenov , _ micromegas_3 : a program for calculating dark matter observables _ , _
comput.phys.commun . _
* 185 * ( 2014 ) 960985 [ http://arxiv.org/abs/1305.0237 [ 1305.0237 ] ] .
a. crivellin , m. hoferichter , m. procura and l. c. tunstall , _ light stops , blind spots , and isospin violation in the mssm _ , _ jhep _ * 07 * ( 2015 ) 129 [ 1503.03478 ] .
ullio and m. valli , _ a critical reassessment of particle dark matter limits from dwarf satellites _ , _ jcap _ * 1607 * ( 2016 ) , no . 07 025 [ http://arxiv.org/abs/1603.07721[1603.07721 ] ] .
a. kounine , _ latest results from the alpha magnetic spectrometer : positron fraction and antiproton / proton ratio _ , presentation at _ `` ams days at cern the future of cosmic ray physics and latest results '' _ , cern , geneva , switzerland , 15 - 17 april 2015 .
o. adriani _ et .
_ , _ measurement of the flux of primary cosmic ray antiprotons with energies of 60-mev to 350-gev in the pamela experiment _ , _ jetp lett .
_ * 96 * ( 2013 ) 621627 .
[ pisma zh .
fiz . * 96 * , 693 ( 2012 ) ] .
collaboration , h. abdallah _ et .
_ , _ search for dark matter annihilations towards the inner galactic halo from 10 years of observations with h.e.s.s _ , _ phys .
* 117 * ( 2016 ) , no . 11 111301 [ http://arxiv.org/abs/1607.08142[1607.08142 ] ] .
collaboration , a. abramowski _ et .
_ , _ search for photon - linelike signatures from dark matter annihilations with h.e.s.s .
_ , _ phys .
* 110 * ( 2013 ) 041301 [ http://arxiv.org/abs/1301.1173[1301.1173 ] ] .
|
0.2cmwe delineate the allowed parameter and mass range for a wino - like dark matter particle containing some higgsino admixture in the mssm by analysing the constraints from diffuse gamma - rays from the dwarf spheroidal galaxies , galactic cosmic rays , direct detection and cosmic microwave background anisotropies .
a complete calculation of the sommerfeld effect for the mixed - neutralino case is performed .
we find that the combination of direct and indirect searches poses significant restrictions on the thermally produced wino - higgsino dark matter with correct relic density .
for @xmath0 nearly the entire parameter space considered is excluded , while for @xmath1 a substantial region is still allowed , provided conservative assumptions on astrophysical uncertainties are adopted .
= 1 tum - hep-1065/16 + ftuam-16 - 38 + ift - uam / csic-16 - 106 + november 2 , 2016 * the last refuge of mixed wino - higgsino dark matter + * m. beneke@xmath2 , a. bharucha@xmath3 , a. hryczuk@xmath4 , + s. recksiegel@xmath2 , and p. ruiz - femena@xmath5 + _ @xmath6physik department t31 , + james - franck - strae 1 , technische universitt mnchen , + d85748 garching , germany + @xmath7 aix marseille univ , universit de toulon , cnrs , cpt , marseille , france + @xmath8department of physics , university of oslo , box 1048 , +
no-0371 oslo , norway + @xmath9national centre for nuclear research , + hoa 69 , 00 - 681 , warsaw , poland + @xmath10departamento de fsica terica and instituto de fsica terica uam - csic , + universidad autnoma de madrid , e-28049 madrid , spain _
|
it is well known that in operational scenarios requiring long - term storage , sources of noise in the machine can lead to detrimental effects on the beam .
examples include rf voltage noise , beam - beam interaction , ground motion , and feedback noise , which all have been subject of studies to mitigate or exploit their impact , as reported in refs .
@xcite recently , a similar concern has raised for numerical noise in self - consistent simulation of high intensity beams .
in fact , the study of space charge effects via self - consistent simulations has become important with the advent of new projects , like the future sis100 synchrotron of the fair project @xcite , and the liu project @xcite for the cern accelerator complex . in these projects
some scenarios require the storage of a high intensity bunched beam for seconds .
the simulation of these operational scenarios significantly raises the computational challenges , provided that they concern the prediction of beam loss , or emittance growth . in order to avoid any artificial noise ,
the first simulation studies were limited to frozen models , which rely on noise - free tracking schemes @xcite . there , the space charge is computed assuming the beam remains frozen , therefore allowing an analytic description of the space charge force .
this approach surely avoids artificial emittance growth , but unfortunately at the expenses of self - consistency , which becomes relevant in certain scenarios e.g. large beam loss , or large beam core growth . on the other hand ,
pic simulations of high intensity beams allow a self - consistent treatment , but require an understanding of the origins and propagation of its numerical noise .
hence , simulation parameters can be set adequately in order to mitigate noise induced artifacts , while keeping the computational load feasible .
the evaluation of pic induced artifacts on particle tracking has been studied by several authors from several points of view , e.g. in the context of a fokker - planck approach , as reported in refs .
@xcite .
differently from previous approaches , we discuss the effect of pic induced noise by following the integration method of a typical beam dynamics code .
we start with an analysis of the noise due to the pic algorithm , and discuss how it propagates via the beam dynamics integration . following this approach ,
the dependence of rms - emittance growth on simulation parameters is derived .
it is also found that the particle s rotation in phase space creates correlations in numerical noise , which enhances artificial emittance growth .
our results are of interest for simulations of millions of turns in circular machines , where the high intensity requires a self - consistent modeling , hence the use of pic algorithms , while at the same time the control of artificial emittance growth .
this paper is organized as follows . in section [ sec : scaling_law ]
we study the origins and properties of static pic noise .
then , in sec . [ ch : singleparticle ] , we model the effect of random pic noise on the dynamics of a single particle , while the effect on the whole ensemble of particles is treated in sec . [ ch : distribution ] . in section [ ch : correlations ] we introduce the concept of correlated and decorrelated numerical noise .
we conclude our studies with a summary in sec .
[ ch : conclusion ] .
the effect of the particle s rotation in phase space on numerical noise is discussed in appendix [ ch : appendix ] , whereas the consequences of the periodic random walk on the diffusion of macro - particles is treated in appendix [ ch : appendix_rw ] .
finally , in appendix [ ch : appendix_sr_tune ] , we discuss the impact of the strength of space charge forces on the excitation of stochastic resonances .
before treating the propagation of numerical noise , we study how the pic scheme is causing numerical noise in a static scenario .
we start by considering a gaussian beam , i.e. the particle s phase space positions are defined randomly according to a gaussian probability density function ( p.d.f . ) , given by @xmath0 where @xmath1 are the transverse standard deviations . in the following discussion ,
we consider round gaussian beams with @xmath2 the standard deviation for the radial coordinate @xmath3 .
the associated radial space charge field at a longitudinal position @xmath4 is given by @xmath5 where @xmath6 is proportional to the gradient of the electric field at @xmath7 .
the transverse electric field of a coasting beam , that determines the self - consistent space charge forces , is computed in long - term simulations most efficiently using a two - dimensional particle - in - cell ( pic ) scheme . to achieve a high computational efficiency , the poisson equation in the transverse plane
is solved on a finite set of @xmath8 grid points and the particle distribution is approximated by @xmath9 macro - particles . since the number of macro - particles @xmath9 is much smaller than the number of physical particles @xmath10 , an artificial granularity of the distribution is introduced , which is one source of numerical noise . to study the impact of the simulation parameters on numerical noise
, we adopt the random start technique ( opposite to the quiet start ) . in the random start approach the positions of @xmath9 macro - particles are randomly initialized according to a p.d.f , as given by eq .
[ eq : gaussdistribution ] for gaussian beams .
therefore , the resulting electric field is not only determined by the p.d.f . itself , but varies according to the random initialization of macro - particles with an amplitude @xmath11 .
this amplitude of fluctuations @xmath11 is identified with the standard deviation of the electric field for multiple random initializations of macro - particles .
previous studies , e.g. refs .
@xcite , have found that the parameters of relevance for a noise analysis of a beam with a fixed size are @xmath12 and @xmath9 , since they fix the number of particles per cell . to find the scaling of @xmath11 on these parameters , we fix one of them and
vary systematically the remaining one .
we find that the simulation data satisfies the scaling @xmath13 where the dependence on the grid resolution , given by the function @xmath14 , is determined by the specific integration method of poisson s equation .
for the multi - particle tracking library micromap @xcite , that incorporates a spectral method , we find that @xmath15{n_g}$ ] for well resolved beams , which is used throughout the article .
in particular , we find that the level of numerical noise can be kept constant , if we choose our simulation parameters such that @xmath16 a dependence of the electric field fluctuation on @xmath17 and @xmath18 is created for non - constant particle distributions , as the number of macro - particles per cell varies spatially . for round gaussian beams this is given by the square root of the p.d.f .
, as we find @xmath19{n_g}}{\sqrt{n_m } } \
e^{-\frac{x^2+y^2}{2\sigma_r^2(s_j ) } } \xi(x , y , s ) \ .\end{aligned}\ ] ] here , @xmath20 is a factor to quantify the standard deviation of the electric field .
it is obtained by calculating the standard deviation in the center of the beam , @xmath21 , for multiple random start initializations for @xmath22 macro - particles and @xmath23 grid points .
we normalize @xmath21 to @xmath24 and @xmath22 , such that @xmath25 is independent of these simulation parameters .
the factor @xmath26 describes the effect due to the bi - linear interpolation of the electric field in between grid points , which causes a grid texture . as an example , the standard deviation for @xmath27 and @xmath28 and a round gaussian beam is given in fig .
[ fig : standard_deviation ] .
the statistics is done with 500 random start initializations .
( 150,190 ) ( -110,-20 ) , for a round gaussian beam .
the exponential spatial dependence of the field fluctuations is superimposed by a grid texture caused by the mesh of the pic solver .
the simulation parameters are given in the text.,title="fig : " ] the scaling law , eq .
[ eq : scaling_law ] , is valid as long as the beam is reasonably resolved , which is the case for @xmath29 grid points within @xmath30 $ ] ( and resp . in @xmath31
$ ] ) of the gaussian particle distribution .
the effects of a low mesh resolution were recently studied in ref .
@xcite , where it is shown that due to grid heating the emittance growth is artificially enhanced . in the studies presented in this paper ,
we consider sufficiently resolved particle distributions , as this is the preferential case in space charge simulations . using the scheme presented in this section ,
we can easily obtain scaling laws for other particle distributions .
in the following , we study the kapchinski - vladimirski ( k - v ) particle distribution @xcite , i.e. a spatially constant distribution of particles in the transverse plane , causing linear space charge forces inside the beam . for this distribution
we find that the electric field fluctuations @xmath32 and @xmath33 inside the beam are constant , since the uncertainty of a pic solver depends on the number of macro - particles per cell .
therefore , the electric field fluctuations inside a k - v beam are given by @xmath34{n_g}}{\sqrt{n_m } } \
\xi(x , y , s ) .\end{aligned}\ ] ] in figure [ fig : fluctuations_kv ] , we show the spatial dependence of the electric field fluctuation @xmath32 for a k - v distribution of size @xmath35 and a rms equivalent gaussian beam @xcite of rms size @xmath36 . for this study we took 1000 samples of electric fields ( random start ) , while the beam was well resolved with a fine grained mesh of @xmath37 grid points . for a k - v distribution of size @xmath35 and a rms equivalent gaussian beam .
inside the beam , i.e @xmath38 $ ] , the field fluctuations for the k - v distributions are approximately constant . ] as seen in fig .
[ fig : fluctuations_kv ] , the maximum electric field fluctuations for the gaussian beam are by a factor @xmath39 larger than for a rms equivalent k - v beam .
this is due to the fact , that in the center of the gaussian beam the density is the double with respect to a rms equivalent k - v beam .
the number of physical particles in the center is therefore by a factor @xmath40 larger for gaussian beams , which enhances the electric field fluctuations .
the results obtained in this section apply for the numerical noise in coasting beam simulations .
however , the same method can also be used to find the dependence of the fluctuations @xmath41 for a bunched beam in the 2.5d scheme , see ref .
in this section , we study the effect of a fluctuating force field on the dynamics of a single particle by use of the random walk theory . later
, these findings can be applied on a large ensemble of particles , such that conclusions on artificial diffusion of a full beam can be drawn .
a particle @xmath42 at the longitudinal position @xmath43 is described by its transverse phase space coordinates @xmath44 .
its dynamics in a linear lattice is given by the hill equations @xmath45 with @xmath46 , @xmath47 the restoring force terms .
in the following , only the @xmath17-plane is discussed , since the @xmath18-plane can be treated equivalently .
a particle @xmath42 with a charge @xmath48 and mass @xmath49 of a high intensity beam is exposed to a force created by the electric field @xmath50 of the beam .
the electric field is calculated in the rest frame of the beam , which is the standard approach in beam dynamics with space charge @xcite , and is used in this form throughout all discussions in this paper .
the effect of the electric self - field on particle dynamics is modeled via space charge kicks @xcite , whose effect is to change the particle coordinates @xmath51 : @xmath52 with @xmath53 the speed of light , and @xmath54 and @xmath55 the relativistic factors . for simplicity , the longitudinal distance between two consecutive space charge kicks , given by @xmath56 , is kept constant during the tracking . by knowing position and charge state of all particles , the resulting space charge field on a particle located at @xmath57 , given by @xmath58 ,
is calculated as a superposition of all coulomb fields .
we refer to this as the exact solution . however , if approximations are used in order to reduce the computational cost of the space charge field calculation , imprecisions are generated .
see e.g. refs .
@xcite for a review . in the following
, the effect of electric field fluctuations is studied in a simplified mathematical model , where the standard deviation is subtracted or added randomly to the exact solution , i.e. @xmath59 here , @xmath60 or @xmath61 is defined randomly with equal probability for any particle index @xmath42 and any longitudinal position index @xmath62 .
this model yields an average of @xmath58 and a standard deviation of @xmath63 for a large number of electric field calculations .
the fluctuations @xmath64 are assumed to be statistically independent of each other , i.e. for every longitudinal position @xmath4 , where a space charge kick is applied , the fluctuation @xmath64 does not depend on the fluctuation at any other longitudinal position @xmath65 for any @xmath66 . throughout this article
, we refer to this kind of fluctuations as _ decorrelated numerical noise_. equation [ eq : hill ] , [ eq : sck ] , and [ eq : e_fluctuations ] are the model for the evolution of a single particle affected by numerical noise in a high intensity beam using the standard integration with the one kick approximation .
we propose that the combined effect of numerical noise and transport can be modeled via an effective map : @xmath67}{m_0c^2\beta^2\gamma^3 } \end{array } \right ) = \nonumber \\ & = \left ( \begin{array}{c } \hat x_{i}(s_{j+1 } ) \\
\hat x^\prime_{i}(s_{j+1 } ) \end{array } \right ) + m \left ( \begin{array}{c}0 \\ \tilde{z}_{ij } \delta x^\prime(x_{i},y_{i},s_j ) \end{array } \right ) , \label{eq : rw_map}\end{aligned}\ ] ] with @xmath68 the strength of the random kick , that is defined by @xmath69 the left hand side of eq .
[ eq : rw_map ] is the usual one step integration , where @xmath70 is the transport matrix of a particle between two consecutive space charge kicks obtained from eq .
[ eq : hill ] .
the coordinates @xmath71 are those of the particle transported noise - free with space charge from @xmath4 to @xmath72 .
the repeated application of eq .
[ eq : rw_mixing ] mixes the noise @xmath73 to both planes @xmath17 and @xmath74 .
later , we model this effect by introducing an integrated effective noise kick . to model the effect of decorrelated numerical noise on @xmath9 macro - particles at @xmath75 space charge kick positions , we define two sets of independent variables @xmath76 with @xmath77 , @xmath78 , in which @xmath79 or @xmath80 ( resp .
@xmath81 or @xmath82 ) defined randomly with equal probability for any index @xmath42 , @xmath62 . as suggested in appendix [ ch : appendix ] , we describe the integrated effect of numerical noise on the phase space coordinates by using an effective model : @xmath83 the additional random elements account for the independence of both planes @xmath17 and @xmath74 , while the normalization factor @xmath84 stems directly from the mathematical model presented in appendix [ ch : appendix ] .
the effect of equation [ eq : kick ] is to create a random walk @xcite : at each longitudinal position @xmath4 , where a space charge kick is applied on single particles , a random kick is added to both phase space coordinates .
we test this ansatz in the following numerical experiment .
a test - particle is initialized at @xmath85 and @xmath86 , while @xmath87 macro - particles are distributed according to a matched gaussian probability density function .
the beam is tracked for 1000 turns in a constant focusing channel of length @xmath88 with a tune set to @xmath89 , while the space charge induced tune shift is set to @xmath90 .
the two - dimensional poisson solver uses a mesh of @xmath91 grid points .
the integration applies one space charge kick per turn , such that @xmath92 .
the tracking is repeated @xmath93 times for the same physical conditions , but with different random initializations of the particle distribution .
due to numerical noise the test particle is found at different final positions . in figure
[ fig : random_walk ] , we plot the distribution of final positions of the test particle after 1000 turns .
the distribution is normalized to the number of particles found in the center of the beam , which is the maximum of the distribution .
+ ( 150,190 ) ( -110,-20 ) , @xmath86 after tracking in a constant focusing channel for @xmath93 different random initializations of macro - particles .
the distribution is normalized to the number of particles found in the center.,title="fig : " ] the results presented in fig . [
fig : random_walk ] show that numerical noise affects both phase space coordinates @xmath17 and @xmath94 in equal measure for a sufficiently large number of space charge kicks . in order to use the random walk model
, we have to specify the spacial dependence of the fluctuations @xmath63 . as a simplified approach we consider @xmath95 with @xmath96 the electric field fluctuation in the center of the beam , which are , only in this first approach , assumed to be identical for any longitudinal position @xmath4
. equation [ eq : fluct_simple ] can be considered as a simplified model for the electric field fluctuations of a gaussian beam , as derived previously , see eq .
[ eq : scaling_law ] .
we now apply the random walk model , as given by eq .
[ eq : kick ] , on a single particle @xmath42 , that is initialized at the phase space position @xmath97 and tracked by using eq .
[ eq : hill ] and [ eq : sck ] . following the standard theory of random walks , the most probable position of the particle after a series of @xmath98 random kicks ,
is given by the position for a noiseless tracking .
figure [ fig : random_walk ] confirms this , as the maximum of the distribution of final positions is in the center .
after @xmath98 random kicks , the variances of the particle s phase space positions @xmath99 scale as @xmath100 each particle can be associated with a single particle emittance @xmath101 with twiss parameters @xmath102 . here
, we discuss the effect of decorrelated numerical noise on the initial single particle emittance @xmath103 in the context of the expectation value of @xmath104 , i.e. the single particle emittance after @xmath98 random kicks . assuming that numerical noise is the dominant cause of emittance growth , we write the expectation value of @xmath104 as a variation of the initial single particle emittance : @xmath105 using equations [ eq : rw_second_order ] and [ eq : spe ]
, we find the scaling @xmath106 thus , in average , the single particle emittance grows over many realizations of space charge kicks due to decorrelated random fluctuations in the space charge field .
in this section , we apply the random walk model on each particle of the beam .
we then find a scaling on simulation parameters for noise induced diffusion . as an application
, we predict the artificial emittance growth for a self - consistent space charge simulation for the sis100 synchrotron at fair @xcite .
the standard deviation of electric field fluctuations @xmath41 is given by eq .
[ eq : scaling_law ] and respectively eq . [ eq : scaling_law_kv ] for any particle at position @xmath17 , @xmath18 , @xmath43 .
since this result applies to any particle , it enables us to study the effect of decorrelated numerical noise on the full ensemble of particles , and in particular on the rms - emittance . in the following , we make use of the standard definition of the rms - emittance in the @xmath17-plane @xmath107 , \end{aligned}\ ] ] while the discussion for the @xmath18-plane is similar . the symbol @xmath108 denotes the average of the particle distribution .
let @xmath109 be the beam emittance resulting from the application of a space charge kick at @xmath4 and transported by @xmath110 to @xmath72 . by using the effective transport map eq .
[ eq : rw_map ] , and with the noise modeling of eq .
[ eq : kick ] , we predict @xmath109 from the beam distribution at @xmath111 . in fact
, from eq .
[ eq : rms_emittance ] it follows @xmath112 ^ 2 , \end{aligned}\ ] ] where the coordinates @xmath113 are those of the particle transported from @xmath4 to @xmath72 in presence of space charge , but without noise . in the following
, we derive the artificial emittance growth averaged over many applications of space charge kicks . for this
we find , that all terms proportional to @xmath114 , or @xmath115 have a minor contribution to the emittance growth , since they are in average zero with a variance proportional to the number of space charge kicks applied .
the same argument holds true for the product of @xmath114 and @xmath116 , since they are statistically independent .
the terms proportional to @xmath117 , and @xmath118 do not average to zero as @xmath119 , and thus they dominate the contribution to the emittance growth . all terms proportional to @xmath120 are much smaller than any other term in eq .
[ eq : rms_emittance_rw ] , and can thus be neglected . finally ,
as we consider matched beams , we use @xmath121 , and therefore we find @xmath122 here , @xmath123 is the square of the emittance of the beam propagated without noise.the right hand side of this equation characterizes the averaged artificial emittance growth in one integration step ; its explicit expression is given by @xmath124 in the following , @xmath41 as given by eq .
[ eq : scaling_law ] and eq .
[ eq : scaling_law_kv ] , is used with the simplifying condition that the fluctuations due to the bi - linear interpolation are set to zero , i.e. @xmath125 .
this ansatz well describes a fine grained mesh .
the average of @xmath126 over @xmath17 and @xmath18 can be evaluated by an integration over the particle distribution function , if the number of macro - particles @xmath9 is large enough to approximate the p.d.f . using the notation @xmath127
, we find for a gaussian beam @xmath128 for a k - v distribution we find respectively @xmath129 using eq .
[ eq : delta_e_long ] , [ eq : fluct_g ] , and [ eq : fluct_kv ] , we find in general @xmath130 with @xmath131 a coefficient that incorporates the type of distribution , with @xmath132 for a kv , and @xmath133 for a gaussian .
we thus find a dependence of the noise induced emittance growth on the number of macro - particles @xmath9 , the number of grid - points @xmath12 , the integration length @xmath110 and the variance @xmath134 of the distribution .
the factor of @xmath133 for a gaussian beam can be understood in the following way : for a k - v particle distribution , macro - particles are equally distributed , and the number of particles in each cell is therefore constant . in contrast to this , macro - particles are mostly located in the center for gaussian beams . therefore , the relative fluctuations of the electric field @xmath135 are smaller , compared to a rms equivalent k - v beam . in this paper
we intend to discuss the effect of pic induced noise in a situation where the beam does not already exhibit emittance growth .
therefore , in the context of eq .
[ eq : expanded ] , we find that @xmath123 , the squared emittance of the beam propagated without noise , will be equal to @xmath136 .
in addition , the treatment of a constant focusing lattice further simplifies the formulas , as we find @xmath137 .
therefore , if the beam has no intrinsic emittance growth , eq .
[ eq : expanded ] becomes @xmath138 and eq .
[ eq : emittance_growth_kv_g ] effectively provides the emittance growth @xmath139 due to pic induced decorrelated noise : @xmath140 the artificial emittance growth may create another disadvantage in terms of an artificial change of tune - shift .
in fact , space charge forces lead to a defocussing of the beam and thus to a change of the betatron tune @xmath141 . for the k - v particle distribution ( un - bunched )
the associated tune shift is given by @xmath142 where @xmath143 is the classical particle radius .
as the initial emittance @xmath144 changes due to pic induced noise , while the number of physical particles @xmath10 remains constant , the associated tune shift changes by @xmath145 by using the expression @xmath146 , with @xmath147 , we find the approximation @xmath148 with the pic induced emittance growth rate @xmath149 as derived in the previous section .
this derivation is valid for gaussian beams as well , and the same scaling is found .
however , the space charge induced tune shift is the double , see e.g. @xcite . in this section
we benchmark the emittance growth predicted by eq .
[ eq : emittance_growth_kv1 ] with simulations performed for the sis100 heavy ion synchrotron @xcite ( without non - linear elements ) .
we use the micromap library @xcite for tracking the particles .
the pic solver @xcite makes use of a mesh of @xmath150 grid points in a box of size @xmath151 . for rms - emittances of @xmath152
the resulting space charge tune shift is @xmath153 and @xmath154 , at tune position @xmath155 and @xmath156 .
the integration step @xmath110 is set to 42 space charge kicks per betatron - wavelength .
the results are presented in fig .
[ fig : sis100_emittance ] , where the red dots show the emittance growth from the simulation , while the blue line is the theoretical prediction from eq .
[ eq : emittance_growth_kv1 ] , i.e. the random walk model applied to the full beam .
due to numerical noise for a simulation with @xmath157 macro - particles ( top ) and @xmath158 macro - particles ( bottom ) for a k - v beam tracked in the sis100 lattice .
the theory developed in this paper ( blue line ) predicts the artificial emittance growth in the simulation ( red dots ) .
the beam and simulation parameters are given in the text .
, title="fig : " ] + due to numerical noise for a simulation with @xmath157 macro - particles ( top ) and @xmath158 macro - particles ( bottom ) for a k - v beam tracked in the sis100 lattice .
the theory developed in this paper ( blue line ) predicts the artificial emittance growth in the simulation ( red dots ) .
the beam and simulation parameters are given in the text .
, title="fig : " ] for the simulations we use a coasting k - v beam .
therefore , we guarantee that space charge forces are linear and thus pic induced numerical noise is the dominant cause of emittance growth . as seen in fig . [
fig : sis100_emittance ] , the theory developed in this section recovers the artificial emittance growth found in the simulation of a complex machine .
for decorrelated numerical noise the random walk model can thus be used in order to a ) understand the effect of noise on single particles , and b ) make predictions on the evolution of the whole ensemble of particles .
equation [ eq : emittance_growth_kv1 ] can be used to find reasonable simulation parameters @xmath12 , @xmath9 and @xmath110 . e.g. considering a simulation with @xmath159 macro - particles , where an artificial emittance growth of @xmath160 is observed , we predict that for a simulation with @xmath161 macro - particles @xmath162 the scaling law , eq .
[ eq : emittance_growth_kv1 ] , can also be used to find an optimal number of macro - particles for a different number of grid points @xmath12 .
in fact , choosing the number of macro - particles such that @xmath163 then the pic induced emittance growth will be the same for all pairs @xmath164 , hence one can choose the most efficient setting .
further , the agreement of theory and simulation allows to extrapolate the results for the sis100 to a storage time of one second , corresponding to @xmath165 turns .
we find , that for the simulation setup used above , we have to use @xmath166 macro - particles , in order to limit artificial emittance growth to @xmath167 . if the physics case requires a higher resolution , such that we have to double the number of grid points @xmath12 , we make use of eq .
[ eq : rw_equivalence ] .
we then find , that to limit artifical emittance growth to @xmath168 for a storage time of one second , we have to use @xmath169 macro - particles .
a change of the integration length @xmath110 changes the slope of artificial emittance growth linearly .
thus , if we double @xmath110 , we have to use the double amount of macro - particles to obtain the same amount of noise induced emittance growth . for the benchmarking , we considered space charge induced tune shifts smaller than the actual design goal for the sis100 .
we remark that for stronger space charge forces , the number of macro - particles has to be further increased , since the fluctuations become stronger , see appendix [ ch : appendix_sr_tune ] . a dedicated survey for the estimation of simulation parameters for certain operational scenarios will be part of future studies .
until now , only decorrelated numerical noise has been considered .
however , the conjoint effect of the pic algorithm and the particle tracker may cause correlation in the numerical noise .
a study dedicated to this effect is presented in the following section .
in the previous section we discussed the effect of decorrelated numerical noise on single particles and the full beam . in this section , we study the case where numerical noise is correlated . if a coasting beam is tracked through a circular lattice , depending on the integration length @xmath110 and the machine tunes @xmath141 and @xmath170 , the particle ensemble returns ( close ) to its initial positions after a certain integer number @xmath171 of space charge kicks .
afterwards , the same sequence of space charge kicks is applied again , such as the electric field fluctuations .
thus , fluctuations of the electric field occur periodically and are no longer random .
as the periodicity of the noise may create resonant effects , we call this a _ stochastic resonance _ @xcite
. these kind of correlations may lead to an enhanced emittance growth that exceeds the estimates of the random walk model . in mathematical terms , correlations in numerical noise are created if @xmath172 where @xmath173 is the intra - kick phase advance in between @xmath4 and @xmath72 and @xmath174 is an integer .
we define @xmath171 as the order of the stochastic resonance . for a constant focusing channel , where @xmath175 is constant for any @xmath62 , eq .
[ eq : sr_first ] can be simplified to @xmath176 with @xmath171 , and @xmath174 co - prime numbers .
when this equation is nearly satisfied , the numerical noise on the first @xmath171 space charge kicks is completely random .
however , for the following space charge kicks , the fluctuations of the electric field occur periodically according to @xmath177 where , for simplicity , we assume the motion in @xmath18 to be frozen .
the recurrence of fluctuations has an important property if @xmath171 is an even number : after @xmath178 space charge kicks the phase advance is @xmath179 , meaning that a particle @xmath42 with initial coordinates @xmath180 is mirrored ( antisymmetry ) on the phase space axes to @xmath181 . in this case the periodicity relation eq .
[ eq : sr_corr ] is modified , and reads @xmath182 we observe that the recurrence of the electric field fluctuations described by eq .
[ eq : sr_corr ] and eq .
[ eq : sr_corr2 ] is limited to a maximum number of applications @xmath183 of the @xmath171 space charge kicks .
afterwards correlations are canceled by accumulated noise on the particle s positions .
the persistence of correlations @xmath183 depends on @xmath171 and on the granularity of the distribution that is controlled by @xmath12 , @xmath9 and @xmath184 . therefore , a scaling law for artificial emittance growth in the presence of a stochastic resonance , analogous to eq .
[ eq : emittance_growth_kv_g ] for decorrelated numerical noise , can only be derived by a dedicated study on the damping of correlations , which is beyond the scope of this paper .
the effect of the first @xmath171 kicks can be modeled by a random walk process , as described in sec .
[ sec : scaling_law_emi ] .
the order of the stochastic resonance @xmath171 fixes the number of random elements in @xmath114 and @xmath185 .
thus , for a total of @xmath75 space charge kicks in a simulation , if @xmath186 we retrieve the random walk .
if instead @xmath171 is a small natural number , correlations in the numerical noise are created .
depending on the number of kicks @xmath75 , the persistence of the correlations @xmath183 , and the order @xmath171 , correlated noise may enhance artificial emittance growth .
this issue is discussed in the context of a periodic random walk in appendix [ ch : appendix_rw ] . in the following ,
we outline the procedure for the two types of fluctuations described by eq .
[ eq : sr_corr ] , and eq . [ eq : sr_corr2 ] .
we start by considering numerical noise with a correlation as described by eq .
[ eq : sr_corr ] .
we define two random variables @xmath187 , for @xmath77 , @xmath188 , in which @xmath189 , and @xmath190 with equal probability . to model the effect of correlated numerical noise , the random elements @xmath114 and @xmath185 are applied periodically @xmath183-times on all particles . when @xmath171 is even , to model the noise with the property of eq .
[ eq : sr_corr2 ] , we define two random variables @xmath187 , for @xmath77 , @xmath191 , in which @xmath189 , and @xmath190 with equal probability . in order to model the anti - symmetric property of the fluctuations , the next @xmath178 random numbers are defined by @xmath192 and @xmath193 , with @xmath194 .
this model is directly suggested by eq .
[ eq : a_elements ] , observing that after @xmath178 kicks the sign of @xmath195 , and @xmath196 inverts , see appendix [ ch : appendix ] .
therefore , @xmath171 kicks of a resonance of an even order @xmath171 have a similar effect on a particle as the first @xmath178 kicks applied twice .
this argument suggests that a resonance of order @xmath178 odd will produce a similar emittance growth as a resonance of order @xmath171 . in the following , we study the properties of the correlations in electric field fluctuations .
for this purpose , a coasting beam is tracked in a constant focusing channel with machine tunes @xmath197 and @xmath198 .
the length of the channel is @xmath199 m. for simplicity , we chose the integration length @xmath92 .
the electric field @xmath200 and @xmath201 at @xmath7 is calculated each turn for a total of 1024 turns , and a spectrum is retrieved by performing a fourier transform of the data . in the corresponding spectrum ,
see fig .
[ fig : spectrum_example ] , the highest peak occurs at the machine tune position .
other frequencies are excited due to a coupling of both planes .
-direction ( top ) and @xmath18-direction ( bottom ) at @xmath7 . for this analysis ,
the beam is tracked in a constant focusing channel with tune @xmath202 and @xmath203 .
peaks occur at the tune frequencies , while a coupling of both planes excites many other smaller local peaks.,title="fig : " ] -direction ( top ) and @xmath18-direction ( bottom ) at @xmath7 . for this analysis ,
the beam is tracked in a constant focusing channel with tune @xmath202 and @xmath203 .
peaks occur at the tune frequencies , while a coupling of both planes excites many other smaller local peaks.,title="fig : " ] since we pursue to study the effect of stochastic resonances in one plane , we have to ensure that we do not have any interference of the other plane .
this can be guaranteed by setting one of the tunes close to an integer value .
if we choose @xmath204 , we decouple sufficiently the @xmath17 and @xmath18 planes , see fig .
[ fig : spectrum_clean ] .
-direction at @xmath7 .
the vertical tune is @xmath205 while the horizontal tune @xmath204 is close to an integer value .
a peak occurs at the tune position , while other frequencies are no longer excited . ] in order to study the effect of the intra - kick phase advance on the artificial emittance growth , we use again the constant focusing channel with the same simulation parameters as above .
we now systematically scan the vertical tune @xmath170 in the range @xmath206 , while keeping @xmath204 fixed .
the initial emittance in the @xmath18-plane is chosen for any machine tune such that we have a round beam ( @xmath207 ) .
for each machine tune a coasting beam is tracked for 60000 turns and the slope of emittance growth per meter is calculated .
the dependence of emittance growth on the machine tune , and thus on the excitation of stochastic resonances , is retrieved and shown in fig .
[ fig : final_scan ] . , while @xmath204 .
the tune @xmath170 is changed systematically in steps of @xmath208 . ] by studying the simplified scenario of a constant focusing channel , we guarantee a constant phase advance in between space charge kicks .
then , stochastic resonances are excited when a multiple of the intra - kick phase advance is an integer of @xmath209 , see eq .
[ eq : sr_second ] . for a realistic model of a machine
the scenario gets more complex , as described by eq .
[ eq : sr_first ] . the emittance growth rates , as presented in fig . [ fig : final_scan ] , confirm the prediction given by the theory developed in this paper . in the absence of stochastic resonances ( decorrelated numerical noise ) ,
we find a growth rate that can be predicted by the random walk model .
the emittance growth rate has a ( local ) peak whenever a stochastic resonance is hit .
the lower the order of the resonance @xmath171 , the stronger we see its effect on the emittance growth rate .
further , we find that if @xmath178 is an odd integer , the growth rate is similar to the one of @xmath171 , e.g. in the case @xmath210 . until now
, the space charge tuneshift was fixed for each simulation .
in appendix [ ch : appendix_sr_tune ] we find that the width of the stopband of a stochastic resonance and the slope of artificial emittance growth are increased for more intense beams .
for long - term tracking studies it is desirable to avoid the occurrence of correlated noise to a ) minimize artificial emittance growth and b ) avoid uncontrollable artifacts , like tails in the distribution . in the following , we present a strategy to ensure that the numerical noise is decorrelated and thus controllable by a proper choice of simulation parameters , as discussed above . in most simulation studies the machine tune
@xmath170 is a multiple of the intra - kick phase advance , i.e. @xmath211 because the integration length @xmath110 is a fraction of the machine length @xmath212 . for simulations close to a machine resonance , i.e. of type @xmath213 for @xmath214 , stochastic resonances due to pic noise
will then be excited simultaneously and will afflict the simulations .
however , this situation can be avoided by choosing an appropriate integration length @xmath110 . as a demonstration of the influence of the integration length on the occurrence of stochastic resonances , the study presented in sec .
[ ch : correlations_eg ] is repeated for an integration length @xmath215 .
the results are presented in fig .
[ fig : shift_scan ] . , while @xmath204 .
the tune @xmath170 is changed systematically in steps of @xmath208 .. the integration length is changed , compared to the blue reference line , by a factor of 0.95 ( red line ) . ] the blue line in fig . [
fig : shift_scan ] is the same as in fig .
[ fig : final_scan ] , where @xmath216 , while the red line corresponds to @xmath217 .
the peaks due to stochastic resonances are shifted for the modified integration length .
the effect of changing the integration length can most easily be understood for the simplified case of a constant focusing channel .
in fact , if the integration length is set to a fraction @xmath218 of the initial one , i.e. @xmath219 , the intra - kick phase advance is changed by a factor @xmath218 .
therefore , the resonance condition , eq . [ eq : sr_second ] , is fulfilled for another tune @xmath220 given by @xmath221 we can thus effectively avoid correlated noise in our working point @xmath170 by changing the integration length , and hence guarantee that single particles are solely affected by white noise . then
, artificial noise effects can be mitigated for an optimal parameter setting obtained using eq .
[ eq : emittance_growth_kv1 ] .
in this paper we developed a model for predicting the effect of white noise on a particle beam by using an equivalent effective field fluctuation acting on all phase space coordinates .
we use simulation results to infer the mathematical modeling of electric field fluctuations on a single particle in the beam , and use this model to develop a treatment of the full beam ensemble .
the inferred scaling law as a function of the simulation parameters enables us to estimate artificial emittance growth .
we find that for specific machine tunes , which recall those of a machine resonance , pic noise becomes correlated .
the effect of these correlations on emittance growth is more dramatic , and we explain this unusual dynamics in terms of stochastic resonances . our studies for the constant focusing channel show that the intra - kick phase advance plays a crucial role on numerical noise as it defines the tunes at which stochastic resonances appear .
this understanding is used to develop a strategy to avoid or minimize correlations in numerical noise , hence avoiding stochastic resonances . the theory developed in this paper
can thus be applied to control unwanted noise effects in tracking simulations of high intensity beams .
last , we remark that in our studies we decoupled the pic induced numerical noise in the transverse planes by setting the horizontal tune @xmath141 close to an integer value .
however , in more realistic applications , the effect of coupling can not be neglected , and certainly will introduce corrections .
investigations on the coupling between the planes will be part of future studies .
the authors acknowledge valuable remarks of prof .
dr . i. hofmann ( gsi ) .
the research leading to these results has received funding from the european commission under the fp7 research infrastructures project eucard-2 , grant agreement no .
in the following , the combined effect of linear tracking and numerical noise is studied . for simplicity , we discuss the constant focusing channel .
let us consider a single particle initialized at position @xmath222 , and track it including the space charge through a constant focusing channel .
due to numerical noise , a fluctuation @xmath223 is induced at any longitudinal position @xmath4 , where a space charge kick is applied . in total @xmath75 space charge kicks
are applied , i.e. the particles are tracked for a distance @xmath224 .
the numerical noise at position @xmath4 affects the particle s phase space coordinates at @xmath225 by @xmath226 we call the first component of this vector @xmath195 and the second @xmath196 , which can be interpreted as random kicks on @xmath227 resp . @xmath228 due to numerical noise at @xmath229 .
since we specialize on a constant focusing channel , we can decompose the transfer matrix @xmath70 into @xmath230 with @xmath231 and @xmath232 using @xmath233 , we simplify eq .
[ eq : a_basic ] to @xmath234 the random elements @xmath195 , @xmath196 can be described by @xmath235 a test - particle initialized at @xmath236 and tracked through the constant focusing channel accumulates the random elements @xmath195 ,
@xmath196 . for @xmath75 space charge kicks the particle coordinates are @xmath237 where @xmath238 are the final coordinates of the test particle for a noise - free tracking .
the random elements @xmath195 , @xmath196 , as derived in eq .
[ eq : a_elements ] , depend on the intra - kick phase advance @xmath239 . for most intra - kick phase advances @xmath239
the cumulated numerical noise remains small , since the contributions @xmath195 resp .
@xmath196 are averaged to zero , i.e. @xmath240 and resp .
we note that @xmath242 ( resp .
@xmath243 ) is a new random variable , which is the result of the summation of the @xmath75 random variables @xmath195 ( resp .
@xmath196 ) . the random variable @xmath242 has a variance of @xmath244 because the random variables @xmath195 are decorrelated . a direct summation of this equation yields @xmath245 ^ 2 , \ ] ] which retrieves the property of the random walk .
therefore , we can substitute each @xmath195 with the random kick @xmath246 with @xmath247 random , to retrieve the same final statistical properties after @xmath75 kicks . the same argument applied to the quantity @xmath243 shows that each @xmath196 can be substituted with @xmath248 with @xmath190 random and statistically independent from the set @xmath114 .
this explains the effective random walk model used in eq .
[ eq : kick ] .
the random walk approach is not applicable anymore , if @xmath249 because a correlation is created .
we show the breaking down of the random walk approach with a numerical experiment . for this
, we define a sequence of random numbers @xmath250 with the correlation @xmath251 which resembles correlated numerical noise in the presence of a stochastic resonance . the cumulated effect for a constant fluctuation @xmath252 for @xmath253 is given by @xmath242 and @xmath243 , as described in eq .
[ eq : a_ab ] .
the averaged cumulated numerical noise @xmath254 and @xmath255 for 100 random configurations of @xmath185 following eq .
[ eq : a_cor ] , is shown in fig . [
fig : sum_a ] and fig .
[ fig : sum_b ] for intra - kick phase advances @xmath256 . for 100 initializations of @xmath257 following eq .
[ eq : a_cor ] .
peaks appear at phase advances @xmath258 , and @xmath259 . ]
+ for 100 initializations of @xmath257 following eq .
[ eq : a_cor ] .
peaks appear at phase advances @xmath260 , @xmath258 , @xmath261 , @xmath259 , and @xmath262 . ] + we observe the cumulation of numerical noise for certain intra - kick phase advances , i.e. @xmath263 , and @xmath264 .
these results can be interpreted in the following way . if the condition for a stochastic resonance , see eq .
[ eq : a_sr ] , is not fulfilled , then the elements @xmath195 , @xmath196 are random and thus the random walk model can be applied . since both parameters are decoupled and are in average equivalent , the model given by eq .
[ eq : kick ] is justified .
however , in the presence of correlations this model is not valid anymore and the noise has to be treated differently , see sec .
[ ch : correlations ] .
to explain the effect of a periodic random walk , we study the following simplified case .
a single particle with the one dimensional coordinate @xmath17 is affected by @xmath75 random kicks of strength @xmath265 in arbitrary units .
we further enforce a periodicity of @xmath171 , i.e. we define a set of random numbers @xmath266 with @xmath267 that is applied @xmath183 times .
the number of repetitions @xmath183 is chosen to fulfill @xmath268 , with @xmath269 constant .
since the averaged position after many random kicks is @xmath270 , the averaged second order moment can be associated with the variance .
the dependence of @xmath271 on @xmath171 is retrieved in a numerical experiment . for this purpose ,
@xmath272 is evaluated for 100 random initializations of @xmath273 for each @xmath274 .
the results are presented in fig .
[ fig : periodic_walk ] .
after @xmath253 random kicks for different periodicities @xmath171 .
the blue line indicates the second order moment for the non - periodic random walk ( without repetition ) . ]
fitting the numerical data , we find for @xmath275 the scaling of the averaged second order moment on @xmath171 @xmath276 for @xmath186 , the averaged second order moment becomes independent of @xmath171 since we apply only @xmath75 kicks in this example .
these results show that for a periodic random walk the diffusion is enhanced .
the smaller the periodicity @xmath171 , the larger is the diffusion .
when applying these results to the situation of a stochastic resonance created by pic tracking , we have to consider that correlations in the numerical noise are washed out due to the numerical noise itself . the scaling , as given in eq .
[ eq : periodic_walk ] , may suggest that the periodic random walk could be an upper bound on the diffusion , because no loss of memory is considered .
if the correlations are completely washed out , the diffusion type will the one of a random walk .
this defines the lower bound on the diffusion , as indicated by the blue line in fig [ fig : periodic_walk ] .
the excitation of stochastic resonances is certainly affected by the strength of the space charge force . to gain a quantitative knowledge
, we repeat the study presented in sec .
[ ch : correlations_eg ] for different beam currents and thus different space charge forces .
the artificial emittance growth is investigated in the vicinity of the third order stochastic resonance , which is presented in fig .
[ fig : scan_detail ] . in the vicinity of the third order stochastic resonance , while the horizontal tune is @xmath204 .
the vertical tune @xmath170 is changed systematically in steps of @xmath208 . the colors represent different beam currents . ] as we learn from fig .
[ fig : scan_detail ] , an increase of the beam , and thus the space charge force , enhances artificial emittance growth and leads to a broadening of the stop - band of the stochastic resonance .
the increase of artificial emittance growth can be explained by the increase of the driving force of the resonance , i.e. of the space charge . at the same time , by increasing the space charge force we enlarge the tune spread of the beam .
therefore , the range of tunes @xmath170 affected by the stochastic resonance grows larger .
1 g. lopez and s. chen , sscl - preprint-539 ( 1993 ) .
y. alexahin nucl .
methods phys .
res . , sect .
a * 391 * , 73 ( 1996 ) .
p. eliasson phys . rev .
beams * 11 * , 051003 ( 2008 ) .
s. sorge , g. franchetti , a. parfenova phys . rev .
beams * 14 * , 052802 ( 2011 ) .
v. lebedev , v. parkhomchuk , v. shiltse , g. stupako sscl - preprint-188 ( 1993 ) . g. dome , cern report no .
84 - 15 ( unpublished ) , pp . 215260 ( 1984 ) .
p. spiller and g. franchetti , nucl .
methods phys .
res . , sect .
a * 561 * , 305 ( 2006 ) ; o.k .
et al . _ , proceedings international particle accelerator conference ipac2014 , dresden , germany , june 15 - 20 , 2014 ( eps - ag , dresden2014 ) , wepro060 ( 2014 )
. j. coupard _
et al . _ , liu technical design report ( tdr ) , cern - acc-2014 - 0337 ( 2014 ) .
g. franchetti , i. hofmann , m. giovannozzi , m. martini , and e. metral , phys . rev .
beams * 6 * , 124201 ( 2003 ) .
a. orzhekhovskaya and g. franchetti proc . of 9th european particle accelerator conference , lucerne , 2004 .
j. struckmeier phys . rev .
beams * 3 * , 034202 ( 2000 ) .
o. boine - frankenheim , i. hofmann , j. struckmeier , and s. appel , nucl .
methods phys .
res . , sect .
a * 770 * , 164 ( 2015 ) .
i. hofmann and o. boine - frankenheim , phys .
beams * 17 * , 124201 ( 2014 ) .
birdsall and a.b .
langdon , _ plasma physics via computer simulation _ , iop , bristol ( uk ) ( 1991 ) .
r.w . hockney and j.w .
eastwood _ computer simulation using particles _ , adam hilger , bristol ( uk ) and new york ( usa ) ( 1988 ) .
g. turchetti _ et al .
phys . j. c * 30 * , 279 - 290 , ( 2003 ) .
a. bazzani _ et al .
_ , 7th int .
conf computational accelerator physics , michigan , usa , 15 - 18 october 2002 .
s. rambaldi _ et al .
methods phys .
res . , sect .
a * 561 * , 223 ( 2006 ) .
see website of g. franchetti , http://web-docs.gsi.de/@xmath277giuliano/ i.m .
kapchinsky , v.v .
vladimirsky , proc .
int . conf . on high energy accelerators , cern , geneva , p. 274
ff ( 1959 ) .
p.m. lapostolle , ieee trans .
* ns-18 * 1101 ( 1971 ) .
sacherer , ieee trans .
* ns-18 * 1105 ( 1971 ) .
f. kesting , g. franchetti , and u. ratzinger , proceedings international particle accelerator conference ipac2014 , dresden , germany , june 15 - 20 , 2014 ( eps - ag , dresden2014 ) , p. 3005 .
m. reiser , theory and design of charged particle beams .
john wiley and sons ( 2008 ) .
g. franchetti , i. hofmann and , g. turchetti , nucl .
methods phys .
res . , sect .
a * 415 * , 450 ( 1998 ) .
g.f . lawler and v. limic ,
_ random walk : a modern introduction _ , vol .
123 , cambridge university press ( 2010 ) .
k. schindl , proceedings of the joint us - cern - japan - russia school on particle accelerators in montreux ( 1998 ) .
l. gammaitoni , p. hnggi , p. jung , and f. marchesoni , rev .
phys . * 70 * , 223 ( 1998 ) .
|
particle - in - cell ( pic ) is the most used algorithm to perform self - consistent tracking of intense charged particle beams .
it is based on depositing macro - particles on a grid , and subsequently solving on it the poisson equation .
it is well known that pic algorithms occupy intrinsic limitations as they introduce numerical noise .
although not significant for short - term tracking , this becomes important in simulations for circular machines over millions of turns as it may induce artificial diffusion of the beam . in this work , we present a modeling of numerical noise induced by pic algorithms , and discuss its influence on particle dynamics .
the combined effect of particle tracking and noise created by pic algorithms leads to correlated or decorrelated numerical noise . for decorrelated numerical noise
we derive a scaling law for the simulation parameters , allowing an estimate of artificial emittance growth .
lastly , the effect of correlated numerical noise is discussed , and a mitigation strategy is proposed .
|
while it is now clear that the cosmic x - ray background ( xrb ) is made by the superposition of many discrete sources , and that the soft ( 0.52 kev ) xrb is mostly produced by agn ( hasinger et al .
1998 , schmidt et al . 1998 ) , the nature of the sources making the energetically dominant hard ( 250 kev ) xrb is still largely unknown .
this is due to the lack , until recent years , of sensitive imaging instruments above 23 kev .
thanks to asca and bepposax , the 2 - 10 kev band is now accessible to surveys ( boyle et al .
1998 , ueda et al .
1998 , giommi et al .
1998 , giommi , fiore & perri 1998 , fiore et al .
1998a , b ) .
the hard x - ray sky poses major problems : no known large class of sources has an emission spectrum matching the kt@xmath3 kev thermal like spectrum of the xrb ; fluctuation analyses of ginga data ( warwick & stewart 1989 ) , and asca ( georgantopoulos et al . 1997 , cagnoni et al . 1998 ) and bepposax ( giommi et al . 1998 , giommi , fiore & perri 1998 ) 210 kev source counts , require 23 times as many sources as expected from rosat counts , if they have the steep power law spectrum typical of rosat sources ( @xmath4 , @xmath5 ) . the leading suggestion to reconcile the source counts , based both on theoretical grounds ( setti & woltjer 1989 , madau et al .
1994 , matt & fabian 1994 , comastri et al . 1995 ) and high energy observations of bright nearby agn ( zdziarski et al 1995 , smith & done 1996 ) , is that heavily obscured agn , emerging strongly at high energies , are the main contributors to the hard xrb .
( these agn are probably not completely invisible at low x - ray energies , i.e. @xmath6 kev , because even if the nuclear emission is completely blocked , different components , like starburts , optically thin gas or scattering of the nuclear radiation , may still be detectable at the 1%10% level , giommi , fiore & perri 1998 , schachter et al .
it is crucial to test these suggestions observationally in a band where the nucleus is directly visible and down to fluxes where the bulk of the hard xrb is produced .
hard x - ray selection would also help in distinguishing between different scenarios .
for instance , if the bulk of the xrb is made by obscured luminous ` quasar 2 ' , following the evolution of type 1 agn , the peak of their activity should be at z=23 and then decrease quickly toward lower redshifts . on the other hand ,
the xrb could be mostly made by a large population of less luminous / active obscured agn , spread in a broader interval of redshifts .
the solution of these problems may have impact also on agn unification schemes .
the bepposax mecs ( boella et al .
1997a , b ) provides a good opportunity to investigate the hard x - ray sky , thanks to its good sensitivity above 5 kev ( 5 - 10 kev flux limit of @xmath7 mcrab in 100 ks ) , and improved point spread function ( psf ) .
we have thus carried out the high energy llarge area survey ( hellas ) .
the survey has been performed in the hard 5 - 10 kev band because this is the band closest to the maximum of the xrb energy density that is reachable with the current imaging x - ray telescopes . including the softer 1.5 - 5 kev range would only increase the background for faint , heavily absorbed sources , thus reducing the chances of their detection .
secondly , the bepposax mecs psf improves with energy , and in the 5 - 10 kev band provides error circles of 1 arcmin , 95% confidence radius , small enough to allow the optical identification of the x - ray sources .
the hellas survey has so far cataloged 180 sources ( fiore et al .
1999 , in preparation ) .
the sky coverage is @xmath8 square degrees at @xmath9 respectively . at the fainter limit we find between 16 and 20 sources deg@xmath10 ( fiore et al .
1998a , b , giommi , fiore & perri 1998 , comastri 1998 ) , implying that about 30 % of the xrb is resolved .
cross - correlations of the hellas sample with catalogs of cosmic sources provides 18 coincidences ( 7 radio - loud agn , 6 radio - quiet agn , 3 clusters of galaxies , 1 cv and 1 normal star ) , suggesting that most of the hellas sources are agn .
however , the 13 agn were discovered in radio , soft x - ray , and optical surveys , and so this sample is biased against highly obscured sources . to remove this bias we have started a program to spectroscopically identify the rest of the hellas sample . in this letter
we report on the first results of this program .
the sources studied have not been selected according to their x - ray or optical properties but solely because of visibility during our observing runs . as a result
, they can be considered representative of the whole hellas sample .
we carried out spectroscopic identification of optical candidates for the hellas sources using the rc spectrograph ( rcsp ) at the kitt peak 4 m telescope for 9 sources , the fast spectrograph at the whipple 60 telescope for 1 source , the hawaii 88 for 2 sources and efosc2 at the eso 3.6 m for 2 sources .
long - slit spectra have been obtained in the 3450 - 8500 @xmath11 range ( rcsp ) or 3800 - 8000 @xmath11 range ( fast , 88 ,
efosc2 ) , with a resolution between 7 and 16 .
between 1 and 7 optical candidates were identified on apm or cosmos scans of the e ( r ) poss plates down to r@xmath12 m , within a conservative ( 95 % ) error - circle radius of 60@xmath13 .
the complete set of spectra will be presented elsewhere ( la franca et al . 1999 in preparation ) .
we identified the x - ray source with the more plausible object in the error box ( if any ) , i.e. either sources with high hard x - ray to optical ratio ( like agn , galactic binaries , clusters and groups of galaxies ) or bright galaxies and stars . unlike previous identification campaigns of asca sources , we decided not to take advantage of correlations with rosat source catalogs or radio catalogs , to avoid selection biases against obscured objects . instead , our basic strategy is to obtain spectra of all candidates down to the chosen r=20 magnitude limit .
however , in 4 cases a few faint candidates still remain to be observed in each error - box .
their magnitude is fainter than that of the candidate identified as an agn , and therefore the identification of the x - ray source with the agn is rather secure . in twelve cases we obtained a unique identification ( only one plausible candidate in the error - box ) .
in one case we found 2 sources in the error - circles which could contribute to the detected hard x - ray flux ( designated a , b in table 1 ) . in one case none of the 6 objects observed with @xmath14 were good candidates for being the x - ray source .
all fields but one ( 0122.1 - 5845 ) have been observed with the vla as part of the nvss survey .
no radio source brighter than 3 mjy is found within any of the hellas error - boxes . given the sources optical magnitude
this implies that all objects can be considered as radio - quiet .
table 1 gives , for each of the identified sources , the position of the optical counterpart and its distance from the x - ray centroid in arcsec , its redshift , r and b magnitudes , optical and x - ray luminosities and the classification .
[ pos ] [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^ " , ] broad emission line quasars with a normal blue continuum , @xmath15 ( @xmath16 , @xmath17z@xmath18 ) , were found in five error - boxes .
broad emission line quasars with a very ` red ' continuum were found in two cases ( @xmath19 , consistent with their b - r color ( see table 1 ) .
their spectra are dominated by starlight , based on the detection of the calcium h and k and mgi absorption features and the small equivalent width of h@xmath20 ( 90 ) observed in 1saxj1353.9 + 1820 . in this case we could estimate a lower limit to the extinction of a@xmath21 , corresponding to log@xmath22 assuming a galactic dust to gas ratio .
so , this quasar appears ` red ' both because the spectrum is dominated by starlight from the host galaxy and because of the large extinction to the nucleus .
narrow emission line galaxies were found in 6 error - boxes . to identify their nature we used the standard line ratios diagnostics ( tresse et al .
1996 , see table 2 ) .
five sources show large [ oiii]@xmath23/h@xmath24 ( @xmath25 ) and/or large [ sii]@xmath26/h@xmath20 ratios .
two sources also show strong [ nev ] lines , which unambiguously identify them as narrow line agn ( schmidt et al .
1998 ) . in all five sources h@xmath20 or h@xmath24
have faint wings of fwzi 30005000 km s@xmath27 , broader than the typical value for agn narrow line region lines .
we then classify these sources as intermediate 1.8 - 1.9 type agn ( osterbrook 1981 ) .
in one error - boxes we found a narrow emission line galaxy , having [ sii]@xmath26/h@xmath20 and [ oii]@xmath28/h@xmath24 ratios indicating a lower excitation than in seyfert 1.8 - 2 galaxies but close to or higher than the value which separate hii galaxies from liner ( tresse et al .
we tentatively classify it as a liner .
figure 1 shows two examples of hellas agn . in summary , 12 of the 14 hellas error - boxes studied contain agns and 1 contains a liner .
we have checked how many of the optical sources lie within 30 arcsec from pspc sources in the wgacat .
we found six coincidences : 0045.7 - 2515 , 1118.2 + 4028 , 1118.8 + 4026 , 1117.8 + 4018 , 1218.9 + 2958 ( also visible in the asca gis field of mark766 ) , and 2226.5 + 2111 .
another 4 sources ( 0122.1 - 5845 , 1353.9 + 1820 , 1519.5 + 6535 , 1134.7 + 7024 ) lie in pspc fields but are not detected in the wgacat .
the surface density of quasars at r=20 ( the magnitude of the faintest of our quasars ) is 2040 deg@xmath10 ( zitelli et al .
1992 ) and therefore the expected number of chance coincidences in our 14 error - boxes is 0.25 - 0.5 . the surface density of galaxies at r@xmath2919 ( the magnitude of the faintest of our seyfert 1.81.9 galaxies and liner ) is about 500 deg@xmath10 ( lilly et al .
tresse et al .
( 1996 ) found that between 8 and 17 % of the galaxies with z@xmath30 host seyfert 1.82 or liner nuclei .
hammer et al .
( 1997 ) found that @xmath31 of the higher redshift and luminosity cfrs galaxies host type 1.82 nuclei .
the surface density of r@xmath32 type 1.82 agn and liner is therefore 4590 deg@xmath10 .
the number of their chance coincidences in the total area covered by 14 error - boxes is so 0.51 .
we note that for the agn this is actually an upper limit , since the distribution of the magnitudes of the type 1.81.9 agn in table 1 is different from a typical galaxy logn - logs , and because two of the type 1.81.9 agn are also pspc sources .
we then conclude that the identification of the quasars and type 1.8 - 1.9 agn is rather secure , while for the liner we can not exclude a chance coincidence .
since for the sources in table 1 we have redshifts , luminosities , optical spectroscopy and x - ray colours , we can start addressing the question of which kind of agn populates the hard x - ray sky .
x - ray hardness ratios can be used to estimate the x - ray spectrum .
assuming an absorbing column equal to the galactic column along the line of sight , produces rather unlikely spectral indices for the 1.8 - 1.9 type agn and the ` red ' quasars ( @xmath33 ) , harder than the @xmath34 of the 3 - 15 kev xrb spectrum and of any known agn . on the other hand , assuming a typical agn power law spectrum of energy index @xmath35 and the source redshift , we obtain the @xmath36 values given in table 1 .
( for the unidentified source we assumed z=0 and so the relative @xmath36 is a lower limit to the true absorbing column . )
the distribution of log@xmath36 is plotted in figure 2 , along with the distribution predicted by the comastri et al .
( 1995 ) syntesis model in the 5 - 10 kev band and at the flux limit of our survey .
the two distributions agree quite well . at a first sight
the lack of heavily absorbed compton thick ( log@xmath37 , i.e. the nuclear emission is completely blocked below 10 kev ) sources in our survey is at variance with the recent findings of maiolino et al .
( 1998 ) and risaliti et al .
however it should be noted that their results have been derived from an optically selected sample of bright , nearby seyfert 1.8 - 2 galaxies , and thus the comparison with our survey is not straigthforward .
the distribution in figure 4 of risaliti et al .
( 1999 ) is the expected one at the flux limit where most of the xrb is resolved .
our survey resolves only a fraction of the xrb and therefore it misses the sources in which large absorption ( log@xmath37 ) reduces the 5 - 10 kev flux below our flux limit . the hard x - ray to optical ratio ( x / o ) of hellas agn is similar to that of nearby seyfert 1 and compton thin seyfert 1.8 - 1.9 galaxies ( from the samples of bassani et al .
1999 , matt et al 1999 ) .
the hellas sources can so be considered as their higher redshift ( up to z=0.4 ) analogs .
it is worth noticing that another intermediate type 1.9 , compton thin quasar has been discovered by almaini et al ( 1995 ) and georgantopoulos et al .
( 1998 ) at even higher redshift ( z=2.35 ) .
conversely , the x / o of the local compton thick seyfert 2 galaxies ( maiolino et al .
1998 ) with @xmath38 is lower by a factor 30 - 100 .
the study of faint compton thick seyfert 2 ( maiolino et al .
1998 ) at z=0.20.4 and v@xmath39 ( which should have @xmath40 ) must therefore await instruments with sensitivity higher than asca and bepposax , like those on board axaf and xmm .
our results , obtained using a purely hard x - ray selected sample and based on new identifications as well as correlation of the hellas sample with catalogs of known sources , provide strong evidence that most of the sources resolved down to a 5 - 10 kev flux level of @xmath41 are agn .
the amount of absorption in these agn is consistent with the prediction of agn synthesis models of the xrb ( comastri et al .
1995 ) in the 5 - 10 kev band and at the above flux limit . however , the source breakdown is rather peculiar .
in addition to five ` blue ' continuum broad line quasars , which dominate optical and soft x - ray surveys , we found 7 `` intermediate '' agn .
these objects are `` intermediate '' both for their optical spectra which pose them between type 1 and type 2 agn and because of their x - ray absorption , which is typically log@xmath36=22.5 - 23 ( estimated from hardness ratios ) . `
red ' quasars have been selected in the past in the radio band ( smith & spinrad 1980 , webster et al .
1995 ) or in soft x - rays ( provided that the absorbing column is not too large , a@xmath42 , log@xmath43 , kim & elvis 1998 ) .
their number density relative to ` blue ' quasars is thought to be a few percent .
in contrast , we found a fraction of @xmath4430% , showing that hard x - ray selection may be a more efficient and less biased way to search for ` red ' quasars .
the fraction of agn showing evidence of extinction / absorption to the nucleus in optical / x - ray ( 7 out of 12 agn ) is @xmath45 ( gehrels 1986 , @xmath46 confidence level ) .
the fraction of intermediate agn + narrow line galaxies in the rosat 0.5 - 2 kev survey of the lockman hole ( @xmath47 , schmidt et al .
1998 ) , and in the shallower cambridge - cambridge rosat serendipity survey ( crss , @xmath48 , boyle et al , 1995a , b ) is [email protected] ( 11 objects out 39 ) and [email protected] ( 12 objects out of 80 ) , respectively . the same fraction in the asca - rosat combined surveys of boyle et al .
( 1998 ) and akiyama et al .
( 1998 ) is [email protected] ( 20 objects out of 61 ) .
we calculated , using both the fisher exact probability test ( siegel 1956 ) and the barnes ( 1994 ) test on the difference of two proportions , the probability that the hellas and the lockman hole , crss and asca - rosat samples differs in the proportion of obscured objects .
this probability is @xmath50 for hellas / lockman hole samples , @xmath51 for the hellas / crss samples and @xmath52 for the hellas / asca - rosat samples , for both statistical tests .
while more hellas identifications will clearly be useful to strengthen this result , the above probabilities already indicate a difference between the hellas sample and at least the two rosat samples . furthermore , while all hellas ` blue ' quasars in pspc fields are detected down to a typical wgacat flux limit of @xmath53 , only three of the six ` intermediate ' agn in pspc fields are detected at this flux limit ( see table 1 and sect 2 ) , again indicating that rosat surveys are more biased against faint highly obscured objects that our survey .
we did not find any luminous ( @xmath54 erg s@xmath27 , highly absorbed ( log@xmath55 ) quasar 2 .
however , one of the ` red ' quasar in our sample has an high x - ray luminosity and may have an absorbing column as high as log@xmath36=23 .
almaini et al .
( 1995 ) discovered an highly obscured quasar at z=2.35 , which has been classified as an intermediate 1.9 agn by georgantopoulos et al .
akiyama et al .
( 1998 ) report the discovery of 2 luminous broad line quasars with a very hard ( possibly absorbed ) x - ray spectrum .
one of our ` blue ' quasar may have absorption in excess of log@xmath36=22 .
although the statistics are not yet good enough to reach firm conclusions , these few positive detections and the absence of even a single detection of a narrow line quasar 2 ( halpern et al .
1998 ) suggest that x - ray absorption in high luminosity objects may be associated more frequantly with intermediate type 1.5 - 1.9 agn , ` red ' quasars , or even normal ` blue ' quasars , rather than with pure type 2 objects .
the bulk of the hard xrb could be made by a large population of seyfert 1.82 like galaxies and moderately absorbed ` red ' and intermediate quasars ( see also kim & elvis 1998 ) .
this is in line with scenarios where the absorption takes place in a starburst region surrounding the nucleus ( fabian et al .
1998 ) . in this case
the amount of the absorbing gas may depend on the nuclear mass ( and luminosity ) , and then the evolution of the type 2 agn luminosity function would also be strongly affected by the nuclear environment and may imply a link between agn evolution and the history of star - formation as measured from optical and sub - mm surveys ( see e.g. madau et al .
1996 , lilly et al .
1998 ) . in this regard ,
it is worth remarking that a sizeable population of low luminosity agn ( m@xmath56 in the range 17 ; 20 ) may exist at very faint optical fluxes ( v=26 - 27 ) in the hubble deep field north ( jarvis & macalpine 1998 ) .
this conclusion will soon be tested by the deep and high spatial resolution surveys that will be performed by axaf and xmm in the next few years .
these deep surveys , together with the shallower but larger area surveys perfomed by asca and bepposax , should be able to discriminate between different scenarios for the xrb and provide information on the connection between agn and galaxy evolution .
we thank the bepposax sdc , soc and occ teams for the successful operation of the satellite and preliminary data reduction and screaning , a. matteuzzi for his work on mecs source position reconstruction , p. massey for his help at the kitt peak 4 m telescope , p. berlind for the whipple 60 spectra of the 1saxj1519.5 + 6535 field , g. c. perola , m. vietri and g. zamorani for a careful reading of the manuscript and useful discussions .
|
hard x - ray selection is the most efficient way to discriminate between accretion - powered sources , such as agn , from sources dominated by starlight .
hard x - rays are also less affected than other bands by obscuration .
we have then carried out the bepposax high energy large area survey ( hellas ) in the largely unexplored 5 - 10 kev band , finding 180 sources in @xmath0 deg@xmath1 of sky with flux @xmath2 .
after correction for the non uniform sky coverage this corresponds to resolving about 30 % of the hard cosmic x - ray background ( xrb ) .
here we report on a first optical spectroscopic identification campaign , finding 12 agn out of 14 x - ray error - boxes studied .
seven agn show evidence for obscuration in x - ray and optical bands , a fraction higher than in previous rosat or asca - rosat surveys ( at a 95 - 99 % and 90 % confidence level respectively ) , thus supporting the scenario in which a significant fraction of the xrb is made by obscured agn .
= -0.5 in x ray : selection background galaxies
|
the orion nebula is the defining blister region ( @xcite ; @xcite ) .
a star cluster ionizes the skin of the molecular cloud , causing an expansion away from the molecular cloud towards us .
the region is in photoionization equilibrium with a density in concert with that of the background photodissociation region ( @xcite ) , and both regions are dusty with high depletions of the refractory elements ( @xcite ) .
bautista , pradhan , & osterbrock ( 1994 ; hereafter bpo ) , bautista & pradhan ( 1995 ; hereafter bp95 ) , and bautista , peng , & pradhan ( 1996 ; hereafter bpp ) have recently suggested that [ ] and [ ] lines are produced by a warm ( @xmath2 k ) region with very high electron density ( @xmath3 @xmath1 ) and a solar iron abundance .
this would be surprising for such a dusty environment , but would have important implications for grain destruction mechanisms . here
we use better spectroscopic data and new photoionization calculations to reexamine these claims .
bp95 s claims were based in part on their analysis of the [ ] intensities reported by osterbrock , tran , & veilleux ( 1992 ; hereafter otv ) and baldwin et al .
( 1991 ; hereafter bfm ) . in these low - resolution spectra the orion [ ] lines
were completely blended with the telluric emission which varies with time and position on the sky . in the case of [ ] 5577 , the telluric emission is much stronger than nebular emission , and no careful subtraction could yield an accurate measurement of this line .
both bfm and otv assigned large uncertainties to their measurement of [ ] 6300 , and did not claim any significant measurement of [ ] 5577 .
we have two sets of observations which dramatically improve on these measurements .
first , we used the faint object spectrograph on the _ hubble space telescope ( hst ) _ with the 0.86 circular aperture and the g570h grating on 1995 october 23/24 ( ut ) .
we made 225 s integrations at positions 1sw ( @xmath4 , @xmath5 ) = ( @xmath6 , 05@xmath723415 ) and x2 ( @xmath8 , 05@xmath723575 ) , roughly 30 wsw and s , respectively , from @xmath9 ori c. the summed spectrum is presented in figure 1 .
line strengths were measured with gaussian fits to the profiles and are presented in table 1 where they are given relative to 6678 and are uncorrected for interstellar extinction .
we note that the _ hst _ spectra are not affected by telluric emission .
the diode glitch between [ ] 5538 and [ ] 5577 in the flat fields does not impact our analysis .
also , a spectrum was taken with the cassegrain echelle spectrograph on the 4 m blanco telescope at ctio , on 12 january 1996 .
this provided an 11 km s@xmath10 resolution spectrum at a time when the earth s motion split the nebular lines from the telluric emission .
the spectra were extracted over a slit width of 1 and length of 12.5 , centered on the bfm position 1 at 37 w of @xmath9 ori c. the line intensities are in table 1 and parts of the spectra are shown in figure 1 .
orion has a spectrum with a strong point to point variation ( e.g. , @xcite ) causing the differences in the two spectra .
the ctio spectrum at the bfm position has the higher ionization .
the @xmath11 spectrum is similar to that reported by otv for another position in the nebula .
table 2 gives reddening corrected intensities relative to h@xmath12 , using bfm s extinction for the ctio data ; for our _ hst _ data , we adopt c(h@xmath13 from the composite h@xmath14/h@xmath12 ratio .
the ctio data in table 2 were renormalized using the 6678/h@xmath12 intensity ratio reported by bfm .
limits to the [ ] ratio @xmath15 are needed to set @xmath16 for an assumed @xmath17 . in neither spectrum was [ ] 5577 detected , but we have obtained a much stricter upper limit .
the observed limits are @xmath18 ( ctio ) and @xmath19 ( _ hst _ ) and the reddening corrected ratios are @xmath20 and @xmath21 .
the superior spectral resolution is the primary reason the ctio spectrum has the more stringent limit .
this limit is very different from the ratio of 22 adopted by bp95 , and shows that warm dense gas does not contribute to the [ ] spectrum .
the ctio [ ] 6716/6731 ratio ( 0.53 ) indicates a @xmath17-insensitive density of @xmath22 @xmath1 ( @xcite ) .
the [ ] 5755/6583 ratio ( 0.018 ) indicates @xmath17 of 10,100 k , fairly independent of @xmath16 in this range .
hst_-derived properties are similar . assuming @xmath23 k , the ctio [ ] ratio limit sets the limit @xmath24 @xmath1 , consistent with the @xmath16 [ ] and much lower than the density deduced by bp95 . assuming @xmath16 [ ] , we find that the [ ] ratio limit sets the limit @xmath25 k , consistent with the [ ] spectrum .
these are all conventional orion region numbers ; we conclude that the [ ] spectrum is consistent with its formation in gas of moderate density ( @xmath26 @xmath1 ) .
the photoionization calculations presented below predict the [ ] ratio to be @xmath27 .
table 3 shows the observed ( extinction corrected ) [ ] line ratios from otv and lowe et al .
( 1979 ) and some predictions .
the claim by bpo that the [ ] lines come from a high - density region in orion is based on a collisional model of the atom , which reproduces the observed optical to near - infrared line ratios at density @xmath28 @xmath1 .
table 3 shows the collisional model reproduced from table 2 of bpp ( @xmath29 @xmath1 and @xmath30 k , using new collision strengths described by bautista & pradhan 1996 , hereafter bp96 ) .
the predictions of the collisional model are sensitive to the adopted collision strengths .
our model i recalculates the line ratios for the same collisional model , using the publicly available collisional data from zhang & pradhan ( 1995 ; hereafter zp ) also used in our photoionization models below .
this shows that the lines 4815 , 5159 , 5262 , and 5334 are stronger by a factor of two than in the bp96-based model in bpp because of different collision strengths .
also shown in table 3 are line ratios we calculated for a density more typical of standard orion models , @xmath31 @xmath1 ( model ii ) . at this density , there is clearly a requirement for some mechanism other than simply collisions to populate the upper levels of the that produce the optical lines .
the alternative to collisional excitation , pumping of [ ] lines by the incident continuum , was considered by bpp ( see also lucy 1995 ) .
the equation used by bpp is correct in the optically thin limit .
however , the ultraviolet lines dominating the pumping are usually optically thick and line self - shielding must be included if energy is to be conserved ( ferland 1992 ) .
the correct formulation of the constant temperature problem outlined in bpp should include the column density as an additional free parameter .
to model the production of the [ ] lines more realistically , we incorporated a 371 level model atom into cloudy ( @xcite , @xcite ) .
collision data for our model atom are from zp , radiative data mainly from quinet , le dourneuf , & zeippen ( 1996 ) and nahar ( 1995 ) , and energy levels from johansson ( 1994 ) .
a cloud is divided into @xmath32 zones , the ionization and @xmath17 are determined self - consistently , and the emission is determined for the local optical depths .
we developed the code to model the broad line regions of quasars ( verner et al .
1995 , 1996 ) , incorporating the physics described by wills , netzer , & wills ( 1985 ) .
we included pumping by line overlap from all @xmath33 emission lines included in cloudy , by line overlap , and by the incident continuum .
we verified that the atom goes to lte in the high - density limit , and that in the low density limit each excitation is followed simply by radiative cascades . a simple limit of our model ( one zone , constant temperature , constant density , no continuum radiation flux , no line pumping , no interactions with other species ) corresponds to the bpo and bpp collisional model , and we can reproduce in this limit the line ratios presented in figure 2 of bpo , using their atomic data . note that figure 2@xmath34 of bpo erroneously compared the predicted @xmath35 ratio with the observed inverted @xmath36 ratio , although their caption for figure 2@xmath34 stated the ratio to be @xmath37
. tables 2 and 3 present results of our photoionization calculations .
model a is calculated with the parameters taken from bfm with improved gas - phase elemental abundances based on results from otv and rubin et al .
( 1991 , 1992 ) : h : he : c : n : o : ne : na : mg : al : si : s : cl : ar : ca : fe @xmath38 , grains , ionization by the central star @xmath9 ori c , and a hydrostatic blister
. results of model a differ from the original bfm model , since the underlying atomic database has been improved considerably .
model b has the same parameters as model a , but the incident continua from the other three trapezium stars are included ( an estimated 70% of the total flux at 2300 is due to @xmath9 ori c ) . in models
a and b , the gas pressure was kept constant .
model c is the same as model b , except that the total ( gas and radiative ) pressure is kept constant ; this reveals some sensitivity of the predicted spectrum to the assumed pressure law .
table 3 highlights the [ ] line ratios obtained from our photoionization models .
a comparison with the observed line ratios and with the line ratios calculated within the collisional model at the same density @xmath31 @xmath1 ( model ii ) indicates that the upper levels of the atom , those that produce the optical lines , can be effectively populated by continuum pumping .
the lower levels that produce the 8617 , 8892 , and 12567 lines are mostly populated by collisions in both photoionization and collisional models .
note therefore the sensitivity of the optical to near - infrared [ ] line ratios to the increased pumping from the other trapezium stars ( model b versus model a ) .
further evidence against high electron densities in the [ ] emitting region in orion comes from infrared observations by lowe et al .
they find that the 12567 line is the only line clearly detected in the infrared spectrum of orion it is actually the strongest [ ] line in the infrared and optical .
our models at @xmath39 @xmath1 predict that the 12567 line is @xmath40 times _ stronger _ than the 8617 line , in good agreement with the observations .
if the 12567 line were 23 times _ weaker _ than the 8617 line as predicted at @xmath41 @xmath1 , it would not have been detected at all .
table 2 includes the relative strength of [ ] 8617 compared to lines of other species , information crucial to abundance analysis .
our photoionization models match the observations in tables 2 & 3 quite well . the strongly depleted fe / h in the gas phase , @xmath42 , is close to the value @xmath43 found from kao observations of the ground state [ ] 22.9@xmath44 m line ( @xcite and in preparation ) .
even with solar abundances , iron is not a major coolant in orion , and so the intensities of [ ] lines scale linearly with abundance .
therefore , if solar fe / h ( bpo , bp95 ) were used instead , the predicted [ ] lines would be 11 times stronger and very discrepant with observations .
bp95 , using [ ] / [ ] line ratios , inferred that the fe / o abundance in the proposed high - density region is close to solar . based on our new upper limit on the [ ] 5577 strength , which is 3.2 times lower than one used by bp95 ,
the fe / o abundance would be higher than 3 times solar , if we adopt their high density .
if iron were depleted , these numbers would be even larger .
both photoionization and collisional models predict that the [ ] and [ ] emission line regions should be nearly coincident .
our upper limit on the density derived from the [ ] line ratio is consistent with the origin of [ ] lines in a region of moderate density , @xmath39 @xmath1 .
\1 . previous measurements of [ ] 5577 were affected by telluric emission . our limit to the line sets limits on the density and temperature consistent with the physical conditions inferred from the other emission lines .
the high - density collisional model for the formation of the [ ] lines is inconsistent with the strongest observed [ ] line at 12567 . in combination with the new upper limit on the [ ] 5577 line , the inferred fe / o ratio in any putative high - density region would be at least 3 times solar , not solar as found by bp95 .
our photoionization models of the [ ] spectrum with parameters close to those found by bfm predict a spectrum in good overall agreement with observations .
the infrared [ ] lines are produced by collisional excitation , and the optical [ ] lines by photon pumping by the ultraviolet continuum in a region of moderate density , @xmath39 @xmath1 .
fe / h must be depleted , @xmath45 , in agreement with independent determinations from kao observations .
the photoionization models have several parameters which can affect the resulting [ ] spectrum , including the flux of the pumping continuum , turbulent velocity , and bulk motions in the gas .
the shape of the ionizing continuum , the pressure law , and gas phase abundances are major parameters which affect the emission of the stronger cno lines .
detailed fitting of the orion nebula spectra will be the subject of a future paper .
we thank stsci for its support through grants go-4385 , go-5748 , go-6056 , go-6093 , and ar-06403.01 - 95a .
rjd and gjf acknowledge nasa / ames research center interchange grants ncc2 - 5008 and ncc2 - 5028 , respectively .
research in nebular astrophysics at the university of kentucky is supported by the national science foundation through award ast 93 - 19034 .
we are grateful to d. osterbrock , l. lucy , a. pradhan and m. bautista for useful discussions .
llll & 5518 & & 0.106 & 5538 & 0.13 & 0.119 & 5577 & @xmath46 & @xmath47 & 5577 sky & 0.044 & & 5755 & 0.17 & 0.322 he
i & 5876 & 3.22 & 3.44 & 6300 & 0.19 & 0.313 & 6300 sky & 0.0071 & & 6312 & 0.49 & 0.565 si ii & 6347 & 0.042 & & 6364 & 0.063 & 0.113 si ii & 6372 & 0.022 & & 6548 & 3.30 & 7.35 h@xmath14 & 6563 & 76 - 113 & 85.2 & 6583 & 11.4 & 21.0 he
i & 6678 & 1.00 & 1.00 & 6716 & 0.52 & 0.865 & 6731 & 0.98 & 1.75 lcccccc hei & 5876 & 1 & 13.7 & 13.3 & 13.3 & 12.9 & 5755 & 1 & 0.763 & 1.06 & 1.06 & 1.97 & 6583 & 1 & 41.6 & 53.1 & 52.8 & 74.1 & 5577 & 1 & @xmath48 & 0.0044 & 0.0043 & 0.0095 & 6300 & 1 & 0.722 & 0.341 & 0.336 & 0.699 & 3727 & 2 & 94 & 188 & 188 & 149 & 5007 & 2 & 343 & 465 & 460 & 379 & 6716 & 1 & 1.85 & 1.63 & 1.62 & 1.67 & 6731 & 1 & 3.47 & 3.19 & 3.17 & 3.57 & 9532 & 2 & 144.5 & 151 & 151 & 150 & 8617 & 3 & 0.0665 & 0.0572 & 0.0584 & 0.0979 cccccccccc 4244 + 4245 & 0.87 & 0.96 & & 1.3 & 0.23 & & 2.1 & 2.4 & 1.3 4277 & 0.64 & 0.60 & & 0.69 & 0.10 & & 0.46 & 0.50 & 0.37 4815 & 0.94 & 0.44 & & 0.96 & 0.21 & & 1.1 & 1.2 & 0.71 5158 + 5159 & 1.3 & 1.0 & & 2.7 & 0.62 & & 1.7 & 1.9 & 1.1 5262 & 0.81 & 0.71 & & 1.4 & 0.24 & & 0.28 & 0.31 & 0.26 5334 & 0.33 & 0.53 & & 0.95 & 0.11 & & 0.22 & 0.24 & 0.21 8617 & 1.00 & 1.00 & & 1.00 & 1.00 & & 1.00 & 1.00 & 1.00 8892 & 0.19 & 0.40 & & 0.40 & 0.34 & & 0.31 & 0.31 & 0.36 12567 & 3.3 & & & 0.39 & 3.6 & & 3.5 & 3.4 & 2.6 baldwin , j. a. , ferland , g. j. , martin , p. g. , corbin , m. r. , cota , s. a. , peterson , b. m. , & slettebak , a. 1991 , , 374 , 580 ( bfm ) balick , b. , gammon , r. h. , & hjellming , r. m. 1974 , , 86 , 616 bautista , m. a. , peng , j. , & pradhan , a. k. 1996 , , 460 , 372 ( bpp ) bautista , m. a. , & pradhan , a. k. 1995 , , 442 , l65 ( bp95 ) bautista , m. a. , & pradhan , a. k. 1996 , , 115 , 551 ( bp96 ) bautista , m. a. , pradhan , a. k. , & osterbrock , d. e. 1994 , , 432 , l135 ( bpo ) erickson , e. f. , rubin , r. h. , haas , m. r. , simpson , j. p. , & colgan , s. w. j. 1989 , baas , 21 , 1156 ferland , g. j. 1992 , , 389 , l63 ferland , g. j. , et al .
1996 , , to be submitted johansson , s. 1994 , private communication lowe , r. p. , moorehead , j. m. , & wehlau , w. h. 1979 , , 228 , 191 lucy , l. 1995 , , 294 , 555 nahar , s. n. 1995 , , 293 , 967 osterbrock , d. e. 1989 , astrophysics of gaseous nebulae and active galactic nuclei ( university science books : mill valley , california ) osterbrock , d. e. , tran , h. d. , & veilleux , s. 1992 , , 389 , 305 ( otv ) peimbert , m. , & torres - peimbert , s. 1977 , , 179 , 217 quinet , p. , le dourneuf , m. & zeippen , c. j. 1996 , , in press rubin , r. h. , dufour , r. j. , & walter , d. k. 1993 , , 413 , 242 rubin , r. h. , erickson , e. f. , haas , m. r. , colgan , s. w. j. , simpson , j. p. , & dufour , r. j. 1992 , in the astrochemistry of cosmic phenomena iau symposium no .
p. d. singh ( kluwer ) , p. 281
rubin , r. h. , simpson , j. p. , haas , m. r. , & erickson , e. f. 1991 , , 374 , 564 tielens , a. g. g. m. , & hollenbach , d. 1985 , , 291 , 722 verner , e. m. , .
ferland , g. j. , korista , k. t. , & verner , d. a. 1995 , baas 27 , 840 verner , e. m. , et al .
1996 , in preparation wills , b. j. , netzer , h. , & wills , d. 1985 , , 288 , 94 zhang , h. l. , & pradhan , a. k. 1995 , , 293 , 953 ( zp ) zuckerman , b. 1973 , , 183 , 863
|
we reexamine the spectroscopic underpinnings of recent suggestions that [ ] and [ ] lines from the orion region are produced in gas where the iron - carrying grains have been destroyed and the electron density is surprisingly high .
our new observations show that previous detections of [ ] 5577 were dominated by telluric emission .
our limits are consistent with a moderate density ( @xmath0 @xmath1 ) photoionized gas .
we show that a previously proposed model of the orion region reproduces the observed [ ] and [ ] spectrum .
these lines are fully consistent with formation in a dusty region of moderate density .
_ astrophysical journal letters _ , 1996 , in press
|
the photon production rates from bremsstrahlung and the @xmath17 processes considering the lpm effects are estimated by using eq.[photrate ] @xcite .
@xmath18 \left[n_f(q_0+p_0 ) \right .
\nonumber \\ & & \left .
( 1-n_f(p_0))\right]~\int \frac{\bf d^2\tilde{p}_\perp}{(2\pi)^2 } 2{\bf \tilde{p}_\perp \cdot\re\tilde{f}(\tilde{p}_\perp ) } \label{photrate}\end{aligned}\ ] ] @xmath19 \label{amy}\end{aligned}\ ] ] @xmath20 \label{amydelta}\end{aligned}\ ] ] @xmath21\left[n_f(q_0+p_0 ) \right .
\nonumber \\ & & \left .
( 1-n_f(p_0 ) ) \right]c_g g(x ) \label{gxprc}\end{aligned}\ ] ] @xmath22 @xmath23 in eq.[photrate ] is the real part of a transverse vector function which consists of the lpm effects due to multiple scatterings .
this can be taken as transverse momentum vector @xmath24 times a scalar function of transverse momentum @xmath25 .
the sign @xmath26 denotes the dimensionless quantities in units of debye mass @xmath27 as defined in @xcite .
the function @xmath28 is determined by the collision kernel ( @xmath29 ) in terms of the amy integral equation ( eq.[amy ] ) and the energy function @xmath30 ( eq.[amydelta ] ) .
this was solved in @xcite using the variational approach .
we simplified this variational approach and extended to finite baryon density case @xcite .
moreover , we reported that the complex lpm effects can be very well reproduced by introducing the photon emission function @xmath31 of dynamical variable @xmath32 @xcite . in terms of the single variable function @xmath31 , the photon emission rates are estimated as shown in eq.[gxprc ] . a figure for @xmath31 function for real photons
will be presented later ( fig.[gxbtgx ] ) .
processes that contribute to virtual photon emission in qgp at @xmath33 order @xcite and the higher order corrections @xcite were reported .
the processes @xmath34 , and @xmath35 contribute to photon polarization tensor at the order of @xmath33 and appear as the one loop processes in htl method . further considering lpm effects , photon emission from qgp as a function of photon mass
was also reported @xcite . for the case of virtual photons , these multiple scatterings modify the imaginary part of self energy as a function of photon energy and momentum both .
it has been shown that the rescatterings modify the self energy at low @xmath36 , more importantly around the tree level threshold , @xmath37 @xcite .
the dilepton emission rates are estimated in terms of the imaginary part of retarded photon polarization tensor , given by @xcite .
in all the calculations that follow , we have used two flavors , three colors and @xmath79=0.30 and t=1.0gev . in eqs.[x0]-[x3 ] we define four dimensionless variables used in the following work . here
one may notice that the variable @xmath80 is the inverse of @xmath32 variable used in @xcite for real photon production .
@xmath81 are defined in the eqs.[itdef],[ildef ] which are obtained by integrating the @xmath82 distributions ( in ref.@xcite there was a 2 factor extra in the definition of @xmath83 as given by @xmath84 ) .
@xmath81 are the quantities required for imaginary part of polarization tensor and the rate equations .
@xmath85 in the above equations superscript @xmath86 represent bremsstrahlung or @xmath3 processes depending on the @xmath87 value used .
the subscripts @xmath88 represent contributions from transverse ( @xmath76 ) or longitudinal @xmath89parts . using the iterations data , we first obtained the integrated values of @xmath40 distributions as given by @xmath81 of eqs.[itdef],[ildef ] .
we define the emission functions @xmath12 in eq.[gbatl ] which are functions of variables @xmath90 .
these @xmath90 variables are given in eqs.[xbt]-[xal ] .
the @xmath12 are obtained from corresponding @xmath81 values by multiplying with @xmath91 coefficient factors which are given in eqs.[cbt]-[cal ] .
therefore , using the @xmath81 obtained from iterations and variational data , we constructed all the four @xmath12 in eq.[gbatl ] .
it should be noted that the quantities @xmath90 and @xmath91 listed in eqs.[xbt]-[cal ] are not definitions .
they are the results of a hard search for dynamical variables hidden in the solutions of amy and agmz equations . versus dynamical variable @xmath92 defined in the text .
the symbols represent the integrated values of @xmath40 distributions of about 500 cases of @xmath93 values .
the different colored symbols represent different photon energies .
, width=377 ] versus dynamical variable @xmath4 defined in the text .
the symbols represent the integrated values of @xmath40 distributions of about 500 cases of @xmath93 values and are transformed by suitable @xmath94 coefficient functions .
the different colored symbols represent different photon energies .
the red curve is an empirical fit given in the text.,width=377 ] versus dynamical variable @xmath5 defined in the text .
the required @xmath95 are the transformation coefficients and empirical fit ( red curve ) are given in the text.,width=377 ] versus dynamical variable @xmath6 defined in the text .
the required @xmath96 are the transformation coefficients and empirical fit ( red curve ) are given in the text.,width=377 ] versus dynamical variable @xmath7 defined in the text .
the required @xmath97 are the transformation coefficients and empirical fit ( red curve ) are given in the text.,width=377 ] @xmath98 @xmath99 figure [ btp2int ] shows the @xmath100 as a function of @xmath92 variable .
all the bremsstrahlung transverse mode data consisting of about 500 points are shown in the figure .
the figure exhibits some trends as a function of @xmath2 values , but still does not bring out the underlying dynamical scale hidden in the data . when plotted with respect to any other variables ,
the representation is much worse and no trend emerges from such plots . in figure [ btgx ] , we show the same data transformed to @xmath101 as a function of @xmath4 variable which is a combination of @xmath92 and @xmath80 variables .
the hidden scale is revealed in this plot , where data for all different @xmath2 values merged into a single curve .
colored symbols represent different photon energies .
this is the transverse mode bremsstrahlung emission function .
importantly , the crosses in the figure represent low @xmath36 data for transverse mode of the @xmath102 process , intentionally plotted over here and will be explained in the next figure .
we have obtained an empirical fits to this data as given by @xmath103 formula in eq.[btgxemp ] .
the need for the empirical fit will become clear later .
@xmath104 similar exercise for transverse @xmath102 is shown in figure [ atgx ] . however , we notice that the @xmath32 variable is different from the previous figure .
further , the @xmath95 that transform the integrated values into @xmath66 functions are very complex .
we have parameterized this data by an empirical function given in eq.[atgxemp ] .
one important thing to notice is for low enough @xmath105 ( possibly the data for @xmath106 ) , the data deviates from general trends .
these low @xmath36 data is shown by crosses in the figure .
it was noticed that the same data when transformed as required in fig.[btgx ] and plotted versus @xmath4 , they are very close to the bremsstrahlung function .
this is shown by crosses in the previous figure ( fig.[btgx ] ) .
therefore , the transverse mode emission functions for the bremsstrahlung and for the low @xmath36 @xmath102 are the same and are given by @xmath101(empirical )
. + the figure [ blgx ] shows the results for bremsstrahlung longitudinal component .
the corresponding @xmath96 coefficient functions and the empirical fits represented by @xmath107 are given in eqs.[cbl],[blgxemp ] .
similarly , the figure [ algx ] shows the results for * aws * longitudinal component .
the corresponding @xmath97 , @xmath108 are given in eqs.[cal],[algxemp ] .
+ versus dynamical variable @xmath32 taken from @xcite , represented by black symbols .
the half value of this @xmath31 in the inverse scale ( that is same as present @xmath80 variable ) is shown by blue symbols .
this inverse variable coincides with the present @xmath4 when @xmath109 .
the results for transverse mode of the virtual photon case versus @xmath4 variable is shown by the red curve ( same as in fig.[btgx]).,width=377,height=377 ] plotted as a function of @xmath110 for a photon energy of 50gev .
the transverse components of bremsstrahlung , @xmath3 , the insignificant contribution from longitudinal parts , and the @xmath111 processes contribution using the eq.[oneloop ] are all shown .
the purpose of the figure is to test our empirical method compared to the results of fig.3 of aurenche _ et .
al.,_@xcite shown by red circles.,width=377 ] contribution from eq.[oneloop ] , the results are shown by colored solid curves.,width=377 ] in figure [ gxbtgx ] , we show the photon emission function in terms of dynamical variable @xmath32 versus @xmath31 defined and shown in ref.@xcite for real photons .
we reproduce that data in the figure .
[ gxbtgx ] represented by black symbols .
the inverse of this @xmath112-variable of @xcite coincides with the present @xmath80 value and is related to the variables of the present study by @xmath113(of ref.@xcite)=@xmath114 .
this is the motivation for choosing the combination of @xmath80 and @xmath92 variables for transverse modes in the present study ; as @xmath115 , the dynamical variables of the transverse parts go over to real photon dynamical variable .
such a need is not there for longitudinal components .
further , the @xmath31 defined in @xcite is two times larger than the definition in the present work for transverse modes ( see factor 2 in @xmath116 in eq.[photrate ] and not in eqs.[itdef],[ildef ] ) .
therefore the rescaled @xmath117 versus @xmath118 ) for real photons is also plotted in the figure by blue symbols . for comparison ,
the results of our present work for @xmath8 versus @xmath4 are also shown in the figure by red curve .
this red curve is the same red curve shown in fig.[btgx ] , given by @xmath119 .
it should be noted that in this figure , the variable for @xmath120 is different for all the three curves .
further , we explained earlier , that for @xmath121 , the transverse component of the @xmath3 data lies very close to the bremsstrahlung curve ( see crosses in the figure [ btgx ] ) . from these results
it is obvious that for small quark momenta and low virtuality of the photons emitted , the transverse part of the virtual photon ( @xmath122 ) emission function and real photon ( @xmath123 ) emission functions are the same as given by , @xmath124 and this @xmath32 for real photons is simply @xmath125 .
further , the other longitudinal components for virtual photons are all negligible .
therefore , for high photon energies and small photon virtualities , the virtual photon emission is identical to real photons emission in the sense that their emission functions are the same .
the imaginary part of retarded photon polarization tensor can be calculated using the @xmath78 integrated values as shown in eq.[impolar ] . in the previous section
, we used the results from variational and iterations methods to obtain the @xmath81 values by integrating the distributions .
we transformed these into @xmath12 functions shown in figs.[btgx]-[algx ] and we fitted these by empirical functions eqs.[btgxemp]-[algxemp ] .
the need for fitting arises , for example , for any @xmath93 values , we can generate the integrated values @xmath126 as given by eq.[ibaemp ] , without solving integral equations . the imaginary part of retarded photon polarization tensor ( in units of @xmath127 and represented by @xmath128 ) is calculated , as in eq.[impolargx ] . note that in this section , no data from variational method or iterations is used .
the empirical fit functions provide the required results circumventing the need to solve the transverse and longitudinal integral equations .
the imaginary part of @xmath129 has @xmath130 factor for longitudinal part and this diverges as @xmath131 .
this trend is seen in the figs.[imself50md1],[imself2md1 ] as @xmath36 increases . in this eq.[impolargx ] , one should note the multiplying factor @xmath132 , in the present work it is equal to @xmath133 .
importantly , one should note the factor @xmath77 in the longitudinal part .
this is because the @xmath69 is already rescaled in eq.[agmz - l - tilde - mod ] . in eq.[impolargx ]
the superscript @xmath134 denotes @xmath135 bremsstrahlung and annihilation depending on the value of the integration variable @xmath87 .
all terms in this equation are calculated and plotted in fig.[imself50md1 ] to show the contribution to polarization tensor .
@xmath136 \otimes \nonumber \\ & & \int \frac{d^2{\bf \tilde{p}_\perp}}{(2\pi)^2}\left [ \frac{p_0 ^ 2+r_0 ^ 2}{2(p_0r_0)^2 } \re{\bf \tilde{p}_\perp.\tilde{f}(\tilde{p}_\perp ) } + \right . \nonumber \\ & & \left .
\frac{1}{\sqrt{\left|p_0r_0\right|}}\frac{q^2}{q^2 } \re \tilde{g}({\bf\tilde { p}_\perp})\right ] \label{impolar}\end{aligned}\ ] ] @xmath136 \otimes \nonumber \\ & & \left(\frac{2}{\pi}\right ) \left [ \frac{p_0 ^ 2+r_0 ^ 2}{2(p_0r_0)^2}\left(\frac{g^i_t \left(x^i_t\right)}{c^i_t}\right)+ \right .
\nonumber \\ & & \left .
\left(\frac{1}{m_d}\right ) \left(\frac{g^i_l \left(x^i_l\right)}{c^i_l}\right ) \right ] \label{impolargx}\end{aligned}\ ] ] the contribution of oneloop processes to imaginary part of the photon polarization tensor is given by equation below ( formula taken from @xcite ) .
we will use this formula to evaluate the contribution of one loop processes and add to our empirical result to obtain the total result for retarded photon polarization tensor .
@xmath137 \label{oneloop}\end{aligned}\ ] ] figure [ imself50md1 ] shows the @xmath128 plotted as a function of @xmath110 for a photon energy of 50gev .
this figure has been compared with the results of fig.3 of @xcite represented by red symbols ( approximately ) for testing our empirical approach .
it can be seen that the empirical approach has predicted rather remarkably well the the results of @xcite over the @xmath110 in the range of @xmath138 .
further , the polarization tensor increases rapidly with @xmath110 beyond 100.0 the bremsstrahlung is insignificant obviously because the @xmath139 is very high .
the transverse component of @xmath3 only contributes , with a small contribution from longitudinal component above @xmath140 .
the @xmath111 processes contribution is calculated using the eq.[oneloop ] .
the modification of the imaginary part of photon polarization tensor due to multiple re - scatterings in the medium has been well discussed in @xcite .
the lpm effects only marginally increase the @xmath128 at low @xmath36 . here
@xmath141 processes also contribute .
importantly , the rescattering corrections smooth out the discontinuity at the threshold @xmath142 . in order to understand how @xmath128 behaves with photon energy , we have calculated this quantity using empirical approach for five photon energies .
these results are shown in figure [ imself2md1 ] .
the dashed curves are the results from present calculations .
when @xmath141 contributions are added to this , we get the solid curves .
the photon emission rates from the quark gluon plasma have been studied as a function of photon mass , considering lpm suppression effects at a fixed temperature .
self - consistent iterations method and the variational method have been used to solve the amy and agmz integral equations .
we obtained the @xmath143 , @xmath144 distributions as a function of photon mass , photon energy and quark momentum .
the integrated values of these distributions for bremsstrahlung and @xmath3 show a very good scaling in terms of different dynamical variables for these two processes .
we defined four new dynamical variables @xmath145 for transverse and longitudinal components of bremsstrahlung and @xmath3 mechanism .
in addition , we defined four new emission functions namely @xmath101,@xmath146,@xmath147@xmath148 , that describe the photon emission for these four components .
we have obtained empirical fits to these emission functions that have been constructed using the exact numerical calculations . in terms of the new emission functions , estimation of the imaginary part photon polarization tensor
reduces to one dimensional integrals . using this empirical approach
, we have calculated the imaginary part of retarded photon polarization tensor as a function of photon energy and mass .
these self energy graphs inlcude the lpm effects due to rescattering in the plasma medium . for small photon @xmath36 values and high photon energies ,
the virtual photon emission function is a single function coinciding with the real photon emission function .
thomas peitzman and markus h. thoma , hep - ph/0111114 . charles gale l. haglin , and [ hep - ph/0306098v3 ] ; charles gale , hep - ph/0512109v2 r. rapp , [ hep - ph/0204003v1 ] j.i .
kapusta , p. lichard , d. seibert , phys .
d 44 , 2774 ( 1991 ) .
r. baier , h. nakkagawa , a. niegawa , k. redlich , z. phys .
c 53 , 433 ( 1992 ) .
e. braaten , r.d .
pisarski , nucl .
b 337 , 569 ( 1990 ) .
p. aurenche , f. gelis , r. kobes and e. petitgirard , phys . rev . *
d54 * 5274 ( 1996 ) [ hep - ph/9604398 ] ; z. phys .
@xmath149,315 ( 1997 ) [ hep - ph/9609256 ] .
p. aurenche , f. gelis , r. kobes and h. zaraket , phys . rev . *
d58 * 085003 ( 1998 ) , [ hep - ph/9804224 ] ; * d61 * 116001 ( 2000 ) [ hep - ph/9911367 ] p. aurenche , f. gelis , and h. zaraket , jhep 0207 , 063 ( 2002 ) ; * d62 * 096012 ( 2000 ) [ hep - ph/0003326 ] peter arnold , guy d. moore and laurence g. yaffe , jhep 11 ( 2001 ) 057 , [ hep - ph/0109064 ] .
peter arnold , guy d. moore and laurence g. yaffe , jhep 12 ( 2001 ) 009 , [ hep - ph/0111107 ] l.d .
landau , i.ya .
pomeranchuk , dokl .
ssr 92 , 535 ( 1953 ) .
landau , i.ya .
pomeranchuk , dokl .
ssr 92 , 735 ( 1953 ) .
migdal , phys .
103 , 1811 ( 1956 ) .
s. v. s. sastry , hep - ph/0208103 .
s. v. s. sastry , phys . rev .
* c67 * , 041901(r ) ( 2003 ) , [ hep - ph/0211075 ] t. altherr , p.v .
ruuskanen , nucl .
b 380 , 377 ( 1992 ) . m.h .
thoma , c.t .
traxler , phys .
d 56 , 198 ( 1997 ) , [ hep - ph/09701354 ] p. aurenche , f. gelis , guy d. moore and h. zaraket , jhep 12 ( 2002 ) 006 , [ hep - ph/0211036 ] .
p. aurenche , f. gelis , and h. zaraket , jhep 05 , 043 ( 2002 ) s.s.v .
suryanarayana dae - brns nucl .
symposium , * 47b * , 448 ( 2004 ) .
|
the landau - pomeranchuk - migdal effects on photon emission from the quark gluon plasma have been studied as a function of photon mass , at a fixed temperature of the plasma .
the integral equations for the transverse vector function ( @xmath0 ) and the longitudinal function ( @xmath1 ) consisting of multiple scattering effects are solved by the self consistent iterations method and also by the variational method for the variable set \{@xmath2 } , considering the bremsstrahlung and the @xmath3 processes .
we define four new dynamical scaling variables , @xmath4 , @xmath5 , @xmath6 , @xmath7 , for bremsstrahlung and * aws * processes and analyse the transverse and longitudinal components as a function of \{@xmath2}. we generalize the concept of photon emission function and we define four new emission functions for massive photon emission represented by @xmath8 , @xmath9 , @xmath10 , @xmath11 .
these have been constructed using the exact numerical solutions of the integral equations .
these four emission functions have been parameterized by suitable simple empirical fits . in terms of these empirical emission functions ,
the virtual photon emission from quark gluon plasma reduces to one dimensional integrals that involve folding over the empirical @xmath12 functions with appropriate quark distribution functions and the kinematic factors . using this empirical emission functions , we calculated the imaginary part of the photon polarization tensor as a function of photon mass and energy .
quark gluon plasma ( qgp ) state is expected to be formed in the relativistic heavy ion collisions . in order to identify the plasma or a de - confined state
, one needs to study the physical processes in quark matter , that can distinctly and conclusively identify this state of matter , such as parton energy loss leading to jet suppression mechanism . in this context , electromagnetic processes such as photons and dilepton emission are also considered as important signals .
photons and dileptons are emitted at various stages during plasma evolution , for an overview one may see @xcite and the references therein . in depth
study of photon emission processes in quark - gluon plasma were reported @xcite including processes also from hot hadron gas @xcite . in the formalism of hard thermal loops @xcite ( htl )
effective theory , the processes of bremsstrahlung @xcite and a crossed process of off - shell annihilation called @xmath13 @xcite contribute to photon emission at the two loop level .
these two processes contribute at the leading order @xmath14 owing to the collinear singularity that is regularized by the effective thermal masses .
higher loop multiple scatterings having a ladder topology also contribute at the same order as the one and two loop processes @xcite .
these higher loop rescatterings , each giving finite decoherent correction to the two loop processes , need to be resummed .
this resummation effectively implements the landau - pomarenchk - migdal ( lpm ) effects @xcite arising due to rescattering of quarks in the medium during the photon formation time .
this results in an integral equation of a transverse vector function for the real photons .
the photon emission rates are suppressed owing to the lpm effects @xcite .
the lpm modification of the photon spectrum is important at very low photon energies or at high photon energies .
for example , the bremsstrahlung radiation is strongly suppressed by almost @xmath15 at very low @xmath16 values , whereas the photon emission from @xmath17 falls strongly for higher @xmath16 values @xcite .
thus the rescattering corrections modify the two loop contributions for bremsstrahlung and @xmath3 processes at opposite ends of the photon spectrum for real photons .
|
ever since the pioneering work of lin & shu ( 1964 , 1966 ) many theoretical models have been proposed to explain the existence of spiral structure in disk galaxies .
many are based upon the lin - shu hypothesis ( e.g. lin , yuan & shu 1969 ; roberts 1969 ; roberts , roberts & shu 1975 ) which describes spiral patterns as density waves , which cause compression of the gas component as it flows through the arms leading to subsequent star formation .
modal theories ( bertin et al .
1989a , 1989b ; bertin & lin 1996 ) are more complex density wave models , which describe disks of galaxies as resonant cavities within which density waves of different modes can co - exist and interfere to produce a range of observed phenomena .
tidal models ( toomre & toomre 1972 ; kormendy & norman 1979 ) set up spiral structure through transient density waves , which are caused by the tidal field of a nearby neighbour .
bars can also potentially drive spiral structure ( sanders & huntley 1976 ) . in this case
the formation of arms is driven by the effect of the bar potential on the dissipative interstellar medium .
finally stochastic self - propagating star formation ( sspsf ) can describe flocculent structure . here
, short irregular arms are produced due to shearing of a part of the disk that has recently formed stars ( gerola & seiden 1978 ) .
sspsf alone can not describe grand - design structure , but some models ( e.g. sleath & alexander 1995 , 1996 ) use sspsf with a weak imposed density wave to describe such global spiral modes .
many of the recent advances in this area have arisen from a study of the atomic and molecular gas component in spiral galaxies , through studies of hi and co line emission . this has permitted the detailed mapping of both the distribution and velocity field of the gas in the spiral arms of nearby galaxies , e.g. m81 ( visser 1980 ) m51 ( tilanus & allen 1991 , 1993 ; rand 1993 ; nakai et al .
1994 ) , m100 ( knapen 1993 ; rand 1995 ) , ngc 3627 ( reuter et al . 1996 ) and ngc 6946 ( regan & vogel 1995 ) .
these studies find streaming velocities of gas through the spiral arms of order a few 10s of km s@xmath0 , and also find offsets between the peaks of the gas density and the old stellar population in the arms and the star formation as revealed by h@xmath1 emission .
all of these findings are in general agreement with the predictions of density wave theories , with the gas being compressed and shocked as it flows into the spiral arm , and subsequently forming stars downstream from the density peak of the spiral arm . however , these studies are only possible in the strongest arms of the nearest spiral galaxies , and hence yield little information on the global importance of density waves in the general population of disk galaxies . here
we undertake the complementary approach of calculating the strength of density waves as mapped out by the old stellar population of a large sample of galaxies , using near - ir k - band ( @xmath2 ) imaging . in this waveband
the old stellar population is observed ( rix & rieke 1993 ) , making this the best observational tracer of the stellar mass distribution .
also , k band measurements are less affected by extinction than measurements in the optical . in a previous paper ( seigar & james 1998 - hereafter paper i )
we described the properties of the bulges , disks and bars of a sample of 45 galaxies . in this paper
we describe how spiral arm structures can be extracted for these galaxies and determine quantitative parameters for them . in section 2
we describe the observations ; section 3 describes the data reduction and extraction of all of the spiral arm parameters used in the analysis ; section 4 then draws on these parameters in discussing , in turn , each of the theoretical models of spiral structure ; in section 5 we describe tests of models of star formation in spiral galaxy disks and in section 6 we summarise our findings .
a sample of 45 face - on spiral galaxies , selected using the nasa / ipac extragalactic database ( ned ) , was observed using the infrared camera ircam3 on the united kingdom infrared telescope ( ukirt ) .
ircam3 utilises a 256x256 array with a pixel size of 0.286 arcsec , and the observations presented here were made in the j and k bands .
the observing dates were 1995 february 5 & 6 , 1995 november 18 , 19 & 20 and 1996 march 22 & 23 .
each galaxy was observed for a total integration time of 1800 seconds in the k - band , and 540 seconds in the j - band .
all measurements presented here are taken from the k band images unless otherwise stated .
for flat - fielding purposes , sky frames were obtained using equal integration times on areas of nearby blank sky .
a dark frame with the same on - chip exposure time was taken both for the galaxy frames and the sky frames .
standard stars , taken from the ukirt faint standards list , were observed throughout all nights to provide photometric calibration .
this sample spans a range of hubble types classified as sa to sd , including barred and non - barred galaxies as from the third reference catalogue of bright galaxy ( de vaucouleurs et al 1991 ; hereafter rc3 ) .
the details of the sample are explained in paper i. the hubble types of the galaxies are listed in column 2 of table 1 .
in this section we describe the technical details of the extraction of quantitative descriptors of spiral structure in the 45 spiral galaxies .
these parameters will then be used in subsequent sections to test theories of spiral structure .
the rgasp package was used to fit ellipses to the isophotes in the reduced galaxy images .
the ellipticity of the outermost ellipses was then adopted and the prof routine , which takes medians around ellipses at increasing radii , was run keeping this ellipticity and position angle constant .
the rebuild facility in rgasp enables a simulated galaxy to be made up using the ellipticity and position angle and the median surface brightness calculated around these ellipses , and subtracting this rebuilt image from the original image removes the elliptically symmetric bulge and disk components .
this procedure leaves an image of the bar and spiral arms . plotting the surface brightness of the fitted ellipses as a function of semi major axis gives light
profiles , which were used to do a one - dimensional bulge - disk decomposition ( see paper i ) .
the images were next deprojected , to remove the effects of galaxy inclination on the derived bar and arm parameters .
each galaxy was rotated so that its major axis was parallel with the y - axis of the cartesian coordinate frame .
the minor axis of the galaxy was then stretched along the x - axis by a factor determined by the ellipticity of the galaxy , resulting in the minor axis becoming equal to the major axis .
the galaxies in the sample are either face - on or nearly face - on , so deprojection effects are small ( typically the ratio of the minor - axis to the major - axis , @xmath3 ) .
each of the deprojected difference images was converted from cartesian to polar coordinates using a method similar to that of elmegreen , elmegreen & montenegro ( 1992 ) .
the cartesian image was mapped onto the new polar grid by calculating the cartesian coordinates for each of the polar coordinate pixels .
the number of counts associated with the calculated x - y position was then put in that polar coordinate position .
the polar frame was divided into 540 steps of @xmath4 in the @xmath5 direction ( i.e. wrapping around the galaxy one and a half times ) , and into 600 in the radial direction where 1 division was equal to @xmath6 and @xmath7 is measured in pixels .
this gave us sufficient oversampling to avoid loss of resolution even in the outer parts of the polar grid , and displaying one and a half turns makes it easier to identify spiral arms as continuous features .
spiral structure is assumed to have the form of a logarithmic spiral , given by @xmath8 where @xmath9 is the number of arms , @xmath5 is the azimuthal angle , @xmath10 is a constant which defines the winding angle , @xmath7 is radius and @xmath11 is a scale length . the pitch angle @xmath12 is given by @xmath13 it can therefore be seen that @xmath14 i.e. the gradient of the arm in a @xmath5 vs @xmath15 plot .
this gradient was calculated by clicking along the ridge of the spiral arm with the cursor and fitting a straight line to these points .
if the pitch angle of the arm changed with radius then we calculated an average pitch angle for that arm . where galaxies had multiple arms the average pitch angle over all the arms was calculated .
the errors in the pitch angle are calculated from the errors in the straight line fit , which is calculated using the residuals between the line and the data points .
this is therefore dominated by the change in pitch angle with radius , in galaxies where this change is significant .
the calculated pitch angles for all of the galaxies in this sample are listed in column 3 of table 1 .
we have defined a new quantity , equivalent angle ( ea ) , which parametrises the strength of arms , and was also used for bars in paper i. it is defined as the angle subtended by a sector of the disk which contains the same amount of flux as does the spiral arm , within the same radial limits .
the advantage of using equivalent angle is that it is unaffected by seeing and other resolution effects unlike , for example , arm - interarm contrast ( see paper i for a more detailed discussion ) .
the polar coordinate images were used for the calculation of ea of spiral arms . using the gradient of the arms ( calculated whilst working out the pitch angle ) the arms were transformed into horizontal features , by shearing columns of pixels according to the calculated gradient . although this is not necessary here , it is necessary for calculating arm cross - sections ( see section 3.6 ) and it also makes the calculation of arm ea easier .
the arm was then divided into strips , each covering a small range in @xmath15 , and the equivalent angle was calculated for each strip , by dividing the arm flux by the flux from the rebuild image in the corresponding strip . in this way it was possible to calculate the equivalent angle as a function of radius for all of the galaxies in the sample .
the overall ea was calculated by taking the mean value of ea over the entire detectable range of the arm .
the cutoff radii for this calculation are taken where the ea falls below @xmath16 .
the overall ea values are listed in column 4 of table 1 .
figure 1 shows how arm ea varies with radius for four typical cases , ngc 1219 , ngc 3512 , ugc 11524 and ic 357 .
the start and end points are defined as in the previous section , except in the case of ic 357 where the starting point is at the end of the bar . using the plots described in the previous section , where the arms are transformed into horizontal features , it is simple to calculate profiles across spiral arms by summing the flux in rows , i.e. along the arms .
figure 2 shows two such cross - sectional profiles ( top row ) for representative galaxies , with their decompositions into symmetric ( middle row ) and antisymmetric ( bottom row ) parts . also shown in figure 2
is a 540@xmath17 cut of ic 1809 .
the profiles can also be used to derive fwhm arm widths , measured in degrees of azimuthal angle , which are listed in column 6 of table 1 .
the errors on these fwhm depend on the errors on the measured gradients ( i.e. pitch angles ) . with the error
typically @xmath185% on the pitch angle it is found that the error on the fwhm is typically @xmath1810% .
this error has the effect of moving the peak of the cross - section by up to 10% and broadening the arm cross - section by 10% and this leads to an error in the calculated symmetries of @xmath195% . dividing these widths into the equivalent angle from column 4 of table 1 gives a good indication of the arm contrast , or degree of non - linearity , averaged along the arm .
this is because ea is a measure of the flux in the arm with respect to the disk and the arm cross - section fwhm is a measure of the width of the arm in which this is contained .
the shock strength is dependent upon the arm strength and we also expect stronger shocks in narrower arms .
therefore the ratio of arm ea to arm fwhm is a plausible indicator of the relative shock strengths in spiral arms .
the decomposition into symmetric and antisymmetric parts was performed using @xmath20 where the first and second terms represent the symmetric and antisymmetric parts , respectively , with respect to the position angle @xmath21 of the major axis of the arm when plotted in @xmath5 vs @xmath15 coordinates ( ohta , hamabe & wakamatsu 1990 ) .
the value of @xmath21 was determined as being the data point that had the highest surface brightness .
if there were two data points near the peak with similar surface brightness , then the value of @xmath21 was taken to be half - way between these two points .
when the value of @xmath21 is changed by up to 2 pixels ( the typical error in determining the value of @xmath21 ) the asymmetries change by @xmath182% . we have measured the dominant modes of spiral structure in these galaxies using a simple fourier technique adapted from the method used by gonzlez & graham ( 1996 ) .
three 10pixel wide strips were taken along the @xmath15 coordinate of the @xmath5 vs @xmath15 frame .
these were binned up to create a one - pixel - wide strip suitable for analysing with a 1-d fast fourier transform ( fft ) routine .
the effect of smearing of the arms for a 10 pixel wide strip is small .
an fft of each of the strips was then taken , resulting in three power spectra , which were summed to increase the signal - to - noise ratio . as the fft is taken over @xmath22 , the power in all modes up to @xmath23 is calculated , although we are only interested in modes up to @xmath24 .
we average all the modes from @xmath25 to @xmath26 and quote the power of individual modes as a ratio with this average , in order to overcome any systematic noise effects ( see section 3.7.2 ) .
the strongest mode for each galaxy in the sample is listed in column 6 of table 1 .
the relative strengths of all the modes are listed in table 2 .
figure 3a shows a histogram of the occurrence of the strongest mode in our sample with a subset in black indicating this for galaxies with near neighbours ( see section 4.4 for a definition of near neighbours ) .
figure 3b shows the same histogram with a subset for the strongly barred galaxies , where a strongly barred galaxy is defined as a galaxy with a bar overall equivalent angle greater than the average for the whole sample ( average bar ea=19.7@xmath17 ) .
these histograms will be discussed in section 4 .
fourier analysis can be subject to effects from noise , which may alter relative mode strengths ( gonzlez & graham 1996 ) .
we have investigated this problem in two ways , the first of which was to add a simulated noise frame to a @xmath5 vs @xmath15 galaxy image , thereby doubling the pixel - to - pixel noise . here , the image of ic 1764 was used .
we found that the mode strengths were changed by @xmath27 even under this significant noise increase .
for the second test , we simulated a strip of constant radius across a @xmath5 vs @xmath15 image by adding together sine waves of different frequencies and introducing noise to this model .
the input mode strengths were typical of those of our images . even for models with pixel - to - pixel noise greater than our noisiest image
, the fft still finds the low order modes to have approximately the same power as was input .
thus it can be assumed that any noise has a small effect on the fourier parameters presented here .
having described the determination of all the necessary parameters from the near - ir images , we will now investigate the implications for various models of the formation of spiral structure in galaxy disks .
this section is organised with each mechanism being discussed in turn , to see whether the new measurements support the mechanism as a strong contributor to the spiral structure .
it should be noted that the various mechanisms discussed are by no means mutually exclusive , and it may be that several or all of them contribute at some level to the structures seen in spiral galaxy disks . a fundamental prediction of the lin
shu hypothesis , and a central feature of the hubble classification scheme , is that galaxies with greater mass concentration in their central bulges should have more tightly wound spiral structure .
this can be demonstrated by the following back - of - envelope calculation .
lin & shu ( 1964 ) derive the following expression for the locus of the spiral pattern , @xmath28}{(2\pi g\mu_{0})}dr\ ] ] where @xmath5 is the azimuthal angle , @xmath29 is the epicyclic frequency , @xmath30 is the fraction of mass in the disk , @xmath31 is the pattern speed and @xmath32 and @xmath33 are the real and imaginary parts , respectively , of the material speed .
using equations 5 and 2 , @xmath34}{2\pi g\mu_{0}}\ ] ] equation 6 contains the desired result , that galaxies with more dominant disks should have spiral structure that is more loosely wound .
roberts et al .
( 1975 ) converted this into a correlation between pitch angle and hubble type , which they assumed to be essentially determined by bulge - to - disk ratio .
they calculated the theoretical winding angle from the bulge - to - disk ratios of a sample of galaxies and plotted this versus hubble type .
the line drawn in figure 5 is their predicted correlation .
roberts et al .
( 1975 ) also state that although the pitch angle has some dependence on the dynamical properties of the disk , e.g. @xmath35 and @xmath36 in equation 6 above , its dependence upon the fraction of mass in the disk is stronger .
the dynamical properties , especially the material speed , seem to correlate more strongly with the luminosity class ( van den bergh 1960 ) of the galaxy than with hubble type .
kennicutt ( 1981 ) found only weak correlations between pitch angle and hubble type , and between pitch angle and bulge - to - disk ratio .
he also found a correlation between pitch angle and the maximum rotational velocities of spiral galaxies .
this seems inconsistent with spiral density wave theories as it shows that arm strength reflects the effects of differential rotation .
kennicutt ( 1982 ) found that arm width increases smoothly with galaxy luminosity , with a slope consistent with broadening by galactic rotation ( i.e. galaxies with larger rotation speeds have broader arms ) .
this is consistent with the luminosity class ( van den bergh 1960 ) being dependent upon dynamical properties of galaxies ( roberts et al .
kennicutt & hodge ( 1982 ) measured pitch angles of 17 galaxies from the sample of roberts et al .
( 1975 ) using h@xmath1 data .
they found that their measured pitch angles correlated well with the roberts et al .
( 1975 ) model pitch angles , but that the model systematically underestimated all the pitch angles .
these authors work was performed using b band and h@xmath1 data which is strongly affected by young stars , and so we have performed this using k band data , which is known to have a significantly different morphology from optical wavebands ( e.g. block et al .
1994 ) .
we can directly test the important prediction of a correlation between mass concentration and winding angle using our images .
the pitch angle ( equation 2 ) quantifies how tightly wound spiral structure is , where small pitch angles imply tightly wound structure . to parametrise the central mass concentration we used the fraction of k - band light in the disk ( @xmath37 ) , calculated by numerical integration over the profiles , from the galaxy centre to beyond the edge of the observed galaxy , for both bulge and disk .
this parameter is listed in column 7 of table 1 .
( we use this parameter rather than the more widely used bulge - to - disk ratio for consistency with lin & shu 1964 . ) a plot of pitch angle versus the fraction of light in the disk ( shown in figure 4 ) shows no correlation of the type predicted by lin & shu ( 1964 ) .
figure 5 shows a plot of pitch angle against hubble type , which also lacks any correlation and is in very clear disagreement with the line derived from the predictions of roberts et al .
( 1975 ) .
there is an even simpler test which demonstrates that spiral arms can not have the fixed form indicated by equation 6 .
the lack of any @xmath5-dependence predicts that the arms in any spiral galaxy should have the same pitch angle at any given radius , in the lin - shu model .
this can be tested by looking at @xmath5 vs @xmath15 images of galaxies , where the pitch angle can be determined directly from the gradient of the arms .
this test is completely independent of the radial variation of dynamical properties .
figure 6 shows a contour plot of the galaxy ngc 2503 , demonstrating that the two arms have very different gradients , and therefore pitch angles , at the same radius .
this occurs in 20 out of the 45 galaxies in our sample , and is also seen clearly in the @xmath5 vs @xmath15 image of m99 presented by gonzlez & graham ( 1996 ) .
of these 20 galaxies , 9 have strong m=1 modes , i.e. a degree of lopsidedness , which can produce variations in pitch angle .
this is seen in galaxies which have a bulge which is offset from the centre of the disk as seen in ugc 3900 ( see figure 8) .
both this and the fact that pitch angle does not correlate with hubble type or b / d ratio seem to suggest that the simplest form of density wave theory is at least incomplete . indeed , it is not clear that a superposition of spiral waves , as postulated by modal theories , can produce this type of azimuthal variation in arm properties , which is most easily reconciled with sspsf models .
we find the lack of any correlation between spiral winding angle and hubble type quite surprising .
in paper i we found that hubble type correlates only weakly with bulge - to - disk ratio ( b / d ) , and we now find no correlation with arm morphology , at least in the near - ir , raising the question of what does determine hubble type .
other recent studies have found that the bulge surface brightness correlates strongly with hubble type ( de jong 1996b ) suggesting that only bulge parameters are important in determining hubble type .
another possibility is that the mass of cold gas plays a key role in determining galaxy type .
this is consistent with arguments of block & wainscoat ( 1991 ) and block et al .
( 1994 ) who find that the population i disk of young stars , which dominates the optical appearance , can be largely decoupled from the distribution of older stars .
this can explain the discrepancy between our findings and the optical study of kennicutt ( 1981 ) .
( note , however , that in section 5.1 we will report reasonably good agreement between arm morphologies for 3 of our galaxies imaged in both the b and k bands , so this decoupling can not be complete . ) we now consider cross sections through spiral arms illustrated for some of the more strongly armed galaxies in figure 2 , and the implications of these for density wave theories .
figure 2 shows that arm cross - sections for strong arms are typically asymmetric at only the 5 - 10% level , which is consistent with density wave theory , assuming that the k - band light traces the stellar mass density and is largely unaffected by extinction and the shocks which perturb the gas density . however , the lin
shu hypothesis predicts that density waves are sinusoidal in nature with a @xmath38 period , leading to a fwhm of @xmath39 for each arm of a 2armed spiral ( c.f .
@xmath40 for a one - armed mode , @xmath41 for 3armed , @xmath42 for 4armed etc ) .
the measured fwhm of the arm cross sections vary between @xmath43 and @xmath44 with typical values in the range @xmath45@xmath41 ( see column 6 of table 1 ) .
thus the arms are almost always narrower than expected for grand design 2armed structure , and in most cases @xmath4625% narrower than expected on the basis of the dominant fourier mode of the galaxy .
the typical error on the fwhm is @xmath1810% and the typical error on the asymmetries is 5% .
the 540@xmath17 cut of ic 1809 also demonstrates that arm cross - sections are not sinusoidal .
there are two possible explanations for this ,
the first of which is that the lin
shu hypothesis may be oversimplistic ( see section 4.2 ) .
the other explanation may be that the arms are narrower due to the effect of young super - giant stars forming near the peak of the density wave .
rix & rieke ( 1993 ) showed that the affect of supergiants is small for m51 , but rhoads ( 1998 ) has shown that supergiant stars contribute up to 30% of the local luminosities , which may have the effect of making arms appear somewhat narrower than predicted by the lin - shu hypothesis .
figure 15 shows k - band arm cross - sections for the galaxies for which we have b - band data .
these cross - sections appear less symmetrical than those shown in figure 2 .
however , the galaxies in figure 2 have strong spiral structure , whereas those in figure 15 have arms that are weaker and the profiles are thus somewhat noisier .
the asymmetries for these less strongly - armed spirals are at about the 25% level .
the narrowness of the arm cross - sections ( shown in figure 2 and described above ) can be explained by modal theory ( bertin & lin 1996 ; bertin et al .
1989a , 1989b ) , where higher order modes can be superposed on the @xmath47 mode resulting in narrower arms . the presence of such multiple modes can be tested using fourier decomposition of the disk structure .
although low - order modes are expected to be dominant , high - order modes should also be present but with weaker amplitudes .
it can be seen from figure 3a that the most common dominant modes in the sample are @xmath25 and @xmath47 , each of which dominates the disk structure of about one - third of the galaxies , whilst the @xmath48 mode accounts for most of the remainder ( a quarter of the total ) .
the @xmath49 mode dominates in only 4 out of 45 cases , and the relative weakness of the @xmath49 mode is probably the strongest result of this analysis .
rix & zaritsky ( 1995 ) performed a fourier analysis of 18 face - on spirals , imaged in the k band .
they found that about half of the galaxies in their sample had a strong two - armed spiral component with an arm - interarm contrast of about 2:1 and consequently strong @xmath47 and @xmath48 modes .
gonzlez & graham ( 1996 ) performed a detailed study of the galaxies m99 and m51 including a fourier analysis of both galaxies .
they looked at the dependence of @xmath25 to @xmath26 modes with radius and found that the @xmath47 mode was the strongest at all radii in the grand - design spiral m51 , and that m99 was dominated by the @xmath47 mode at most radii .
this was found to be true in both optical and k - band images of m99 .
analysis of a sample of seven galaxies in the k@xmath50-band ( @xmath51 ) reported by block et al .
( 1994 ) finds structure dominated by the @xmath25 and @xmath47 modes , with the lack of higher modes being attributed to damping of such modes in the old stellar disk by the inner lindblad resonance ( ilr ) .
we can therefore conclude that the frequent occurrence of the low order @xmath25 and @xmath47 modes is in agreement with predictions from modal theories , as is the suppression of the @xmath49 mode .
the dominance of the @xmath48 mode in 9 out of 45 cases , however , can not be explained by modal theories , which predict that such modes should be damped by the ilr .
the arm profiles of the galaxies ngc 1219 and ngc 3512 ( shown in figure 1 ) show some evidence for modulation ( bertin & lin 1996 ; bertin et al .
1989a , 1989b ) , i.e. peaks and troughs in the radial light profiles of arms of spiral galaxies .
these authors interpret modulation as the result of interference between different wave packets within the disk of the galaxy .
our measurement of the same effect in k - band images demonstrates that this is not simply due to extinction , but represents a true modulation in arm strength . of the 16 galaxies with sufficiently long and strong arms to test clearly for this effect , 4 show clear modulation and a further 6 show some indication of such effects .
these modulation effects are superimposed on arm profiles generally showing an increase in ea at small radii , which then turns over and decreases .
this implies that in the inner disk , the scale length of the arms is greater than that of the disk ( i.e. the arm surface brightness falls off less rapidly than the disk surface brightness ) and in the outer disk the scale lengths of arms are smaller than disk scale lengths .
this is in agreement with the result of gonzlez & graham ( 1996 ) who found a peak in the arm - interarm contrast at intermediate radii .
bars may be responsible for the driving of spiral structure ( sanders & huntley 1976 ) but sellwood & sparke ( 1988 ) have shown that that bar driving is only important for the most strongly - barred spiral galaxies .
in addition , it is thought that bar forcing should not extend far outside the region where the bar potential is strong ( sanders & tubbs 1980 ) .
we have tested models of bar - driven spiral structure by looking for a correlation between bar strength and arm strength .
the result of this can be seen in figure 7 , which shows a plot of arm strength in the inner part of the disk against bar strength .
there is no correlation .
it would therefore appear that bars do not strongly affect the strength of spiral arms , even in the central regions of disks .
figure 3b shows enhancement of the @xmath49 and @xmath48 modes for galaxies that we have classed as strongly barred , relative to the sample as a whole .
a kolmogorov - smirnov ( ks ) test showed that this result is not significant .
if this result is confirmed by analysing larger samples of galaxies in the same way , then it would be surprising as bar - driven models predict strong 2-armed spiral structure ( e.g. sanders & huntley 1976 ; sanders 1977 ) .
the relation between the morphology of bars and arms has been discussed by sellwood & sparke ( 1988 ) , who claim that bars and arm structures may have different pattern speeds .
if this were the case then galaxies would sometimes have arms starting ahead of the bar ends , although their simulations show that this should not happen very often .
they find that in 100 time steps of their model there are two steps for which the ends of bars and the beginnings of arms do not meet .
we have observed this behaviour in 2 out of 15 strongly barred galaxies . our image of ugc 3900 ( figure 8) shows quite evidently that its arms do not start at the end of the bar , and ic 568 shows a similar morphology .
the qualitative agreement between the appearance of these galaxies and the simulations is very good , but it may happen somewhat more frequently than predicted by sellwood & sparke ( 1988 ) .
kormendy & norman ( 1979 ) proposed an alternative explanation to spiral density waves for the maintenance of global spiral structure . noting that the ` spiral winding problem ' is only actually a problem in the differentially rotating parts of disks
, they postulated that grand - design structure may only occur in one of two situations .
firstly , some galaxies may have grand - design spiral structure only within the radius within which solid body rotation occurs , and thus completely avoid the ` spiral winding problem ' .
kormendy and norman then claim that galaxies exhibiting grand - design spiral structure outside this radius could be undergoing tidal interactions with other nearby galaxies , or be strongly barred . in these cases
the driving field from the neighbour or bar maintains spiral structure even in the regions of the disk subject to differential rotation ( although note our conclusion in section 4.3.1 ) .
studying a sample of 54 galaxies , they found that 45 exhibited global spiral patterns , and of these , only two clearly show global spiral structures in the differentially - rotating part of the disk , with no bars or neighbours to provide a driving field .
all others fitted into the proposed model .
observational support for this view has been provided by elmegreen , seiden & elmegreen ( 1989 ) who observed that , generally , grand - design spiral galaxies are found in richer environments than flocculent spirals ( i.e. grand - design spiral structure galaxies are more likely to have undergone a tidal interaction with another galaxy in their history ) .
we now test these results with our data .
kormendy & norman ( 1979 ) define a tidally - significant neighbour galaxy as one within 35 galaxy diameters and 1 magnitude ( v - band ) of the spiral galaxy in question .
we have relaxed this somewhat to 6 diameters because few of the galaxies in our sample meet the more stringent criterion , which is , in any case , somewhat arbitrary , particularly since there is time for galaxies to move apart between the time of interaction and the development of spiral structure .
we performed a search for near neighbours using the digitised sky survey and diameters ( b@xmath52 ) from the rc3 .
the apparent visual magnitudes of the near neighbours are comparable to those of the sample galaxies .
figure 9 shows a histogram of the distribution of arm strengths ( in terms of equivalent angle ) in the whole sample with the galaxies with near neighbours indicated in black .
it can be seen that 4 out of 7 galaxies with arm equivalent angle greater than @xmath53 have near neighbours whereas only 12 out of 45 galaxies in the sample have companions . whilst a ks test shows that this difference is not statistically significant due to the small numbers involved ,
there is some indication here of a relation between arm strength and tidal interactions .
this conclusion is strengthened by figure 3a which shows that the presence of near neighbours enhances the @xmath47 mode , with this being the dominant mode in 7 out of 12 galaxies with nearby neighbours , compared to only 7 out of 33 galaxies with no apparent neighbours .
figure 10 shows a histogram of the distribution of the total azimuthal angle over which the strongest arm of each galaxy in our sample can be detected , both for the whole sample and the galaxies with near neighbours ( black ) .
this figure shows no evidence of near neighbours enhancing global modes which would be represented by longer spiral arms subtending larger azimuthal angles .
it therefore seems that the presence of nearby galaxies may affect the strength of arms and enhance the @xmath47 mode , but they have no effect on the angular extent over which arms can be detected .
this provides some support for the model presented by kormendy and norman ( 1979 ) , but they also require driving by bars to be efficient at generating and maintaining global arm structures , and we find no evidence for this effect ( section 4.3.1 ) . kormendy & norman ( 1979 ) claim that a dominant bulge will prevent the formation of global spiral arms in galaxies due to the differential rotation caused by a strong central mass concentration .
we can test this by determining the azimuthal angle that arms subtend as a function of bulge - to - disk ratio .
this is plotted in figure 11 , which shows a hint of a deficit of global arm structures for galaxies with dominant bulges , i.e. no points lying in the upper right - hand corner of the plot .
this is consistent with the predictions of kormendy & norman ( 1979 ) .
toomre ( 1964 ) claimed that galaxies with heavier disks are more unstable to perturbations .
galaxies with heavier disks should have brighter disk central surface brightnesses and if these ` heavy disks ' were more unstable we would expect to obtain larger arm equivalent angles for them .
figure 12 shows a plot of arm equivalent angle against disk central surface brightness .
no correlation is shown but there does seem to be a lack of low - surface - brightness galaxies with large arm ea .
in this section we try to distinguish between the two main theories for the occurrence of star - formation in spiral galaxy disks .
the first of these is the large scale shock scenario ( roberts 1969 ) , which predicts that as gas moves through the density wave it becomes shocked and star formation results . in this case star formation rate
depends on the strength of the shock and thus should correlate with measured arm strength .
however , elmegreen & elmegreen ( 1986 ) claim that density waves can not trigger star formation because their contrast above the disk is not high enough .
they suggest that although density waves are responsible for organising material in spiral arms , some other mechanism must be responsible for forming stars .
sleath & alexander ( 1995 , 1996 ) suggest that this mechanism is stochastic self propagating star formation ( gerola & seiden 1978 ) . in this type of model ,
no strong correlation would be expected between arm strength and star formation rate in the disk .
we have used the iras @xmath54 and @xmath55 fluxes to calculate the far - infrared luminosity for the galaxies in our sample for which iras data was available .
the far - infrared luminosity can be used as a quantitative indicator of star - formation rates ( e.g. spinoglio et al .
this was divided by the k - band luminosity of the galaxies to compensate for differences in their overall size , and then compared with their overall arm strengths in ea normalised to the arm fhwm to see if shocks in the arms are triggering the formation of stars .
the far - infrared luminosity in terms of the @xmath54 and @xmath55 flux is given by lonsdale et al .
( 1985 ) as @xmath56 where @xmath57 is the far - infrared luminosity in solar units , @xmath58 is the distance to the galaxy in mpc , @xmath59 is the @xmath54 flux in jy and @xmath60 is the @xmath55 flux in jy .
figure 13 shows a plot of the far - infrared luminosity normalised to k - band luminosity against the overall ea to arm fwhm ratio .
the correlation coefficient is 0.79 at a significance level of 99.8% .
the error bars here are calculated from the errors on the iras fluxes found using ned .
ic 568 ( denoted by a hollow circle ) was found to have a starburst nucleus on the basis of a very red nuclear colour ( paper i ) , and this undoubtedly contributes to the high fir luminosity .
it was omitted from the calculation of the correlation coefficient given above . if the large scale shock scenario ( roberts 1969 ) is correct a strong correlation between arm strength ( or the strength of shocks in the arms ) and star formation rate would be expected .
knapen & beckman ( 1996 ) find both star
formation rates and star formation efficiencies are enhanced in the arms compared to the interarm regions of m101 and this supports the large scale shock scenario .
figure 13 shows that we also find a good correlation , in the expected sense .
additional b band images were obtained for the galaxies ngc 5478 , ugc 6958 and ugc 8939 .
the dominant source of blue light is young bright stars and so these data can be used to investigate the dependence of star formation on the underlying spiral density perturbation .
the b - band images were taken at the saao 1 m elizabeth telescope with a 512x512 tek8 ccd , and were reduced and analysed in an essentially identical manner to that for the k - band data .
here we compare the arm profiles , cross - sections and pitch angles derived from k - band images with the same parameters from b - band data table 3 shows a comparison for the overall arm strengths between the k band data and b band data . for the strongly armed galaxies ( ngc 5478 and ugc 6958 ) the k band arms are stronger than the b band arms
however , for ugc 8939 ( which has weak arms ) the b band arms are stronger than the k band arms . in our search for near neighbours we found that ugc 8939 has two near neighbours at a distance of about 2 galaxy diameters and with similar surface brightnesses to ugc 8939
. it may be that the companion galaxies are affecting the star - formation rate in ugc 8939 , thus causing its b band arms to appear stronger than its k band arms .
if the large scale shock scenario ( roberts 1969 ) and the lin - shu hypothesis ( lin & shu 1964 , 1966 ) are correct we expect gas to approach the arm from the concave side , assuming arms are trailing and that the area under consideration is within corotation .
the gas will then be compressed and , if the perturbation is strong enough , shocked at the centre of the arm , leading to subsequent star formation .
thus we should see star formation on the leading , convex edges of spiral arms , within the corotation radius .
observationally , this should result in b - band arms leading k - band arms . in order to test this
we have compared the azimuthal angles at the start of the arms , and the pitch angles , for the b - band and k - band images .
( note that the image of ugc 6958 was first flipped to make the arms spiral outwards in a clockwise direction , consistent with the other two galaxies . )
table 4 shows the results of this analysis .
figure 14 also shows a plot of the difference between the azimuthal angle of the k - band arm and that of the b - band arm for ugc 8939 as radius increases .
it can be seen that , generally , the b - band arms lead k - band arms by an average azimuthal angle of @xmath61 at low radii increasing to 10s of degrees at high radii .
this result and table 4 show that k - band arms are more tightly wound than b - band arms for ugc 8939 .
similar plots for ugc 6958 and ngc 5478 ( not shown ) , whilst noisier , show some tendency for k - band arms to have smaller pitch angles than b band arms .
this confirms the result of kennicutt ( 1981 ) who used b - band and r - band optical data .
this therefore seems to confirm the large scale shock scenario .
however , there is another possible explanation .
visible images show dust lying preferentially on the trailing edges of arms , which would also result in b - band arms appearing to lead k - band arms as a result of the increased extinction at b. however , the fact that dust lanes appear on the trailing edges of arms is in itself evidence for the large scale shock scenario ( roberts 1969 ) since dust lanes are conventionally taken as tracers of shocks .
figure 15 shows a comparison of b- and k - band arm cross - sections .
in general it can be seen that the fwhm of the cross - sections of b - band arms are much smaller than those of k - band arms
. there could be a number of factors contributing towards this . in the b - band ( or any visible waveband ) spiral arms
are defined by the narrow regions where stars are formed and this may be contributing to the low fwhm of b - band arms .
in addition , b - band arms are frequently observed to bifurcate into smaller , narrower arms .
finally , blue light is strongly affected by dust which lies predominantly on one side of spiral arms and would therefore have the effect of making b - band arms appear narrower than k - band arms .
we now summarise the implications of these observations for the main models of the formation and stability of spiral structure .
no one theory fits all of the observations , but a combination can explain most or all of the features we have observed .
this section summarises these conclusions .
simple density wave theories ( e.g. lin & shu 1964 , 1966 ) are undermined in three ways .
firstly , we do not get the correlation between the fraction of light in the disk and pitch angle which is predicted by lin & shu ( 1964 ) and roberts et al .
secondly , lin & shu ( 1964 ) predict that pitch angle should be the same for arms at the same radius within a given galaxy .
we find that this is not the case .
finally , we have found that the fwhm of arm cross - sections are somewhat narrower than predicted by simple density wave theories .
more complicated density wave theories , e.g. modal theory ( bertin et al .
1989a , 1989b ; bertin & lin 1996 ) , are supported by modulation effects that we have found in arm strength as a function of radius , and which appear to be fairly common .
bar driven models of spiral structure ( sellwood & sparke 1988 ; sanders & huntley 1976 ) are not supported by our observations , as we find no correlation between arm strength and bar strength , even in the inner part of the disk .
tidal effects from near neighbours seem to have some effect .
we find that 4 out of 7 galaxies with arm ea greater than @xmath53 have near neighbours whereas only 12 out of 45 galaxies in whole sample have near neighbours .
we also find that in the fourier analysis of galaxies with near neighbours , the even low - order modes are dominant , especially the @xmath47 mode .
we have also discussed factors affecting star formation in the disks of spiral galaxies .
we have found evidence both for and against the large scale shock scenario ( roberts 1969 ) .
firstly , we find a correlation between arm contrast and the normalised star formation rate , in agreement with the predictions of the large scale shock scenario .
secondly , we find that b - band arms are more loosely wound and narrower than k - band arms , confirming the predictions made by density wave theories ( roberts 1969 ) .
however , this can also be interpreted as an extinction effect because dust lies on the trailing edges of arms .
finally , and possibly most significantly , we find no correlation between the arm properties of this sample of galaxies and their classified hubble type . combining this with the poor correlation between hubble type and k - band b / d ratio ( paper i )
, it would appear impossible to allocate a hubble type to these galaxies on the basis of these k - band observations , and the morphology of the old stellar population varies surprisingly little between hubble types sa and sd .
we speculate that the determining parameter for hubble type is cold gas content , which controls the star formation rate and the distribution of the young stars , which dominate the optical appearance .
we thank paul lynam for the b - band images , taken at the 1 m elizabeth telescope at saao .
we also wish to acknowledge debra elmegreen and frederick chromey for useful suggestions , and john porter for his comments on an early draft of this paper .
the united kingdom infrared telescope is operated by the joint astronomy centre on behalf of the u.k . particle physics and astronomy research council .
this research has made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
we thank the referee for many helpful suggestions which improved the content and presentation of this paper .
figure 2 : arm cross - sections ( top ) , symmetric components ( central ) and antisymmetric components ( lower ) .
leading edges ( i.e. the convex sides ) of arms are plotted on the right .
a 540@xmath17 cut of ic 1809 is also shown ( bottom ) .
figure 3 : the occurrence of the strongest fourier modes found in the disks of the entire sample compared with ( a ) galaxies with companions within 6 galaxy diameters shaded in black ; ( b ) galaxies with bars stronger than the average overall bar strength shaded in black .
figure 6 : the galaxy ngc 2503 in polar coordinates .
@xmath62 is plotted in the x direction where @xmath7 is measured in pixels .
@xmath5 is plotted in the y direction in units of degrees , so the plot represents a turn and a half around the galaxy .
the arms appear straight ( with a constant gradient ) in this plot and their pitch angles can be measured directly from their gradients .
ngc 2503 demonstrates that pitch angle is not the same even for arms within the same galaxy .
figure 8 : greyscale unsubtracted and difference images of the galaxy ugc 3900 showing that its arms start ahead of the end of its bar and that its bulge is offset from the centre of the disk .
the axes are labelled in units of pixels , where 1 pixel = 0.286 arcsec .
figure 13 : far - infrared bolometric luminosity normalised to k - band magnitude versus arm contrast .
the point represented by a hollow circle is ic 568 , which we believe has a starburst nucleus that probably dominates its fir luminosity .
the line is an error weighted regression , ignoring ic 568 .
|
we have imaged a sample of 45 face - on spiral galaxies in the k - band , to determine the morphology of the old stellar population , which dominates the mass in the disk .
the k - band images of the spiral galaxies have been used to calculate different characteristics of the underlying density perturbation such as arm strengths , profiles and cross - sections , and spiral pitch angles .
contrary to expectations , no correlation was found between arm pitch angle and hubble type , and combined with previous results this leads us to conclude that the morphology of the old stellar population bears little resemblance to the optical morphology used to classify galaxies .
the arm properties of our galaxies seem inconsistent with predictions from the simplest density wave theories , and some observations , such as variations in pitch angle within galaxies , seem hard to reconcile even with more complex modal theories .
bars have no detectable effect on arm strengths for the present sample .
we have also obtained b - band images of three of the galaxies . for these galaxies
we have measured arm cross - sections and strengths , to investigate the effects of disk density perturbations on star formation in spiral disks .
we find that b - band arms lead k - band arms and are narrower than k - band arms , apparently supporting predictions made by the large scale shock scenario , although the effects of dust on b - band images may contribute towards these results .
galaxies : spiral - galaxies : structure - galaxies : fundamental parameters
|
the special features of turbulence in the presence of rotation have attracted the interest of many authors @xcite . relying on some experiments ,
the methods of study used have ranged from analytic approaches to numerical simulations @xcite .
the central theme in rotating fluids is the effect of the coriolis force , which induces anisotropy ( there is a preferred direction , that of the rotation axis ) .
this anisotropy is extreme in the limit of fast rotation , which actually forces the flow to become two - dimensional ( proudman - taylor theorem ) @xcite . in this work
we apply perturbation theory to the randomly forced navier - stokes equation with coriolis force as a model for the turbulent regime of a rotating fluid .
the perturbative study of the ordinary randomly forced navier - stokes equation , in combination with the renormalization group ( as an improvement of perturbation theory ) , has a long tradition @xcite .
the addition of the coriolis force , induced by the rotation of the fluid , introduces one additional parameter , the angular velocity @xmath0 or , in dimensionless form , either the rossby or ekman numbers ( in addition to the reynolds number ) @xcite .
let us focus on the ekman number , @xmath1 , that depends on the viscosity @xmath2 and a scale @xmath3 , roughly associated with the size of the fluid system .
the ekman number gives the relative importance of the viscosity and coriolis forces .
we will _ assume _ henceforth that for small @xmath0 ( @xmath4 ) the turbulence is isotropic and the only relevant parameter is the viscosity .
in this limit , the results of the study of the ordinary randomly forced navier - stokes equation hold ( the random force is always assumed isotropic ) . for larger @xmath0 we will encounter new features .
in fact , the only restriction on perturbative correction terms is that they respect the basic _ symmetry _ of the equations , in our case , the axial symmetry about the rotation axis .
we will see that perturbation theory generates new terms fulfilling these symmetry constraints .
therefore , one must find the complete set of allowed terms that can arise in perturbation theory
. we will determine all the terms that can be represented by the components of an axisymmetric `` viscosity tensor '' and , in particular , the ones that arise at first order in perturbation theory .
the part of this `` viscosity tensor '' that is pair - antisymmetric in the indices plays a significant role ; however , it does not lead to dissipation , and therefore , is not truly viscous . as in homogeneous and isotropic turbulence , we assume that the physical region of study is sufficiently far from the surfaces , where the boundary conditions are imposed , for them not to have any _ direct _ effect , except the presence of the scale @xmath3 .
in contrast to ordinary turbulence , this condition only implies that we can have homogeneous turbulence but , due to rotation , it can not be isotropic .
it is pertinent to mention here that the possibility of anisotropic forced turbulence , and precisely with axial symmetry , has already been considered @xcite . in this reference , however , the authors assume that the breakdown of isotropy occurs through a random force whose two - point correlation function depends on the anisotropy vector @xmath5 .
they derive a renormalized force proportional to second and fourth powers of @xmath5 . in our case
, we will see that the first perturbative correction is linear in @xmath6 , like the coriolis force itself .
this paper is organized as follows . in section [ sec : basic+perturbation ] we introduce the randomly forced hydrodynamical equations with rotation .
we assume that the fluid is incompressible and show how to formulate them as a problem of homogeneous but anisotropic incompressible turbulence .
fourier analysis of the turbulent velocity field is used to organize the perturbation expansion @xcite in subsection [ sec : fourier ] .
we introduce in subsection [ sec : response ] the linear response function . unlike the isotropic case
, the coriolis term leads to a _ non - symmetric _ linear response function matrix . in subsection
[ sec : perturbative ] we define the non - linear response function and present its perturbative expansion ( slow rotation ) .
we also compute the first order perturbative correction to the response function , which allows the identification of the ( rotation independent ) effective shear viscosity ( proportional to the cube of the reynolds number ) , and a new anisotropic force . in section [ sec : effective - tensor ] we write down the most general axially symmetric `` viscosity tensor '' , as the existence of a preferred direction , singled out by the fluid rotation , requires the introduction and use of axisymmetric tensors .
this rank four `` viscosity tensor '' expresses the proportionality between the fluid stress tensor and the rate of strain tensor . in isotropic and homogeneous incompressible turbulence ,
the viscosity tensor depends only on one parameter , the fluid shear viscosity . in the case of rotating turbulence , and for slow rotation , we find that the axisymmetric `` viscosity tensor '' depends on two parameters : the molecular shear viscosity ( coming from the isotropic terms of the `` viscosity tensor '' ) and a new one , that arises from the anisotropic terms in the `` viscosity tensor '' . we also show that this new parameter can be identified as the coefficient of the anisotropic force calculated perturbatively ( in the previous section [ sec : perturbative ] ) .
having thus established the equivalence between the perturbatively corrected randomly forced navier - stokes equation with coriolis force on the one hand , and a ( newtonian ) rotating incompressible fluid with an effective axisymmetric `` viscosity tensor '' on the other , we proceed , in section [ sec : effective ] , to discuss some physical consequences of the new terms in the perturbed fluid equations . in subsection
[ sec : axial - force ] we consider the quasi - local force induced by the anisotropic components of the `` viscosity tensor '' and show that it is proportional to the cube of the reynolds number , and that it does not lead to dissipation . in subsection
[ sec : waves ] we study the dynamical effects of this force on the propagation of inertial waves .
we end by discussing our results and proposing further work on the problem of rotating turbulence . in appendix
[ sec : app - diag ] we introduce the diagrammatic representation of the exact navier - stokes equation and the diagram encoding the first order correction to the response function and in appendix [ sec : app - perturbation ] we present the technical details needed to carry out the perturbative calculation in the slow rotation limit .
we start from the hydrodynamical equations for a fluid with density field @xmath7 , velocity field @xmath8 , pressure @xmath9 , and molecular shear ( tangential ) and bulk kinematic viscosities @xmath2 and @xmath10 , respectively .
we assume that the fluid is rotating with constant angular velocity @xmath11 along the @xmath12 axis and that it is subject to an isotropic random forcing _ per unit mass _ @xmath13 .
the mass and momentum conservation equations are @xmath14 { \vec \nabla}({\vec \nabla \cdot { \bf u } } ) -2{\bf \omega \wedge u } - { \bf \omega \wedge ( \omega \wedge x ) } + { \bf f } \ ; , \label{moment1}\end{aligned}\ ] ] where @xmath15 is the number of space dimensions
. the dimension of space @xmath15 will be kept as a free variable , although when we consider rotation - dependent expressions , these must be evaluated for @xmath16 .
the momentum equation ( [ moment1 ] ) is supplemented with a random stirring force that leads to a statistical distribution for the velocity field and can be used to model turbulent flows just as is done for isotropic randomly stirred ( non - rotating ) turbulence @xcite . regarding the random force spectrum and statistics , we take a gaussian random force that is white in time , for simplicity , but we allow for ( translation invariant ) spatial correlations .
so we can write @xmath17 where the angular brackets denote an average over the random force realizations . the spectral function for @xmath18
will be specified below .
we assume that the fluid is incompressible so that the density field is constant ( @xmath19 ) and @xmath20 . under this condition
we need only consider the equation for the conservation of momentum ( [ moment1 ] ) and write @xmath21 + \nu \nabla^2 { \bf u } - 2{\bf \omega \wedge u } + { \bf f } \ ; .
\label{u - field1}\end{aligned}\ ] ] notice that the force per unit mass @xmath13 will be taken solenoidal as well , that is @xmath22 , in order to avoid having a random component in the pressure . in the absence of random stirring particular solutions of eq .
( [ u - field1 ] ) are well known ( for an incompressible fluid ) : it admits plane wave solutions , called _ inertial waves _ @xcite .
these are exact solutions of the non - linear equations , but superposition does not hold
. they may have a role in the transition to turbulence @xcite . in subsection
[ sec : waves ] we will study how the perturbative corrections modify the propagation of inertial waves . we now proceed to eliminate the gradient term of eq .
( [ u - field1 ] ) by making use of the incompressibility condition @xcite .
we define the generalized pressure as @xmath23 . by taking the divergence of the previous equation we can solve for @xmath24 to obtain @xmath25 \ ; , \label{p*}\ ] ] so that the pressure @xmath24 can be eliminated from eq .
( [ u - field1 ] ) by writing @xmath26 \ ; .
\label{grad - p*}\ ] ] we can write the navier - stokes equation as follows : @xmath27 = \nu \nabla^2 { \bf u } - { \cal p}(2{\bf \omega \wedge u } ) + { \bf f } \ ; , \label{u - field2}\end{aligned}\ ] ] where , following standard practice , we have introduced the constant @xmath28 in front of the advective term for book - keeping purposes @xcite ( @xmath28 will be useful when carrying out the perturbation expansion and is to be set to one afterwards ) .
the projection operator @xmath29 is given by @xmath30 and ensures that the non - linear and coriolis terms are solenoidal .
( [ u - field2 ] ) , if @xmath31 is solenoidal so is @xmath32 and vice versa . unlike eqs .
( [ moment1 ] ) or ( [ u - field1 ] ) , eq .
( [ u - field2 ] ) is translation invariant .
that is , the centrifugal term in ( [ moment1 ] ) clearly distinguishes the origin @xmath33 as a special point ; but as we have seen in eq .
( [ grad - p * ] ) we can include this term into the generalized pressure and eliminate @xmath24 from the equation .
this yields eq .
( [ u - field2 ] ) in which a preferred _ direction _ ( but no preferred point ) is singled out by the angular velocity .
this latter equation is invariant under translations , hence , we can make use of the fourier transform .
since in fourier space @xmath34 eq .
( [ u - field2 ] ) only contains vectors orthogonal to @xmath35 and we may refer to this equation as the _ transverse _ navier - stokes equation .
we choose the random force spectrum ( [ noise-1 ] ) as follows : @xmath36 where @xmath37 , @xmath38 is a measure of the amplitude of the random force and the real exponent @xmath39 characterizes the random force spectrum @xcite . when @xmath40 the velocity correlations produce an energy spectrum resembling the kolmogorov spectrum .
the random force acts _ isotropically _ on the fluid , thus , whatever anisotropies emerge at large scale must be due to the coriolis force . in the following sections we analyse the nature of the anisotropies by solving eq .
( [ u - field2 ] ) in perturbation theory . in what follows we use the fourier transformed equation of motion ( [ u - field2 ] ) .
our convention for the fourier transform is given by @xmath41 we have introduced a wave - number cut - off @xmath42 , so that the integral over @xmath43 is restricted to the values @xmath44 . the inverse of this cut - off , @xmath45 , can be associated with the dissipation ( kolmogorov ) scale , and we assume that @xmath46 @xcite . in order to write the navier - stokes eq .
( [ u - field2 ] ) in wave - number representation we transform the @xmath31 field according to ( [ fourier ] ) , apply convolution to the non - linear term , and invert the fourier transform to obtain @xmath47 \int_{p < \lambda } \frac{d^d \vec p}{(2\pi)^d}\ , \int_{-\infty}^{\infty } \frac{d\omega'}{2\pi}\ , u_j(\vec k - \vec p,\omega - \omega ' ) u_k(\vec p,\omega ' ) \nonumber \\ & + & f_i(\vec k,\omega ) \ ; .
\label{fourier - u2}\end{aligned}\ ] ] this equation , but without the coriolis term , is a familiar expression in turbulence research @xcite .
( [ fourier - u2 ] ) can be iterated to any desired order in @xmath28 and will serve as the starting point for constructing the perturbation expansion , which is considered in the following subsection .
if we `` shut - off '' the non - linear terms ( proportional to @xmath28 ) in ( [ fourier - u2 ] ) we can identify the ( inverse ) linear response function , which will be used in carrying out the perturbative calculation .
thus , setting @xmath48 in eq .
( [ fourier - u2 ] ) we have @xmath49 where @xmath50 is the linear velocity field . the previous equation can be written as @xmath51^{-1}{\bf u}^{(0)}(\vec k,\omega ) = { \bf f}(\vec k , \omega ) \ ; , \label{g - inv1}\ ] ] which defines the linear ( inverse ) response function matrix , and allows one to solve for the linear velocity field @xmath52 . in the next subsection
we calculate the perturbative corrections to the linear inverse response function due to the existence of non - linearities and random forcing , and establish their effect on the physical parameters ( _ e.g. , _ viscosity ) describing the transverse navier - stokes equation . in this subsection
we proceed to expand eq .
( [ fourier - u2 ] ) by iteration , in powers of the non - linear coupling ( see the diagrammatic representation , fig . [ fig1 ] ) . in this way
one can calculate the perturbative corrections to the response function to any desired order in @xmath28 .
we do not present all the details of such expansion as these can be found in review articles @xcite and textbooks @xcite .
the calculation of the non - linear response function , @xmath53 $ ] , requires the correction matrix , @xmath54 $ ] , which is defined by a recursion formula : @xmath55 \equiv [ g_0 ] + [ g_0 ] [ m ] [ g ] = [ g_0 ] + [ g_0 ] [ m ] [ g_0 ] + [ g_0 ] [ m ] [ g_0 ] [ m ] [ g_0 ] + \cdots \ ; , \label{gg}\ ] ] which implies @xmath56}^{-1 } = [ g_0]^{-1 } - [ m ] \label{inverse } \ ; .\ ] ] in appendix [ sec : app - diag ] we present the diagrammatic representation of the exact transverse navier - stokes equation , as well as the first order correction to the response function .
the diagram of eq .
( [ gg ] ) with the correction @xmath54 $ ] , calculated to first order of perturbation theory in the random force amplitude , is presented in fig .
when written in components , the correction matrix @xmath54 $ ] is given by @xmath57\ , i_{rjn}(\vec k , \omega ) \ ; , \label{m}\ ] ] where the factor @xmath58 is combinatorial ( see fig .
[ fig2 ] ) , and the function @xmath59 is defined by @xmath60 \nonumber \\ & & \times ( 2d ) \
, \vert \vec p \vert^{-y}\ , [ g_0(\vec p,\omega')]_{ra}{\cal p}_{ab}({\bf p } ) [ g_0 ( -\vec p,-\omega')]_{bs}^t\ , [ g_0(\vec k - \vec p , \omega-\omega')]_{jl } \ ; , \label{loopint}\end{aligned}\ ] ] ( where @xmath61 means the transposed matrix ) .
the integral over @xmath62 would be divergent at @xmath63 , so it is cut off at the low wave - number given by the inverse of the system size scale , @xmath64 .
the evaluation of eq .
( [ loopint ] ) also involves an integration over the frequency @xmath65 , which has been carried out for _ slow rotation _
( see appendix [ sec : app - perturbation ] ) .
as we are interested in the late time and large scale limits we have computed @xmath66 only for @xmath67 and up to second order in @xmath43 @xcite .
the integration over wave - number @xmath62 can be decomposed into an integration over modulus and angles .
the integrand must be expanded , up to second order in @xmath68 , as this is sufficient to obtain the effective viscosity [ see eq .
( [ diagonal ] ) ] .
the calculation of @xmath59 is straightforward but tedious and its technical details are presented in appendix [ sec : app - perturbation ] . according to eq .
( [ inverse ] ) , the matrix @xmath54 $ ] provides the order @xmath69 correction to the inverse response function . to obtain @xmath54 $ ] ( for @xmath67 and in the limit @xmath70 ) , we carry out the integral of eq .
( [ loopint ] ) and substitute into eq .
( [ m ] ) : @xmath71 \label{one - loop - g - inv - isotropicb } \ ; , \end{aligned}\ ] ] where @xmath72 is the surface area of the unit sphere in @xmath15 dimensions , @xmath42 and @xmath64 are , respectively , the upper and lower wave - number cut - offs , @xmath73 , and @xmath74 the levi - civita tensor for @xmath16 . in arriving at this final form
we have assumed slow rotation , that is , we have kept the rotation dependent terms up to @xmath75 in the calculation .
this is explained in appendix [ sec : app - perturbation ] and we will come back to this point in section [ sec : discussion ] .
we can compare the linear inverse response function ( [ inverse0 ] ) to its first order correction @xmath76 . in order to do so , we set @xmath77 in eq .
( [ inverse0 ] ) to obtain @xmath78^{-1}_{ij } & = & \nu k^2 \ , \delta_{ij } + 2\epsilon_{nmj } \omega_m \ , { \cal p}_{in}({\bf k } ) \ ; , \label{inverse0c - projected}\end{aligned}\ ] ] which is split into an isotropic part ( proportional to @xmath2 ) and an anisotropic part ( proportional to @xmath0 ) . the first part of @xmath54 $ ] ( [ one - loop - g - inv - isotropica ] ) corrects the molecular shear viscosity @xmath2 , exactly as in the absence of rotation @xcite .
the non - linear inverse response function must take the same form as @xmath79^{-1}_{ij}$ ] , with @xmath2 replaced by @xmath80 , the effective viscosity @xcite .
the value of @xmath80 can be obtained by making use of eqs .
( [ inverse ] ) and ( [ one - loop - g - inv - isotropica ] ) , so that @xmath81 if we multiply eq . ( [ eff - vis-1 ] ) by @xmath82 and make use of the fact that @xmath29 is idempotent , we obtain @xmath83 which shows that the isotropic term of @xmath54 $ ] , eq .
( [ one - loop - g - inv - isotropica ] ) , renormalizes the molecular shear viscosity @xmath2 .
we must distinguish three cases , namely , @xmath84 , @xmath85 , and @xmath86 . if @xmath85 , naive perturbation theory fails .
given that @xmath87 , in the case @xmath84 , the term @xmath88 dominates over @xmath89 , so we could take the limit @xmath90 ( the @xmath62 integral is convergent in the lower limit ) .
in contrast , in the case @xmath91 , the term @xmath89 dominates and we instead neglect the contribution proportional to @xmath88 . in consequence ,
the actual expansion parameter is @xmath92 or @xmath93 , according to whether @xmath94 or @xmath86 , respectively
. we will take @xmath95 .
the number @xmath93 can be identified with the cube of the reynolds number , on account of the interpretation of @xmath3 as the system size scale , and that the dissipation rate is proportional to the amplitude of the random force @xmath96 @xcite . in particular , in @xmath16 ,
@xmath97 the most interesting case is @xmath98 , which yields the kolmogorov energy spectrum .
the second part of the correction ( [ one - loop - g - inv - isotropicb ] ) has the same structure as the anisotropic term on the right - hand side of eq .
( [ inverse0c - projected ] ) , on considering that the first term within the square brackets vanishes when it acts on @xmath99 .
however , it does not correspond to a renormalization of @xmath11 , owing to the additional dependence on @xmath100 , namely , the @xmath101 factor . in fact
, this factor is adequate for a `` viscosity term '' that depends on @xmath11 ( anisotropy ) .
therefore , it suggests the introduction of an anisotropic `` viscosity '' , which we address in the following section .
in the previous section we have seen that the term ( [ one - loop - g - inv - isotropicb ] ) induces no corrections to @xmath2 . in this section
we give a physical interpretation of this anisotropic contribution , and show that it arises from an effective `` viscosity tensor '' for the rotating fluid with mixed symmetry in its indices . for a newtonian fluid the linear relation between
the rate of strain and stress tensors involves a rank four viscosity tensor , so that we can write @xcite @xmath102 as both @xmath103 and @xmath104 are symmetric tensors the only symmetries of the `` viscosity tensor '' are the following : @xmath105 and @xmath106 . from the above symmetries
we conclude that the `` viscosity tensor '' has @xmath107 independent components ( in @xmath16 ) .
we write @xmath108 as a sum of a pair - symmetric ( s ) and a pair - antisymmetric ( a ) part as follows @xmath109 so that @xmath110 has the same symmetries as @xmath108 plus @xmath111 , and @xmath112 has the same symmetries as @xmath108 plus @xmath113 .
there are @xmath114 pair - symmetric and @xmath115 pair - antisymmetric independent components .
the presence of axial symmetry ( induced by the coriolis term ) reduces the number of independent components of both @xmath116 and @xmath117 .
the most general axisymmetric tensor ( in @xmath16 ) can be constructed from @xmath118 , @xmath119 , and @xmath74 as follows @xmath120 we observe that the number of independent components has been reduced from @xmath114 to @xmath121 for the pair - symmetric term and from @xmath115 to @xmath122 for the pair - antisymmetric part .
the symmetry arguments used in deducing ( [ nus],[nua ] ) are analogous to those used in the theory of elasticity @xcite .
the coefficient functions @xmath123 and @xmath124 can be written as a series in @xmath125 : @xmath126 in the limit of slow rotation ( linear order in @xmath0 ) these coefficients reduce to their constant ( @xmath127 ) value and , furthermore , the terms proportional to @xmath128 and @xmath129 vanish at this order .
we can then write for the `` viscosity tensor '' @xmath130 the pair - symmetric part is isotropic and we can identify @xmath131 and @xmath132 , where @xmath10 and @xmath2 denote the molecular bulk and shear kinematic viscosities , respectively ; the pair - antisymmetric part , proportional to the coefficient @xmath133 , is identified below .
incompressibility means that @xmath134 , leaving only @xmath2 , in the absence of rotation .
this is the molecular viscosity in the original navier - stokes equation ( [ moment1 ] ) , which gets renormalized , becoming an effective ( rotation independent ) viscosity , as seen in the previous subsection [ sec : perturbative ] .
however , it is to be expected that perturbation theory , at sufficiently high order , generates a dependence of @xmath135 on @xmath0 , such that one would be led to consider a rotation dependent viscosity . at the order we are working , the effect of rotation is to generate a rotation dependent anisotropic `` viscosity '' of the type ( [ nua-1 ] ) , as will be demonstrated below . the anisotropic `` viscosity '' ( [ nua-1 ] ) is associated to a stress tensor @xmath136 , which yields the following viscous force @xmath137 in wave - number representation this can be written as @xmath138 where we have made use of the symmetry within pairs of indices of the tensor @xmath139 .
in order to be consistent with the transverse navier - stokes equation , we must project eq .
( [ ufield1 ] ) by means of @xmath140 @xmath141 we now substitute eq .
( [ nua-1 ] ) into the previous expression to obtain @xmath142 \ ,
u_q = -{\beta_1}\ , \omega_n \ , k^2 \epsilon_{mqn } { \cal p}_{im}({\bf k } ) \ , u_q \ ; , \label{proj - u - field2}\end{aligned}\ ] ] where we have taken into account the antisymmetric properties of the levi - civita tensor and the fact that @xmath143 .
we now compare this equation with the first order perturbative correction of the linear inverse response function .
the force given by eq .
( [ proj - u - field2 ] ) agrees identically with the force provided by the anisotropic part of the correction @xmath144 , namely , the product of its expression in ( [ one - loop - g - inv - isotropicb ] ) and @xmath99 , once we choose the coefficient in ( [ nua-1 ] ) to be @xmath145 we have thus shown that the contribution ( [ one - loop - g - inv - isotropicb ] ) to the response function can be understood as arising from the anisotropic `` viscosity '' tensor ( [ nua-1 ] ) .
since we are considering the case @xmath86 , we can neglect the term @xmath146 .
the expansion parameter , proportional to @xmath147 , must again be identified with the cube of the reynolds number .
we postpone this identification to the following section .
if we set @xmath16 in the expression for @xmath133 , we obtain @xmath148
in this section we focus on the new force @xmath149 that appears in the fluid equation of motion due to first order perturbative corrections , and discuss some possible physical implications . here
we use the results derived in previous sections to `` correct '' the transverse navier - stokes equation with the newly generated ( first order in @xmath0 ) terms that arise at large scales due to rotation . in coordinate space the force ( [ ancorr ] ) becomes @xmath150 \right ] \ ; . \label{pseudo}\end{aligned}\ ] ] in deriving this equation we have made use of the incompressibility condition and the antisymmetry properties of the levi - civita tensor .
this is a quasi - local ) and its derivatives . ]
force ( per unit volume ) in which the angular velocity couples to the laplacian of the fluid velocity and to the vorticity @xmath151 . the physical character of the force @xmath152 can be revealed by writing the _ effective _ equation of motion in coordinate space .
we have @xmath153 where @xmath80 is the rotation independent effective kinematic viscosity , [ see eq .
( [ nu ] ) ] , and @xmath154 .
the magnitude of the correction linear in @xmath0 ( the new force @xmath155 ) can be determined by comparing it with either the coriolis force or the viscosity correction .
we obtain for the ratio to the coriolis force : @xmath156 , assuming that the scale of spatial velocity fluctuations is @xmath45 , that is , @xmath157 .
hence , the expansion parameter can be identified with the cube of the reynolds number ( as in the previous section ) times @xmath158 . on the other hand ,
the ratio between the correction linear in @xmath0 and the correction linear in @xmath80 is @xmath159 .
this was to be expected , since @xmath160 measures the relative strength of the viscosity and coriolis force .
we conclude this subsection by providing an important property of the new force . in general , for a stress tensor associated with a pair - antisymmetric `` viscosity tensor '' ,
the power is given by @xcite @xmath161 but , as @xmath162 we conclude that @xmath163 ; therefore , this stress tensor , @xmath164 , does not lead to dissipation and is not truly viscous .
this implies that the name `` viscosity tensor '' is not appropriate , and we have only introduced it by analogy with the truly viscous pair - symmetric tensor @xcite . as already mentioned , rotating incompressible fluids support wave solutions that are exact solutions of the non - linear equations @xcite .
although we have derived the quasi - local force from the randomly forced equations ( that describe turbulence ) , it is of interest to see how this force affects these wave solutions .
let us consider wave solutions of the form @xmath165 as single mode plane waves of the linearized equation ( [ newmodelu ] ) ( the incompressibility condition annihilates the advective term ) but without the forcing term @xmath166 we can write this equation in components as follows @xmath167 we are free to choose @xmath168 and @xmath169 .
this choice implies that @xmath170 ( incompressibility condition ) , so that the waves are transverse . from our choice for @xmath35 the only non - vanishing components of @xmath140
are @xmath171 and @xmath172 , which are equal to one .
we can write the @xmath173 and @xmath174 components [ the @xmath175 component of ( [ comps - ii-2 ] ) is identically null ) ] as follows @xmath176 these equations yield the ( complex ) frequency of the plane waves ( dispersion relation ) @xmath177 where @xmath178 is the angle between @xmath35 and @xmath11 , so that @xmath179 , with @xmath180 and @xmath181 . by making use of the dispersion relation ( [ disp - ii ] )
it is easy to see that the waves are circularly polarized and transverse @xmath182 wave packets are not solutions of the non - linear equations and can only be considered for small wave amplitudes .
the group velocity @xmath183 of a wave packet is ( for vanishing viscosity ) @xmath184 and the phase velocity @xmath185 ( for vanishing viscosity ) @xmath186 the `` standard textbook '' result is recovered by taking the limit @xmath187 @xcite . from this calculation
we see that the new term in eq .
( [ nsp - ii-2 ] ) not only changes the wave frequency and phase velocity , but also the group velocity . moreover , the group velocity is no longer perpendicular to the phase velocity @xmath188 the quasi - local force would cause the energy transport to not be perpendicular to the phase velocity and a small fraction of the energy in the waves to be transported parallel to the wave vector .
we have applied perturbation theory to a homogeneous incompressible viscous fluid subject to solid body rotation and isotropic random forcing . at small scales , we assume that only the molecular shear viscosity and the coriolis force are needed in writing down the _ transverse _ navier - stokes equation ( [ u - field2 ] ) .
our ( first order ) perturbative results demonstrate that ( i ) anisotropic components of the effective `` viscosity tensor '' are dynamically generated at large scales by the combined interplay of the coriolis force , the random forcing term , and the inherent non - linearity of the navier - stokes equation , ( ii ) the molecular shear viscosity @xmath2 gets corrected in the same manner as for isotropic randomly stirred turbulence @xcite .
these perturbative results are corroborated by a symmetry principle . by making use of the axial symmetry of a rotating fluid
, we have constructed the most general `` viscosity tensor '' that is invariant under such symmetry . the preferred direction singled out by
the angular velocity @xmath11 breaks the isotropy and leads to new terms in the `` viscosity tensor '' absent in the isotropic case @xcite .
we have also determined and described the role of the first order perturbative new term in the effective fluid equation of motion .
it acts as a quasi - local force and , like the coriolis force , is not dissipative and does no net work on the fluid .
most importantly , we find that this quasi - local force affects the propagation of inertial waves in rotating fluids . for small wave amplitudes a fraction of the wave energy is transported in the same direction as the phase velocity .
the perturbative calculation is developed as a double expansion : in both the amplitude of the random force and the angular velocity .
the actual dimensionless expansion parameters turn out to be the cube of the reynolds number and this number times the inverse of the ekman number , respectively . for simplicity ,
we have restricted ourselves to the computation of the lowest order in both : that is , first order in the random force amplitude and linear order in @xmath11 .
this is sufficient to generate two , out of the total of eight tensor terms ( seven , on account of incompressibility ) allowed by axial symmetry [ see eqs .
( [ nus ] ) and ( [ nua ] ) ] .
we conjecture that , at first order in the random force amplitude , all the remaining tensor terms are generated for higher powers in @xmath0 : @xmath189 , @xmath190 , and @xmath129 at quadratic order , @xmath191 at cubic order , and finally @xmath192 at quartic order . of course , as increasing powers of @xmath0
are taken into account , the coefficient functions ( [ beta ] ) must be expanded out to the order of @xmath0 being investigated .
the higher order terms will be needed to study the effects of fast rotation and to track the onset of the bi - dimensionalization of the fluid @xcite .
an important step in this direction will be provided by a complete renormalization group ( rg ) analysis of the large scales properties of a rotating incompressible fluid . in order to carry out this
rg analysis one may need to calculate the ( perturbative ) corrections to the non - linear coupling term @xmath28 , as well as the random force amplitude @xmath96 , and combine these two with the response function calculation presented in this paper .
once in hand , the rg fixed points can be determined and the corresponding asymptotic behavior of the rotating fluid deduced , allowing us to compute quantities such as the scale - dependent reynolds and ekman numbers , among others .
we hope to report on these developments elsewhere .
the authors thank juan prez - mercader and alvaro domnguez for discussions and for reading an earlier version of this paper .
one of us ( dh ) acknowledges correspondence with arjun berera covering a wide range of topics in turbulence theory .
our work is supported by grant bfm2002 - 01014 and the work of jos gaite is further supported by a ramn y cajal contract , both of the ministerio de ciencia y tecnologa .
in this section we include a diagrammatic representation of the exact transverse navier - stokes equation [ see eq .
( [ fourier - u2 ] ) ] and of the recursion relation for the first order response function [ see eq .
( [ gg ] ) ] .
we have followed the convention introduced in ref .
@xcite .
the exact equation can be represented as where the thick lines represent @xmath193 , the open circles represent @xmath194 , and the thin lines represent @xmath195 .
this graphic representation implies that the linear velocity field @xmath196 is given by a thin line attached to an open circle , as we have @xmath197 . in this diagrammatic representation
we can also depict the recursion relation for the response function @xmath198 , eq .
( [ gg ] ) , given by the first order correction @xmath54 $ ] , [ eq . ( [ m ] ) ] , as follows where the solid circle represents the random force average @xmath199 introduced in eq .
( [ unoise ] ) , and the double thin lines represent the first order response function @xmath53 $ ] .
the large distance and long time renormalizability of the transverse navier - stokes equation ( [ u - field2 ] ) implies that the corrected response function @xmath200 must have the same mathematical structure as its linear counterpart @xmath201 .
@xmath200 has , therefore , the same frequency and wave - number dependence as @xmath201 and contains the same number of parameters .
the renormalization of the response function yields the ( rotation independent ) effective viscosity @xmath80 . in this appendix
we outline the major points in the calculation of the first order correction to the response function @xcite .
( [ g - inv1 ] ) defines the linear inverse response function , which is given by @xmath51^{-1}= \left ( \matrix { a - b & h & 0 \cr f & a+b & 0 \cr c & d & a } \right ) \ ; , \label{inverse0}\ ] ] where the matrix entries are : @xmath202 , @xmath203 , @xmath204 , @xmath205 , @xmath206 , and @xmath207 . in the limit of vanishing rotation ( @xmath208 )
we recover an isotropic diagonal matrix for the linear response function and its inverse @xcite @xmath51^{-1}_{ij } = ( -i \omega + \nu k^2 ) \
, \delta_{ij } \ ; , \ ; \ ; \ ; [ g_0(\vec k,\omega)]_{ij } = ( -i \omega + \nu k^2)^{-1 } \ , \delta_{ij } \ ; . \label{diagonal}\ ] ] given the matrix form of the inverse linear response function ( [ inverse0 ] ) we can write for the linear response function @xmath209 = \frac{1}{a^3 - ab^2 -afh}\ , \left ( \matrix { a^2 + ab & -ah & 0 \cr -af & a^2-ab & 0 \cr -ac - bc+df & -ad+bd+ch & a^2 - b^2 -fh } \right ) \ ; . \label{freeuresp}\ ] ] the first order correction to @xmath200 is given in subsection [ sec : perturbative ] [ see eqs .
( [ gg ] ) , ( [ inverse ] ) , ( [ m ] ) , and ( [ loopint ] ) ] . we have restricted ourselves to the limit of slow rotation ( linear order in @xmath0 ) , and therefore need to expand both @xmath201 and @xmath210 to this order .
we point out that @xmath210 is already linear in @xmath0 [ see eq .
( [ inverse0 ] ) ] . to linear order in @xmath0 the linear response function
is given by @xmath209 = \frac{1}{a^2}\ , \left ( \matrix { a+b & -h & 0 \cr -f & a - b & 0 \cr -c & -d & a } \right ) + o(\omega^2 ) = \frac{1}{a}\ , \left ( \matrix { 1 & 0 & 0 \cr 0 & 1 & 0 \cr 0 & 0 & 1 } \right ) + \frac{1}{a^2}\ , \left ( \matrix { b & -h & 0 \cr -f & -b & 0 \cr -c & -d & 0 } \right ) + o(\omega^2 ) \ ; .
\label{linear - omega - go}\ ] ] the first step of this calculation is to carry out the frequency integration ( over @xmath65 ) in eq .
( [ loopint ] ) . from eq .
( [ inverse0c - projected ] ) we see that the viscosity @xmath2 is the coefficient of the @xmath101 term , and to obtain the effective viscosity we can set the frequency @xmath211 to zero from the outset .
once we make @xmath67 the integral over @xmath65 can be computed by means of the calculus of residues @xcite .
we close the contour in the lower half plane , where there are in general three simple poles , which coalesce into one double pole in the limit of slow rotation .
we obtain ( keeping up to linear terms in @xmath0 ) : @xmath212_{rt}[g_0(-\vec p,-\omega')]_{sq } [ g_0(\vec k - \vec p,-\omega')]_{jl } & = & \delta_{rt}\,\delta_{sq}\ , \delta_{jl } \frac{(1 + \frac{\vec p \cdot \vec k}{p^2})}{(2\nu p^2)^2 } + \delta_{rt}\,\delta_{sq } \frac{(1 + \frac{2\vec p
\cdot \vec k}{p^2})}{(2\nu p^2)^3}[\omega(\vec k -\vec p)]_{jl } \nonumber \\ & + & \delta_{rt}\,\delta_{jl } \frac{(1 + \frac{\vec p \cdot \vec k}{p^2})}{(2\nu p^2)^3}[\omega(-\vec p ) ] _ { sq } + \delta_{sq}\,\delta_{jl } \frac{(2 + \frac{3\vec p \cdot \vec k}{p^2})}{(2\nu p^2)^3}[\omega(\vec p ) ] _ { rt } \nonumber \\ & + & o(k^2 ) \ ; . \label{between - step}\end{aligned}\ ] ] we have carried out an expansion in @xmath37 and only kept up to linear order , as eq . ( [ m ] ) is already linear in @xmath100 , and we only need to compute the @xmath101 term .
the previous equation was evaluated assuming a positive molecular shear viscosity coefficient ( @xmath213 ) .
a change in sign will change the location of the poles and modify the frequency integration .
we have introduced the matrix @xmath214 $ ] defined as follows @xmath215_{mk } = -2\epsilon_{ijk } \omega_j \ , { \cal p}_{im}({\bf p } ) \ ; , \ ] ] which is already linear in the angular velocity . if we make use of this intermediate result ( [ between - step ] ) and substitute it into eq .
( [ loopint ] ) , we are left with the integration over the wave - number @xmath62 . from eqs .
( [ inverse ] ) and ( [ inverse0c - projected ] ) we can see that the renormalization of the molecular viscosity @xmath2 requires that we expand the first order correction @xmath54 $ ] up to second order in the wave - number @xmath43 . the factor in square brackets in eq .
( [ m ] ) is already linear in @xmath100 , so that the integral ( [ loopint ] ) only needs to be expanded to first order in @xmath100 .
it is important to note that this integral depends on @xmath43 not only through the integrand but also through its limits of integration.this is because all wave - numbers appearing in ( [ loopint ] ) must remain within the set of wave - numbers to be integrated over ( in this case we must ensure that both @xmath62 and @xmath216 belong to this set ) @xcite .
this means we must integrate over the _ intersection _ of the domains @xmath217 and @xmath218 , where @xmath64 is the lower cut - off . to first order in @xmath43 the second inequality can be written as @xmath219 , where @xmath178 is the angle between @xmath43 and @xmath62 ( @xmath220 ) .
there are two cases to consider : ( i ) if @xmath221 the intersection of the two intervals can be expressed as the difference of intervals @xmath222 - [ 1/l,1/l + k\cos \theta]$ ] and ( ii ) if @xmath223 the intersection can be written as @xmath222 - [ \lambda + k\cos \theta , \lambda]$ ] .
this means that the complete wave - number integration , valid up to @xmath224 , can be written as @xmath225 where @xmath226 is the surface element of the unit sphere in @xmath15 dimensions .
we also present the following projection operator product expansions that prove to be useful when handling the intermediate steps of the computation @xmath227 where @xmath228 finally , various identities needed for the angular integrations are collected here .
let @xmath72 represent the surface area of the unit @xmath15-sphere and @xmath229 denote a unit vector in the @xmath230-th direction .
the only angular integrations required are of the following types @xmath231 the angular integration of the product of an odd number of unit vectors over the unit @xmath15-sphere vanishes identically .
if we make use of the previous results , we obtain the first order correction as written in ( [ one - loop - g - inv - isotropica],[one - loop - g - inv - isotropicb ] ) .
greenspan , _ the theory of rotating fluids _ ( cambridge university press , 1968 , reissued by breukelen press , 1990 ) . c. cambon and l. jacquin , _ spectral approach to non - isotropic turbulence subjected to rotation _ , j. fluid mech . ,
* 202 * ( 1989 ) 295 .
l. jacquin , o. leuchter , c. cambon and j. mathieu , _ homogeneous turbulence in the presence of rotation _ , j. fluid mech . ,
* 220 * ( 1990 ) 1 .
b. dubrulle and l. valdettaro , _ consequences of rotation in energetics of accretion disks _ , astron .
* 263 * ( 1992 ) 387 , and references therein .
s. veeravalli , _ an experimental study of the effects of rotation on turbulence _ , center for turbulence theory , annual research briefs ( 1990 ) . c. cambon , n.n .
mansour and f.s .
godeferd , _ energy transfer in rotating turbulence _ , j. fluid mech . , * 337 * ( 1997 ) 303 .
godeferd and l. lollini , _ direct numerical simulations of turbulence with confinement and rotation _ , j. fluid mech . ,
* 393 * ( 1999 ) 257 .
d. forster , d.r .
nelson and m.j .
stephen , _ large - distance and long - time properties of a randomly stirred fluid _ , phys .
a * 16 * ( 1977 ) 732 , and references therein .
v. yakhot and s.a .
orszag , j. sci .
, * 1 * ( 1986 ) 3 ; phys . rev . lett .
* 57 * ( 1986 ) 1722 .
mccomb , _ the physics of fluid turbulence _ ( oxford university press , oxford , 1990 ) .
j. busa , m. hnatich , j. honkonen and d. horvath , _ stability of kolmogorov scaling in anisotropically forced turbulence _ ,
e * 55 * ( 1997 ) 381 .
s. chandrasekhar , _ hydrodynamic and hydromagnetic stability _
( dover , new york , 1981 ) . l.d .
landau and e.m .
lifshitz , _ fluid mechanics _ ( pergamon press , oxford , 1987 ) .
smith and s.l .
woodruff , _ renormalization - group analysis of turbulence _ ,
rev.fluid mech .
* 30 * ( 1998 ) 275 .
batchelor , _ introduction to fluid dynamics _
( cambridge university press , 1967 ) .
landau and e.m .
lifshitz , _ theory of elasticity _
( pergamon press , oxford , 1986 ) .
|
large scale features of a randomly isotropically forced incompressible and unbounded rotating fluid are examined in perturbation theory . at first order in both the random force amplitude and the angular velocity
we find two types of modifications to the fluid equation of motion .
the first correction transforms the molecular shear viscosity into a ( rotation independent ) effective viscosity .
the second perturbative correction leads to a new large scale non - dissipative force proportional to the fluid angular velocity in the slow rotation regime .
this effective force does no net work and alters the dispersion relation of inertial waves propagating in the fluid .
both dynamically generated corrections can be identified with certain components of the most general axisymmetric `` viscosity tensor '' for a newtonian fluid .
# 1 pacs : 47.27-i , 47.32.-y
|
the kinematical ( `` jeans '' ) analyses of near - spherical structures using the spherical jeans equation for stationary systems : @xmath2 where @xmath3 is the 3d density distribution of the tracer population , suffer from the fact that this one equation involves two unknowns : the mass distribution @xmath4 and the velocity anisotropy @xmath5 ( we make use of the symmetry of spherical systems yielding @xmath6 ) . in some cases of simple anisotropy profiles
@xmath7 , the jeans equation can be inverted to yield the radial dispersion as a single integral over radii of @xmath8 times some kernel , and one can insert this solution into the equation that links the line - of - sight velocity dispersion to the radial one , to obtain after some algebra , the line - of - sight dispersion as a single integral over @xmath8 and some other kernel . in what follows , we attempt to lift the mass / anisotropy degeneracy by adding a second equation , namely the 4th order jeans equation for constant anisotropy ( okas 2002 ) @xmath9 where we express the line of sight _ kurtosis _ as a double integral over @xmath8 and @xmath10 and a third kernel ( the details are given in okas & mamon 2003 ) .
we apply our formalism to the nearby coma cluster , for which we ve extracted from ned the positions and line - of - sight velocities of 967 galaxies , among which 355 early - type galaxies whose kinematics we analyze here ( we omit spiral galaxies , which are expected to have a non negligible mean radial velocity caused by their infall and subsequent transformation at cluster pericenter ) . our mass model includes 1 ) a stellar component , which we find is well fit by an nfw ( navarro et al .
1996 ) model ; 2 ) a dark matter component , which we assume to be a generalized nfw model with arbitrary inner slope : @xmath11 where @xmath12 is the scale radius of the dark matter ( in general different from that of the luminous distribution ) , and @xmath13 for the nfw model .
the profiles differ by their inner slope @xmath14 but have a common outer limiting behavior of @xmath15 ; and 3 ) an isothermal gas component , taken from the rosat x - ray observations by briel et al .
( 1992 ) . using 7 radial bins at @xmath16 , we jointly fit our four parameters , mass @xmath17 within the virial radius @xmath18 , velocity anisotropy @xmath19 , inner slope @xmath20 of the dark matter density profile , and dark matter concentration @xmath21 . contrary to the case
when only velocity dispersion is studied , the minimization procedure now converges ( except for a concentration / inner slope degeneracy ) .
the best fits are shown in figure [ allplots ] .
the best fit parameters are @xmath22 ( corresponding to dark matter virial radius @xmath23 mpc ) , @xmath24 , @xmath25 and @xmath26 with @xmath27 .
in other words , our best fit models are isotropic and the generalized dark matter model of equation ( [ rhoofr ] ) provides excellent fits to the data .
our best fit radial profiles for each half - integer value of @xmath20 are shown in figure [ profs ] .
the best fit density profiles are indistinguishable beyond @xmath28 , and the outer mass profiles are very close to the infall estimates of geller , diaferio & kurtz ( 1999 ) .
the density profiles are 50% more concentrated than in structures of comparable mass within cosmological @xmath0-body simulations in a @xmath29cdm universe ( bullock et al .
interestingly , the @xmath30 profiles have nearly constant slope for the nfw case ( @xmath13 ) .
our analysis differs from previous kinematical analyses ( e.g. merritt 1987 ; den hartog & katgert 1996 ; carlberg et al . 1997 ; van der marel et al . 2000 ; biviano & girardi 2003 ; katgert et al .
2004 ) , as 1 ) we have , for a single cluster , a larger sample of galaxies , which , given their early morphological type , should be in dynamical equilibrium in the cluster potential ; 2 ) we remove pairs from the computation of the velocity moments ; 3 ) we include kurtosis in the analysis ; 4 ) we model the dark matter distribution using a generalized formula inspired by the results of cosmological @xmath0-body simulations ; 5 ) we include hot gas . in comparison to studies based upon stacking of many clusters , our analysis of the coma cluster benefits from not having to introduce errors in any stacking procedure , and from a cleaner removal of interlopers .
on the other hand , the analyses of stacked clusters have the advantage of averaging out particular inhomogeneities of individual clusters such as coma , expected in hierarchical scenarios of structure formation and observed in coma , both in projected space ( fitchett & webster 1987 ; mellier et al . 1988 ; briel et al . 1992 ) and velocity space ( colless & dunn 1996 ; biviano et al .
in our jeans analysis of the coma cluster , we assumed that the cluster was spherical , with no substructure nor streaming motions , and that projection effects were not serious .
we test ( details in sanchis , okas & mamon 2004 ) these assumptions using an @xmath29cdm cosmological @xmath0-body ( tree - code ) simulation , run by ninin ( 1999 ) , with @xmath31 particles , and with @xmath32 , @xmath33 , @xmath34 , and @xmath35 ( details in hatton et al .
we analyze each of the 10 most massive halos along 3 axes : the major axis , an orthogonal one and an intermediate one .
figure [ mom0 ] shows the radial profiles of line - of - sight velocity moments ( mean , dispersion , skewness and kurtosis ) along the principal axis .
these profiles display statistically significant radial variations caused by substructures with radial streaming motions .
moreover , the moments vary considerably from halo to halo and for the 3 different projections of a given halo . note also that the projected mean velocity differs substantially between all the particles in the projected cylinder and the subset within the virial sphere .
figure [ fits ] shows the results of our fits from our jeans analysis ( _ open symbols _ ) in comparison with our fits to the 3d data ( _ crosses _ ) and table [ stats ] summarizes the statistical accuracy of our fitting procedure . the virial mass is typically underestimated by @xmath36 , while the concentration parameter is overestimated by @xmath37 and the ratio @xmath38 underestimated by @xmath39 .
hence , our kinematical analysis yields good results _ on average _ , but with a fairly large dispersion in accuracies ( with the most inaccurate estimates for halos with non zero mean line of sight velocities at some radii ) .
halo 4 , which departs the most from the 3d measures of concentration and anisotropy , happens to have important variations in its projected mean velocity profile , whereas nearly all other halos have normal mean velocity profiles .
this suggests that the joint dispersion / kurtosis jeans analysis is adequate for structures with fairly constant projected mean velocity profiles . given the large dispersion on the concentration parameter , our high value of @xmath40 for coma is now consistent with the 1.5 times lower value expected from cosmological @xmath0-body simulations
|
we provide the first direct lifting of the mass / anisotropy degeneracy for a cluster of galaxies , by jointly fitting the line of sight velocity dispersion and kurtosis profiles of the coma cluster , assuming an nfw tracer density profile , a generalized - nfw dark matter profile and a constant anisotropy profile .
we find that the orbits in coma must be quasi - isotropic , and find a mass consistent with previous analyses , but a concentration parameter 50% higher than expected in cosmological @xmath0-body simulations .
we then test the accuracy of our method on realistic non - spherical systems with substructure and streaming motions , by applying it to the ten most massive structures in a cosmological @xmath0-body simulation .
we find that our method yields fairly accurate results on average ( within 20% ) , although with a wide variation ( factor 1.7 at 1@xmath1 ) for the concentration parameter , with decreased accuracy and efficiency when the projected mean velocity is not constant with radius .
|
users and computers in the internet face a wide range of security risks . of particular concern ,
are _ epidemic risks _ , such as those propagated by worms and viruses .
epidemic risks depend on the behavior of other entities in the network , such as whether or not those entities invest in security solutions to minimize their likelihood of being infected .
our goal in this paper is to analyze the strategic behavior of agents facing such epidemic risks .
the propagation of worms and viruses @xcite , but also many other phenomena in the internet such as the propagation of alerts and patches @xcite or of routing updates @xcite , can be modeled using epidemic spreads through a network . as a result , there is now a vast body of literature on epidemic spreads over a network topology from an initial set of infected nodes to susceptible nodes .
however , much of that work has focused on modeling and understanding the propagation of the epidemics proper , without considering the impact of network effects and externalities .
recent work which did model such effects has been limited to the simple case of two agents , i.e. a two - node network .
for example , reference @xcite proposes a parametric game - theoretic model for such a situation . in the model
, agents decide whether or not to invest in security and agents face a risk of infection which depends on the state of other agents .
the authors show the existence of two nash equilibria ( all agents invest or none invests ) , and suggest that taxation or insurance would be ways to provide incentives for agents to invest ( and therefore reach the `` good '' nash equilibrium ) .
however , their approach does not scale to the case of @xmath0 agents , and it does not handle various network topologies connecting those agents .
our work addresses precisely those limitations . specifically , we study a network of interconnected agents , which are subject to epidemic risks and which can decide whether or not to invest some amount to self - protect and deploy security solutions .
we make three contributions in the paper .
first , we introduce a general model which combines a model for the propagation and dynamics of epidemics among networked agents , as well as an economic model for agents which captures network effects and externalities .
second , we introduce a local mean field ( lmf ) model , which extends the standard mean - field approximation to take into account the correlation structure on local neighborhoods .
third , we solve the lmf model and we derive explicit asymptotic results for general sparse random graphs . ,
for which we obtain asymptotic results .
our main result is that we can explicitly identify the impact of network externalities on the adoptability and deployment of security features . in other words ,
we identify both the economic and network properties that determine the adoption of security technologies .
a similar result is well - known in public economics : in an economy with externalities , the equilibrium outcomes is generally inefficient , i.e. the agents invest too little in self - protection relative to the socially efficient level .
this fact seems very relevant to the situation observed on internet , where under - investment in security solutions and security controls has long been considered an issue .
security managers typically face challenges in providing justification for security investments , and in 2003 , the president s national strategy to secure cyberspace stated that government action is required where `` market failures result in under - investment in cybeysecurity '' @xcite .
the main reason for this situation is that the possible loss can not be avoided completely by a self - protection investment : a residual ( indirect ) risk remains .
this risk depends on the behavior of other agents .
those who invest in self - protection incur some cost and in return receive some individual benefit through the reduced individual expected loss .
but part of the benefit is public : the reduced indirect risk in the economy from which everybody else benefits .
this feature is captured by our simple model of epidemic risks and moreover , we are able to quantify the impact of the underlying network structure linking the agents .
the rest of the paper is organized as follows . in section [ sec : model1 ] , we describe our model for epidemic risks with network effects and externalities . in section [ sec : lmf ] , we introduce our local mean field model ( lmf ) and state asymptotic results that can be obtained with lmf . in section [ sec : netext ] , we use the lmf model to examine both cases when agents invest in strong security solutions ( which perfectly protect the agents deploying them against propagated risks ) and in weak solutions .
we find known phenomena such as free riders and tipping points @xcite .
we also observe counter - intuitive phenomena , such as increasing the quality of the security technology can result in a decreased adoption of that technology in the network . in section [ sec : discussion ] , we discuss our results and conclude the paper . the internet has become a fundamental component of modern economies , and it provide services , starting with connectivity , that are strategic to companies , governments , families and individual users , and in general to the well functioning of modern life . as a result , it has become crucial to the various entities ( operators , enterprises , individuals , ... ) that deliver or use internet services to protect the internet infrastructure and the services delivered through that infrastructure ( or the ability of the infrastructure to deliver those services ) against risks .
there are typically four options that individuals or organizations can take in the face of risks and the associated damages : 1 ) avoid the risk , 2 ) retain the risk , 3 ) self - protect and mitigate the risk , and 4 ) transfer the risk .
option 1 involves preventing any action that could involve risk , and it is clearly not realistic for the internet . option 2 involves accepting the loss when it occurs .
option 3 involves investing in methods to reduce the impact of the risk and the severity of the damages .
option 4 involves transferring the risk to another willing party through contract or hedging .
most entities have so far chosen , or are only aware of the possibility of , a mix of options 2 and 3 . as a result ,
these entities have been busy investing in people and devices to identify threats and develop and deploy coutermeasures . in practice , this has led to the development and deployment of a vast array of systems to detect threats and anomalies ( both malicious such as intrusions , denial - of - service attacks , port scanners , worms , viruses , etc . , and non - intentional such as overloads from flash crowds ) and to protect the network infrastructure and its users from the negative impact of those anomalies , along with efforts in the area of security education in an attempt to minimize the risks related to the human factor @xcite . in parallel ,
most of the research on internet security has similarly focused on issues related to option 3 , with an emphasis on algorithms and solutions for threat or anomaly detection , identification , and mitigation .
unfortunately , self protecting against risk or mitigating risk does not eliminate risk .
there are several reasons for this .
first , there do not always exist fool - proof ways to detect and identify even well defined threats ; for example , even state of the art detectors of port scanners and other known anomalies suffer from non - zero rates of false positives and false negatives @xcite .
furthermore , the originators of threats , and the threats they produce , evolve on their own and in response to detection and mitigation solutions being deployed , which makes it harder to detect and mitigate evolving threat signatures and characteristics @xcite .
other types of damages caused by non - intentional users , such as denial of service as a result of flash crowds , can be predicted and alleviated to some extent but not eliminated altogether .
finally , eliminating risks would require the use of formal methods to design provably secure systems , and formal methods capture with difficulty the presence of those messy humans , even non malicious humans , in the loop @xcite . in the end ,
despite all the research , time , effort , and investment spent in internet security , there remains a residual risk : the internet infrastructure and its users are still very much at risk , with accounted damages already reaching considerable amounts of money and possible damage even more daunting ( e.g. @xcite , @xcite for a discussion on worm damage and conference web site for an opinion on damage cost estimation . )
the question then is how to handle this residual risk .
one way to handle residual risk which has not been considered in much detail yet is to use the fourth option mentioned above , namely transfer the risk to another willing entity through contract or hedging . a widely used way to do
this is through insurance , which is one type of risk transfer using contracts . in practice
, the risk is transferred to an insurance company , in return for a fee which is the insurance premium .
insurance allows individuals or organizations to smooth payouts for uncertain events ( variable costs of the damages associated with security risks ) into predictable periodic costs . using insurance to handle security risks in the internet
raises several questions : does this option make sense for the internet , under which circumstances ? does it provide benefits , and if so , to whom , and to what extent ?
our goal in this paper is to consider those questions .
the economic approach considers that a limit to insurability can not be defined only on the characteristics of the risk distribution , but should take into account the economic environment .
we take this approach in the paper .
a revoir in this section , we consider the case of economic agents ( namely , agents that attempt to optimize some kind of utility function ) subject to epidemic risks , when insurance is not available .
we describe a model for this case and describe the relevant results , obtained in @xcite , on the impact of network externalities on the adoption of self - protection . in the next section ,
we augment this model to include an insurance market .
we then solve that model to obtain our main results for this paper .
our models in this and the next section include two components , so to speak : one component describes the economic model of the agents , the other component describes the spread of the security risk ( worm or malware spread , virus attack ... ) among the agents .
we consider simple one - period probabilistic models for the risk , in which all decisions and outcomes occur in a simultaneous instant .
we model agents using the classical expected utility model , where agents attempt to maximize a utility function @xmath1 .
we assume that agents are rational and that they are risk averse , i.e. their utility function is concave ( see proposition 2.1 in @xcite ) .
risk averse agents dislike mean - preserving spreads in the distribution of their final wealth .
we denote by @xmath2 the initial wealth of the agent . the _ risk premium _
@xmath3 is the maximum amount of money that one is ready to pay to escape a pure risk @xmath4 , where a pure risk @xmath4 is a random variable such that @xmath5=0 $ ] .
the risk premium corresponds to an amount of money paid ( thus decreasing the wealth of the agent from @xmath2 to @xmath6 ) which covers the risk ; hence , @xmath3 is given by the following equation : @xmath7 = \ee[u[w+x]]$ ] .
each agent faces a potential loss @xmath8 , which we take in this paper to be a fixed ( non - random ) value .
we denote by @xmath9 the probability of loss or damage .
there are two possible final states for the agent : a good state , in which the final wealth of the agent is equal to his initial wealth @xmath2 , and a bad state in which the final wealth is @xmath10 .
if the probability of loss is @xmath11 , the risk is clearly not a pure risk .
the amount of money @xmath12 the agent is ready to invest to escape the risk is given by the equation : @xmath13 + ( 1-p)u[w ] = u[w - m].$ ] we clearly have @xmath14 thanks to the concavity of @xmath1
. we can actually relate @xmath12 to the risk premium defined above : @xmath15.\end{aligned}\ ] ] we first recall the model and the main results proved in @xcite .
the model has two levels : one describes the spread of the attack among the agents and the other the economic model for the agents .
we consider simple one - period probabilistic models for the risk , in which all decisions and outcomes occur in a simultaneous instant .
the model for the spread of the attack is an elementary epidemic model .
agents are represented by vertices of a graph .
then any infected agent contaminates neighbors independently of each others with probability @xmath16 .
there is no recovery .
to completely describe the epidemic , we need to specify the initial condition : the set of agents that are active before the contagion process takes place .
this initial state is related to the economic decision of agent concerning his investment in security and this decision depends on the behavior of other entities in the network .
hence we need to briefly describe the model for self - protection .
an agent can invest some amount in self - protection , which in practice would reflect an investment in antivirus or anomaly detection solutions .
if an agent decides to invest in self - protection , we say that the agent is in state @xmath17 ( as in safe or secure ) .
if the agent decides not to invest in self - protection , it is in state @xmath0 ( not safe ) .
if the agent does not invest , its probability of loss is @xmath18 . if it does invest , for an amount which we assume is a fixed amount @xmath19 , then its loss probability is reduced and equal to @xmath20 . in state @xmath0 ,
the expected utility of the agent is @xmath21+(1-p^n)u[w]$ ] ; in state @xmath17 , the expected utility is @xmath22+(1-p^s)u[w - c]$ ] . using the definition of risk premium , we see that these quantities are equal to @xmath23 $ ] and @xmath24 $ ] , respectively .
therefore , the optimal strategy is for the agent to invest in self - protection only if the cost for self - protection is less than the threshold @xmath25-\pi[p^s].\end{aligned}\ ] ] we describe now our model for the epidemic risk .
agents are represented by vertices of a graph .
we assume that an agent in state @xmath17 has a probability @xmath26 of direct loss and an agent in state @xmath0 has a probability @xmath27 of direct loss with @xmath28 .
then any infected agent contaminates neighbors independently of each others with probability @xmath29 if the neighbor is in state @xmath17 and @xmath30 if the neighbor is in state @xmath0 , with @xmath31 .
special cases of this model are examined in @xcite , where @xmath32 , and in @xcite , where agents in state @xmath17 are completely secure and can not be infected , i.e. @xmath33 .
let @xmath34 be a graph on a countable vertex set @xmath35 .
agents are represented by vertices of the graph . for @xmath36
, we write @xmath37 if @xmath38 and we say that agents @xmath39 and @xmath40 are neighbors .
the state of agent @xmath39 is represented by @xmath41 ; agent @xmath39 is infected ( respectively healthy ) iff @xmath42 ( respectively @xmath43 ) .
we now describe the fundamental recursion satisfied by the vector @xmath4 .
we first introduce the following sequences of independent identically distributed ( i.i.d . )
random variables ( r.v . ) : * @xmath44 bernoulli r.v .
with parameter @xmath26 ; * @xmath45 bernoulli r.v . with parameter @xmath27 ; * @xmath46 bernoulli r.v . with parameter @xmath29 ; * @xmath47 bernoulli r.v . with parameter @xmath30 .
let @xmath48 if agent @xmath39 is in state @xmath17 and @xmath49 otherwise .
we define @xmath50 .
the variable @xmath51 models the direct loss : if @xmath52 there is a direct loss for agent @xmath39 , otherwise there is no direct loss for agent @xmath39 .
we also define @xmath53 .
the variable @xmath54 models the possible contagion from agent @xmath40 to agent @xmath39 : if @xmath55 , there is contagion otherwise there is no contagion .
then the fundamental recursion satisfied by the vector @xmath56 is @xmath57 in order to completely specify our model , we still need to define how to choose the variables @xmath58 , i.e. whether agent @xmath39 invests in self - protection ( corresponding to @xmath48 ) or not ( @xmath49 ) . first , note that the probability of loss for agent @xmath39 is given , depending on whether or not it invests in self protection , by @xmath59\label{eq : ps}% p^s_i & = & 1-\ee\left [ ( 1-a^s_i)\prod_{j\sim i}(1-b^s_{ji}x_j)\right ] , \mbox { or , } \\
p^n_i&:= & \ee[x_i|d_i=0 ] .
\label{eq : pn } % p^n_i & = & 1-\ee\left [ ( 1-a^n_i)\prod_{j\sim i}(1-b^n_{ji}x_j)\right].\end{aligned}\ ] ] in view of ( [ eq : optic ] ) , the best response of agent @xmath39 is given by : @xmath60-\pi_i[p^s_i]),\end{aligned}\ ] ] where @xmath61 and @xmath62 are given by ( [ eq : ps ] ) and ( [ eq : pn ] ) .
our model is defined by the graph @xmath63 ( which topology is arbitrary ) and the set of equations ( [ eq : rec],[eq : ps],[eq : pn],[eq : d ] ) . in the rest of this paper , we will make a simplifying assumption : we consider a heterogeneous population , where agents differ only in self - protection cost and potential loss
. the cost of protection should not exceed the possible loss , hence @xmath64 .
the cost @xmath65 and the potential loss @xmath66 are known to agent @xmath39 and varies among the population .
hence we model this heterogeneous population by taking the sequence @xmath67 as a sequence of i.i.d .
random variables independent of everything else .
so far , we have not yet specified the underlying graph .
we will consider random families of graphs @xmath68 with @xmath69 vertices and give asymptotic results as @xmath69 tends to infinity . in all cases ,
we assume that the family of graphs @xmath68 is independent of all other processes .
in this section , we introduce our local mean field ( lmf ) model .
it extends the standard mean - field approximation by allowing to model the correlation structure on local neighborhoods .
it can be shown that the lmf gives the exact asymptotic behavior of the process @xmath4 as the number of vertices tends to infinity for sparse random graphs with asymptotic given degree distribution @xmath70 ( see @xcite for a definition ) .
a rigorous proof of this fact can be found in @xcite for a particular case of the model described in section [ sec : stepid ] .
we will not attemp to give a general proof here .
the main tool is the notion of local weak convergence @xcite . since the graphs we are considering can be considered locally to be like trees ( with high probability ) , we first examine the case where @xmath71 is a tree with nodes @xmath72 and a fixed root @xmath73 . for a node @xmath39 ,
we denote by @xmath74 the generation of @xmath39 , i.e. the length of the minimal path from @xmath73 to @xmath39 .
also we denote @xmath75 if @xmath39 is a children of @xmath40 , i.e. @xmath76 and @xmath40 is on the minimal path from @xmath73 to @xmath39 . for an edge @xmath38 with @xmath75 ,
we denote by @xmath77 the sub - tree of @xmath78 with root @xmath39 when deleting edge @xmath79 from @xmath78 .
we have a family of trees @xmath80 and we run the epidemic model according to equation ( [ eq : rec ] ) with the same variables @xmath81 on each tree .
hence the epidemics on the various subtree of @xmath78 are coupled thanks to these random variables .
we say that node @xmath39 is infected from @xmath77 if the node @xmath39 is infected in @xmath77 .
we denote by @xmath82 the corresponding indicator function with value @xmath83 if @xmath39 is infected from @xmath77 and @xmath84 otherwise .
a simple induction shows that the recursion ( [ eq : rec ] ) becomes : @xmath85 if the tree @xmath78 is finite , we can compute all the @xmath82 recursively starting from the leaves with @xmath86 for any leaf @xmath8 . as
a consequence ( and this is the main difference with ( [ eq : rec ] ) which makes the model on a tree tractable ) , the random variables @xmath87 with @xmath88 in the right - hand term of ( [ eq : rect ] ) are independent of each others and independent of the @xmath89 . for any node @xmath90
, we just defined @xmath82 and the family @xmath91 is a tree - indexed process called a recursive tree process ( rtp ) .
consider now the case where @xmath78 is a galton - watson branching process with offspring distribution @xmath92 .
the tree @xmath78 is now possibly infinite but it is still possible to define an invariant rtp on @xmath78 .
one way to construct it consists in defining a rtp for each finite depth-@xmath93 tree and then show that these rtps converge to an invariant rtp as the depth @xmath93 tends to infinity @xcite .
we first introduce the recursive distributional equation ( rde ) : @xmath94 where @xmath95 has distribution @xmath92 , @xmath96 , @xmath97 where @xmath98 is a bernoulli r.v . with parameter @xmath99 ,
@xmath100 and @xmath87 are i.i.d .
we also assume that the random variables @xmath98 , @xmath101 , @xmath102 , @xmath103 , @xmath104 and @xmath87 are independent of each others .
note however that @xmath105 and the @xmath106 s are not independent of each others .
rde for rtp plays a similar role as the equation @xmath107 for the stationary distribution of a markov chain with kernel @xmath108 , see @xcite .
the following result ( proved in appendix [ ap : rde ] ) solves the rde .
[ prop : rde ] for @xmath109 , the rde ( [ eq : rde ] ) has a unique solution : @xmath100 is a bernoulli random variable with parameter @xmath110 , the unique solution in @xmath111 $ ] of @xmath112 where @xmath113 $ ] is the generating function of the distribution @xmath92 .
moreover the function @xmath114 is non - increasing in @xmath115 . as a consequence
, we see that it is possible to construct an invariant version of the rtp on the tree @xmath78 where for each @xmath116 , the sequence @xmath117 is a sequence of i.i.d .
bernoulli random variables with parameter @xmath118 , see @xcite .
our lmf model is characterized by the connectivity distribution @xmath70 but the underlying tree @xmath78 as to be slightly modified compare to previous section : if we start with a given vertex then the number of neighbors ( the first generation in the branching process ) has distribution @xmath119 but this is not true for the second generation .
let @xmath78 be a galton - watson branching process with a root which has offspring distribution @xmath119 and all other nodes have offspring distribution @xmath92 given by @xmath120 for all @xmath121 .
[ rem : er ] note that if @xmath119 is the poisson distribution with parameter @xmath122 which is the asymptotic degree distribution for erdos - renyi graph @xmath123 , then @xmath92 is also poisson with mean @xmath122 .
we now explain how to define the lmf based on the analysis made in previous section .
clearly , the crucial point in recursion ( [ eq : rect ] ) is the fact that the @xmath82 can be computed `` bottom - up '' . however a node can also be infected from its parent and @xmath82 is not a good approximation of the real process @xmath41 .
indeed the only node for which previous analysis gives an approximation of the process @xmath4 is for the root and the @xmath82 s encode the information that the root is infected by an agent in the subtree of @xmath78 `` below '' @xmath39 .
hence we define @xmath124 where @xmath0 has distribution @xmath119 , @xmath105 and @xmath125 are the same as in ( [ eq : rde ] ) and the @xmath87 s are i.i.d .
bernoulli r.v . with parameter @xmath110 ,
i.e. satisfying the rde ( [ eq : rde ] ) with @xmath95 having distribution @xmath92 .
we now show how to get quantitative results from our lmf .
the goal of section [ sec : netext ] is to derive such results for various cases .
we consider a family of random graphs on @xmath69 vertices @xmath68 and the associated process @xmath126 satisfying the equations of our model on @xmath68 .
we assume that our family of random graphs converges locally to a tree as described in previsous section .
this property is true for sparse random graphs @xcite .
it can be shown that the process @xmath127 is asymptotically equivalent to the process defined on the tree , i.e. the corresponding lmf model described in previous section @xcite .
hence we restict our analysis to the lmf model and the quantities computed here correspond to the asymptotic values of the corresponding quantites for the process @xmath127 for large values of @xmath69 .
let @xmath115 be the fraction of the population investing in self - protection
. then by symetry , the random variables @xmath58 are i.i.d .
bernoulli r.v . with parameter @xmath115 .
thanks to the results of the previous section , we can compute the law of the @xmath41 s . from this law
, we can compute the corresponding probability of loss depending on the choice made to invest or not .
then one has to check self - consistency : the fraction of the population for which the best - response consists in investing in self - protection should be @xmath115 .
hence to solve our lmf model , we need to solve the following fixed point equation : @xmath128\\ \label{f1}&=&1-\ee\left [ ( 1-a^n)\prod_{i=1}^n(1-b^n_iy_i)\right],\\ \nonumber p^{s,\gamma } & = & \ee[x(d)|d=1]\\ \label{f2}&=&1-\ee\left [ ( 1-a^s)\prod_{i=1}^n(1-b^s_iy_i)\right],\\ \label{f3}c^\gamma & = & ( p^{n,\gamma}-p^{s,\gamma})\ell+\pi[p^{n,\gamma}]-\pi[p^{s,\gamma}],\\ \label{f4}\gamma & = & \pp(c\leq c^\gamma),\end{aligned}\ ] ] where the distribution of @xmath129 is given by ( [ eq : x ] ) or equivalently the @xmath82 are i.i.d .
bernoulli r.v . with parameter @xmath110 given by proposition [ prop : rde ] .
let @xmath130 be a solution of this fixed point equation .
then we have the following interpretations : @xmath130 is the fraction of the population investing in self - protection , @xmath131 is the probability of loss for an agent not investing in self - protection and @xmath132 is the probability of loss for an agent investing in self - protection .
hence the average probability of loss is @xmath133 = \gamma^ * p^{s,\gamma^*}+(1-\gamma^*)p^{n,\gamma^*}.\end{aligned}\ ] ] the outcome of rational behavior by self - interested agents can be inferior to a centrally designed outcome . by how much ?
the price of anarchy , the most popular measure of the inefficiency of equilibria , is defined as the ratio between the worst objective function value of an equilibrium of the game and that of an optimal outcome ( possibly centralized in which case it will not be described by the model introduced above ) . in our setting , the cost incurred to agent @xmath39 is @xmath134 if it invests in security and @xmath135 otherwise .
so for a given equilibrium , we can compute the total cost incurred to the population .
the price of anarchy is the ratio of the largest ( among all equilibria ) such cost divided by the optimal cost .
the price of anarchy is at least @xmath83 and a value close to @xmath83 indicates that the given outcome is approximately optimal .
we refer to @xcite for an introduction to the inefficiency of equilibria ( in particular chapter 17 ) .
we show in the next section how to compute this price of anarchy .
we next use our lmf model to compare the following situations : * case 1 : strong protection .
if an agent invest in self - protection , it can not be harmed at all by the actions or inactions of others : @xmath33 ( this is as in @xcite ) * case 2 : weak protection . investing in self - protection does not change the probability of contagion : @xmath32 ( as in @xcite ) in both cases , agents that invest in self - protection incur some cost and in return receive some individual benefit through the reduced individual expected loss . but part of the benefit is public , namely the reduced indirect risk in the economy from which everybody benefits .
hence , there is a negative externality associated with not investing in self - protection , namely the increased risk to others .
we analyze our model on a large sparse random graph @xmath136 on @xmath69 nodes @xmath137 , where each potential edge @xmath79 , @xmath138 is present in the graph with probability @xmath139 , independently for all @xmath140 edges . here
@xmath141 is a fixed constant independent of @xmath69 .
this corresponds to the case of the erds - rnyi graph which has received considerable attention in the past @xcite . as explained in section [ sec : lmf ] , our analysis is not restricted to this class of graphs , but it is simpler in this case since the degree distribution @xmath119 is a poisson distribution with mean @xmath122 ( see remark [ rem : er ] ) . in this case , the fixed point equation for @xmath110 in proposition [ prop : rde ] becomes : @xmath142 then the equations ( [ f1 ] ) and ( [ f2 ] ) are given by : @xmath143 for simplicity , we drop the risk adverse condition , so that @xmath144 and we assume that costs for the self - protection are the same for all agents and equal to @xmath19 , and the possible losses are also the same and equal to @xmath8
. then we have @xmath145 recall that an agent decides to invest in self - protection iff @xmath146 .
the monotonicity of @xmath147 in @xmath115 is crucial and it depends on the value of the parameters @xmath148 .
we first consider case 1 where @xmath33 , so that @xmath149 and @xmath150 .
then by proposition [ prop : rde ] , @xmath151 is non - increasing and the fixed point equation ( [ f1],[f2],[f3],[f4 ] ) has a unique solution . in the vicinity of @xmath152.,width=170 ] in this case , as @xmath115 the fraction of agents investing in self - protection increases , the incentive to invest in self - protection decreases .
in fact , it is less attractive for an agent to invest in self - protection , should others then decide to do so . as more agents invest , the expected benefit of following suit decreases since there is a reduction in the negative externalities which translates into a lower probability of loss .
hence there is a unique equilibrium point which is a nash equilibrium .
however , there is a wide range of parameters for which the nash equilibrium will not be socially optimal because agents do not take into account the negative externalities they are creating in determining whether to invest or not .
indeed it is easily shown that at least for @xmath153 , the price of anarchy is strictly larger than one ( see figure [ fig : pa1 ] ) .
[ prop : fr ] the fixed point equation ( [ f1],[f2],[f3],[f4 ] ) reduces to @xmath154 it has a unique solution .
the price of anarchy is given by @xmath155 where @xmath110 is the unique solution of @xmath156 see appendix [ ap : fr ] for a proof .
we now consider case 2 where @xmath32 , so that @xmath157 is non - decreasing .
the analysis of this case is described @xcite ( see proposition 5 ) .
the situation is quite different from the results we derived for case 1 above . in particular , we can have two nash equilibria involving everyone or no one investing in security .
when there are two nash equilibria , the socially optimal solution is always for everyone to invest : each agent will find that the cost of investing in self - protection will be justified if it does not incur any negative externalities and society will be better off as well .
we have @xmath158 and * if @xmath159 , then there is only one nash equilibrium where every agent invest in self - protection ; * if @xmath160 , then there is only one nash equilibrium where no agent invest in self - protection ; * if @xmath161 , then both nash equilibria are possible . the price of anarchy is given by : @xmath162 if we take @xmath163 , then we have @xmath164 and @xmath165 solution of @xmath166 .
so that we have @xmath167 as a function of @xmath168 , with @xmath152 and @xmath163.,width=170 ] figure [ fig : price ] shows the value of @xmath169 as a function of @xmath168 .
note that typically @xmath170 so that the price of anarchy can be substantially larger than one .
we have shown that both situations with strong or weak protections exhibit externalities and that the equilibrium is not socially optimal : therefore , there is a market failure .
however there are several important differences to understand between strong and weak protections before trying to resolve this market failure . in case 1 , the situation is similar to the free - rider problem which arises in the production of public goods . if all then agents invest in self - protection , then the general security level of the network is very high since the probability of loss is zero . but a self - interested agent would not continue to pay for self - protection since it incurs a cost @xmath19 for preventing only direct losses that have very low probabilities .
when the general security level of the network is high , there is no incentive for investing in self - protection .
this results in an under - protected network .
note that in this case , if the cost for self - protection is not prohibitive , there is always a non - negligible fraction of the agents investing in self - protection . in case 2 , the situation is quite different since no agent at all invests in self - protection . even if a small fraction of agents does invest , and
so raises the general level of security of the network , it is not sufficient for the benefit obtained by investing in self - protection for a new agent to be larger than the cost of self - protection .
these facts seem very relevant to the situation observed in the internet , where under - investment in security solutions and security controls has long been considered an issue .
security managers typically face challenges in providing justification for security investments , and in 2003 , the president s national strategy to secure cyberspace stated that government action is required where `` market failures result in under - investment in cybersecurity '' @xcite .
it shows the power of our basic model to note that these interesting and very relevant phenomena emerge from our analysis .
note also that these phenomena correspond to two extreme values of the parameter @xmath29 , namely case 1 corresponds to @xmath171 and case 2 corresponds to @xmath172 . hence taking @xmath163 and fixing all other parameters
, we have a family of models indexed by @xmath29 , denoted simply @xmath16 in what follows , which varies continuously between the two cases .
recall that @xmath16 is the probability of contagion when the agent invests in self - protection .
if @xmath173 , the agent is completely secure whereas for @xmath174 , agents have the same probability of contagion whatever their choices to invest or not in self - protection .
hence @xmath16 can be interpreted as the inverse of the quality of the technology used for self - protection .
first note that when @xmath173 , the technology is perfect since there is no possible loss .
we are in the situation of case 1 and we see that due to purely economic reasons , the technology is under - deployed in the network because people free - ride the benefit of the technology .
consider now the case of an arbitrary @xmath16 .
figure [ fig : adopt ] shows the adoption curves for different values of @xmath16 .
this curve shows the fraction of the population investing in security technology as a function of its cost ( normalized by the loss ) .
other parameters are @xmath175 and @xmath176 .
we observe some counter - intuitive phenomena .
first for a fixed price , increasing the quality of the security technology can lead to a decrease of its adoption in the population ! here is a qualitative interpretation of how this arises : when the technology is not very good , propagation of the epidemic is possible even if the agent uses the technology
. then agents have to pool their efforts in order to compensate for the weakness of the technology .
in other words , a large number must invest in self - protection in order to have an acceptable level of security .
but when the technology becomes better , then agents that did invest in it start to step down from the group of investors and choose to free - ride .
second there is a barrier for choosing self - protection ( except when @xmath173 ) .
namely for a fixed @xmath16 , we see that there is a range for the parameter @xmath19 ( close to @xmath177 ) such that the population is trapped in state @xmath0 whereas for the same values of the parameters , the situation where a large fraction of the population is investing would be a sustainable equilibrium point .
there is a possibility of tipping or cascading : inducing some agents to invest in self - protection will lead others to follow suit .
the curves of figure [ fig : adopt ] allow us to quantify the minimal number of agents to induce in order to trigger a large cascade of adoption .
10 d. aldous and a. bandyopadhyay .
a survey of max - type recursive distributional equations .
1047 - 1110 , 2005 .
d. aldous and j.m .
the objective method : probabilistic combinatorial optimization and local weak convergence . ,
springer , vol .
110 , pp . 1 - 72 , 2004 .
j. bolot and m. lelarge . a new perspective on internet security using insurance .
j. bolot and m. lelarge .
cyber - insurance as an incentive for it security . ,
e.g. coffman jr . , z. ge , v. misra .
network resilience : exploring cascading failures within bgp .
r. durrett random graph dynamics , 2006 .
a. ganesh , l. massoulie , d. towsley .
the effect of network topology on the spread of epidemics . ,
miami , fl , march 2005 . c. gollier . .
mit press , 2004 .
h. kunreuther and g. heal .
interdependent security : the case of identical agents .
, 26(2):231249 , 2003 .
m. lelarge and j. bolot .
network externalities and the deployment of security features and protocols in the internet . , annapolis , md , jun . 2008 .
m. lelarge and j. bolot . a local mean field analysis of security investments in networks . ,
seattle , aug .
t. moscibroda , stefan schmid and roger wattenhofer .
when selfish meets evil : byzantine players in a virus inoculation game .
, 3544 , 2006 .
n. nisan , t. roughgarden , e. tardos and v.v .
vazirani ( eds ) .
algorithmic game theory .
, 2007 .
m. vojnovic and a. ganesh . on the race of worms , alerts and patches . ,
fairfax , va , nov . 2005 .
white house .
`` national strategy to secure cyberspace '' , 2003 .
available at whitehouse.gov/pcipb . c. zou ,
w. gong , d. towsley .
code red worm propagation modeling and analysis . ,
washington , dc , nov 2002 .
recall that the rde is given by : @xmath94 where @xmath95 has distribution @xmath92 , @xmath96 , @xmath97 where @xmath98 is a bernoulli r.v .
with parameter @xmath99 , @xmath100 and @xmath87 are i.i.d . copies .
let @xmath178 , then we have @xmath179\\ & & + ( 1-\gamma)(1-\pp(a^n=0))\ee\left [ \pp(b^n_ky_k=0)^{n^*}\right],\end{aligned}\ ] ] and the first part of proposition [ prop : rde ] follows .
we define : @xmath180 so that @xmath118 is solution of the fixed point equation @xmath181 . by taking the derivate of @xmath182 in @xmath183
, we see that @xmath184 is a non - decreasing concave function .
note that @xmath185 and @xmath186 .
so that for @xmath187 , there exists a unique solution to the fixed point equation @xmath181 . if @xmath188 , we have @xmath189 and @xmath190 .
then if @xmath163 , the fixed point equation has a unique solution @xmath191 and if @xmath192 , then @xmath193 and the fixed point equation has still an unique solution .
we now prove that the function @xmath114 is non - increasing . by taking the derivate of the function @xmath194
, we see that this function is non - increasing in @xmath115 ( while @xmath183 is fixed ) . then for @xmath195
, we get @xmath196 and the claimed monotonicity of @xmath118 follows .
recall that the fixed point equation for @xmath110 is : @xmath197 consider now that the cost @xmath19 and loss @xmath8 are random variables such that the function @xmath198 is continuous , then equation ( [ f4 ] ) is @xmath199 since the function @xmath118 is non - increasing , we see that the right - hand side of the first line is a non - increasing function in @xmath115 , hence there exists a unique solution @xmath130 to this fixed point equation .
if we take a sequence of distributions such @xmath200 tends to a constant , we see that the solution @xmath130 is such that @xmath201 and the first part of proposition [ prop : fr ] follows . note that we have @xmath202 .
so for a fixed @xmath115 , the average cost incured to the population is @xmath203 .
now for @xmath204 , we have @xmath205 , so that the average cost is just @xmath19 and the last part of proposition [ prop : fr ] follows .
|
getting agents in the internet , and in networks in general , to invest in and deploy security features and protocols is a challenge , in particular because of economic reasons arising from the presence of network externalities .
our goal in this paper is to model and investigate the impact of such externalities on security investments in a network . specifically , we study a network of interconnected agents subject to epidemic risks such as viruses and worms where agents can decide whether or not to invest some amount to deploy security solutions .
we consider both cases when the security solutions are strong ( they perfectly protect the agents deploying them ) and when they are weak .
we make three contributions in the paper .
first , we introduce a general model which combines an epidemic propagation model with an economic model for agents which captures network effects and externalities .
second , borrowing ideas and techniques used in statistical physics , we introduce a local mean field ( lmf ) model , which extends the standard mean - field approximation to take into account the correlation structure on local neighborhoods .
third , we solve the lmf model in a network with externalities , and we derive analytic solutions for sparse random graphs of agents , for which we obtain asymptotic results .
we find known phenomena such as free riders and tipping points .
we also observe counter - intuitive phenomena , such as increasing the quality of the security technology can result in a decreased adoption of that technology in the network . in general , we find that both situations with strong and weak protection exhibit externalities and that the equilibrium is not socially optimal - therefore there is a market failure . insurance is one mechanism to address this market failure . in related work
, we have shown that insurance is a very effective mechanism @xcite and argue that using insurance would increase the security in a network such as the internet . [ network problems ] [ modeling techniques ]
|
millimeter - wavelength continuum surveys of the galactic plane provide the most efficient way to find molecular clumps that are the likely formation sites of massive stars and star clusters .
the development of detector arrays has made blind surveys of large areas possible .
such surveys bypass the need for selection based on the presence of embedded stars or star clusters , infrared sources , masers , or radio continuum emission .
in particular , they can locate molecular clumps before stars form , providing vital information on the initial conditions of star formation .
these surveys can also provide valuable constraints on the physical properties of the clumps , especially masses and mean densities , when combined with distance information .
galaxy - wide surveys are essential for measuring the impacts of the environment on clump properties and star formation activity .
do clump properties vary with galactocentric distance , or with location with respect to spiral arms ?
do they depend on the level of nearby star formation activity ? answering these questions in our galaxy will provide the essential `` ground truth '' required for the analysis of distant galaxies where individual clouds and clumps are not resolved and only galaxy - wide average quantities can be measured .
surveys of the galaxy in low - lying rotational transitions of the co molecule @xcite have identified the locations of large molecular clouds and traced the transition from atomic gas in the outer galaxy to molecular gas in the inner galaxy .
surveys in co have detected a strong preference for gmcs to form along spiral arms @xcite .
recent improved measurements of kinematic distances ( e.g. * ? ? ?
* ) and accurate distances from vlbi @xcite have refined our view of the spiral structure of the milky way .
this new increase in data now allows a more complete investigation of the relationship between spiral density waves and star formation .
locating the early stages of star formation is quite difficult , and often relies on the serendipitous location of a cold , dark cloud which appears as an irdc .
while dark clouds were known from optically obscured regions , the `` discovery '' of irdcs from the msx ( mid - course space experiment ) satellite was presented by @xcite and @xcite .
irdcs have now been extensively cataloged with msx @xcite and glimpse 8 @xmath14 data @xcite . in the past decade , compelling evidence has emerged which suggests that irdcs are the pre - cursors to massive stars , and therefore , stellar clusters ( see , e.g. @xcite ) .
irdcs , however , require the favorable viewing condition of being in front of a bright , mid - ir background for detection .
millimeter wave surveys , however , provide an efficient means of identifying cold , dense dust throughout the galaxy independent of the galactic background , and thus may be seen at large distances .
moreover , the millimeter - wave emission is optically thin , allowing the properties of the cloud to be determined .
millimeter wave surveys such as bgps are essential for understanding massive star and cluster formation on a galactic scale , from the inner to the outer galaxy .
studies of nearby clouds have demonstrated that star formation is far from uniform over molecular clouds , but is concentrated in unusually dense regions called clumps or cores , which fill a small fraction of the area of most clouds @xcite .
following @xcite and @xcite , we use the term ` core ' to refer to a very dense region destined to form an individual star or small multiple star system and the term ` clump ' to refer to a region likely to form a group or cluster of stars .
while clumps forming massive clusters may be even denser than cores forming low mass stars @xcite , in general , clumps are larger and have a lower mean density @xcite . to delineate clumps from their parent clouds as traced in co , line emission from molecules which are high - density tracers , such as , h@xmath15co , cs , hcn , etc .
are extremely useful .
spectral lines also provide excellent diagnostics of line - of - sight motions , temperatures , and densities .
however , the interpretation of these gas tracers is complicated by variations in tracer abundances caused by freeze - out onto grains , sublimation caused by star formation , and subsequent complex chemical processing @xcite .
interpretation is further complicated by uncertainties in optical depths and excitation conditions and the impacts of radiation fields and shocks .
these complications make the derivation from such molecular line data of column densities , masses , and other physical properties of star - forming clumps very difficult .
a further limitation of spectral line surveys is that the tracers of denser gas tend to be weak , and multi - element receiver arrays have only a modest number of elements , making blind surveys very time - consuming .
instruments are beginning to be developed to address this limitation , such as nrao s k - band focal plane array and jcmt s harp - b @xcite .
in contrast , focal - plane arrays containing hundreds of individual bolometers sensitive to millimeter and sub - millimeter ( sub - mm ) radiation are now available .
this advance in technology enables blind surveys of the galactic plane , providing a uniform inventory of massive star - forming and starless clumps .
catalogs resulting from these surveys will provide the data base for subsequent observations in tracers such as hard - to - excite molecular lines and higher resolution studies with instruments like carma and alma .
a significant advantage of continuum observations of dust at long wavelengths is the low optical depth of the dust @xcite , permitting a straight - forward estimate of the mass of the emitting region .
submillimeter observations of galactic sources have been conducted by a variety of different groups focusing on small regions .
the submillimeter common - user bolometer array ( scuba ; @xcite ) on the 15-m james clerk maxwell telescope mapped many significant regions in the galactic plane over its 8 year lifetime ( see @xcite for a summary ) , but never completed a contiguous survey of the galactic plane .
the bolocam instrument has been used on the 10.4-m caltech submillimeter observatory ( cso ) to map nearby large molecular clouds at 1.1 mm and identify hundreds of cores @xcite .
mambo on the 30-m iram telescope has been used at 1.2 mm to map both small , nearby cores @xcite and more distant regions of more massive star formation @xcite .
none of these surveys covered a significant fraction of the galactic plane .
this situation has begun to change dramatically .
recent wide - area surveys have been carried out in the submillimeter bands 250 @xmath14 - 500 @xmath14 by the balloon - borne large aperture submillimeter telescope ( blast ) @xcite .
this work provides an unprecedented view of a wide range of star forming environments and demonstrates the richness of submillimeter datasets , creating great anticipation for the imminent results from instruments aboard _
, particularly the spire camera . from the ground , the atlasgal survey @xcite , using the laboca instrument at 870 with 192 resolution on the apex telescope in chile , has covered 95 square degrees in the galactic plane , @xmath16 and @xmath17 with @xmath18
. the survey will eventually encompass @xmath19 with @xmath20 .
the bgps is the first millimeter survey of a substantial fraction of the galactic plane in the northern hemisphere .
the bgps maps in the first quadrant overlap with the vla galactic plane survey ( vgps ) in hi and radio continuum @xcite , the _ spitzer_-glimpse and glimpse - ii fields @xcite , mipsgal @xcite , the bu - fcrao @xmath21co @xmath22 galactic ring survey @xcite , and atlasgal @xcite , among others .
the bgps coverage will also overlap with the upcoming _ herschel _ higal @xcite and scuba-2/jcmt plane surveys @xcite .
the provides a long - wavelength dust continuum data point , complementing the spectral energy distribution provided by other surveys . by covering the both the galactic center and anti - center regions with a single instrument
, a consistent comparison over the widest possible range of galactocentric radius is possible .
the bgps gives a view of massive star and cluster formation throughout the galaxy , and provides an important finder chart for future high - resolution observations with facilities such as carma and alma .
this paper presents the imaging data from the bolocam galactic plane survey ( bgps ) .
the source catalog is described in a companion paper ( @xcite ; hereafter r10 ) .
the outline of this paper is as follows .
section [ sec : observations ] describes the instrument and the observations .
the data analysis is described in sections [ sec : astrometry ] ( astrometry ) , [ sec : mapping ] ( the map - making algorithm ) , and [ sec : calibration ] ( flux calibration ) .
key discussions are in section [ sec : photometry ] , which explains the effect of the data processing on the accurate recovery of extended structure , and in section [ sec : fluxcomparison ] , which compares the bgps flux with other surveys , including a discussion of a possible systematic offset .
the final image data products and the public release are described in section [ sec : finalmaps ] .
we conclude in section [ sec : discussion ] with a brief discussion of the broad features of the bgps in comparison with other surveys .
we used bolocam to survey more than 170 square degrees of the northern galactic plane .
bolocam is the facility 144-element bolometer array camera mounted at the cassegrain focus of the 10.4-m mirror of the cso on the summit of mauna kea .
we used the filter configuration with a band center of 271.1 ghz ( hereafter 1.1 mm ) and fractional bandwidth @xmath23 ( 46 ghz ) .
the detectors are silicon nitride micromesh absorbers with ntd germanium thermistors , operated at a temperature of @xmath24250 mk .
the bolocam array field - of - view is 75 , with individual detectors having nearly gaussian beams of 31 fwhm .
the spacing between the individual detector beams on the sky is 38 , so the focal plane is not instantaneously fully sampled .
the bolocam instrument is described in greater detail in @xcite .
the maps presented here were acquired during six separate observing sessions at the cso over the course of two years between june 2005 and september 2007 .
the observing epochs are given in table [ tab : observing ] .
bolocam observations were typically scheduled when @xmath25 . between observing epochs ,
bolocam was removed from its mount at the re - imaged cassegrain focus and stored warm .
thus the flux calibration and pointing model were re - computed for each epoch to allow for variations in the instrument and optics .
the final data products were aligned to the well - constrained pointing model created from the epoch v and vi data ( section [ sec : astrometry ] ) .
we found that the flux calibration did not in fact differ significantly between epochs ; see section [ sec : fluxcalibration ] for more details .
figure [ fig : coverage ] shows the coverage of the .
the survey is contiguous over the range @xmath26 . towards the cygnus x spiral arm , the coverage
was flared to @xmath20 for @xmath3 .
in addition , cross - cuts to @xmath27 were made at @xmath28 and 31 .
the total area of this section is 133 square degrees . with the exception of the increase in latitude ,
no pre - selection criteria were applied to the coverage in this region .
in addition to the contiguous region , four targeted regions in the outer galaxy were observed : ic1396 ( 9 square degrees , @xmath5 , @xmath29 ) , a region towards the perseus arm ( 4 square degrees centered on @xmath7 , @xmath8 near ngc7538 ) , w3/4/5 ( 18 square degrees , @xmath30 ) and gem ob1 ( 6 square degrees , @xmath10 ) .
the total area of good coverage for the bgps is 170 square degrees .
our basic observing strategy was to raster scan a field by moving the primary mirror of the cso .
this was done to modulate the astrophysical signal faster than fluctuations in atmospheric opacity .
each field was scanned in a pattern with alternating rasters along lines of constant @xmath31 , followed by a series of rasters along lines of constant @xmath32 .
fields were observed several times with such raster scans to improve signal - to - noise ( s / n ) .
thus each point in the final map of a field was observed multiple times with the array moving in different directions with respect to the astrophysical emission , helping to separate it from atmospheric emission ( see section [ sec : mapping ] for further details on this process ) . in epochs i and ii , the fundamental observing block was a region @xmath33 , covered by 23 rasters along each of @xmath31or @xmath32 . starting with epoch iii , the fundamental block
was changed to @xmath34 for increased mapping efficiency ( i.e. , less time spent in turnarounds at the end of a raster ) .
each @xmath34 block was covered using either 23 rasters along lines of constant @xmath32 , or 67 rasters along lines of constant @xmath31 . in both cases ,
the spacing between adjacent rasters was 162 .
the total time for observing the entire block was 39 ( 48 ) minutes for rasters along constant @xmath32 ( @xmath31 ) .
the data were electronically sampled at 10 hz along the scan direction , slightly higher than the nyquist rate for the scan speed of 120s@xmath12 .
these fundamental observing blocks were stitched together to make 29 large , contiguous maps .
these images define the `` fields '' used for comparison .
each `` field '' includes all observations covering a region between 1 and 6 square degrees .
because the bolocam detector beams do not fully overlap on the sky , if the array is scanned along detector rows , gaps will be left in the resulting map . to ameliorate this
, we began using in epoch ii a field rotator to adjust the rotation angle of the array with respect to the scan direction so that as the scan proceeded the full extent of the fov in the direction orthogonal to the scan was sampled by at least one detector .
this is shown in figure [ fig : array ] . achieving a uniform sampling along the scan - orthogonal direction
is further complicated by a number of missing bolometers .
a simulation was performed to determine the optimal angle to rotate the array with respect to the scan direction to account for both the effects of the beam spacing and missing bolometers . during the turnaround following a raster , the field rotator
is adjusted to this optimum angle for the subsequent raster . without the field rotator
, the coverage shows variations of 100% from pixel to pixel in a single raster of a field . with the rotator ,
this is reduced to @xmath35 .
we constructed an absolute reference system for the bgps by observing bright quasars and blazars near the galactic plane with small raster maps .
the sources were chosen from the sma submillimeter calibrator list , since they have positions reliably determined from interferometric measurements and are point sources at the scale of the bolocam beam .
the distribution of absolute pointing sources over the sky is shown in figure [ fig : pointingcalibrators ] .
pointing observations were performed approximately once every two hours over each night .
the elevation offsets showed a deviation from zero which was empirically well - modeled by a quadratic function of altitude .
no systematic deviation was observed for the azimuth offsets .
the derived model is shown in figure [ fig : pointingmodel ] .
an rms scatter of @xmath36 for the model over the entire observed range was achieved , with somewhat worse scatter in altitude than in azimuth .
in addition to the model for the pointing center , it is necessary to empirically determine the actual projected pattern of the array on the sky , and measure the rotation angle between the focal plane and the sky coordinate system .
this is done by making observations which track all detectors across a bright source ( e.g. a planet ) and making maps from each bolometer individually . in our first data
release , the projected positions of the detectors on the sky are made using the nominal positions without correction for optical distortion .
this results in a slightly reduced effective resolution ( 33 instead of 31 ) .
the effective beam size was measured from fully - sampled beam - mapping observations of planets , and is a good description of the blurring of the beam for the portions of a field observed by all the detectors . within about one fov of the field edges
, however , only a subset of the array observes the field , and so the effective beam size varies slightly at these positions , and may be asymmetric .
because this distortion is asymmetric , it can also result in a small pointing offset .
however , this effect is strongly mitigated by the low number of hits at field edges : this data is generally flagged out .
the residual offset is found to be within the overall pointing model error , but caution should be used when measuring source locations at the field edges in the first data release .
future releases will have the distortion correction applied , removing this effect and improving the effective resolution .
a master pointing field was constructed from observations taken in the epoch with the best - constrained pointing model for that field ( either epoch v or vi ) for every field observed by the .
relative alignment between the master and other observations of the same field was performed by finding the peak of the cross - correlation between images .
the cross - correlation peak was fit with a gaussian profile and the difference between the gaussian peak and the image center was used as the pointing offset .
the offset may be determined accurately to within the error in finding the peak of the gaussian , typically @xmath37 . to create the final image , all observations of a field
were merged into a single timestream with these pointing offsets applied .
this method of alignment makes use of all the information available in the maps , and avoids the ambiguities inherent in using extracted sources to align fields , as the bgps sources are rarely point - like .
it further avoids the slightly larger effective beam and loss of peak flux density which would result if the maps of individual observations were combined using the pointing model alone , where each individual observation would be co - added with the 6 rms uncertainty of the model . in some fields with few sources , particularly those in the range @xmath31=65
to 75 , there was not enough signal to acquire a pointing offset using cross - correlation . in these fields ,
the pointing model was used directly , and so the above mentioned effects affect these fields . all scuba 850 and 450 @xmath14 data has been re - processed in a uniform manner and made publicly available as the scuba legacy catalog ( * ? ? ?
* hereafter slc ) .
these maps allow a cross - check of the accuracy of the bgps pointing model .
we have applied the same cross - correlation procedure used to obtain the relative alignment of bgps observations to compare the bgps to the slc 850 @xmath14 images . @xcite claim a typical pointing accuracy for scuba of @xmath38 , with offsets as large as @xmath39 occurring occasionally .
no further adjustment to the nominal jcmt pointing model was performed in producing the slc maps , and thus we assume these numbers are typical .
there are a number of differences in the bolocam and slc maps which might lead to difficulties in comparing their astrometry .
in particular , different scan strategies were used ( scuba observations were typically taken in jiggle - map mode , which tends to remove extended structure ) , and of course there is the difference in wavelength and beam size between the instruments .
nevertheless , the by - eye morphological comparison between slc 850 @xmath14 and bgps sources is generally excellent , as exemplified in figure [ fig : scubapointingcomparison - a ] .
a cross - correlation technique was used in order to minimize these limitations , since it does not require the establishment of a single position for an object .
the slc has the advantage of having observations over a large portion of the plane , allowing the pointing to be checked ( in spots ) for deviations as a function of position . to test this ,
21 of the 29 bgps fields were used which overlapped with the slc . for these fields , we find @xmath40 , @xmath41 , nearly consistent with zero , and an rms dispersion about the mean bgps position of 62 , consistent with the errors derived for the bolocam pointing model .
the offset for each field is shown in figure [ fig : scubapointingcomparison - b ] .
we have _ not _ corrected the bolocam pointing based on the slc or any external data set , but have produced our final maps with the internally derived model described in section [ sec : pointingmodel ] .
at wavelengths near 1 mm , the emission from the atmosphere dominates any astrophysical emission . for an atmospheric zenith optical depth of @xmath42 = 0.1 , the sky brightness temperature is around 30 to 40 kelvin at typical zenith angles .
this background loading determines the instrument calibration ( see section [ sec : fluxcalibration ] ) , but the primary time - variable signal is due to few percent fluctuations in atmospheric opacity .
the conversion @xmath43 from jy beam@xmath12to kelvin for bolocam is @xmath44 where @xmath45 is boltzmann s constant and @xmath46 is the effective collecting area of the telescope .
thus atmosphere fluctuations are 10 s of jy beam@xmath12 , as compared to the typical brightness of galactic sources , which lie in the range of 0.1 - 1 jy beam@xmath12 .
the essential signal - processing problem to be solved by the mapping algorithm , then , is the estimation of the astrophysical emission in the presence of the fluctuating atmosphere signal , without _ a priori _ knowledge of either . for this work , a custom pipeline was developed to address issues specific to the bgps . it is based on , but significantly improves , the method of @xcite for generating maps and characterizing their properties , and is similar in many respects to the methods used by other bolometer cameras , e.g. , @xcite ( mustang / gbt ) or @xcite ( sharc-2/cso ) .
it also incorporates some of the technical developments of @xcite developed for bolocam data at 2.1 mm .
a key element of the reduction is the iterative estimation of the atmospheric fluctuations and the astrophysical signal .
the atmospheric model is developed from a set of principal components of the bolometer signal time series under the assumption that the bulk of the correlated signal is atmospheric .
after subtraction of the atmosphere model , the astrophysical emission is estimated , then subtracted , and the process is repeated .
we assume the raw timestream data @xmath47 for each bolometer ( indexed by @xmath48 ) at discrete time points ( indexed by @xmath49 ) can be written as @xmath50 where @xmath51 is the astrophysical signal , @xmath52 is the median atmospheric fluctuation seen by all detectors , @xmath53 are terms correlated between bolometers that are distinct from @xmath52 , @xmath54 are non - random signals due to the instrument itself , @xmath55 are long timescale fluctuations not modeled by @xmath52 or @xmath53 , and @xmath56 is irreducible gaussian noise due to photon fluctuations and detector noise .
( each of these terms is discussed in greater detail in the following sections . ) in the presence of purely gaussian noise , the maximum likelihood map @xmath57 can be obtained by minimizing the goodness - of - fit statistic @xmath58 here @xmath47 is the concatenated data from all bolometers ( of dimension @xmath59=number time samples of all detectors ) .
@xmath60 provides the mapping between a given sample in the time - ordered data and the pixel in the sky map to which it corresponds .
@xmath60 is a matrix of dimension @xmath61number of map pixels .
each element of @xmath60 is either 1 or 0 .
note that we are considering the data @xmath47 and the map @xmath57 as vectors ( the mapping to a two - dimensional map being implicit in the matrix @xmath60 ) .
@xmath62 is an @xmath63 matrix which is the inverse of the covariance matrix @xmath64 of the time domain noise @xmath56 , @xmath65 where @xmath66 denotes an ensemble average over many realizations of the random timestream noise process @xmath67 .
further detail about the way in which @xmath62 was estimated is given in section [ sec : noise ] .
the result of the @xmath68 minimization is @xmath69 thus , the mapping from data to map is a linear operator , @xmath70 . for compactness ,
equation [ eq : lsmap ] will be written as @xmath71 = m$ ] .
note that the mapping operator is not invertible .
however , given a map @xmath57 and an observing matrix @xmath60 , there exists a linear operation which makes predictions about the observed timestreams , namely @xmath72 ; this will be denoted @xmath73 = d$ ] . because the model @xmath74 in equation [ eq : chisquared ] only includes the celestial contribution to the observed data @xmath47 , direct use of equation [ eq : lsmap ] will produce a map containing artifacts due to the unmodeled components @xmath52 , @xmath53 , @xmath54 , and @xmath55 .
one approach would be to include templates for these terms as additional rows in @xmath60 and proceed with a simultaneous fit .
in general , however , the correct forms to use are not known _ a priori_. the goal , then , is to produce a time series for each bolometer which as closely as possible approximates @xmath75 so that we may produce the best estimate of the astrophysical signal @xmath76 = s + n$ ] .
we proceed iteratively , estimating each term in equation [ eq : timedomainmodel ] in the order of its relative strength . in the following ,
we denote the best estimate of a time series @xmath77 in the @xmath78 iteration as @xmath79 . in the first iteration
, we estimate the largest signal , @xmath52 , according to the method in section [ sec : atmospheremodel ] , and assume @xmath80 .
we then form @xmath81 this is the best estimate of signal plus irreducible noise at iteration step @xmath82 , and is made into a map @xmath83 = \tilde{m}^{(n)}\ ] ] the current best map @xmath84 is then deconvolved to provide a relatively low - noise , smooth map from which to generate a timestream @xmath85 = \tilde{s}^{(n)}_{it}\ ] ] where @xmath86 represents the deconvolution operation . at this stage , we can subtract both @xmath87 and @xmath88 from the original data , and estimate @xmath89 , @xmath90 , and @xmath91 .
the iterative process begins again with equation [ eq : signalestimate ] .
an example time series showing the successive removal of these models is shown in figure [ fig : iterativemapping ] .
we discuss the estimation of each of the various terms @xmath52 , @xmath53 , @xmath54 , and @xmath55 in the following sections .
it is useful at each step to produce a `` residual map '' @xmath92\ ] ] this residual map serves as a visualization of the progress of the iterative method , since in a perfect process it would be map of @xmath56 .
since atmosphere fluctuations are the largest signal in the raw data , they are modeled first . by construction , the bolocam beams
as they pass through the dominant layer of atmospheric water vapor a few km above the telescope are still highly overlapped , and therefore sample a nearly identical region of atmosphere .
the simplest model for the atmosphere fluctuations makes use of this fact by assuming that the largest correlated signals between detectors are due to to the atmosphere fluctuations ; this is the term @xmath52 in equation [ eq : timedomainmodel ] .
we construct a model of the atmosphere fluctuations which are seen in common by all detectors , which we denote the `` median atmosphere model '' . in each iteration @xmath82
, we construct the non - atmosphere - fluctuation part of the time series as @xmath93 the median atmosphere model is then simply @xmath94\ ] ] each bolometer s timestream is then fit to the median atmosphere model , i.e. , we find @xmath95 such that @xmath87 most closely approximates @xmath96 .
this defines the relative gains of the detectors as @xmath97 the @xmath98 converge fairly rapidly .
( see the discussion in section [ sec : fluxcalibration ] . )
the simple common mode model of equation [ eq : meanatmospheremodel ] does not remove all of the signal correlated at zero time lag between the bolometers .
some of this remaining correlated signal may also be atmosphere .
@xcite found for the 2.1 mm bolocam data that the subtraction of a second order spatial polynomial across the array accounted for nearly all of their residual noise ; however , the 1.1 mm data seem to have additional correlated components that are not completely described by such a model . therefore , to further remove the atmosphere ( and any other correlated instrument noise ) without a detailed physical model , we use a principal component analysis ( pca ) as given in @xcite .
while this method was developed for point source detection in deep , high - redshift surveys , the technique works equally well for this survey , with the caveat that it removes large - scale structure on the size of the array fov .
this effect is mitigated by the iterative process ( the @xcite analysis did not iterate ) .
figure [ fig : pca_graphical ] shows an example of the modes removed by the pca model .
these are a graphical visualization of the terms @xmath53 in equation [ eq : timedomainmodel ] .
most of the instrument error signals have characteristic features which aid in their removal . in particular , unlike the atmosphere emission and correlated signals described above , these signals are _ not _ correlated between detector channels .
the error signals include the following : 1 .
pickup from the 60 hz ac power .
this appears as narrow lines in the power spectral density ( psd ) of the data .
the second harmonic of 60 hz is aliased via beating against the 130 hz bolometer bias frequency into 10 hz with sidebands split at @xmath99 hz .
it is removed by replacing components at these frequencies in the fourier transform with gaussian noise matched to the local mean amplitude .
spikes in voltage due to cosmic ray strikes on the bolometers ( `` glitches '' ) .
an example of a glitch in the time series is shown in figure [ fig : glitches ] , along with the extracted distribution .
microphonic pickup due to vibrations of the receiver .
the microphonic effect is due to a change in the capacitive coupling of the readout wires to the circuit ground , which is converted by the high impedance of the detectors into a measurable signal , much like a condenser microphone .
the most noticeable microphonics occur at the end of each scan when the telescope is turning around and the field rotator is adjusting .
this leads to broad spikes in the time series , whose long decay must be removed from the data , particularly during the beginnings of scans .
fortunately , most of the ac powerline pick - up occurs at frequencies where there is no astrophysical signal , given the beam size and scan speed ( c.f .
figure 1 of @xcite ) . thus this error
is dealt with by first notch filtering at the line frequencies and then low - pass filtering the data .
because of the low correlation with astrophysical signal , this step is performed only once and is not iterated .
both glitches and microphonic pickup from the scan turnarounds are degenerate with astrophysical signal and must be estimated as part of the iterative process .
glitches are identified as large excursions from the rms level after subtraction of the best atmosphere and bright source model , and the data there is excluded in subsequent maps .
the turnaround microphonics are explicitly modeled as decaying exponentials at the beginnings and ends of scans .
residuals from this model may also be removed by the pca cleaning .
after iteratively subtracting the astrophysical model , each scan also has a fifth order polynomial fit in time removed to deal with the longest time scale modes ( @xmath55 in equation [ eq : timedomainmodel ] ) .
these modes are not correlated between detectors .
they correspond to spatial scales much larger than those to which bolocam remains sensitive after subtraction of the pca model , so their subtraction produces maps with fewer `` stripes '' , without appreciably degrading the signal . due to the large volume of data generated by the survey , it was necessary to develop new tools to quickly visualize the data and ensure data quality .
a common tool in radio astronomy is the so - called `` waterfall '' plot , which is an image with frequency on the x - axis , time on the y , and an intensity proportional to the interferometer visibility amplitude or phase .
we have used a variant on this to visualize the bolocam data by displaying bolometer number ( related to the bolometer position in the focal plane ) on the x - axis and time on the y , with the intensity given by the bolometer s response ( in jy ) at that time .
an example of this is shown in figure [ fig : flagger ] .
anomalies such as the glitch shown can be detected manually ( with practice ) and interactively flagged out .
an automated flagger was also created that flags out outlier data on a per - map - pixel basis . in order to make a robust measurement of the variance of the fluxes assigned to each pixel , we used the median average deviation ( mad ) over the data in the pixel and rejected high and low outliers at the 3-sigma level . pixels with too little data to compute a deviation , i.e. those with @xmath100 data points , were also flagged out - these scan - edge pixels are the dominant contribution to the total number of flagged data points .
the fraction of the data flagged by all methods was about 0.08% .
the timestream data is made into a spatial map using the pointing data corresponding to each time point for each bolometer .
the data is weighted by inverse variance across a single scan and then drizzled into a map with @xmath101 pixels using a nearest - neighbor algorithm .
the nearest - neighbor matching allows the map to be returned to a timestream in the same manner , but with the s / n improved by averaging over all hits on a given pixel .
it has the disadvantage that it accentuates the unevenness of the cross - scan sampling , since each timestream point s value is assigned to an area much smaller than the beam . in principle
this can be addressed by various gridding algorithms , and may be a part of future releases of the bgps data . the resulting map is then subjected to a maximum - entropy clean algorithm @xcite using a specified kernel to produce a deconvolved image which is mapped back into the timestream to be subtracted ( equation [ eq : maptotime ] ) .
this has several advantages over using the original map directly , including rejecting artifacts smaller than the kernel and decreasing the noise of the resulting timestream .
it is also better than a `` threshold '' method , which tends to produce discontinuities and include noise outliers . in some fields , flux from some sources
spread out over the course of the iterative process using certain kernel sizes .
the cause of this artifact is not yet well - understood , but using a different kernel size was an effective workaround and did little to change the properties of the final map .
the standard kernel size was a 144 fwhm gaussian ; a 216 kernel was used when artifacts were encountered .
simulations indicate that this change in kernel has a negligible effect on the flux of the source ; see section [ sec : fluxcalibration ] .
we also investigated whether the choice of deconvolution kernel produced a noticeable effect on the recovered flux .
figure [ fig : deconvolution ] demonstrates the effects of using various kernel sizes .
the best results , as determined by the flux recovered in simulations and examination of the residual maps , were produced by a kernel smaller than the beam but larger than the pixel size used .
the weighting of the time stream data used to produce a map is given in equation [ eq : noisecovariance ] .
if the noise were uncorrelated between timestream points , then @xmath102 would be diagonal .
if the noise _ is _ correlated between time samples , but is stationary , i.e. , the noise properties are time - translation invariant @xmath103 and the weighting can be done using the fourier transform of the correlation , which is diagonal : @xmath104 where @xmath105 is the power spectrum of the noise @xmath67 . finally , if the noise is gaussian , then the covariance matrix provides all the necessary statistical information about the noise .
in practice , few of these assumptions hold for the data : the noise is not stationary , due to atmosphere variations , and the unsubtracted astrophysical signal causes the residuals from which we estimate the noise to be non - gaussian .
further , the true remaining correlations between detectors due to unsubtracted atmosphere are difficult to estimate .
our approach , then , is to use a non - optimal but reasonable weighting , estimating the noise from the integral over the power spectrum of the residual timestream for each bolometer and each raster separately .
this accounts for the non - stationarity but ignores the correlations , which are small after the pca subtraction . the residual maps @xmath106 ( equation [ eq : residualmap ] ) form the basis for the noise estimation in the map domain , and also provide a way of estimating the systematic error resulting from imperfect subtraction of bright sources .
these maps are produced for each region from the data residuals after removing the terms @xmath51 , @xmath52 , @xmath53 , @xmath55 , and @xmath54 in equation [ eq : timedomainmodel ] .
an example of a residual map is shown in figure [ fig : deconvolution ] . in an ideal case , the residual map @xmath106 represents a map of a realization of the underlying irreducible noise @xmath56 . however , the imperfect estimation of the signal results in `` ghosts '' of bright sources .
these features are a guide to the number of iterations necessary and an estimate of the remaining systematic error . in constructing the catalog ,
r10found that smoothed versions of the residual maps provide a reliable means of determining the pixel - to - pixel error which accounts for the local variations in the noise , including those due to such artifacts ( c.f . r10 figure 2 ) . because observing conditions varied widely during the survey , the rms noise level varies over the bgps . however , the noise within a given @xmath33 field is fairly uniform in the absence of very bright sources .
the noise level varies between 11 and 53 mjy beam@xmath12 in the inner galaxy .
we show the variation of the depth as a function of galactic longitude in figure [ fig : noisevslongitude ] .
this noise level interacts with the source density to produce the completeness of source extraction .
the variation of completeness as a function of galactic longitude , determined by simulations of the source extraction , is shown in r10figure 9 .
the absolute flux calibration is derived from observations of mars and uranus ( the `` primary calibrators '' ) .
the millimeter - wave flux of these planets is known to @xmath107 @xcite .
the fluxes from these models were extracted using the jcmt s fluxes program with a central frequency 271.1 ghz and bandwidth 46.0 ghz ( table [ tab : colorcorrections ] ) .
calibrations at ( sub)millimeter wavelengths are strongly affected by atmospheric opacity corrections .
further , the detector responsivity of bolocam s bolometers is a non - linear function of the mean atmospheric loading . to address both of these problems and relate observations of the primary calibrators to observations of the bgps fields ,
we make use of the following relation .
the calibration @xmath108 , referenced to the detectors , is given by @xmath109 = { \cal r}(\tau ) \eta a \exp{(-\tau ) } \delta\nu\ ] ] where @xmath110 is the bolometer responsivity ( [ v / w ] ) , @xmath111 is the system optical efficiency , @xmath60 is the effective telescope collecting area , @xmath112 is the bandwidth , and @xmath42 is the line - of - sight , in - band atmosphere opacity . under the assumption that the only power variation on the detectors is due to the power from the atmosphere ( i.e. that astronomical sources are faint relative to the atmosphere ) , which may be parametrized by @xmath42 , @xmath113 is a single - valued function of @xmath42 .
we have used the measured potential difference across the detector thermistor @xmath114 as a proxy for @xmath42 , since the bolometer resistance is a single - valued function of loading .
this quantity is monitored continuously for all observations .
note that this calibration curve folds in both the effects of changing atmospheric transmission and the changes in the detector response with optical loading .
the calibrator field observations were obtained and reduced differently from the science fields .
calibrators were observed in @xmath115 radius fields with two sets of observations , scanned in orthogonal directions . in these small fields ,
the iterative process tended to diverge with more iterations because of instabilities in the deconvolution procedure .
therefore , the calibration curve was determined from non - iteratively - mapped calibrator maps .
a future work will explore applying a more consistent application of the processing to both calibrator and science fields .
a fit to the observed values ( @xmath114 vs. @xmath113 ) for the primary calibrators was performed for each epoch separately .
the agreement between epochs was good , so a single combined calibration was used for all data .
the calibration curve is shown in figure [ fig : calibrationcurves ] .
the resulting error on the calibration curve fit is less than @xmath116 ( statistical ) over the observed range of @xmath42 .
the above flux density calibration only accounts for the average calibration of bolometer volts to janskys .
it is also necessary to account for the variation of bolometer response across the focal plane ( often referred to as `` flat - fielding '' ) .
we do this by monitoring the response to the atmosphere emission in all bolometers ( astrophysical signal is negligible in comparison to the sky except towards sgrb2 ) .
being in the near field , the atmosphere is common to all detectors , and thus this serves as a common relative reference to calibrate out variation in the individual detector responsivities and optical efficiencies .
( for more detail regarding the properties of atmospheric noise above mauna kea as inferred from the bolocam data , see @xcite . )
the change in relative response with loading for a typical detector is shown in figure [ fig : calibrationcurves ] .
the actual application of the method to the time series of the bolometer data is shown in figure [ fig : flatfield ] , where it is clear that correlation of the timestreams is indeed improved by the application of a single multiplicative factor ( the @xmath98 of equation [ eq : relativegains ] ) .
the bolocam point spread function ( psf , or `` beam '' ) is measured using the planets uranus and neptune , which are nearly point sources for bolocam . by stacking all observations of planets , we obtain a high s / n profile of the beam ; this is shown in figure [ fig : psf ] . the main lobe is fit to a gaussian profile .
this yields the effective beam fwhm of 33 , corresponding to a sigma of 142 and equivalent tophat radius of 198 .
the corresponding solid angle is @xmath117 steradians .
the calibration from jy beam@xmath12 into surface brightness ( mjy sr@xmath12 ) is obtained once the beam area is determined .
for the bgps , the conversion factor is @xmath118 for the standard 72 pixel size .
the uncertainty in the mean beam diameter is @xmath119 , leading to a corresponding @xmath120 uncertainty in the beam area ( and in the surface brightness calibration ) .
this factor is in addition to the uncertainty in the best fit @xmath114 versus @xmath113 curve . due to the sidelobes of the beam
, aperture photometry may not capture all of the flux from a point source if the aperture is too small .
r10 used a standard set of apertures , and the multiplicative correction factors to be applied to fluxes obtained from these apertures for a point source are given in table [ tab : aperturecorrections ] .
the correction is only significant for apertures @xmath121 , beyond which effectively all of the flux is included . for extended objects ,
the situation is more complicated because of the effects of the mapping algorithm .
this is discussed further in the next section . to do accurate and meaningful photometry on the bolocam maps
requires a good estimate of the noise ( section [ sec : noise ] ) but also an understanding of the spatial filtering imposed by the observing strategy and the cleaning and mapping of the data .
the additional complexities of defining `` sources '' for the complicated morphologies actually observed are discussed in @xcite .
here we restrict our discussion to the limitations on performing photometry imposed by the data reduction .
the fundamental feature of the data is that there is a degeneracy between large spatial scale astrophysical emission and the atmosphere and other signals with long spatial or temporal variations .
consequently , there are limits to the ability of the algorithm to separate the signal @xmath51 from the atmosphere @xmath52 and time - correlated terms @xmath53 in equation [ eq : signalestimate ] .
inevitably , some of the signal which should be present in @xmath51 is mixed into these terms and is thus not present in the final map @xmath57 . because @xmath51 is generated from a deconvolution algorithm , it has non - zero mean by construction , but importantly , it does _ not _ contain an unbiased estimate of the astrophysical signal on all spatial scales : large spatial scales are preferentially attenuated .
the use of a constant factor to convert to surface brightness implicitly assumes that the response to all angular scales larger than the beam is the same .
in fact , the data reduction process acts differently on sources of different angular extent . in performing photometry , the flux density in an aperture should in fact
be written as @xmath122 where @xmath123 is the flux density in each pixel in jy beam@xmath12 , @xmath124 is the beam solid angle in steradians , and @xmath125 is a correction factor depending on the true angular size of the source .
the factor @xmath126 can deviate significantly from unity for extended sources , and as @xmath127 is generally unknown , this leads to the largest systematic uncertainty in obtaining photometric fluxes . to determine the factor @xmath125 ( the flux recovered as a function of angular scale of the source ) , we performed a series of simulations .
gaussian sources with a range of peak flux densities and sizes were inserted on top of a background consisting of real bgps time series data with the astrophysical source model removed .
the simulated time series were then processed identically to the real data .
the recovered fluxes were measured in ellipses that included single well - separated sources .
figure [ fig : pca_filter ] shows the fraction of flux recovered in the map as a function of source size for a range of different pca components subtracted in the cleaning .
the final reduction used 13 pca components because this produced a reasonable compromise between attenuating extended structure and cleaning the atmospheric contribution . note that this is a substantially more aggressive cleaning than that used in the reduction of the bolocam observation of the _ spitzer _ cores - to - disks fields @xcite , largely because our mean integration time per pixel was @xmath24seconds , whereas theirs was @xmath24minutes .
the recovered flux does depend on the source size , but importantly does _ not _ depend on the source strength except for the strongest sources ( e.g. , sgrb2 ) .
while the precise effect of the flux reduction requires a full simulation of the data reduction pipeline to ascertain , the flux reduction effect can be reproduced with a relatively simple prescription which is intuitively motivated . because the atmosphere model @xmath52 and pca model @xmath53 remove from the timestream `` signals '' which are correlated spatially across the array at the same time , this effectively acts as a spatial filter on the map , removing spatial scales comparable to the array fov .
we have found that a simple `` brick wall '' filter in fourier space which nulls all modes below 0.1 inverse arcminutes ( spatial scales larger than 10 ) reproduces very well the attenuation seen in the simulation .
the comparison between the flux attention produced by this filter and that found from simulation is shown in figure [ fig : pca_filter ] .
the implication is that the bgps is not sensitive at all to scales larger than 10 .
the width of the bolocam passband is such that the effective band center , or equivalently , the flux density referred to a fixed band center , will change somewhat depending on the source spectral energy distribution ( sed ) .
atmospheric transmission variations may also affect the effective passband .
figure [ fig : bandpass ] shows the passband .
bolocam is largely insensitive to variations in the strong water absorption lines at 183 and 325 ghz .
we compute color corrections due to changes in the source spectrum in appendix [ app : colorcorrections ] .
note that we have made no correction for the contamination of the continuum flux densities by emission from lines .
the bolocam passband was specifically constructed to exclude the strong @xmath128 emission line at 230 ghz .
we can estimate the fractional contribution of a line to the continuum as @xmath129 where @xmath130 is the line strength in kelvin , @xmath131 is the line shape ( of width @xmath132 in km s@xmath12 ) , @xmath133 is the bolocam passband ( with equivalent bandwidth @xmath112 and bandcenter @xmath134 ) , and @xmath43 is given by equation [ eq : jyperk ] .
the approximation holds for lines in the center of the band .
the scaling values are chosen to be typical of the line - to - continuum actually seen .
thus any given line is likely to at most contribute a few percent to the measured flux , but in the most extreme star forming environments ( e.g. , sgr b2 ) we may expect that lines may contribute a substantial fraction of the total measured flux density due to the integrated effect of the lines . for an example , see the lower panel of figure [ fig : bandpass ] . for completeness
, we note that towards hii regions , free - free emission may represent a non - trivial fraction of the millimeter flux density in addition to the thermal dust continuum .
this effect is of course best assessed on a case - by - case basis by comparing the bolocam maps with radio surveys , e.g. , the vgps @xcite .
comparison of the bolocam flux densities is easiest for an instrument with the same or similar bandpass , since the additional complication of the source spectral index is avoided .
for this reason , we choose to make an initial comparison of the accuracy of our calibration with the mambo and simba instruments , since their passbands at 1.2 mm are quite close to that of bolocam .
mambo operates on the iram 30-meter diameter telescope on pico veleta in spain , and simba operated on the 15 m sest antenna prior to that facility s closure .
an important future cross - check will be comparison of the bgps flux density calibration with scuba 850 @xmath14 or atlasgal 870 @xmath14 results .
the study of the cygnus - x region by @xcite ( hereafter m07 ) presents one of the largest fields in the galactic plane that has been mapped with mambo which overlaps with the bolocam survey fields .
m07 used both the 37 element mambo and 117 channel mambo-2 instruments .
the fwhm beam size for the observations was 11 .
the mambo beam model resulting from the chopping - scanning observing strategy is described in the appendix of @xcite .
a detailed description of the mambo data reduction is given in @xcite .
m07 used uranus and mars as their fundamental flux calibration , with an estimated uncertainty in the flux of 20% .
the effective passband of the images presented by m07 is about 90 ghz centered at 260 ghz and thus includes the bright j = 21 co transition ( see figure [ fig : bandpass ] ) .
a survey of 2 square degrees towards @xmath135 was conducted by @xcite ( hereafter m09 ) .
these images have 22resolution , and unlike the m07 survey , which used chopping and scanning simultaneously , the simba data were obtained in fast scan mode only . to compare the bgps images to the m07 and m09 images
, we first obtained the data in fits image format .
the m07 data is publicly available .
the m09 data was kindly provided by h. kirk ( private communication ) .
all data were re - gridded to a common projection and coordinate system grid with 4 pixels using the program montage .
we then converted to @xmath136 in a given pixel based on the beam and pixel sizes .
finally , the bolocam map was convolved with either an 11 ( m07 ) or 22(m09 ) fwhm gaussian and the m07 and m09 maps are convolved with a 33 fwhm gaussian so that the effective resolution of the two pairs ( bgps - m07 , bgps - m09 ) is the same .
the convolution preserves the total flux in the map , but changes the aperture in which that flux appears .
this results in maps which , were the calibration and observed features in each map identical , can be compared on a pixel - by - pixel basis .
the plots of pixel values versus each other are shown in figure [ fig : fluxcomparison ] .
the bgps values are indeed correlated with both m07 and m09 over a range of about two orders of magnitude , but with a large amount of scatter .
we attribute the scatter to both noise but primarily to the difference in the handling of extended structure between the different surveys .
however , the average value of the correlation coefficient deviates significantly from 1 in comparing to both m07 and m09 , with a multiplicative factor of 1.21 to 1.51 required to make the bgps fluxes match the other surveys ( or 0.66 to 0.83 to make the other surveys agree with the bgps ) .
this factor also depends on the range over which the fit is done , as given in table [ tab : fluxcomparison ] . including all pixels down to 3 @xmath136 produces systematically larger required correction factor .
using the average value ( to account for variation of the factor with brightness ) yields a factor of 1.35 . for a dust spectral index of @xmath137 ,
the 1.2 mm measurements should have produced a flux that is about 12% lower than that measured by bgps ( see table [ tab : colorcorrections ] ) .
assuming this is a typical spectral index for these clumps , * we therefore note that one must multiply by a factor of 1.35 @xmath138 1.12 = 1.5 ( with a range of plus or minus 0.15 ) the version 1.0 images and catalog released through ipac to obtain good agreement with these surveys .
* the source of this discrepancy remains unclear .
it is unlikely to be due to line contamination in the mambo / simba bands .
line contamination could increase the continuum flux by at most a few percent for mambo ( see equation [ eq : linecontamination ] ) , while being negligible for bolocam . for most sources , it is also not obvious that the discrepancy is due to filtering of extended structure .
the filtering does not reach the requisite 50% needed to explain the difference until the source fwhm reaches 230(figure [ fig : pca_filter ] ) .
however , the simulations were run on gaussian sources , and it is still not certain whether the filter function differs for other source shapes .
the effective filter functions of the simba and mambo surveys are also not well understood .
work is ongoing to thoroughly investigate all potential sources of this discrepancy with further more realistic simulations , a new treatment of atmosphere removal and application of distortion corrections .
in addition , the processing for bright flux densitycalibrators in the current pipeline is different from the processing of the science fields , and attempts are being made to process the different observation types with an identical pipeline .
this analysis will be discussed in a future work .
the final maps are produced by co - adding all observations of a given region .
the maps were processed in 29 separate pieces .
the maps are made in galactic coordinates using a plate carr ( fits header car ) projection .
( this is the same projection used by the -glimpse and bu - fcrao grs surveys . ) by placing them all on the same equatorial grid , mosaicking the maps together is straightforward .
as these maps are near the coordinate system equator , the difference between truly equiareal pixels and the pixels used is at most 0.4% , even 5out of the plane ( as for ic1396 ) .
the pixel size is 72 , chosen to be much smaller than the bolocam beam .
the maps are written to standard fits files .
an example of the fits header for the bgps images is given in appendix [ app : fits_header ] .
all processed maps are available through ipac .
ipac provides a cutout service for the images as well as a searchable version of the catalog provided in @xcite .
the version 1.0 release includes the following images , each covering the same regions : 1 . calibrated maps for all regions observed ( map ) .
the maps are standard astrophysical maps subject to the caveats previously mentioned , the most important being the spatial filtering function ( section [ sec : noise ] ) .
maps of integration time per pixel ( nhitsmap ) .
the nhitsmap shows how many recorded data points from the timestream have been assigned to each pixel ; it is proportional to the dwell time per pixel ( each ` hit ' is 0.1s ) .
maps of the residual time series ( noisemap ) .
the residual map is the map of the model timestream subtracted from the data timestream .
examples of each are given in figure [ fig : sampleimages ] .
future releases may contain additional data products .
the bgps has observed a 170 square degree area at an effective resolution of a gaussian psf of 33 fwhm ( corresponding to an area of @xmath117 ster ) .
the positional uncertainty of the maps is 6 rms .
the maps are calibrated into jy beam@xmath12 using mars as a primary flux standard , and accounting for atmosphere opacity variations in real time .
the bgps is well suited for studying source structure smaller than 3 arcminutes in size .
the processing of the maps removes more than 90% of flux for features with extents larger than 59 , and attenuates the aperture flux of structures extending to 38 by 50% .
we have compared our flux densities to those of m07 and m09 and find that we need to multiply our flux densities up by a factor of 1.5 to match .
we have investigated possible sources for the discrepancy , but do not understand the source of the difference . for now , we recommend multiplying the flux densities in the maps presented here and in the r10 catalog by the factor @xmath139 to obtain consistency with other data sets .
we are continuing to explore the source of this discrepancy , particularly in the calibration and filtering stages of the data pipeline .
the latest data releases and updates on the discrepancy can be found on the bgps ipac website .
maps from the entire bgps are presented in figure [ fig : bgpsmontage ] .
figures [ fig : scubapointingcomparison ] and [ fig : montd ] show the generally excellent morphological agreement between maps and those at other ( sub)millimeter wavelengths .
filamentary structures with aspect ratios up to the maximum spatial dynamic range of the survey ( @xmath140 ) are present throughout the galactic plane .
the most crowded fields consist of frothy and clumpy structure .
the majority of sources are at least moderately resolved .
we suspect that higher resolution observations will resolve bgps clumps into clusters of protostellar objects , as has been seen in serpens @xcite , omc 1 @xcite , and s255n @xcite where bolometer observations at low resolution have been complemented by interferometer observations at high resolution .
figure [ fig : multiwavelength ] shows a comparison of the bgps with _ spitzer_-glimpse 8 @xmath14 @xcite and vgps 20 cm continuum images @xcite . in figure
[ fig : multiwavelength ] , irdcs appear as the bright white clumps , often filamentary , near the mid - plane .
bgps sources are clearly associated with some of the hiiregions detected by the vgps ; a zoom in of this region is shown in figure [ fig : hii ] .
the comparison with 8 @xmath141 m irdcs is particularly striking , with a high degree of correspondence in their morphologies .
detection of an irdc requires that the cold , dense cloud be on the near - side of a bright mid - ir background , whereas bgps sources are subject only to a sensitivity limit .
comparison of the glimpse and bgps data ( see figure [ fig : irdc ] ) show that some bgps sources are clearly associated with irdcs , a few are associated with ir - bright features ( likely due to the association of free - free emission in the millimeter and pah emission at 8 @xmath141 m ) , and some have no clear ir - counterpart .
the bgps sources which are neither bright nor dark at 8 @xmath141 m are probable candidates for `` irdc - like '' clumps ( cold , dense , potential proto - clusters ) on the far - side or outer edges of the galaxy .
thus , the bgps could provide candidate pre - clusters independent of the favorable viewing angle of being on the near - side of bright mid - ir emission .
an exploration of this population , as well as a detailed analysis of the physical properties and star formation tracers of irdcs corresponding to bgps sources may be found in @xcite .
it is worth noting that many features in the bgps maps have no corresponding counterpart in the _ spitzer _ or radio image .
there are clear maxima in the emission filling factor in the galactic center and in the region from @xmath142 to @xmath143 , the line of sight through the densest part of the molecular ring
. additional areas of high source density are seen at @xmath144 and @xmath145 , as shown in figure [ fig : fillingfactor ] .
this result is in contrast to co filling factors which are @xmath99 for @xmath146 @xcite . in parts of the plane where only one co emission line is seen along any given line of sight ( @xmath147 )
, the integrated emission is much lower , but there are still detections of co in every pixel of the @xcite map in the latitude range @xmath148 .
this behavior is consistent with the interpretation that the bgps preferentially detects denser material than co. the connection of the millimeter emission with star forming material awaits the detailed comparison to other star formation tracers .
bgps clumps are , however , associated with well known star forming regions : the brightest clumps visible in our maps are , , , , , and .
these associations suggest that other bright sources of millimeter emission will prove to be massive star- or cluster - forming regions . in some cases
, the distance to bgps clumps can be determined by matching galactic ring survey @xmath21co and 1.1 mm continuum morphology @xcite . however , in other cases the association between the co and 1.1 mm data is not clear , either because of confusion or because the bgps sources are too compact to identify in @xmath21co morphology .
heterodyne follow - up observations using dense gas tracers ( nh@xmath149 , n@xmath15h@xmath150 , hco@xmath150 , and cs ) are being conducted to provide radial velocity , density , chemistry , and temperature measurements @xcite .
we have detected 8454 clumps in the surveyed fields @xcite , providing a sample of clumps suitable for multi - wavelength and high - resolution studies with existing telescopes and future facilities .
this sample is free of the usual biases associated with observing only regions with signposts of star formation , such as hii regions or masers .
the survey depth can be converted into an estimate of limiting mass sensitivity via standard estimates , i.e. , under the assumption of a temperature , opacity , and distance . using the dust opacity from @xcite ( @xmath151 ) , a gas - to - dust mass ratio of 100 , the bolocam bandcenter from table [ tab : colorcorrections ] , and the beam area as given in section [ sec : fluxcalibration ] the mass sensitivity
can be written as @xmath152 at the coldest temperature expected ( 10 k ) , the mass increases by a factor of 2.9 from the mass assuming 20 k. while sensitivity limits will bias the survey against distant , low - mass objects , initial results suggest that the bolocam sources lie at a range of distances , and thus range from cores to clumps @xcite .
we have begun to explore the variation of properties with galactocentric radius with detailed analyses of the bgps galactic center @xcite and anti - center @xcite data .
millimeter - wavelength thermal dust emission reveals the repositories of the densest molecular gas , ranging in scale from cores to whole clouds . by pinpointing these regions ,
the bgps allows the connection of this gas to nascent and ongoing star formation to be explored .
_ facilities : _
we would like to acknowledge the staff and day crew of the cso for their assistance .
the bgps project is supported by the national science foundation through nsf grant ast-0708403 .
j.a . was supported by a jansky fellowship from the national radio astronomy observatory ( nrao ) .
the first observing runs for bgps were supported by travel funds provided by nrao .
support for the development of bolocam was provided by nsf grants ast-9980846 and ast-0206158 .
team support was provided in part by nsf grant ast-0607793 to the university of texas at austin .
we recognize and acknowledge the cultural role and reverence that the summit of mauna kea has within the hawaiian community .
we are fortunate to conduct observations from this mountain .
if an experiment has finite bandwidth ( @xmath153 ) , to report a source surface brightness at a single frequency , one must assume a source spectrum .
the power detected from that source is assumed to be @xmath154 ( here @xmath111 and @xmath155 are the optical efficiency and throughput of the instrument , @xmath156 is the nominal ( assumed ) surface brightness of the source , and @xmath133 is the bandpass transmission normalized to @xmath157 at its peak . )
bolocam is single - moded , meaning that @xmath158 .
the effective band center @xmath134 is defined implicitly by the equation @xmath159 in calculating the color corrections and effective band centers , we consider two types of source spectra .
the most physically motivated is a `` greybody '' spectrum parametrized as @xmath160 } ) b_{\nu}(t)\ ] ] where @xmath161 is the planck function , @xmath162 is the frequency - dependent emissivity , and @xmath163 is the frequency at which the optical depth @xmath164 reaches unity . for a frequency
much lower than the blackbody peak and in the optically thin limit , this reduces to a simple power law @xmath165 the color correction @xmath166 is defined by the relative change in the intensity which would be measured at the same effective band center if the source had a different spectrum @xmath167 than that assumed for calculating the band center ; thus @xmath168^{-1 } \\ & \equiv & k^{-1 } i_0(\nu_c)\end{aligned}\ ] ] we calculate the effective band centers and color corrections for bolocam in table [ tab : colorcorrections ] . the fiducial spectrum for quoting the bandcenter ( chosen because of its closeness to mean galactic properties ;
e.g. @xcite ) is taken to be that of equation [ eq : greybody ] with parameter values @xmath169 k , @xmath170 , and @xmath171 ghz ( 100 @xmath14 ) . the bandcenters and color corrections relative to this fiducial spectrum
are also computed for various values of @xmath172 in equation [ eq : powerlaw ] . over the frequency range of the bolocam passband ,
a power law with @xmath173 is a very good approximation to the assumed greybody .
below is a sample fits header from the version 1.0.2 release of the bgps images described in this paper .
simple = t / written by idl : sun apr 19 17:59:29 2009 bitpix = -32 / bits per pixel naxis = 2 / number of axes naxis1 = 1519 / axis length naxis2 = 651 / axis length extend = f / file may contain extensions origin = ' noao - iraf fits image kernel july 2003 ' / fits file originator iraf - tlm= ' 13:33:03 ( 13/04/2009 ) ' / time of last modification date = ' 2009 - 04 - 13t19:33:03 ' / creation utc ( cccc - mm - dd ) date of fits header lonpole2= 180.000000000 /lonpole
latpole2= 0.00000000000 /latpole decfwhm = 14.4000 /deconvolution kernel fwhm meandc = 2.06882 /mean dc level stddc = 0.499487 /std .
dc level units = ' jy / beam ' / ppbeam = 27.7012530153 /pixels per beam calib_0 = -3.26472e-15 / 0th coefficient for flux cal ( see methods paper calib_1 = 0.398740 / 1st coefficient for flux cal calib_2 = 3.32002 / 2nd coefficient for flux cal ctype1 = ' glon - car ' / coordinate type ctype2 = ' glat - car ' / coordinate type equinox = 2000.00 / equinox of ref . coord .
cd1_1 = -0.00199999986216 / degrees / pixel cd2_2 = 0.00199999986216 / degrees / pixel crpix1 = 766.378744566 / reference pixel in x crpix2 = 345.530969032 / reference pixel in y crval1 = 6.00002238819 / galactic longitude of reference pixel crval2 = -7.44007626964e-05 / galactic latitude of reference pixel pv2_1 = 0.00000000000 /projection parameter 1 protitle= ' bolocam galactic plane survey ' / contact = ' john bally ' /[email protected] maptype = ' map ' / bunit = ' jy / beam ' /units in map bmaj = 0.00988889 / bmin = 0.00988889 / bpa = 0 / bgpsiter= 50 /iteration number bgpsnpca= 13 /number of pca components subtracted bgpsvers= ' 1.0 ' /bgps processing version number waveleng= 1.12000 /mm ( avoids co 2 - 1 ) comment fits ( flexible image transport system ) format is defined in ' astronomy comment and astrophysics ' , volume 376 , page 359 ; bibcode 2001a&a ... 376 .. 359h comment made by the bolocam galactic plane survey ( bgps ) pipeline comment described in aguirre et al 2009 ( not yet published ) comment bgps data was taken at the caltech submillimeter observatory
comment pixel coverage is in the nhitsmap file ( each hit represents .1s dwell ti comment pixel weighting is in the weightmap file comment flag counts are in the flagmap file comment deconvolved model is in the model file history putast : apr 7 09:22:26 2009 world coordinate system parameters written history dates and observation numbers included : history 060616_ob3 history 060621_o20 history 060621_o23 history 060622_o15 history 070701_o27 history 070702_o14 history 070704_o16 history 070715_ob5 history 070706_o22 history 070707_o31 history 070708_o15 history 070708_o16 history 070709_o15 history 070709_o16 wcsdim = 2 cdelt1 = -0.00199999986216 cdelt2 = 0.00199999986216 ltv1 = -64 .
ltv2 = -20 .
ltm1_1 = 1 .
ltm2_2 = 1 .
wat0_001= ' system = image ' wat1_001= ' wtype = car axtype = glon ' wat2_001= ' wtype = car axtype = glat ' end .... top : the position and ellipticity of the detectors in the bolocam focal plane as mapped to the sky .
the size of the pixel corresponds to its fwhm ; it is clear that the beams do not completely overlap on the sky .
the relative response to source is shown by the greyscale , with the darkest being the largest response .
the path of the center of each pixel as the array is scanned from left to right is shown for a particular orientation of the array relative to the scan direction . in this case
, the sampling of the sky orthogonal to the scan direction is poor .
bottom : the same as the top , but with the array rotated to its optimal angle relative to the scan direction .
it is clear that the sampling orthogonal to the scan direction is now much more uniform.fig:array0.7 and @xmath55 models . from top top bottom :
raw time series 2 .
after removal of residual 60hz signal 3
. exponential decay function at scan turnarounds subtracted 4 . a polynomial ( with astrophysical source rejection )
is fit to remove large - scale variations across the scan . ] .
the bar in the bottom right indicates a 10 length . _
top 4 rows _ : a grid of maps of each of the first 20 pca components displayed at the same scale .
it is clear that ncg 7538 , the object imaged , has emission in each component .
there are varying levels of noise in each component , with the first and second being the most obvious atmospheric components .
most of the other noise components are probably detector noise correlated among a subset of the bolometers .
the streaks are residuals of the scan - turnaround noise that was not removed by the exponential model fit . _
bottom 4 rows _ : the same figure , but after 20 iterations .
the figure is the breakdown of the atmosphere remainder ( i.e. the raw data minus the astrophysical model ) into eigenfunctions .
very little astrophysical flux remains at any level of correlation , though there is some at large ( few arcminute ) scales . ] .
the bar in the bottom right indicates a 10 length . _
top 4 rows _ : a grid of maps of each of the first 20 pca components displayed at the same scale .
it is clear that ncg 7538 , the object imaged , has emission in each component .
there are varying levels of noise in each component , with the first and second being the most obvious atmospheric components .
most of the other noise components are probably detector noise correlated among a subset of the bolometers .
the streaks are residuals of the scan - turnaround noise that was not removed by the exponential model fit . _ bottom 4 rows _ : the same figure , but after 20 iterations .
the figure is the breakdown of the atmosphere remainder ( i.e. the raw data minus the astrophysical model ) into eigenfunctions .
very little astrophysical flux remains at any level of correlation , though there is some at large ( few arcminute ) scales . ]
left : an illustration of a `` glitch '' in a single bolometer timestream ( red ) , likely due to a cosmic ray strike .
note the acausal ringing due to the application of the downsampling filter .
the time series for a physically adjacent bolometer is shown in blue , and a bolometer which does not pass over the source in green .
the pca atmosphere estimation has already been subtracted from the data .
right : the distribution of glitch amplitudes flagged and removed from the data in the @xmath174 field with their corresponding rate of occurrence .
positive - going glitches are the solid histogram and negative - going the dashed .
the negative - going glitches are likely the ringing from positive - going glitches .
both distributions are well - described by a power law above @xmath175 jy beam@xmath12 ; below this detection threshold , glitches may contribute to the overall noise.fig:glitches0.8 the effect of deconvolution on the iterative process can be seen in its effect on the residuals in these images of sgr b2 , all displayed at the same scale . from left to right : map , residual , model .
top to bottom kernel size : 144 , 216 , 312 , 72 .
the final version of the pipeline uses 144 , which has the result of leaving no flux in the residual at the location of sgr b2 , and does not `` dig a hole '' in the residual map , as the 72 kernel does .
the 72 kernel also lacks the noise - rejection features of the larger kernels , as can be seen in the bottom right panel.fig:deconvolution1.0 top : the bolocam 1.1 mm bandpass ( thick line ) .
also shown is the atmospheric transmission at mauna kea for 1 mm of precipitable water vapor ( thin line ) .
the grey curve is the mambo-2 bandpass ( taken from http://www.mpifr-bonn.mpg.de/div/bolometer/mambo_parameters_020430.html ) .
note that the bolocam bandpass is well away from the strong water absorption features at 183 and 325 ghz .
bottom : the effect of line contamination is illustrated with the integrated line intensities from the @xcite survey of sgr b2 , who found that 22% of the flux density in one pointing was due to line emission .
note that the bolocam passband rejects @xmath13 of the @xmath176co flux .
however , as shown from the @xcite survey , so@xmath15 and ch@xmath149oh lines can be strong contributors to line flux in the passband due to their broad width .
other lines lying in the bolocam passband include cs(@xmath177 ) and ( @xmath178 ) ( 245 and 293 ghz ) , hcn(@xmath179 ) ( 265 ghz ) and hco@xmath180 ( 267 ghz ) .
fig : bandpass1.0 to @xmath181 in this figure and the following , the brightest sources , e.g. sgr b2 , sgr a , and sources near @xmath144 and @xmath145 , appear to be saturated , but this is only a display artifact .
the astrophysical sources are always much fainter than the atmosphere ( which is within the dynamic range of the detectors ) and therefore do not saturate .
the noise is more pronounced from @xmath182 to @xmath183 because this region was observed less . ] to @xmath184 .
g34.3 + 0.15 , w 51 , w 43 , w 49 , and m 17 appear to be saturated , but this is only a display artifact .
the @xmath185 region through the 4 - 8 kpc molecular ring and approximately the termination of the galactic bar is particularly rich in clumps .
] is extended to @xmath188 .
bottom : a zoom - in of the region near dr21 , showing at left the data from the m07 iram study of this region , and at right the bgps map , in which the filamentary nature of the emission is more apparent .
flux densities from the bgps and m07 are compared in section [ sec : fluxcomparison ] . ]
is extended to @xmath188 .
bottom : a zoom - in of the region near dr21 , showing at left the data from the m07 iram study of this region , and at right the bgps map , in which the filamentary nature of the emission is more apparent .
flux densities from the bgps and m07 are compared in section [ sec : fluxcomparison ] . ] , @xmath8 .
top : bgps , middle : _ spitzer_-glimpse 8 @xmath14 ( 19 ) , bottom : vgps 20 cm continuum ( 40 ) .
irdcs appear in the _ spitzer _ negative image as wispy white regions against the grey emission background . while some of the bolocam sources are apparent as irdcs or strong infrared or radio sources , many lack any apparent association with features at these wavelengths .
the box indicates the zoom - in region in figures [ fig : hii ] and [ fig : irdc ] . ]
tab : observing i & 2005 jul 03 & 2005 jul 09 + ii & 2005 sep 05 & 2005 sep 12 + iii & 2006 jun 02 & 2006 jun 30 + iv & 2006 sep 03 & 2006 sep 19 + v & 2007 jul 01 & 2007 jul 25 + vi & 2007 sep 04 & 2007 sep 09
|
we present the bolocam galactic plane survey ( bgps ) , a 1.1 mm continuum survey at 33 effective resolution of 170square degrees of the galactic plane visible from the northern hemisphere .
the bgps is one of the first large area , systematic surveys of the galactic plane in the millimeter continuum without pre - selected targets .
the survey is contiguous over the range @xmath0 , @xmath1 . towards the cygnus x spiral arm ,
the coverage was flared to @xmath2 for @xmath3 .
in addition , cross - cuts to @xmath2 were made at @xmath4 3 , 15 , 30 and 31 .
the total area of this section is 133 square degrees . with the exception of the increase in latitude ,
no pre - selection criteria were applied to the coverage in this region .
in addition to the contiguous region , four targeted regions in the outer galaxy were observed : ic1396 ( 9 square degrees , @xmath5 , @xmath6 ) , a region towards the perseus arm ( 4 square degrees centered on @xmath7 , @xmath8 near ngc7538 ) , w3/4/5 ( 18 square degrees , @xmath9 ) and gem ob1 ( 6 square degrees , @xmath10 ) . the survey has detected approximately 8400 clumps over the entire area to a limiting non - uniform 1-@xmath11 noise level in the range 11 to 53 mjy beam@xmath12 in the inner galaxy .
the bgps source catalog is presented in a companion paper @xcite .
this paper details the survey observations and data reduction methods for the images .
we discuss in detail the determination of astrometric and flux density calibration uncertainties and compare our results to the literature .
data processing algorithms that separate astronomical signals from time - variable atmospheric fluctuations in the data time - stream are presented .
these algorithms reproduce the structure of the astronomical sky over a limited range of angular scales and produce artifacts in the vicinity of bright sources . based on simulations
, we find that extended emission on scales larger than about 59 is nearly completely attenuated ( @xmath13 ) and the linear scale at which the attenuation reaches 50% is 38 .
comparison with other millimeter - wave data sets implies a possible systematic offset in flux calibration , for which no cause has been discovered .
this presentation serves as a companion and guide to the public data release through nasa s infrared processing and analysis center ( ipac ) infrared science archive ( irsa ) .
new data releases will be provided through ipac - irsa with any future improvements in the reduction .
the bgps provides a complementary long - wavelength spectral band for the ongoing atlasgal and _ herschel_-spire surveys , and an important database and context for imminent observations with scuba-2 and alma .
|
stochastic equicontinuity typically captures the key difficulty in weak convergence proofs of estimators with non - differentiable objective functions .
precise and elegant methods have been found to deal with cases where the data dependence structure can be described by mixing conditions ; see @xcite for an excellent summary
. mixing assumptions are convenient in this context because they measure how events generated by time series observations rather than the observations themselves relate to one another and therefore also measure dependence of functions of such time series .
the downside to these assumptions is that they can be hard to verify for a given application .
@xcite describes alternatives and considers parametric classes of functions that behave like mixingales , but his results come at the expense of lipschitz continuity conditions on these functions and rule out many applications of interest . in this paper
i give simple and easily verifiable conditions under which objective functions of econometric estimators are stochastically equicontinuous when the underlying process is a stationary time series of the form @xmath0 here @xmath1 is a sequence of iid copies of a random variable @xmath2 and @xmath3 is a measurable , possibly unknown function that transforms the input @xmath4 into the output @xmath5 .
the stochastic equicontinuity problem does not have to be parametric and no continuity conditions are needed .
the class allows for the construction of dependence measures that are directly related to the stochastic process and includes a large number of commonly - used stationary time series models .
the next section provides several specific examples . in the following
, @xmath6 denotes @xmath7 and @xmath8 and @xmath9 are outer probability and expectation , respectively ( see * ? ? ?
limits are as @xmath10 .
let @xmath11 be the empirical process evaluated at some function @xmath12 . here
@xmath12 is a member of a class of real - valued functions @xmath13 . in econometric applications
, @xmath13 is typically a parametric class @xmath14 , where @xmath15 is a bounded subset of @xmath16 , although no parametric restriction on @xmath13 is necessary in the following .
define a norm by @xmath17 .
an empirical process is said to be _ stochastically equicontinuous _
( see , e.g. , * ? ? ?
* ) on @xmath13 if for all @xmath18 and @xmath19 , there is a @xmath20 such that @xmath21 as mentioned above , proving stochastic equicontinuity is often the key difficulty in weak convergence proofs .
the next four examples illustrate typical applications .
[ ex : quantilo ] @xcite measure the directional predictive ability of stationary time series @xmath22 with the quantilogram , a normalized version of @xmath23 with @xmath24 and @xmath25 , where @xmath26 is the @xmath27-quantile of the marginal distribution of @xmath22 .
let @xmath28 and @xmath29 . under the null hypothesis of no directional predictability , we have @xmath30 for all @xmath25 .
let @xmath31 be the sample @xmath27-quantile and replace population moments by sample moments to obtain @xmath32 , the sample version of @xmath33 .
apart from a scaling factor , the asymptotic null distribution of the sample quantilogram can be determined through the decomposition @xmath34 if the distribution of @xmath35 is smooth , the delta method can be used to control the first term on the right and , under dependence conditions , an ordinary central limit theorem applies to the second term .
further , we have @xmath36 whenever @xmath37 ( see example [ ex : quantiloc ] below ) . hence , we can take @xmath38 , where @xmath15 is a compact neighborhood of @xmath26 , and
as long as holds , the third term on the right - hand side of the preceding display converges to zero in probability because in large samples @xmath39 [ ex : huber ] robust location estimators can often be defined implicitly as an @xmath40-estimator @xmath41 that nearly solves @xmath42 in the sense that @xmath43 .
popular examples include the median with @xmath44 and huber estimators with @xmath45 for some @xmath46 .
add and subtract in @xmath43 to see that stochastic equicontinuity implies @xmath47 the limiting behavior of @xmath48 can then again be determined through the delta method and a central limit theorem .
[ ex : dominance ] when comparing two stationary time series @xmath49 and @xmath50 , @xmath51 is said to ( weakly ) stochastically dominate @xmath52 over @xmath15 if @xmath53 for all @xmath54 .
let @xmath55 and @xmath56 .
the null of stochastic dominance can be expressed as @xmath57 .
@xcite study weak convergence of the rescaled sample equivalent @xmath58 to the supremum of a certain gaussian process when the null hypothesis is satisfied with @xmath59 .
this convergence follows from the continuous mapping theorem as long as [ ex : dominance_i ] @xmath60 is a totally bounded pseudometric space , [ ex : dominance_ii ]
@xmath61 converges in distribution for every finite set of points @xmath62 in @xmath15 , and @xmath63 is stochastically equicontinuous ; see , e.g. , @xcite .
condition can be shown to hold if @xmath15 is bounded and @xmath5 has lipschitz continuous marginal distribution functions .
condition can be verified with the help of a multivariate central limit theorem .
[ ex : cquantreg ] @xcite develop bahadur representations for quantile regression processes arising from a linear latent variable model with outcome @xmath64 and covariate vector @xmath65 .
denote the random censoring time by @xmath66 .
only @xmath65 , the random event time @xmath67 , and the associated censoring indicator @xmath68 are observed .
let @xmath69 and @xmath70 . as @xcite point out in their remark 3.2 , a key condition for the validity of their bahadur representations under dependence is stochastic equicontinuity of @xmath63 .
stochastic equicontinuity can not hold without restrictions on the complexity of the set @xmath13 ; see , e.g. , @xcite . here ,
complexity of @xmath13 is measured via its bracketing number @xmath71 , the smallest number for which there are functions @xmath72 and functions @xmath73 ( not necessarily in @xmath13 ) such that @xmath74 and @xmath75 for all @xmath76 .
in addition , some restrictions are required on the memory of the time series . for processes of the form
, the memory is most easily controlled by comparing @xmath5 to a slightly perturbed version of itself ( see * ? ? ? * ) .
let @xmath77 be an iid copy of @xmath78 , so that the difference between @xmath5 and @xmath79 are the inputs prior to period @xmath80 .
assume the following : [ as : coupling ] let @xmath13 be a uniformly bounded class of real - valued functions with bracketing numbers @xmath81 .
then there exists some @xmath24 and @xmath82 such that [ as : coupling1 ] @xmath83 and [ as : coupling2 ] @xmath84 for any given @xmath85 .
assumption [ as : coupling ] is a short - range dependence condition .
proposition [ prop : gmc ] below presents a device to establish this condition and shows that assumption [ as : coupling ] often imposes only mild restrictions on the dependence structure . at the end of this section , i provide a detailed discussion of how to verify this assumption for examples [ ex : quantilo][ex : cquantreg ] . because @xmath13 is assumed to be uniformly bounded , the bounding functions @xmath86 can be chosen to be bounded as well .
hence , in view of lemma 2 of @xcite , the exact choice of @xmath87 is irrelevant , for if assumption [ as : coupling ] holds for some @xmath87 , then it holds for all @xmath82 .
assumption [ as : coupling ] and a complexity requirement on @xmath13 given by a bracketing integral imply a strong form of stochastic equicontinuity .
the following theorem ( the proof of which is found in the appendix ) is similar to @xcite s ( @xcite ) theorem 2.2 with their mixing condition replaced by assumption [ as : coupling ] .
it implies via the markov inequality .
[ l : strongequi ] suppose that assumption [ as : coupling ] holds and @xmath88 for some @xmath89 and an even integer @xmath90 .
then for every @xmath91 , there is a @xmath20 such that @xmath92 as a referee points out , assumption [ as : coupling ] is not primitive and therefore the application at hand determines how difficult it is to verify this assumption .
suitable sufficient conditions can be obtained , e.g. , if @xmath93 and @xmath94 can be uniformly bounded above by constant multiples of @xmath95 for some @xmath96 .
since the expectation operator is a smoothing operator , these bounds can hold even if @xmath12 and @xmath86 are not lipschitz or , more generally , hlder continuous . assumption [ as : coupling ] is then satisfied as long as the geometric moment contraction ( gmc ) property of @xcite holds , i.e. , there is some @xmath97 and @xmath98 such that @xmath99 .
time series models with the gmc property include , among many others , stationary ( causal ) arma , arch , garch , arma - arch , arma - garch , asymmetric garch , generalized random coefficient autoregressive , and quantile autoregressive models ; see @xcite and @xcite for proofs and more examples .
the problem in examples [ ex : quantilo][ex : cquantreg ] and in a variety of other applications is the appearance of one or more indicator functions that cause kinks or discontinuities in the objective function .
the following result ( a generalization of proposition 3.1 of * ? ? ? * ) combines the gmc property and smoothness conditions on the distribution of the underlying stochastic process to generate the kinds of bounds needed for assumption [ as : coupling ] when indicator functions are present .
i discuss the result in the examples below . here and in the remainder of the paper ,
if @xmath5 is vector - valued and has a subvector @xmath35 , then the corresponding subvector of the perturbed version @xmath100 is denoted by @xmath101 . [ prop : gmc ] suppose @xmath102 has the form and @xmath3 takes values in @xmath103 .
assume that uniformly in @xmath104 , @xmath105 is lipschitz in @xmath106 on an open interval containing @xmath107 .
suppose @xmath108 has the gmc property .
if either [ prop : gmc_single ] @xmath109 is a singleton , or [ prop : gmc_func ] there is a measurable function @xmath110 with @xmath111 and @xmath112 is bounded , or [ prop : gmc_indep ] @xmath113 and @xmath114 are independent conditional on @xmath115 and @xmath112 is bounded , then @xmath116 and @xmath117 for some @xmath24 and all @xmath82 . before concluding this section , the next four examples illustrate how to apply and how to verify assumption [ as : coupling ] in practice .
[ ex : quantiloc ] suppose @xmath118 is lipschitz on an open interval containing @xmath15 .
take a grid of points @xmath119 and let @xmath120 .
given a @xmath121 , we can then find an index @xmath122 such that @xmath123 , where i used the fact that @xmath124 .
moreover , by stationarity @xmath125 which is bounded above by a constant multiple of @xmath126 due to lipschitz continuity . hence ,
if @xmath127 for all @xmath76 , we can choose bracketing numbers with respect to @xmath128 of order @xmath129 as @xmath130 ( see * ? ? ?
* ; * ? ? ?
* ) and the bracketing integral converges , e.g. , for @xmath131 and @xmath132 . by the same calculations as for @xmath133 , all @xmath134 satisfy @xmath135 as @xmath136 and therefore @xmath36 if @xmath37 .
in addition , suppose the gmc property holds . because @xmath137 uniformly in @xmath138 by love s @xmath139 inequality , apply proposition [ prop : gmc ] twice with @xmath140 and @xmath141 to see that assumption [ as : coupling ] is also satisfied .
in both cases we can take @xmath142 , @xmath143 , and @xmath144 ( say ) .
the same reasoning applies to @xmath86 .
nearly identical arguments as in the preceding example yield stochastic equicontinuity for the median . for the huber estimator ,
take the grid from before and note that we can find a @xmath122 such that @xmath145 . a routine argument ( * ? ? ?
* ; * ? ? ?
* example 19.7 , pp .
270271 ) yields bracketing numbers of order @xmath146 as @xmath130 ; the bracketing integral is finite , e.g. , for @xmath131 and @xmath147 .
assumption [ as : coupling ] can be verified via the bound @xmath148 and holds trivially .
[ ex : dominancec ] replace @xmath149 in example [ ex : quantiloc ] with @xmath150 .
a slight modification of the arguments presented there to account for the fact that the distribution functions of @xmath51 and @xmath52 are not identical yields stochastic equicontinuity of @xmath63 .
@xcite also consider a more general case with @xmath151 for @xmath152 . here
@xmath150 is unobserved and @xmath153 has to be estimated by some @xmath154 .
let @xmath155 and @xmath156 . if @xmath157 , the asymptotic null distribution of the test statistic @xmath158 can be derived from an application of the continuous mapping theorem in @xmath159 under appropriate conditions on @xmath160 , the behavior of the first term on the right follows from the functional delta method
; see lemma 3 and 4 of @xcite .
the asymptotic properties of the second term are the same as those of @xmath63 from the simple case where @xmath150 is observed directly .
verifying stochastic equicontinuity of @xmath161 to control the third term is more involved .
let @xmath162 be a bounded neighborhood of @xmath163 and assume the conditional distribution functions of @xmath164 given @xmath165 , @xmath152 , are lipschitz continuous .
the bounded set @xmath166 can then be covered by balls with suitably chosen radius @xmath167 and centers @xmath168 , @xmath169 , so that the bracketing integral in theorem [ l : strongequi ] converges ( * ? ? ?
* lemma 1 ) . here
we take @xmath170 and @xmath171 where @xmath172 is euclidean norm . for fixed @xmath122 and @xmath173 ,
define a function @xmath174 such that the first term on the right - hand side of the preceding display equals @xmath175 .
assume @xmath5 has the gmc property . by the reverse triangle and cauchy - schwartz inequalities , @xmath176 has this property as well
. apply proposition [ prop : gmc ] with @xmath177 , @xmath178 , @xmath179 , and @xmath180 to establish @xmath181 .
the same holds for the remaining indicator functions with appropriate choices of @xmath174 .
conclude @xmath182 for which verifies assumption [ as : coupling ] .
further , write @xmath183 so that for all @xmath184 we can bound @xmath185 , @xmath54 and @xmath186 , by a constant multiple of @xmath187 view the two suprema in the display as suprema over @xmath188 for @xmath152 and apply proposition [ prop : gmc ] twice with @xmath189 , @xmath190 , @xmath191 to see that assumption [ as : coupling ] holds as well . following ( * ? ? ?
* condition c1 ) , assume there is some fixed @xmath46 with @xmath192 .
add and subtract to see that @xmath193 does not exceed a constant multiple of @xmath194 if @xmath195 has the gmc property , then the first term is of size @xmath196 by @xcite s ( @xcite ) lemma 2 .
proposition [ prop : gmc ] can be used to establish the same order of magnitude for the second term as long as the conditional distribution of @xmath64 given @xmath65 satisfies the smoothness condition stated in the proposition .
in addition , @xcite assume that @xmath64 and @xmath66 are independent conditional on @xmath65 .
the size of the third term is then also @xmath196 by proposition [ prop : gmc ] .
this verifies assumption [ as : coupling ] . an argument similar to the one provided in example [ ex : dominancec ] establishes assumption [ as : coupling ] .
this follows from a simple modification of @xcite s ( @xcite ) proof of their theorem 2.1 .
the proof requires three steps : replace their @xmath197 with @xmath122 and their @xmath198 with @xmath199 ; then apply lemma [ l : moments ] below instead of @xcite s ( @xcite ) lemma 3.1 in the derivation of their inequality ( 3.5 ) to deduce @xmath200 and use this in ( i ) instead of their inequality ( 3.5 ) .
another application of the lemma establishes the required analogue of their inequality ( 3.5 ) used in ( ii ) .
the same inequality can also be applied in ( iii ) .
the other arguments remain valid without changes .
[ l : moments ] let @xmath201 for some @xmath89 and suppose that assumption [ as : coupling ] holds . for all @xmath202 , all @xmath203 , and
every even integer @xmath90 we have @xmath204 where @xmath205 depends only on @xmath206 , @xmath207 , and @xmath27 .
the inequality remains valid when @xmath208 is replaced by @xmath86 for any given @xmath209 .
let @xmath210 .
assume without loss of generality that @xmath211 for all @xmath212 ; otherwise rescale and redefine @xmath205 . define @xmath213 and note that @xmath214 for all @xmath215 and all @xmath203 because @xmath5 and @xmath100 are identically distributed . for
fixed @xmath216 , @xmath217 , and @xmath218 , consider integers @xmath219 so that @xmath220 .
since @xmath221 and @xmath222 are stationary , repeatedly add and subtract to see that @xmath223 in particular , the last summand on the right - hand side is zero because @xmath224 and @xmath225 are independent and @xmath226 and @xmath225 are identically distributed . for a large enough @xmath227 and some @xmath228 , assumption [ as : coupling ] and distributional equivalence of @xmath229 and @xmath230
imply @xmath231 .
hlder s inequality then bounds the first term on the right - hand side of the preceding display by @xmath232 where the reciprocals of @xmath87 , @xmath233 , and @xmath234 sum to @xmath80 . proceeding similarly to @xcite , another application of the hlder inequality yields @xmath235 whenever @xmath236 and similarly @xmath237 whenever @xmath238 .
suppose for now that @xmath239 .
if @xmath240 , take @xmath241 and @xmath242 . decrease the resulting exponent of @xmath243 from @xmath244 to @xmath245 so is bounded by @xmath246 . if @xmath216 and @xmath247 , the factor @xmath248 is not present in , but we can still choose @xmath241 and @xmath249 to obtain the same bound .
identical arguments also apply to each of the other summands in .
hence , we can find some @xmath250 so that @xmath251 in fact depends on @xmath122 , but this does not disturb any of the subsequent steps .
now replace ( a.2 ) in @xcite by the inequality in the preceding display .
in particular , replace their @xmath252 with @xmath253 and their @xmath254 with @xmath255 .
the rest of their arguments now go through without changes .
the desired result for @xmath86 follows mutatis mutandis : simply define @xmath256 , repeat the above steps , and invoke assumption [ as : coupling ] in place of assumption [ as : coupling ] .
consider the second term in the preceding display .
denote the distribution function of @xmath113 conditional on @xmath115 by @xmath261 . by assumption
, we can always find a large enough @xmath262 such that for all @xmath263 the set @xmath264 is contained in the interval on which @xmath261 is lipschitz .
apply the markov inequality , the gmc property , conditional independence , and continuity to write @xmath265 the first term in satisfies @xmath266 for the same reasons .
take @xmath267 and combine these bounds via the love @xmath139 inequality to establish the desired result .
assertion also follows because the above arguments remain valid with @xmath268 . for , we have @xmath269 and hence @xmath270 .
the second term in is then no longer present , which removes the boundedness restriction on @xmath112 .
finally , note that all of the above remains valid when strong inequalities are replaced by weak inequalities .
|
in this paper i provide simple and easily verifiable conditions under which a strong form of stochastic equicontinuity holds in a wide variety of modern time series models .
in contrast to most results currently available in the literature , my methods avoid mixing conditions .
i discuss several applications in detail .
|
silicon carbide ( sic ) is a wide - band - gap semiconductor with excellent physical , electronic and mechanical properties@xcite such as high thermal conductivity , high breakdown field , low density , high saturation velocity , high mechanical strength , and stability at high temperature .
these exceptional features make sic a promising candidate to replace silicon in electronic devices operating in high - power , high - frequency , and high - temperature regimes.@xcite in the last years , sic nanostructures ( like nanospheres,@xcite nanosprings,@xcite nanowires@xcite and nanotubes@xcite ) have been successfully synthesized , and several theoretical and experimental works @xcite have been performed to investigate their structural and electronic properties .
the unique features of sic combined with quantum - size effects make the sic nanostructures interesting materials for nanotechnology applications .
for instance , sic nanowires and nanotubes have been considered as candidates for hydrogen storage nanodevices@xcite and for building blocks in molecular electronic applications.@xcite in particular , silicon carbide nanowires ( sic nws ) have excellent field emission properties,@xcite high mechanical stability and high electrical conductance,@xcite and could be used as nanoscale field emitters or nanocontacts in harsh environments .
some optical and electronic properties of semiconductors may be modified by the presence of defects .
the most common defects in sic are vacancies , interstitials , antisites and clusters .
these defects are mainly formed during the growth process and ion implantation of dopants .
vacancies and interstitials of c and si in 3c- , 4h- and 6h - sic bulks have been thoroughly investigated in theoretical and experimental works.@xcite these investigations have showed that the c and si vacancies are electron and hole traps,@xcite and that c and si interstitials have higher mobility than vacancies , although the mobility of point defects in sic is reduced as compared to another semiconductors ( like silicon).@xcite the high mobility of carbon interstitials favors the formation of carbon - interstitial clusters . using
_ ab initio _ methods , gali et al.@xcite systematically investigated small clusters of carbon interstitials and antisites in 3c- and 4h - sic bulks , and verified that the formation of carbon aggregates is energetically favored . in spite of some theoretical studies on carbon aggregates in sic
bulk,@xcite the investigation of carbon clusters in sic nws is very scarce .
motivated by the lack of studies on c aggregates in sic nws and by their energetically favorable formation in sic bulk , in this work we performed an _ ab initio _ study of small carbon clusters in sic nws .
we considered hydrogen - passivated sic nws grown along the [ 111 ] and [ 100 ] directions and examined clusters with interstitial and antisites carbon atoms .
the formation energies of these clusters were determined as a function of the c concentration .
we calculated the binding energies of some aggregates , and our results indicate that the formation of carbon clusters in sic nanowires is energetically favored . besides [ 111 ] and [ 100 ] sic nws ,
the carbon clusters were also investigated in 3c - sic bulk , in order to compare the effect of c defects in sic bulk and nw .
in this work we present _ ab initio _ calculations based on density functional theory@xcite ( dft ) carried out by using the siesta code.@xcite we used local spin - density approximation@xcite ( lsda ) for the exchange - correlation functional and norm - conserving fully - separable pseudopotentials@xcite to treat the electron - ion interactions .
the kohn - sham orbitals were expanded using a linear combination of numerical pseudoatomic orbitals@xcite and a double - zeta basis set with polarization functions@xcite ( dzp ) was employed to describe the valence electrons . the 3c - sic bulk , and the [ 111 ] and [ 100 ] sic nws , were modeled within the supercell approach , with 128 , 232 and 279 atoms , respectively .
the sic nws were constructed from the 3c - sic structure and the dangling bonds of their surfaces were saturated with hydrogen atoms . due to the periodic boundary conditions ,
a vacuum region of about 10 was used to avoid interactions between a nw and its image
. we have considered axial lengths ( along the nw growth direction ) of about 15.2 and 13.2 for the [ 111 ] and [ 100 ] sic nws , respectively .
the geometries were optimized using the conjugated gradient scheme , within a force convergence criterion of 0.05 ev@xmath0 .
the brillouin zone was sampled by using 2 special * k * points for 3c - sic bulk , and 1 special * k * point for [ 100 ] and [ 111 ] sic nws .
we verified the convergence of our total - energy results with respect to the number of special * k * points using up to 9 * k * points for 3c - sic bulk and 4 * k * points for [ 100 ] and [ 111 ] sic nws .
we examined c clusters in 3c - sic bulk and in hydrogen - passivated sic nws grown along the [ 111 ] and [ 100 ] directions.@xcite we considered sic nws with diameter of about 10 , constructed from the 3c - sic structure.@xcite for [ 100 ] sic nws , two kinds of wires were studied : carbon - coated [ 100 ] sic nw and silicon - coated [ 100 ] sic nw , whose surfaces are , respectively , carbon and silicon terminated .
figure [ fig1 ] presents the cross - section view of the structural models of carbon - coated [ 100 ] sic nw [ fig .
[ fig1](a ) ] , silicon - coated [ 100 ] sic nw [ fig .
[ fig1](b ) ] and [ 111 ] sic nw [ fig .
[ fig1](c ) ] . note that the [ 111 ] sic nw surface has the same number of si and c atoms .
cross - section view of the structural models of ( a ) c - coated [ 100 ] sic nw , ( b ) si - coated [ 100 ] sic nw and ( c ) [ 111 ] sic nw .
carbon , silicon and hydrogen atoms are represented by filled , empty and small - empty circles , respectively . ]
optimized geometries of the carbon defects investigated in 3c - sic bulk and in hydrogen - passivated [ 100 ] and [ 111 ] sic nws .
( a ) c@xmath1 , ( b ) c c , ( c ) c
c@xmath1 , ( d ) c si , ( e ) ( c c)@xmath2 , ( f ) 2(c c)@xmath3 , ( g ) [ ( c c)@xmath1 + ( c c)@xmath3 ] , and ( h ) 2(c c)@xmath5 . ] eight c defects were investigated@xcite in the present work ( see fig .
[ fig2 ] ) : c@xmath1 ( carbon antisite ) , c c ( @xmath6 split of c interstitial + c site ) , c
c@xmath7 ( @xmath6 split of c interstitial + c antisite ) , c si
( @xmath6 split of c interstitial + si site ) , ( c c)@xmath2 , 2(c
c)@xmath3 , [ ( c c)@xmath1 + ( c c)@xmath3 ] , and 2(c c)@xmath1 . in order to examine the energetic stability of the c defects in the 3c - sic bulk and in the [ 100 ] and [ 111 ] sic nws
, we used the grand canonical potential at @xmath8 k ( as suggested in the refs . and )
@xmath9 here @xmath10 is the total energy of the considered structure , @xmath11 is the chemical potential of atomic specie @xmath12 ( @xmath13 si , c or h ) and @xmath14 is the number of atoms @xmath12 in the system . in the thermodynamic equilibrium,@xcite @xmath15 and @xmath16 where @xmath17 is the formation heat of the sic bulk . employing diamond structure@xcite for bulk phases of si and c , and zincblend structure@xcite for sic , we found @xmath18 ev , which is in good agreement with the experimental value @xmath19
ev.@xcite the upper limits of @xmath20 and @xmath21 are @xmath22 and @xmath23 , respectively .
hence , the potential fluctuations @xmath24 and @xmath25 are restricted to @xmath26 this range defines the c - rich ( or si - poor , where @xmath27 and @xmath28 ) and si - rich ( or c - poor , where @xmath29 and @xmath30 ) limits . in table
[ tab1 ] we present the formation energies , relative to the pristine structures , of c defects in 3c - sic bulk , and in hydrogen - passivated [ 100 ] and [ 111 ] sic nws .
the two values presented in each column correspond to si - rich ( or c - poor ) and c - rich limits . for 3c - sic bulk at stoichiometric condition we found that the formation energies of c@xmath1 , c c , c
c@xmath1 , ( c c)@xmath2 , 2(c c)@xmath1 and @xmath31(c c)@xmath1 + ( c c)@xmath3@xmath32 $ ] are 3.49 , 7.38 , 6.87 , 8.79 , 9.51 and 11.55 ev , respectively
. these results are in good agreement with those obtained by gali et al . in ref . .
[ cols="<,<,<,<,^,<,^,<,^ " , ] [ tab2 ] we next examine the energy gain upon the formation of c clusters through the combination of interstitial c atoms . in this case , the energy gain was determined by comparing the total energies of the separated systems ( @xmath33 ) and the total of a given ( stoichiometrically equivalent ) c cluster ( @xmath34 ) .
we define the binding energy of the c cluster as @xmath35 .
our calculated binding energies are summarized in table [ tab2 ] .
the results presented in each column correspond to the stoichiometric condition . in 3c - sic bulk or sic nw with a carbon antisite ( c@xmath7 defect )
, a carbon interstitial can be captured by c site , si site or c antisite .
the last process is more favorable than the first ( second ) one by 4.00 ev ( 4.60 ev ) in 3c - sic bulk , 3.93 ev ( 4.68 ev ) in c - coated [ 100 ] sic nw , 4.95 ev ( 4.65 ev ) in si - coated [ 100 ] sic nw , and 3.62 ev ( 3.98 ev ) in [ 111 ] sic nw .
considering two carbon interstitials , the formation of carbon clusters is energetically favored in 3c - sic bulk and in both [ 100 ] and [ 111 ] sic nw . in fact , the aggregates ( c c)@xmath36 and 2(c c)@xmath3 are more stable than two isolated c c defects by 5.51 and 2.88 ev in c - coated [ 100 ] sic nw , and by 6.11 and 3.61 ev in [ 111 ] sic nw .
still , in si - coated [ 100 ] sic nw the energy gain is 13.6 ev to form 2(c
c)@xmath1 from isolated ( c c ) and c@xmath1 defects , and 3.7 ev to form 2(c
c)@xmath1 from two isolated c c@xmath37 . in 3c - sic bulk
these energy gains are 10.2 and 2.2 ev , respectively . here
we can infer that , in general , the formation of c clusters is quite likely in sic nws , in particular for thin si - coated [ 100 ] sic nws .
it is worth to note that , in order to get a complete picture of the formation of c clusters , the c diffusion mechanism is an important issue , however , it is beyond the scope of the present work .
( color online ) electronic band structure and the projected density of states ( pdos ) of si - coated [ 100 ] sic nw .
( a ) pristine nw , and with the presence of defects , ( b ) c - si , ( c ) 2(c - c)@xmath37 , and ( d ) c c@xmath1 . in the pdos diagrams ,
the shaded regions correspond to the total density of states ; in the electronic band structure diagrams , solid lines ( dotted lines ) represent the spin - up ( spin - down ) electronic channels .
the red dashed lines correspond to the position of the fermi level and the zero energy corresponds to the valence band maximum . ] focusing on the electronic properties , we find that similar to the 3c - sic bulk phase,@xcite c@xmath37 is an electrically inactive defect in sic nws .
c@xmath1 does not introduce states within the fundamental band gap . on the other hand
, the formation of c clusters gives rise to electronic states lying within the energy bandgap .
figure [ pdos - nano100si](a ) presents the electronic band structure and the projected density of states ( pdos ) of the pristine si - coated [ 100 ] sic nw .
the pristine system exhibits an energy bandgap ( at the @xmath38 point ) of 1.98 ev , calculated within our dft - lsda approximation , where the highest occupied ( lowest unoccupied ) states are mostly composed by c 2p ( si 3p ) orbitals .
the formation of c clusters give rise to localized electronic states within the bandgap , as depicted in figs .
[ pdos - nano100si](b)[pdos - nano100si](d ) .
c si defect gives rise to an occupied ( empty ) state at 1.2 ev ( 1.5 ev ) above the valence band maximum [ fig [ pdos - nano100si](b1 ) ] .
the dispersionless character of those states , along the @xmath38x direction ( _ i .
parallel to the nw growth direction ) , indicate that those defects states are localized around c si . indeed , the pdos diagram [ fig . [ pdos - nano100si](b2 ) ] indicate that there is a significant contribution from the interstitial c atom to the occupied defect state , while the nearest neighbor si atom contributes to the formation of the lowest unoccupied defect state . at the c
- rich limit , the 2(c
c)@xmath1 defect is the most favorable one .
its electronic band structure is depicted in fig .
[ pdos - nano100si](c1 ) .
we find an occupied state at 1 ev above the vbm , and three unoccupied states lying within an energy interval of 2.1 and 2.6 ev above the vbm .
those defect states are mostly localized around the interstitial c atoms [ fig .
[ pdos - nano100si](c2 ) ] .
finally , for c c@xmath37 ( energetically most likely at the stoichiometric condition ) we find the formation of spin - unpaired states within the nearby the fermi level , figs .
[ pdos - nano100si](d1 ) and [ pdos - nano100si](d2 ) .
those defect states a localized around c interstitial atoms . in summary , different from c@xmath37 defects , we verify that the formation of c clusters in si coated [ 100 ] sic nws gives rise to deep states within the bandgap , which may acts as a trap to the electronic carriers .
we performed an _ ab initio _ investigation of small carbon clusters in sic bulk and nws .
we examined clusters with carbon interstitials and antisites in 3c - sic bulk and in hydrogen - passivated [ 100 ] and [ 111 ] sic nws .
we observed that the composition of the sic nanowire surface strongly influences the formation of the carbon clusters .
in fact , c@xmath37 is the energetically most favorable configuration in 3c - sic bulk , and in c - coated [ 100 ] and [ 111 ] sic nws , but is not expected to occur in si - coated [ 100 ] sic nw . the energetically most stable defect in si - coated [ 100 ]
sic nw is c si at si - rich limit , c
c@xmath1 at stoichiometric conditions and 2(c - c)@xmath1 under c - rich conditions . comparing the total energies of the c
c and c c@xmath1 defects ( in 3c - sic bulk and in [ 100 ] and [ 111 ] sic nws ) , we find that c c@xmath1 is more stable than c c at stoichiometric and c - rich limits .
this finding is in accordance with the larger relaxation required to put two c atoms in a c site than to put two c atoms in a si site .
however , we observed that c
c becomes energetically more stable than c
c@xmath37 for low c concentration .
further total energy calculation indicate that the formation of carbon clusters in 3c - sic bulk and in both [ 100 ] and [ 111 ] sic nws is energetically favored .
finally , our electronic band structure calculations indicate that ( i ) similar to the 3c sic bulk phase , c@xmath37 defects are electrically inactive , while ( ii ) in si coated [ 100 ] nws the c clusters gives rive to deep ( localized ) levels within the nw bandgap .
this work received financial support from the brazilian agencies cnpq and fapemig .
jmm wishes to thank i. s. santos de oliveira for fruitful technical discussions .
all calculations were performed using the computational facilities of cenapad / sp .
|
using first - principles methods we performed a theoretical study of carbon clusters in silicon carbide ( sic ) nanowires .
we examined small clusters with carbon interstitials and antisites in hydrogen - passivated sic nanowires growth along the [ 100 ] and [ 111 ] directions .
the formation energies of these clusters were calculated as a function of the carbon concentration .
we verified that the energetic stability of the carbon defects in sic nanowires depends strongly on the composition of the nanowire surface : the energetically most favorable configuration in carbon - coated [ 100 ] sic nanowire is not expected to occur in silicon - coated [ 100 ] sic nanowire .
the binding energies of some aggregates were also obtained , and they indicate that the formation of carbon clusters in sic nanowires is energetically favored .
|
photospheric convection and the coronal magnetic field play a key role in heating the solar corona . for the magnetically confined ( closed ) regions of the corona
it has long been suggested that small heating events , dubbed _ nanoflares _ , continuously deposit energy at the small scales and can contribute a substantial fraction of the total heating @xcite . the slow photospheric motions ( with a timescale @xmath0 8 minutes , magnitude @xmath1 , and correlation scale @xmath2 ) transfer energy from the photosphere into the corona shuffling the footpoints of the magnetic field lines . the work done by the denser photospheric plasma on the magnetic field lines footpoints injects energy into the corona , mostly as magnetic energy .
the perturbations generated at the photospheric level propagate along the loop at the alfvn speed . in a coronal loop
the alfvn velocity associated to the strong axial magnetic field @xmath3 is @xmath4 ( @xmath5 is the average mass density ) , and considering a typical loop length @xmath6 we obtain for the alfvn crossing time @xmath7 . the crossing time is therefore significantly smaller than the photospheric timescale : @xmath8 .
furthermore as magnetic pressure largely exceeds plasma pressure the plasma @xmath9 is small ( @xmath10 ) .
because of the _ fast alfvn timescale _ and _ low _ @xmath11 the dynamics of the magnetically confined solar corona are generally approximated as a _ quasi - static evolution through a sequence of equilibria _ , with instabilities leading the system from an equilibrium to the next one through relaxation @xcite . within this framework
many works have studied the relaxation dynamics in detail ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
on the other hand it has been shown that in some instances the validity of that picture is not valid , e.g. , in reduced magnetohydrodynamic ( mhd ) simulations of the parker model for coronal heating @xcite .
in fact that approximation is attained neglecting the velocity and plasma pressure in the mhd equations whose solution is then bound to be a static force - free equilibrium .
but the self - consistent evolution of the plasma pressure and velocity , although _ small _ compared with the dominant axial magnetic field @xmath3 ( and therefore very close in first approximation to the equilibrium solution of a uniform axial magnetic field ) , does not bind the system to force - free equilibria and allows the development of turbulent nonlinear dynamics with formation of field - aligned current sheets where a significant heating occurs .
most numerical simulations of a simple model coronal loop in cartesian geometry threaded by a strong magnetic field shuffled at the top and bottom plates by photospheric motions have used as boundary velocity an incompressible field with all wavenumbers of order @xmath12 excited @xcite : in real space this corresponds to distorted vortices with length - scale @xmath13 one next to the other filling the whole photospheric plane ( * ? ? ?
* figures 1 and 2 )
. this configuration does not give rise to instabilities .
the system transitions smoothly from the linear to the nonlinear stage where integrated physical quantities like energies and dissipation fluctuate around a mean value in what is best described as a statistically steady state and the energy deposited at the small scales is approximately in the _ nanoflare _ range for the numerous heating events .
this disordered vortical forcing mimics a uniform and homogeneous convection and the resulting coronal dynamics give rise to a basal background coronal heating within the lower limit of the observational constraint . with this kind of forcing the system
is not able to accumulate a significant amount of magnetic energy to be subsequently released in more substantial heating events like _ microflares _ and _ flares_. it is therefore pivotal to implement different kinds of photospheric forcings to understand the role of convective motions in coronal _ heating _ and the physical mechanisms and conditions for a significant _ storage _ of magnetic energy and its _
shearing or twisting the field lines might appear as the most straightforward way to make the system accumulate energy . to this end , in recent work @xcite we have implemented a 1d velocity forcing with a sinusoidal _ shear flow _ at the boundary , with @xmath14 , wavenumber @xmath15 , spanning the whole photospheric plane ( @xmath16 is the cross - section length ) . initially
the magnetic field that develops in the coronal loop is a simple map of the photospheric velocity field , i.e. , @xmath17 , with its intensity growing linearly in time .
a sheared magnetic field is known to be subject to tearing instability @xcite , in fact magnetic energy accumulates until a tearing instability sets in , magnetic energy is released and the system transitions to the nonlinear stage . on the other hand
continuing the simulation we found that , once the system has become fully nonlinear the dynamics are fundamentally different : _ the magnetic shear is not recreated_. once fluctuations are present , the orthogonal magnetic field ( @xmath18 and @xmath19 ) is organized in magnetic islands with @xmath20 and @xmath21 points , the nonlinear terms do not vanish any more and energy can be transported efficiently from large to small scales where it is dissipated . in the nonlinear stage
the dynamics are very similar whether the forcing velocity is a shear flow or made of disordered vortices , and magnetic energy is not stored efficiently , therefore larger releases of energy are not possible .
a shear photospheric flow can give rise to a sheared coronal magnetic field only in the unlikely condition that no relevant perturbations are present in the corona , or for very strong shear flows with velocity higher than the typical photospheric velocity of @xmath22 , when the perturbations naturally present in the corona can be neglected .
all photospheric forcings described so far fill the entire photospheric plane .
spatially localized velocity fields , like a single vortex that does not fill the entire plane , might be able to induce a higher storage of magnetic energy in the corona .
a vortex twists the magnetic field lines , and the resulting helical magnetic structure is kink unstable and widely used to model coronal loops @xcite . in these studies a twisted magnetic field lines structure
is used as initial condition , but this is assumed to have been induced by photospheric motions shuffling their footpoints , while the field lines are actually line - tied to a _
motionless photosphere_. to understand the dynamics of this photospherically driven system we have performed numerical simulations applying localized vortices at the photospheric boundary .
@xcite performed boundary forced simulations , but they stop just after kink instability sets in . in this paper
we continue the simulations for longer times .
this allows us to understand the dynamics of the system both when initially only an axial uniform magnetic field is present and a smooth ordered flux - tube with twisted field lines gets formed and kink instability sets in , and the later dynamics when the localized boundary vortex twists the footpoints of a disordered magnetic field where magnetic fluctuations and small scales are already present and field lines are no longer smooth .
furthermore , highly twisted magnetic structures , such as _ flux ropes _ , are broadly used to model the initiation of coronal mass ejections ( cmes , e.g. , see @xcite , and references therein ) . to advance our understanding of eruption initiation
it is therefore important to understand if and under which conditions photospheric motions can self - consistently generate flux ropes @xcite , or if these structures can only be advected into the corona from sub - photospheric regions via emerging flux .
the paper is organized as follows . in [ sec : gebc ] we describe the basic governing equations and boundary conditions , as well as the numerical code used to integrate them . in [ par3 ]
we discuss the initial conditions for our simulations and briefly summarize the linear stage dynamics more extensively detailed in @xcite .
the results of our numerical simulations are presented in [ sec : ns ] , while the final section is devoted to our conclusions and discussion of the impact of this work on coronal physics .
we model a coronal loop as an axially elongated cartesian box with an orthogonal cross section of size @xmath16 and an axial length @xmath23 embedded in an homogeneous and uniform axial magnetic field @xmath24 aligned along the @xmath25-direction .
any curvature effect is neglected .
the dynamics are integrated with the equations of rmhd @xcite , well suited for a plasma embedded in a strong axial magnetic field . in dimensionless form
they are given by : @xmath26 where @xmath27 and @xmath28 are the velocity and magnetic fields components orthogonal to the axial field and @xmath29 is the plasma pressure .
the gradient operator has components only in the perpendicular @xmath30-@xmath31 planes @xmath32 while the linear term @xmath33 couples the planes along the axial direction through a wave - like propagation at the alfvn speed @xmath34 .
incompressibility in rmhd equations follows from the large value of the axial magnetic fields @xcite and they remain valid also for low @xmath9 systems @xcite such as the corona . furthermore @xcite have recently performed fully compressible simulations of a similar cartesian coronal loop model , showing that the inclusion of thermal conductivity and radiative losses , that transport away the heat produced by the small scale energy dissipation , keep the dynamics in the rmhd regime . to render the equations nondimensional ,
we have first expressed the magnetic field as an alfvn velocity [ @xmath35 , where @xmath5 is the density supposed homogeneous and constant , and then all velocities have been normalized to the velocity @xmath36 , the order of magnitude of photospheric convective motions .
lengths and times are expressed in units of the perpendicular length of the computational box @xmath37 and its related crossing time @xmath38 . as a result ,
the linear terms @xmath33 are multiplied by the dimensionless alfvn velocity @xmath39 , where @xmath40 is the alfvn velocity associated with the axial magnetic field .
we use a computational box with an aspect ratio of @xmath41 , which spans @xmath42 our forcing velocities have a linear scale of @xmath43 that corresponds to the convective scale of @xmath44 in conventional units , thus the box extends @xmath45 .
lccccr + run & @xmath46 & @xmath47 & @xmath48 + + a & 512 x 512 x 208 & @xmath49 & @xmath50 + b & 256 x 256 x 192 & @xmath51 & @xmath52 + the index @xmath53 in the diffusive terms ( [ eq : eq1])-([eq : eq2 ] ) is called _ dissipativity _ and for @xmath54 these correspond to so - called hyperdiffusion ( e.g. , * ? ? ?
* ) . for @xmath55 standard diffusion ( @xmath56 )
is recovered and in this case the kinetic and magnetic reynolds numbers are given by : @xmath57 where @xmath58 is the speed of light , and numerically they are given the same value @xmath59 . in the simulations presented in this paper we use hyperdiffusion with @xmath60 .
hyperdiffusion is used because the implemented boundary velocity forcings and the magnetic flux tubes induced initially are localized to a small area of the computational box , and the dynamics would be dramatically diffusive with standard diffusion at a reasonable resolution ( see next section [ par3 ] for a more detailed discussion ) .
our parallel code rmh3d solves numerically equations ( [ eq : eq1])-([eq : eq3 ] ) written in terms of the potentials of the orthogonal velocity and magnetic fields in fourier space , i.e. , we advance the fourier components in the @xmath30- and @xmath31-directions of the scalar potentials . along the @xmath25-direction ,
no fourier transform is performed so that we can impose non - periodic boundary conditions ( [ par3 ] ) , and a central second - order finite - difference scheme is used . in the @xmath30-@xmath31 plane , a fourier pseudospectral method is implemented .
time is discretized with a third - order runge - kutta method .
for a more detailed description of the numerical code see @xcite .
magnetic field lines are line - tied to the top and bottom plates ( @xmath61 and @xmath23 ) that represent the photospheric surfaces . here
we impose , as boundary condition , a velocity field that convects the footpoints of the magnetic field lines . along the @xmath30 and @xmath31 directions
periodic boundary conditions are implemented . all simulations ( see table [ tbl ] )
employ a circular vortex applied at the top plate @xmath62 .
the velocity potential for this vortex is centered in the interval @xmath63 $ ] of linear extent @xmath64 and vanishes outside : @xmath65 \sin^2\ ! \left
[ 4\pi \left ( y - \frac{1}{8 } \right ) \right ] \nonumber \\ & & \mathrm{for } \quad x , y \in \mathit{i } , \quad \mathrm{and } \quad \varphi = 0 \quad \mathrm{for } \quad x , y \notin \mathit{i}. \label{eq : f0}\end{aligned}\ ] ] the velocity is linked to the potential by @xmath66 and its components are:@xmath67 \sin\ ! \left [ 8\pi \left ( y - \frac{1}{8 } \right ) \right ] \label{eq : f01}\\ & & u_y^l\ ! \left ( x , y \right ) = -\frac{2}{\sqrt{3}}\ , \sin\ ! \left [ 8\pi \left ( x - \frac{1}{8 } \right ) \right ] \sin^2\ ! \left [ 4\pi \left ( y - \frac{1}{8 } \right ) \right ] \label{eq : f00}\end{aligned}\ ] ] in the interval _
i _ and vanish outside . .
_ below : _ plot of the velocity @xmath31-component as a function of @xmath30 at @xmath68 .
[ fig : fig1 ] ] as shown in figure [ fig : fig1 ] equations ( [ eq : f0])-([eq : f00 ] ) describe a counter - clockwise vortex centered in the middle of the plane @xmath62 and has circular streamlines , with a slight departure from a perfectly circular shape toward the edge of the interval _
i_. averaging over the surface _
i _ the velocity rms is @xmath69 , the same value of the boundary velocity fields used in our previous works @xcite . in all simulations
a vanishing velocity is imposed at the bottom plate @xmath61 : @xmath70 at time @xmath71 we start our simulations with a uniform and homogeneous magnetic field along the axial direction @xmath72 .
the orthogonal component of the velocity and magnetic fields are zero inside our computational box @xmath73 , while at the top and bottom planes the vortical velocity forcing is applied and kept constant in time .
we briefly summarize and specialize to the case considered in this paper the linear stage analysis described in more detail in @xcite .
in general for an initial interval of time smaller than the nonlinear timescale @xmath74 , nonlinear terms in equations ( [ eq : eq1])-([eq : eq3 ] ) can be neglected and the equations linearized . for simplicity we will at first neglect also the diffusive terms and consider their effect later in this section .
the solution during the linear stage with a generic boundary velocity forcing @xmath75 , and @xmath76 , ( respectively at the top and bottom planes @xmath62 and @xmath77 ) is given by : @xmath78 where @xmath79 is the alfvn crossing time along the axial direction @xmath25 .
the magnetic field grows linearly in time , while the velocity field is stationary and the order of magnitude of its rms is determined by the boundary velocity profile .
both are a _ mapping _ of the boundary velocity field @xmath75 . for a generic forcing
the solution ( [ eq : lin1])-([eq : lin2 ] ) is valid only during the linear stage , while for @xmath80 when the fields are big enough the nonlinear terms can not be neglected .
nevertheless there is a singular subset of velocity forcing patterns for which the generated coronal fields ( [ eq : lin1])-([eq : lin2 ] ) have a vanishing lorentz force and the nonlinear terms vanish exactly .
this subset of patterns is characterize by having the _ vorticity constant along the streamlines _ @xcite . in this case
_ magnetic energy grows _ quadratically in time until some _ instability _ eventually sets in .
two kind of velocity patterns can be identified : a ) 1d patterns with their streamlines all parallel to each other , like a _ shear flow _ , or b ) a radial pattern with circular streamlines , like a _ circular vortex_. since in the linear stage the coronal fields are a mapping of the boundary velocity ( [ eq : lin1])-([eq : lin2 ] ) , a shear flow induces a sheared magnetic field subject to tearing instabilities @xcite , while the vortical flows considered in this paper twist the field lines into helices subject to kink instabilities .
the vortex ( [ eq : f01])-([eq : f00 ] ) is not perfectly circular as the streamlines depart from an exact round shape toward the edge ( figure [ fig : fig1 ] ) , but as we show in [ sec : runa ] field line tension adjusts the induced coronal orthogonal field lines in a round shape .
so far we have neglected the diffusive terms in the rmhd equations ( [ eq : eq1])-([eq : eq3 ] ) . in [ sec : ns ] we show that for this kind of problem the use of hyperdiffusion is crucial , otherwise the dynamics are dominated by diffusion .
overlooking this numerical fact can result in misleading conclusions @xcite , upon which we will comment in
[ sec : ns ] . here
we want to understand the diffusive effects on the linear dynamics , i.e. , when nonlinear terms are negligible or artificially suppressed by a low numerical resolution .
we now consider the effect of standard diffusion ( case @xmath55 in eqs .
( [ eq : eq1])-([eq : eq2 ] ) ) on the solutions ( [ eq : lin1])-([eq : lin2 ] ) : these are the solutions of the linearized equations obtained from ( [ eq : eq1])-([eq : eq2 ] ) retaining also the diffusive terms . in the _ linear regime _
, as the magnetic field grows in time ( [ eq : lin1 ] ) , the diffusive term ( @xmath81 ) becomes increasingly bigger until diffusion balances the magnetic field growth , and the system reaches a saturated equilibrium state .
including diffusion the magnetic field evolves as @xmath82 , \label{eq : diff1}\ ] ] i.e. , for times smaller than the diffusive timescale @xmath83 equation ( [ eq : lin1 ] ) is recovered with the field growing linearly in time , while for times bigger than @xmath83 the field asymptotes to its saturation value .
the diffusive timescale associated with the reynolds number @xmath84 is @xmath85 where @xmath86 is the length - scale of the forcing pattern , that for the pattern ( [ eq : f0])-([eq : f00 ] ) is given by @xmath87 where @xmath16 is the orthogonal computational box length . the total magnetic energy @xmath88 and ohmic dissipation rate @xmath89
will then be given by @xmath90 ^ 2 , \label{eq : diff2 } \\ & & j = \frac{1}{re}\ , \int_v\ !
\mathrm{d}^{^3}\hspace{-.4em } x \ \mathbf{j}^2 = j^{sat } \ , \left [ 1 - \exp \left ( - \frac{t}{\tau_r } \right ) \right]^2 .
\label{eq : diff3}\end{aligned}\ ] ] for times smaller than the diffusive timescale @xmath83 both quantities grow _ quadratically _ in time , while for @xmath91 they asymptote to their saturation value @xmath92 and @xmath93 : @xmath94 magnetic energy saturates to a value proportional to the square of both the reynolds number and the alfvn velocity , while the heating rate saturates to a value that is proportional to the reynolds number and the square of the axial alfvn velocity .
even though we use grids with @xmath95 points in the x - y plane , the timescales associated with ordinary diffusion are small enough to affect the large - scale dynamics , inhibiting the development of instabilities and nonlinearity .
the diffusive time @xmath96 at the scale @xmath97 associated with the dissipative terms used in equations ( [ eq : eq1])-([eq : eq2 ] ) is given by @xmath98 for @xmath55 the diffusive time decreases relatively slowly toward smaller scales , while for @xmath60 it decreases far more rapidly . as a result for @xmath60
we have longer diffusive timescales at large spatial scales and diffusive timescales similar to the case with @xmath55 at the resolution scale .
numerically we require the diffusion time at the resolution scale @xmath99 , where n is the number of grid points , to be of the same order of magnitude for both normal and hyper - diffusion , i.e. , @xmath100 then for a numerical grid with @xmath101 points that requires a reynolds number @xmath102 with ordinary diffusion we can implement @xmath103 ( table [ tbl ] ) , removing diffusive effects at the large scales and allowing , if present , the development of kink instabilities and nonlinear dynamics . ) and kinetic ( @xmath104 ) energies as a function of time ( @xmath79 is the axial alfvn crossing time ) .
the dashed curves show the time evolution of magnetic energy if the system were unperturbed [ eq . ( [ eq : diff2 ] ) ] , or nonlinearity suppressed , e.g. , by numerical diffusion ( see [ par3 ] ) .
the case @xmath105 corresponds to the linear case with no diffusion , attained ( at the large scales ) with the implementation of hyperdiffusion .
[ fig : fig2 ] ]
in this section we present the results of the numerical simulations summarized in table [ tbl ] .
simulations a and b have the same parameters , but simulation b employs a lower resolution to achieve a very long duration . in all simulations
the vortical velocity pattern ( [ eq : f0])-([eq : f00 ] ) is applied at the top plate @xmath62 , and a vanishing velocity at the bottom plate @xmath61 .
initially no perpendicular magnetic or velocity field is present inside the computational box @xmath106 , and the system is threaded only by the constant and uniform field @xmath72 .
the computational box has an aspect ratio of @xmath41 , with @xmath107 and @xmath108 .
we present here the results of run a , a simulation performed with a numerical grid of @xmath109 points , and hyperdiffusion coefficient @xmath110 with diffusivity @xmath60 .
the alfvn velocity is @xmath111 , corresponding to a nondimensional ratio @xmath112 .
the total duration is @xmath113 axial alfvn crossing times @xmath79 .
figures [ fig : fig2]-[fig : fig3 ] show the temporal evolution of the total magnetic and kinetic energies @xmath114 the total ohmic dissipation rate @xmath115 and @xmath116 , the power injected from the boundary by the work done by convective motions on the field lines footpoints ( see equation ( [ eq : pflux ] ) ) , along with some saturation curves for magnetic energy ( [ eq : diff2 ] ) .
additionally figure [ fig : fig4 ] shows snapshots of the magnetic field lines of the orthogonal component @xmath28 and electric current @xmath117 , the leading order component in rmhd ordering @xcite , at selected times in the mid - plane @xmath118 . )
dissipation rate and the integrated poynting flux @xmath116 ( the injected power ) versus time .
viscous dissipation is negligible respect to the ohmic contribution .
inset shows the ohmic dissipative peak corresponding to the development of kink instability .
[ fig : fig3 ] ] the circular vortical velocity field ( [ eq : f01])-([eq : f00 ] ) applied at the top boundary ( @xmath119 ) initially induces velocity and magnetic fields in the computational box that follow the linear behavior given by equations ( [ eq : lin1])-([eq : lin2 ] ) , i.e. , they are a mapping of the velocity at the boundary with the magnetic field increasing linearly in time ( figure [ fig : fig4 ] , times @xmath120 and @xmath121 ) . in the linear stage ( @xmath122 )
magnetic energy is well - fitted ( figure [ fig : fig2 ] ) by the linear curve ( [ eq : diff2 ] ) in the limit @xmath123 , i.e. , in the absence of diffusion ( indeed in this limit the curve can be obtained directly from the linear equation ( [ eq : lin1 ] ) ) .
this is because we are using hyperdiffusion that effectively gets rid of diffusion at the large scales .
two other magnetic energy linear diffusive saturation curves are drawn for @xmath124 and @xmath125 , typical reynolds numbers used in our previous simulations with standard diffusion @xmath55 and orthogonal grids with respectively @xmath126 and @xmath127 grid points ( see , e.g. , * ? ? ?
their saturation level is very low compared to the magnetic energy values when kink instability develops ( @xmath128 ) and in the following nonlinear stage .
this is because the vortex and induced magnetic field occupy only a limited volume elongated along z at the center of the x - y plane : at these scales diffusion dominates with these resolutions using standard diffusion .
+ + + + + + + + for this reason the use of hyperdiffusion is crucial to study this problem , otherwise diffusion dominates and a balance between the injection of energy from the boundary and its numerical removal by diffusion is reached very soon , inhibiting the development of kink instability and nonlinear dynamics .
this diffusive _ linear _ regime was reached in previous simulations by @xcite , where four similar vortices were applied at the boundary .
therefore their conclusion that nonlinear dynamics or instabilities ( not to mention turbulence ) can not develop in such physical systems is simply a numerical issue : this can be overcome adopting hyperdiffusion as we have done here or , alternatively , implementing grids with much higher resolutions that require impractically large numerical resources .
the localized boundary vortex ( shown with a colored contour in figure [ fig : fig5 ] ) generates a mostly poloidal magnetic field confined to the axial volume in correspondence of the vortex , resulting in helical field lines for the total magnetic field ( figure [ fig : fig5 ] , time @xmath129 ) . outside this volume
the poloidal field vanishes and only the axial field @xmath3 is present .
ampre s law then guarantees that the total net current is zero .
as shown in figure [ fig : fig4 ] in the linear stage ( @xmath120 , @xmath121 ) there is a stronger up - flowing current concentrated in the middle , and a weaker ring - shaped down - flowing current distributed at the edge of the flux - tube .
this magnetic configuration is well known to be kink unstable , and is similar to the nc ( null current ) force - free model studied by @xcite .
the main differences are that their axial field @xmath3 is not uniform , dropping by @xmath130 outside the flux tube , and that the field lines are line - tied to a motionless photosphere .
they performed a linear stability analysis of this configuration finding that there is a critical axial loop length @xmath131 beyond which the system is unstable and has a constant growth rate @xmath132 .
they also examined other equilibria with net current finding a similar qualitative behavior , with variations for the critical length and growth rates .
@xcite found that for the nc case the ratio of the axial critical length over the cross - length of the flux - tube is @xmath133 . in the case
considered here the ratio of the axial length ( @xmath108 ) over the cross - length of the flux tube ( the extent of the boundary vortex @xmath134 , equation ( [ eq : f0 ] ) ) is @xmath135 , therefore it is fully in the unstable region .
of course at a given length ( beyond the critical length ) there is also a critical twist beyond which the configuration is unstable . in our simulations
the system is continuously forced at the boundary , and in the linear stage the twist grows linearly in time ( from equation ( [ eq : lin1 ] ) , as the twist is proportional to @xmath136 ) , thus such a critical twist is certainly attained . in our case
the `` equilibrium '' solution is not static but is given by the linear solution ( [ eq : lin1 ] ) , indicated here with @xmath137 , with the magnetic field growing linearly in time while mapping the boundary vortex .
thus we compute the perturbed magnetic energy as @xmath138 we find that in the linear stage this quantity grows exponentially in time , obtaining for the perturbed magnetic field a growth rate @xmath132 , as @xcite for their nc equilibrium model .
this growth rate is also confirmed by the fact that kink instability sets in at @xmath128 ( figures [ fig : fig2 ] and [ fig : fig3 ] ) and @xmath139 .
as mentioned in [ par3 ] the forcing boundary vortex departs from an exact circular shape at its edges where its vorticity is not exactly constant along the streamlines , thus there is a small lorenz force for the resulting magnetic field ( [ eq : lin1 ] ) .
this small difference in the linear field acts as a perturbation .
additionally @xcite found out that configurations with zero net current are unstable to the _ internal _ kink mode ( opposed to the global kink mode for configurations with a net current ) , for which magnetic perturbations and the radial displacement of the plasma column are confined within the original flux tube .
this is found also in our simulation as shown in figure [ fig : fig4 ] at the onset of the nonlinear stage at @xmath140 , when the plasma displaces inside the flux tube toward its edge where a strong current sheet forms .
the internal kink mode releases almost 90% of the accumulated energy around time @xmath141 ( figure [ fig : fig2 ] ) in correspondence of the big ohmic dissipative peak shown in figure [ fig : fig3 ] .
the released energy is @xmath142 , in the micro - flare range ( the factor to convert energy into dimensional units , given our normalization choice discussed in [ par3 ] , is @xmath143 , i.e. , @xmath144 ) . as a result of the kink instability magnetic
reconnection occurs ( figure [ fig : fig4 ] , @xmath145 ) and the magnetic field lines get substantially unwind as shown in figure [ fig : fig5 ] ( times @xmath129 and @xmath146 ) with field lines twisting only @xmath147 after the instability . in summary , during the linear stage
, the transition to and the first phase of the nonlinear regime , the analysis of @xcite is fully confirmed also for the photospherically driven case considered here : the system forced by a circular vortex is unstable to an _ internal kink mode _ , releases most of the stored magnetic energy and magnetic reconnection untwists the field lines .
linear calculations @xcite show that similar dynamics are expected also for different configurations with different aspect ratios and magnetic guide field values , except for those that fall below the instability threshold .
the phenomenology described so far is also in agreement with that of three - dimensional simulations with a realistic geometry @xcite .
in particular strong nonlinearities persist right after the instability occurs ( @xmath148 and @xmath146 ) , when the system can not be described as a constant-@xmath149 force - free state .
an inverse cascade of magnetic energy is observed , as the orthogonal magnetic field acquires longer scales and the overall volume occupied by twisted field lines increases , as shown in figure [ fig : fig4 ] just before ( @xmath148 ) and after ( @xmath150 ) the instability . in @xcite
this corresponds also to an inverse cascade of magnetic helicity , corresponding in the rmhd case to an inverse cascade of the square potential @xmath151 ( see the end of this section and our discussion in 5 for more about this quasi - invariant analogous to helicity in rmhd ) . ) and kinetic ( @xmath104 ) time - averaged energy spectra as a function of the orthogonal wavenumber @xmath152 the inset shows the kinetic energy spectrum of the boundary velocity vortex [ equations ( [ eq : f0])-([eq : f00 ] ) ] applied at the top plate @xmath119 . [
fig : fig7 ] ] on the other hand , at later times the dynamics are certainly surprising when , in the fully nonlinear stage , fluctuations created by the kink instability are present in the corona .
for @xmath153 magnetic energy increases steadily , while kinetic energy remains small ( figure [ fig : fig2 ] ) .
this is in contrast to all our previous simulations with space - filling boundary motions , either distorted vortices @xcite or shear flows @xcite , when in the nonlinear regime a _ magnetically dominated statistically steady state _ was reached where integrated quantities would fluctuate around an average value ( with velocity fluctuations smaller than magnetic fluctuations ) . in our case
ohmic dissipation @xmath89 and the integrated poynting flux @xmath116 do reach a statistically steady state ( figure [ fig : fig3 ] ) .
the integrated pointing flux @xmath154 is the power entering the system at the boundaries as a result of the work done by photospheric motions on the footpoints of magnetic field lines ( @xmath75 is the photospheric forcing velocity ) .
but in contrast to our previous results , here the power does not balance on the average the dissipation rate , its average is slightly higher resulting in the magnetic energy growth shown in figure [ fig : fig2 ] . fluctuate around a mean value , the first two modes increase steadily showing that an inverse cascade occurs . [ fig : fig8 ] ] in physical space the dynamics are surprising in two ways . _ first _ , after the kink instability , even though we continue to stir the field lines footpoints with the same vortex , no further kink instability develops .
analogously to the shear flow case @xcite , once the system transitions to the nonlinear stage the magnetic fluctuations generated during the instability do not have a vanishing lorentz force .
in fact around @xmath155 , at the end of the big dissipative event , the topology of the orthogonal component of the magnetic field is characterized by circular , but _ distorted _ , field lines ( figure [ fig : fig4 ] ) .
naturally the lorentz force does not vanish now and the vorticity is not constant along the streamlines . _ nonlinear terms do not vanish _ as they do during the linear stage for @xmath156 . when they vanish magnetic energy can be stored , without getting dissipated , into an ordered flux - tube with helical field lines ( figure [ fig : fig5 ] , @xmath157 ) , and matching perfectly round orthogonal magnetic field lines ( figure [ fig : fig4 ] , @xmath158 ) .
but now nonlinearity continuously _ transfers _ energy from large to small scales where it is dissipated . in physical space
small scales are not uniformly distributed , but they are organized in field - aligned current sheets . these , once formed during the onset of the nonlinear stage , persist throughout the subsequent dynamics ( as shown in figures [ fig : fig4 ] and [ fig : fig6 ] ) , with the energy cascade continuously feeding them .
_ second _ , the photospherical vortical motions do not give rise to an orderly helical flux - tube as in the linear stage ( figure [ fig : fig5 ] , @xmath129 ) .
however , magnetic field lines get twisted , but in a disordered way ( figure [ fig : fig5 ] and [ fig : fig4 ] , @xmath159 ) .
a new phenomenon occurs : on longer timescales the magnetic field acquires longer spatial scales ( figure [ fig : fig4 ] ) , the volume where field lines are twisted increases ( figure [ fig : fig5 ] ) , while the current exhibits always a small filling factor occupying a small fraction of the volume ( figure [ fig : fig6 ] ) .
to better understand these phenomena we need to investigate the energy dynamics in fourier space .
we consider the spectra in the orthogonal @xmath30-@xmath31 plane integrated along the @xmath25 direction .
as they are isotropic in the fourier @xmath160-@xmath161 plane we compute the integrated 1d spectra , so that for the total magnetic energy @xmath88 we obtain : latexmath:[\[\begin{aligned } e_m & = & \frac{1}{2 } \int\limits_0^l\ !
\mathrm{d}z\ , \iint\limits_0^{\quad \
! \mathrm{d}x\,\mathrm{d}y\ \mathbf{b}_{_{\!\perp}}^2 \nonumber \\ & = & \frac{1}{2 } \int\limits_0^l\ ! \mathrm{d}z\ \ell^2 \sum_{\mathbf{k } } where @xmath53 indicates the _ shell _ in @xmath163-space with wavenumber @xmath164 included in the range @xmath165 , and @xmath166 is the maximum wavenumber admitted by the numerical grid ( corresponding to the smallest resolved orthogonal scale ) .
erg , in the _ micro - flare _ range .
[ fig : fig9 ] ] the _ time averaged _ magnetic and kinetic energy spectra as a function of wavenumber are shown in figure [ fig : fig7 ] , the inset shows the spectrum of the boundary vortex kinetic energy ( see equation ( [ eq : f0 ] ) ) .
photospheric motions therefore inject energy at wavenumbers between 2 and 7 ( see equation ( [ eq : pflux ] ) ) , the system is magnetically dominated and the power - laws exhibited at higher wavenumbers , in the inertial range , are similar to those obtained with previous space - filling boundary forcings @xcite , with the spectrum of magnetic energy much steeper than that of kinetic energy .
however the time - average of the low - wavenumber modes hides an interesting dynamics .
figure [ fig : fig8 ] shows the first five magnetic energy modes as a function of time . while modes with wavenumbers @xmath167 after the kink instability fluctuate around a mean value , the first two modes @xmath168 grow steadily with mode @xmath55 becoming prevalent .
this shows that an _ inverse cascade _ takes place . while the _ direct cascade _ transfers energy from the injection scale toward small scales ( current sheets ) where energy is dissipated , analogously the inverse cascade transfers energy toward the large scales ( modes 1 and 2 ) where _ no dissipative process
_ is at work and consequently _ energy accumulates_. in physical space this process gives rise to the large scales that the magnetic field acquires in the orthogonal direction , shown in figures [ fig : fig4 ] and [ fig : fig5 ] , discussed previously . in the rmhd system with boundary conditions as we apply here there is no strict invariant known to follow an inverse cascade , such as magnetic helicity in 3d mhd or the square of the vector potential in 2d mhd @xcite .
rmhd resembles the 2d mhd case in the sense that though the square of the vector potential is not conserved , the terms violating conservation arise only from the boundaries in the axial direction .
a dynamical magnetic inverse cascade mechanism is therefore still active , impeded only by the inputs coming from photospheric motions at the boundary , and this explains the accumulation of magnetic energy at the largest transverse scales .
the simulation described in the previous section ( run a ) has a duration of @xmath169 , but this time span leaves undetermined the behavior of the low wavenumber modes over longer time scales .
indeed these modes keep growing , as shown in figure [ fig : fig8 ] , resulting in a _ steady growth _ of total magnetic energy , shown in figure [ fig : fig2 ] . to understand the long - time dynamics of the system , we have performed another simulation , run b , with the same physical parameters of run a , but half the orthogonal resolution ( table [ tbl ] ) , extending the duration up to @xmath170 .
] figure [ fig : fig9 ] shows that over longer times the energy of the system is prevalently in mode 1 , i.e. , the largest possible scale .
but this mode does not grow indefinitely and over these much longer time - scales it reaches a statistically steady state , fluctuating around its mean value . the largest energy fluctuations shown in figure [ fig : fig9 ] result in energy drops of @xmath171 , that in dimensional units correspond to a micro - flare with @xmath172 , releasing about twice the amount of energy released by the kink instability around @xmath173 in run a ( compare with figure [ fig : fig2 ] ) .
notice that the kink instability does not appear in figure [ fig : fig9 ] because the sampling time interval for the modes is too long in run b , but it is clearly shown in the r.m.s . of the energies ( not shown ) and in the dissipation rate ( see inset in figure [ fig : fig10 ] )
. comparing figure [ fig : fig9 ] with figure [ fig : fig10 ] , where the ohmic dissipation rate @xmath89 is shown as a function of time , displays another interesting result .
the large and sharp energy drops shown in figure [ fig : fig9 ] correspond to large dissipative peaks in figure [ fig : fig10 ] , e.g. , at times @xmath174 and @xmath175 , but it is also possible to have equally large but more gradual energy drops , e.g. , between times @xmath176 and @xmath177 , without a corresponding single large dissipative peak but rather a cluster of smaller peaks .
in physical space we have already seen in run a that initially the inverse cascade corresponds to a perturbed magnetic field that occupies an increasingly larger volume ( figure [ fig : fig4 ] ) until all the field lines in the box get twisted ( figure [ fig : fig5 ] ) . in run
b we observe that successively , once the computational box has been filled with perpendicular magnetic field , the rising amplitude of modes 1 corresponds to an increase of the magnetic field intensity , while the fluctuations in the energy mode are due to magnetic reconnection events .
in fact due to the periodic boundary conditions in @xmath30 and @xmath31 the same system repeats indefinitely along these directions . when the orthogonal magnetic field reaches the boundary it starts to interact with the neighboring structures ( i.e. with itself coming from the other side ) .
the magnetic energy drops in mode 1 correspond to magnetic reconnection events that make the system oscillate between the different possible configurations with energy contained at the ( large ) scales of mode 1 shown in figure [ fig : fig11 ] ( there is no preferred orthogonal direction for the system at this scale ) .
while the periodic boundary conditions limit the interactions of large - scale twisted magnetic structures it is clearly shown that interaction with such other magnetic copies of itself is one of the ways in which the accumulated energy can be released .
further possibilities and the dynamics of these interactions will be the subject of future works .
in this paper we have investigated the dynamics of a closed coronal region driven at its boundary by a localized photospheric vortex .
such small vortical motions with scales typical of photospheric convection ( @xmath178 km ) have been recently observed in the photosphere @xcite , and can induce relevant dynamics in the solar corona @xcite .
a `` straightened out '' closed region of the solar corona is modeled as an elongated cartesian box where the top and bottom plates mimic the photosphere , and the dynamics are integrated with the reduced mhd equations @xcite , well suited for a plasma threaded by a strong axial magnetic field .
the initial condition consists simply of a uniform axial magnetic field .
its field lines are originally straight and its footpoints are line - tied at both ends in the top and bottom photospheric plates .
the photospheric vortex drags the field lines footpoints twisting the magnetic field lines ( figure [ fig : fig5 ] , @xmath157 ) . even though the vortex that we employ is not perfectly circular ( figure [ fig : fig1 ] , equations ( [ eq : f0])-([eq : f00 ] ) ) in the linear stage the field lines tension straightens out in a round shape the orthogonal magnetic field lines ( figure [ fig : fig4 ] , @xmath179 and @xmath158 ) , in this way the lorentz force vanishes in the planes and the system is able to accumulate energy .
the small departure from a round shape ( at its edge ) of the boundary vortex introduces a small perturbation in the coronal field .
the system is then unstable to the internal kink mode ( figure [ fig : fig4 ] , @xmath158 ) , and releases about @xmath180 of the accumulated energy in a dissipative event ( figures [ fig : fig2 ] and [ fig : fig3 ] ) .
the energy released in this event is of the order of a _ micro - flare _ with @xmath181 erg .
these results are in agreement with those of @xcite , that consider similar initial conditions , performs a refined linear analysis , but does not employ a boundary forcing , i.e. , the field lines are line - tied to a motionless photosphere . therefore the initial _ linear stage _ and the development of the kink instability are in agreement with previous works that have always employed large - scale smooth fields with no broad - band fluctuations as initial conditions , both in the case of field lines line - tied to a motionless photosphere @xcite and with a boundary driver @xcite .
on the other hand in the solar corona perturbations are continuously injected from the lower atmospheric layers .
numerical simulations ( e.g. , * ? ? ?
* ) confirm that especially in closed regions , where waves can not escape toward the interplanetary medium , broad - band magnetic fluctuations of the order of a few percent of the strong axial magnetic field ( not infinitesimal perturbations as classically used in instabilities studies ) are naturally present . therefore in order to gain a first insight of the coronal dynamics when the magnetic field is already structured , i.e. , there are finite magnetic fluctuations ( small but not infinitesimal ) with small scales and current sheets , we continue the simulation after kink instability develops .
in fact right after kink instability the magnetic energy is small ( figure [ fig : fig2 ] ) , with @xmath182 , but it is already structured with current sheets ( figure [ fig : fig4 ] , @xmath183 ) and a broad band spectrum ( figure [ fig : fig7 ] ) .
the boundary vortex continues to twist the magnetic field lines , but in a disordered way ( figure [ fig : fig5 ] , @xmath184
@xmath185 ) .
the presence of an already structured magnetic field allows nonlinear dynamics to develop : once current sheets and small scales are present , an energy cascade continues to feed them , as shown by the energy spectra in figure [ fig : fig7 ] .
therefore current sheets do not disappear , and the continuous transfer of energy from the large to the small scales prevents the field lines to increase their twist beyond @xmath147 .
the twist remains approximately constant in the nonlinear stage as shown in figure [ fig : fig5 ] ( @xmath184 @xmath185 ) .
furthermore because the current is now concentrated in thin current sheets ( figures [ fig : fig4 ] and [ fig : fig6 ] ) kink instabilities do not develop .
we had already observed a similar behavior in our previous simulations that employed space - filling boundary drivers . in particular when the field lines were _ sheared _ by a 1d boundary forcing @xcite the coronal field was sheared only in the linear stage , but after that a multiple tearing instability developed and in the coronal field magnetic fluctuations and current sheets were formed , the continuous shearing motions at the boundary were not able to recreate a sheared coronal field and further instabilities were not observed .
but in the simulations presented in this paper a new phenomenon occurs .
although the field lines twist is approximately constant in the nonlinear stage , the volume where field lines are twisted increases , and the magnetic field acquires larger scales ( figures [ fig : fig4 ] and [ fig : fig5 ] ) . besides a _
direct _ cascade that transfers energy from the large to the small scales where it is dissipated in current sheets , an _ inverse _ cascade takes place , transferring energy from the injection scale toward larger scales , where no dissipation takes place and energy can accumulate .
the analysis of the magnetic energy modes ( figures [ fig : fig8 ] and [ fig : fig9 ] ) shows indeed that on long time - scales most of the energy is stored at the largest possible scale ( mode 1 ) .
the inverse cascade is able to store a significant amount of magnetic energy .
although magnetic helicity is not defined in rmhd , the integral of the magnetic square potential @xmath151 is approximately conserved ( see discussion in last paragraph of [ sec : runa ] ) .
the inverse cascade of magnetic energy corresponds also to an inverse cascade of the square potential , as clearly shown in figure [ fig : fig4 ] where the field lines are the contour of @xmath151 ( and @xmath186 ) . in future compressible simulations we expect to observe for magnetic helicity ( well defined in 3d mhd ) dynamics similar to those shown here for magnetic energy , i.e. , an increase of magnetic helicity ( injected from the boundary ) and its inverse cascade , in analogy to the the inverse helicity cascade observed by @xcite . because of the periodic boundary conditions along @xmath30 and @xmath31 the system is virtually repeated along these directions .
when the field lines get twisted in the entire computational box , this twisted structure interacts with these neighboring twisted structures .
this interaction is the only condition that limits the growth of magnetic energy , giving rise to impulsive magnetic reconnection events , that now is not inhibited by the circular topology of the orthogonal magnetic field lines of a single structure .
these events make the system oscillate between the many possible configurations with energy in mode 1 ( two of these are shown in figure [ fig : fig11 ] ) .
the associated energy drops shown in figure [ fig : fig9 ] are also in the _ micro - flare _ range with @xmath187 erg , twice the value of the energy released initially by the kink instability .
although in the presented simulations the generated magnetic structures interact only with similar structures repeated by the periodic boundary conditions along @xmath30 and @xmath31 , we can infer that the interaction of a single twisted magnetic structure with other magnetic structures can give rise to similar release of energy .
a more general investigations of the interaction between twisted magnetic structures is under way to understand under which conditions the interaction leads to energy storage and/or release , and to determine quantitatively these properties .
previous simulations that employed a space - filling photospheric forcing @xcite were not able to accumulate a significant amount of energy to be successively released in micro or larger flares .
those photospheric motions , that mimic a uniform and homogeneous convection , give instead rise to a basal background coronal heating rate in the lower range of the observational constraint ( @xmath188 erg@xmath189s@xmath190 ) and a million degree corona @xcite . in the case of a space - filling boundary driver
we had also observed that the inverse cascade is inhibited ( see * ? ? ?
* ) for typically strong dc magnetic fields .
an inverse cascade is possible only for weak guide fields ( see * ? ? ?
* ) , a condition applicable only to limited regions of the corona .
we conclude that in presence of line - tying and a strong guide field , inverse cascade can be a good mechanism to store energy , but only if the boundary motion is localized in space as the vortex used here , and not space - filling .
subsequently the interaction of this magnetic structures with others can release the accumulated energy . in general photospheric motions will be a superposition of approximately homogenous space - filling convective motions and localized vortical and also shearing motions ( e.g. , see * ? ? ?
* for a localized shear case ) . while the space - filling motions give rise to a basal background coronal heating ( e.g. , * ? ?
* ) , localized motions can give rise to higher impulsive releases of energy in the micro - flare range and above , contributing to coronal heating while increasing the temporal intermittency of the energy deposition and of its associated radiative emissions .
in future works we will consider cases with localized motions superimposed to a homogenous space - filling convection - mimicking velocity field to determine , among other things , how stronger the localized velocity has to be respect to the background motions in order to develop dynamics similar to those presented in this paper . as mentioned in the introduction , highly ( and orderly ) twisted magnetic structures , such as _ flux ropes _ , are used to initiate solar eruptions ( e.g. , see @xcite for a recent application , and the reviews by @xcite for further examples of this model ) .
kink - like instabilities developing in these flux - ropes give rise to an explosive dynamics leading to the formation of a cme .
we have shown that kink - unstable flux ropes are not formed in the corona by boundary vortical motions , unless a very strong vortex is applied and the coronal magnetic fluctuations can then be neglected . therefore , although flux ropes can be formed in the complex dynamics in and around a prominence region @xcite , given the ubiquitous presence of magnetic fluctuations in the solar corona , the development of kink - like instabilities may be strongly limited . while the dynamics of the induced cme can be a good approximation , we conclude that such models offer a poor model of the initiation process for which more realistic models are called for @xcite
. generally speaking , in a realistic 3d geometry one might expect that the growth of energy in the transverse field leads to an inflation and rise of a magnetic loop due to the curvature , which we have neglected here .
this effect was included by @xcite , who showed that twisting the footpoints of a curved flux rope leads to its gradual expansion and the system rises to larger solar radii . in our simulations
the twist does not increase ( the overall field lines twist is limited to @xmath191 ) , remaining roughly constant in the nonlinear stage .
it is left to future work to understand under which conditions such a system , including curvature , has dynamics similar to those of @xcite , or whether different dynamics are possible ( see also * ? ? ?
* ) , and how the dynamics develop in a 3d geometry considering small or large - scales photospheric vortices . this work was carried out in part at the jet propulsion laboratory under a contract with nasa .
this research supported in part by the nasa heliophysics theory program nnx11aj44 g , and by the nsf solar terrestrial and shine programs ( ags-1063439 & ags-1156094 ) , by the nasa mms and solar probe plus projects .
simulations have been performed through the nasa advanced supercomputing smd awards 11 - 2331 and 12 - 3188 .
amari , t. & aly , j. j. 2009 , in iau symp .
257 , universal heliophysical processes , ed .
n. gopalswamy & d. f. webb ( cambridge : cambridge univ . press ) , 211 amari , t. , aly , j. , luciani , j. , mikc , z. , linker , j. 2011 , , 742 , l27 amari , t. & luciani , j. f. 2000 , , 84 , 6 amari , t. , luciani , j. f. , aly , j. j. , & tagger , m. 1996 , , 466 , l39 amari , t. , luciani , j. f. , mikc , z. , & linker , j. 1999 , , 518 , l57 baty , h. 2001 , , 367 , 321 baty , h. , & heyvaerts , j. 1996 , , 308 , 935 berger , m. a. 1997 , , 102 , 2637 bhattacharjee , a. , ng , c. s. , & spangler , s. r. 1998 , , 494 , 409 biskamp , d. 2003 , magnetohydrodynamic turbulence ( cambridge : cambridge univ .
press ) bonet , j. , mrquez , i. , snchez almeida , j. , cabello , i. & domingo , v. 2008 , , 687 , l131 bonet , j. , et al .
2010 , , 723 , l139 brandenburg , a. , & matthaeus , w. h. 2004 , phys .
e , 69 , 56407 brandt , p. , scharmer , g. , ferguson , s. , shine , r. & tarbell , t. 1988 , nature , 335 , 238 browning , p. k. , gerrard , c. , hood , a. w. , kevis , r. , & van der linden , r. a. m. 2008 , , 485 , 837 chen , p. f. 2011 , living rev .
solar phys . ,
8 , 1 dahlburg , r. b. , einaudi , g. , rappazzo , a. f. , & velli , m. 2012 , , 544 , l20 dahlburg r. b. , liu j. , klimchuk j. a. , nigro g 2009 , , 704 , 1059 dmitruk , p. , & gmez , d. o. 1997 , , 484 , l83 dmitruk , p. , & gmez , d. o. 1999 , , 527 , l63 dmitruk , p. , gmez , d. o. , & matthaeus , w. h. 2003 , phys .
plasmas , 10 , 3584 einaudi , g. , & velli , m. 1999 , phys .
plasmas , 6 , 4146 einaudi , g. , velli , m. , politano , h. , & pouquet , a. 1996 , , 457 , l113 furth , h. p. , killeen , j. , & rosenbluth , m. n. , 1963 , phys .
fluids , 6 , 459 georgoulis , m. k. , velli , m. , & einaudi , g. 1998 , , 497 , 957 gerrard , c. l. , arber , t. d. , & hood , a. w. 2002 , , 387 , 687 gerrard , c. l. , hood , a. w. , & brown , d. s. 2004 , sol . phys . , 222 , 79 heyvaerts , j. , & priest , e. r. 1984 , , 137 , 63 hood a. w. , browning p. k. , van der linden , r. a. m. 2009 , , 506 , 913 kadomtsev , b. b. , & pogutse , o. p. 1974
, sov . phys .
jetp , 38 , 283 2009 , in american geophysical union , fall meeting 2009 , abstract # sm42b-03 2010 , in american astronomical society , aas meeting # 216 , abstract # 302.05 , bulletin of the american astronomical society , 41 , 847 lionello , r. , velli , m. , einaudi , g. , & mikc , z. 1998 , , 494 , 840 low , b. c. 2001 , , 106 , 25141 mikc , z. , schnack , d. d. , & van hoven , g. 1990 , , 361 , 690 montgomery , d. 1982 , phys .
t , 2 , 83 panasenco , o. , martin , s. f. , & velli , m. 2013 , , in press parker , e. n. 1972 , , 174 , 499 parker , e. n. 1988 , , 330 , 474 parker , e. n. 1994 , spontaneous current sheets in magnetic fields ( new york : oxford univ . press ) pontin , d. i. , wilmot - smith , a. l. , hornig , g. , galsgaard , k. 2011 , , 525 , a57 rappazzo , a. f. , & velli , m. 2011 , phys .
e , 83 , 065401(r ) rappazzo , a. f. , velli , m. , & einaudi , g. 2010 , , 722 , 65 rappazzo , a. f. , velli , m. , einaudi , g. , & dahlburg , r. b. 2007 , , 657 , l47 rappazzo , a. f. , velli , m. , einaudi , g. , & dahlburg , r. b. 2008 , , 677 , 1348 strauss , h. r. 1976 , phys .
fluids , 19 , 134 taylor , j. b. 1974 , , 33 , 1139 taylor , j. b. 1986 , rev . mod .
phys . , 58 , 741 trk , t. , panasenco , o. , titov , v. , miki , z. , reeves , k. k. , velli , m. , linker , j. , & de toma , g. 2011 , , 739 , l63 velli , m. , & liewer , p. 1999 , space sci .
, 87 , 339 velli , m. , lionello , r. , & einaudi , g. 1997 , sol .
phys . , 172 , 257 wedemeyer - bhm , s. , scullion , e. , steiner , o. , rouppe van der voort , l. , de la cruz rodriguez , j. , fedun , v. , erdlyi , r. 2012 , nature , 486 , 505 wilmot - smith , a. l. , pontin , d. i. , hornig , g. , yeates , a. r. 2011 , , 536 , a67 yeates , a. r. , hornig , g. , wilmot - smith , a. l. 2010 , , 105 , 85002 zank , g. p. , & matthaeus , w. h. 1992 , j. plasma phys . , 48 , 85
|
we present simulations modeling closed regions of the solar corona threaded by a strong magnetic field where localized photospheric vortical motions twist the coronal field lines .
the linear and nonlinear dynamics are investigated in the reduced magnetohydrodynamic regime in cartesian geometry . initially the magnetic field lines get twisted and the system becomes unstable to the internal kink mode , confirming and extending previous results . as typical in this kind of investigations , where initial conditions implement smooth fields and flux - tubes ,
we have neglected fluctuations and the fields are laminar until the instability sets in . but previous investigations indicate that fluctuations , excited by photospheric motions and coronal dynamics , are naturally present at all scales in the coronal fields .
thus , in order to understand the effect of a photospheric vortex on a more realistic corona , we continue the simulations after kink instability sets in , when turbulent fluctuations have already developed in the corona . in the nonlinear stage
the system never returns to the simple initial state with ordered twisted field lines , and kink instability does not occur again
. nevertheless field lines get twisted , but in a disordered way , and energy accumulates at large scales through an inverse cascade .
this energy can subsequently be released in micro - flares or larger flares , when interaction with neighboring structures occurs or via other mechanisms . the impact on coronal dynamics and cmes initiation
is discussed .
|
the undistorted field of view of a given vlbi data set is usually limited by two main effects : bandwidth smearing and time - average smearing @xcite
. the narrower the individual frequency channels and the smaller the integration time , the larger the unaberrated field of view .
data generated by vlbi correlators are comprised of a set of measurements of the complex visibility as a function of frequency ( or delay ) and time .
most continuum vlbi data sets are delivered to the astronomer with relatively narrow frequency channels ( @xmath1 mhz ) and short integration times ( @xmath2 secs ) .
for example , a typical @xmath3 cm evn data set , in its original form , boasts a field - of - view , @xmath4 , in excess of @xmath5 arcminute .
the same evn data set , having for example being averaged in frequency over @xmath6 mhz , has a field - of - view of @xmath7 milliarcseconds ( mas ) .
this reduction in the field - of - view by a factor in excess of two orders of magnitude is often considered to be unimportant .
the aim of this paper is to show , by illustration , that this presumption may no longer be valid . )
we also detect compact structure in the arcsec - scale jet ( previously labelled in vla maps of the source as features `` jet '' and `` c '' . ]
in the early 1980s , the off - line computer resources available to most astronomers were ill - equipped to deal with extremely large and cumbersome vlbi data sets .
not surprisingly , one of the major goals of vlbi _ data reduction _ was to severely average continuum data at the earliest possible stage in the analysis process ( as soon as fringe - fitting corrections had been applied ) .
today , the processing power and data storage volumes enjoyed by the vast majority of vlbi astronomers is @xmath2 orders of magnitude greater than the shared systems used previously .
nevertheless , the custom of excessive data averaging continues .
this practice severely limits the natural field of view of vlbi images , and is often inappropriate - especially in the era of high sensitivity observations . in fig .
[ fig1 ] we present a @xmath8 cm evn - only map of the gravitational lens , 0957 + 561 a , b .
the data were calibrated in the usual manner . to avoid bandwidth smearing ,
the data were held in the form of 28 independent but contiguous 2 mhz channels .
the data were averaged in time in a baseline dependent manner with integration times ranging from 2.5 seconds ( on the longest projected baselines ) to 30 seconds ( on the shortest projected baselines ) .
[ fig1 ] clearly shows the two main , compact components , 0957 + 561 a , b and a very faint compact source known as g@xmath9 , lying about 1 arcsec to the north of b. all three components have been detected by previous vlbi campaigns .
however , the real excitement relates to the fact that we have detected and imaged two low brightness temperature features ( @xmath10 k ) , that are associated with compact regions or hot - spots in the singly imaged arc - second scale jet that dominates vla maps of this source .
there must be many more cases where similar low surface brightness emission goes undetected , simply because a wide - field approach to the data analysis is not pursued .
vlbi observations of such emission could improve our understanding of large - scale jet physics , distinguishing between various hot - spot models and allowing a comparison between the properties of the jet ( e.g. flow velocity ) and the intergalactic medium on pc and kpc scales .
a wide - field approach has also been applied by @xcite to @xmath3 cm evn observations of the starburst galaxy m82 .
previous vlbi observations had focussed on the brightest snr ( 41.95 + 575 ) , but by following a wide - field approach to the data analysis @xcite have been able to generate exquisite images of 4 other compact snr in the field , of which the faintest has a peak flux of @xmath11 mjy / beam .
a sub - section of the entire 1 arcminute field is shown in fig .
[ fig2 ] . @xcite
have also re - analysed `` vintage '' evn data of m82 from epoch 1986 . by employing a wide - field approach to the re - analysis
they have been able to measure expansion rates for one of the snr and place upper - limits on two others .
this work is a superb example of the type of information `` gain '' one can easily acquire through the general application of wide - field techniques to vlbi data .
the advances in vlbi hardware described elsewhere in this volume suggest that for a global vlbi array r.m.s .
noise levels of @xmath12jy / beam will be attainable by the end of the millennium .
looking further ahead ( 10 - 15 years ) there is every reason to believe that @xmath13jy / beam noise levels will be achievable . at this level of sensitivity , the radio sky becomes a very crowded place . at the @xmath0jy level one may expect to encounter 1 source
every few arcseconds ( see @xcite ) . we can expect to reach these levels of sensitivity within the next 10 - 20 years ,
by which time wide - field vlbi imaging may have evolved into a standard vlbi processing route in order to avoid source confusion . even at the level of a few mjy
a wide - field approach may pay dividends .
@xcite have embarked on a survey of faint mjy sources located within a few degrees of a bright , compact radio source , 1156 + 295 .
two of the faint sources surveyed are separated by only a few arcminutes on the sky . even although the observations had not been set up with wide - field imaging specifically in mind , it was still possible to produce tapered images of both sources simultaneously from a single @xmath14 cm global vlbi data set ( see fig .
[ fig3 ] ) .
we note that this wide - field approach permits a very _ direct _ measurement of the relative astrometric positions of radio sources in the same undistorted field of view .
currently the main limits on wide - field imaging are set by the maximum data rate that can be generated by vlbi correlators . in other words ,
the limitation is exactly the reverse of the situation in the early 1980 s , where the bottleneck was associated with off - line data processing facilities .
another restriction is the size of the primary beam of the larger vlbi antennas , especially phased arrays ( e.g. wsrt & vla@xmath15 ) . in this case
at least , `` small is beautiful '' .
this latter point should be borne in mind when the next generation of large telescope arrays are being designed ( e.g. ska ) , especially with regard to their possible incorporation within existing vlbi networks .
|
we discuss the technique of wide - field imaging as it applies to very long baseline interferometry ( vlbi ) . in the past vlbi data sets
were usually averaged so severely that the field - of - view was typically restricted to regions extending a few hundred milliarcseconds from the phase centre of the field .
recent advances in data analysis techniques , together with increasing data storage capabilities , and enhanced computer processing power , now permit vlbi images to be made whose angular size represents a significant fraction of an individual antenna s primary beam .
this technique has recently been successfully applied to several large separation gravitational lens systems , compact supernova remnants in the starburst galaxy m82 , and two faint radio sources located within the same vla first field .
it seems likely that other vlbi observing programmes might benefit from this wide - field approach to vlbi data analysis . with the raw sensitivity of global vlbi set to improve by a factor 4 - 5 over the coming few years
, the number of sources that can be detected in a given field will rise considerably .
in addition , a continued progression in vlbi s ability to image relatively faint and extended low brightness temperature features ( such as hot - spots in large - scale astrophysical jets ) is also to be expected .
as vlbi sensitivity approaches the @xmath0jy level , a wide - field approach to data analysis becomes inevitable .
techniques : interferometric , image processing , , methods : data analysis 95.75.-z , 95.75.mn , 98.58.mj
|
single - photon sources are critical resources for quantum - enhanced technologies , such as quantum metrology @xcite , quantum communications @xcite , and photonic quantum computation @xcite .
heralded single - photon sources ( hspss ) , first demonstrated by hong and mandel @xcite in 1986 , have been of great interest in quantum information applications .
heralded single photons may be created by a photon - pair producing nonlinear optical process such as spontaneous parametric downconversion ( spdc ) or spontaneous four - wave mixing ( sfwm ) ; a single photon s presence is `` heralded '' by the detection of its twin photon . although hspss are not on - demand sources due to the probabilistic nature of the pair generation process , they are easier to design and implement compared to single emitter sources @xcite such as single atoms , ions , and semiconductor devices , which typically require vacuum and/or cryogenic systems . moreover , it is possible to improve a hsps to be pseudo - on - demand by using temporal or spatial multiplexing techniques @xcite . in the context of quantum information applications , there are essentially three key metrics for an hsps : indistinguishability , heralding efficiency , and source brightness .
indistinguishable photons , i.e. , pure and identical photons , are desirable because they can exhibit multi - photon interference @xcite that is central to many quantum information applications . in order to permit heralding of pure single photons , the spdc or sfwm
two - photon state source must be factorable .
heralding efficiency , the probability of a single - photon state , conditional on a heralding detector signal , corresponds to the purity of a heralded state in the photon - number basis , and thus is a very important factor to scale up quantum information protocols .
lastly , source brightness is often characterized as the photon generation rate per unit pump power ( e.g. , cps / mw ) .
a bright source can relax pump power requirements , and lead , e.g. , to higher success probability for multi - photon protocols .
it is a challenging task to optimize these three metrics simultaneously , as there are usually trade - offs between them .
for example , a typical spdc source does not generate spectrally factorable photon pairs and requires narrowband spectral filtering for high purity and indistinguishability .
however , the spectral filtering significantly reduces source brightness and heralding efficiency . in order to realize a practical high - quality hsps , several experiments @xcite have demonstrated highly nondegenerate spdc . for high brightness of heralded single photons , it is important to use not only a highly efficient spdc source but also efficient trigger detectors . in this sense , it is advantageous to generate trigger photons within the range of the high sensitivity wavelength ( 600 - 800 nm ) for off - the - shelf , efficient si avalanche photodiodes ( si - apds ) . on the other hand , for high total system heralding efficiency
, it is preferable to generate photons in the telecommunication band for lower - loss propagation through optical fibers .
thus , such nondegenerate spdc is useful to achieve high brightness and heralding efficiency in practice .
however , to the best of our knowledge , until now there has been no source based on highly nondegenerate spdc that simultaneously achieved high purity , due to the presence of strong spectral correlations between the photons .
here we propose and demonstrate an hsps based on highly nondegenerate , spectrally factorable spdc which can simultaneously attain high purity , high heralding efficiency , and a moderately high brightness .
a group velocity matching ( gvm ) condition in a periodically - poled potassium titanyl phosphate ( ppktp ) crystal allows us to produce intrinsically factorable photon pairs at @xmath1800 nm and @xmath11590 nm , which is in the telecommunication l - band .
our spectrally factorable spdc source can generate pure heralded single photons without the narrowband filtering which typically reduces the brightness and heralding efficiency .
we characterize the purity of the heralded single photons using several techniques , which consistently reveal high state purity .
finally , we discuss the possibility of further improvements to our source .
the basic configuration of our hsps is depicted in fig .
[ ppktp ] .
laser pulses with a wavelength of @xmath1530 nm are used to pump a ppktp crystal with crystal length @xmath2 and poling period @xmath3 , designed so that it collinearly produces signal - idler photon pairs at @xmath1800 nm and @xmath11590 nm in order to obtain practical benefits of higher detection efficiency by si - apds ( at @xmath1 800 nm ) and high optical - fiber transmission ( at @xmath11590 nm ) . in general , spdc has the highest conversion efficiency when the pump , signal , and idler modes satisfy energy and momentum conservation ( i.e. , phase - matching condition ) : @xmath4 @xmath5 where @xmath6 and @xmath7 are angular frequency and the wavenumber of the pump ( @xmath8 ) , signal ( @xmath9 ) , and idler ( @xmath10 ) modes , respectively . for simplicity , we assume here that the fields which participate in the spdc process are in the form of plane waves ; in appendix b we present a version of the two - photon state , where the pump is a more realistic gaussian beam , and where the signal and idler photons are collected by gaussian - beam modes , leading to single - mode optical fibers ( smfs ) . under the plane - wave approximation for the three interacting waves ,
the two - photon joint spectral state ( omitting a normalizing factor ) is given by @xmath11 where @xmath12 represents the joint spectral amplitude ( jsa ) , and @xmath13 denotes a photon - pair state with signal and idler frequencies @xmath14 and @xmath15 , respectively .
for the specifics of our experiment involving collinear downconversion with a pump which is not too tightly focused , the jsa is well approximated by @xmath16 . here , @xmath17 is the pump spectral envelope function , assumed to be well described by a gaussian function : @xmath18 \\ & = \exp\left[{-\frac{(\omega_s + \omega_i ) ^2}{\sigma_p^2 } } \right ] , \label{pumpenvelope}\end{aligned}\ ] ] where @xmath19 is the frequency detuning from the signal ( idler ) central frequency @xmath20 : @xmath21 , and @xmath22 .
@xmath23 denotes the pump spectral bandwidth .
the joint distribution @xmath24 has a maximum value when energy - conservation condition shown in eq .
is satisfied : @xmath25 .
the phase - matching function @xmath26 in a nonlinear crystal is given by @xmath27 while @xmath24 is maximized according to the energy conservation , the distribution @xmath26 is governed by dispersion in a nonlinear crystal . in our highly nondegenerate spdc
, @xmath28 can be approximated by the first - order dispersion : @xmath29 where @xmath30 is the inverse group velocity .
the phase - matching function has its maximum value for the case that @xmath31 is equal to the dispersion parameter @xmath32 .
thus , in order to obtain a desirable ( i.e. factorable ) jsa , one can choose appropriate parameters @xmath23 , @xmath2 , and @xmath33 . ; ( b ) phase - matching function @xmath34 , projected to the fiber collection modes ( see appendix b ) ; ( c ) fiber - coupled jsi @xmath35 ; ( d ) dispersion parameter @xmath32 ; the dashed line shows where @xmath36 .
( e ) heralding efficiencies versus signal collection beam waist for pump beam and idler collection beam waists of 220 @xmath37 m and 120 @xmath37 m , respectively . ] here we show how to make a factorable jsa , i.e. , so that it may be written as @xmath38 in order to generate a pure heralded single - photon state .
we use type - ii phase - matching : pump , signal , and idler fields are polarized respectively along the @xmath39 , @xmath40 , and @xmath39 crystal axes for a ppktp crystal .
( note that while our simulations incorporate the full bi - axial angle - dependence of the refractive indices , because the crystal used in our experiment is cut parallel to the crystallographic axes , with propagation of the three waves along the @xmath41 direction , the signal and idler refractive indices are approximately equal to @xmath42 and @xmath43 , respectively ) .
figure [ predictions ] ( d ) shows @xmath33 plotted as a function of the signal and idler wavelengths .
we can see that @xmath33 is approximately zero for wide ranges of signal and idler wavelengths around 800 nm and 1590 nm , respectively . for @xmath36
, one can obtain a factorable jsa by adjusting the pump bandwidth to be much larger than the phase - matching bandwidth @xcite : @xmath44 , where @xmath45 is a coefficient used for the approximation @xmath46 .
figure [ predictions ] ( a - c ) show examples of @xmath24 , @xmath34 , and joint spectral intensity ( jsi ) @xmath35 for @xmath475 thz , @xmath48 mm , and @xmath49 0.4 thz . note that our simulations presented in fig .
[ predictions ] take into account the projection of the two - photon joint amplitude on the fiber - collection modes ( see appendix b ) .
we can clearly see the factorability of @xmath35 shown in fig .
[ predictions ] ( c ) . using schmidt decomposition @xcite
, we estimate the expected attainable purity with this spdc source to be 97% .
in terms of attainable purity , utilizing spdc with @xmath50 is more advantageous compared to ones with @xmath51 that have been demonstrated in several experiments @xcite . for the @xmath51 case , peripheral lobes from the sinc phase - matching function appear along the anti - diagonal ( @xmath52 ) direction , limiting the single - photon purity to @xmath185% . for @xmath53 with a broadband pump spectrum ,
the peripheral lobes are widely distributed along @xmath54 , and thus degrade the purity to a lesser extent .
this gvm spdc source can also achieve high heralding efficiency .
it has been analyzed @xcite and demonstrated experimentally @xcite that a high heralding efficiency can be realized by using large pump beam waist sizes , relative to the collection modes . as discussed in @xcite , roughly speaking , a larger beam waist size has a narrower wavevector distribution , which makes a two - photon state strongly entangled in spatial mode . in addition , the collection beam profile of signal and idler modes may be chosen to eliminate spatial - spectral entanglement , i.e. , spatial chirp , of each spdc mode . in this case , the signal and conjugate idler spatial mode can each be efficiently coupled into an smf .
our numerical calculations based on the theory described in appendix b predict this proposed spdc source can achieve heralding efficiency up to 95% by optimizing the beam waist sizes of the three interaction modes ( see fig .
[ predictions ] ( e ) ) .
nm ) , smf : single - mode optical fiber , pbs : polarizing beam - splitter , hwp : half - wave plate , fpc : fiber polarization controller , apd : avalanche photodiode , ld : laser diode .
, title="fig : " ] nm ) , smf : single - mode optical fiber , pbs : polarizing beam - splitter , hwp : half - wave plate , fpc : fiber polarization controller , apd : avalanche photodiode , ld : laser diode .
, title="fig : " ]
figure [ setup ] illustrates a schematic diagram of our experimental setup for generation and characterization of the proposed spdc and heralded single photons .
we used the second harmonic ( central wavelength 521 nm , pulse duration 160 fs ) of a mode - locked yb - doped fiber laser ( pulse repetition rate 100 mhz ) as a pump source .
for the second harmonic generation , we used a lithium tri - borate ( lbo ) crystal with noncritical phase - matching condition to achieve the wavelength conversion without spatial walk - off .
the pump beam is focused in the middle of a 20-mm long ppktp crystal with a poling period of 28.275 @xmath37 m , producing signal - idler photon pairs at 775 nm and 1590 nm , respectively .
wide - band ( @xmath55 = 20 nm ) interference filters ( if ) and a dispersive prism are used for suppression of the pump beam . for photon counting measurements we used si- ( ingaas- ) apds for signal ( idler ) modes , with detector efficiencies of 60% ( 19% ) . in this spdc configuration
, there is a slight group velocity mismatching between pump and signal modes ( @xmath56 ) ; we chose this for the given pump wavelength of 521 nm so that the idler wavelength is within the sensitive range of the ingaas - apds .
for the pump wavelength , perfect gvm is satisfied when the signal - idler photon pair is generated at 762 nm and 1647 nm , as shown in fig .
[ predictions ] ( d ) .
according to the numerical simulations shown in fig . [ predictions ] ( e ) , we determined a pump beam waist size of 220 @xmath37 m and spdc collection beam waist sizes of 125 @xmath37 m and 120 @xmath37 m for the signal and idler modes . using the method proposed in @xcite , we observed smf coupling efficiencies of @xmath57% and @xmath58% for the signal and idler modes , respectively@xmath59 .
these coupling efficiencies are higher than in previous demonstrations @xcite using highly nondegenerate spdc sources , and correspond to the attainable heralding efficiencies assuming lossless collection optics .
the source brightness is another key metric for an hsps . in our souce
we estimate the photon - pair generation efficiency as @xmath60 , accounting for the collection optics loss and smf coupling and spd efficiencies .
this value is @xmath110@xmath61 higher than those used for recent 8-photon experiments @xcite , which generate pure single photons by narrowband spectral filtering .
although many nondegenerate spdc sources utilizing the highest @xmath62 nonlinear susceptibility have significantly higher brightness @xcite , those need narrowband spectral filtering to attain state purity , which also significantly reduces brightness and heralding efficiency .
in order to characterize the heralded single - photon purity , we performed three different measurements .
first , in order to map the jsi , i.e. , @xmath63 of our spdc source , we used a frequency - resolved optical parametric amplification ( opa ) technique @xcite ( see fig . [ setup ] ( a , b ) ) . for this measurement ,
in addition to the pump laser for the spdc process , we used a laser diode ( ld ) whose wavelength is stabilized ( linewidth @xmath64 20 mhz ) and tunable around the idler central wavelength . the wavelength - stabilized beam is coupled to the collection smf for the idler mode , propagating backwards along the optical path of the idler photons .
the backward propagating beam is reflected by a curved mirror and subsequently focused into the ppktp crystal to seed the opa process ; the corresponding signal spectral mode in the jsi is then amplified .
in other words , only the `` cross - section '' of the jsi at the idler seed wavelength is amplified .
thus , the jsi can be reconstructed by measuring the change of the amplified signal spectrum as the seed wavelength is scanned . with our 1-mw seed power ,
the amplification gain over the spdc process is approximately @xmath65 , which is sufficient so as to observe the amplified signal spectra by a standard ccd - based spectrometer .
compared to the methods relying on coincidence detection of spdc photon pairs , this opa - based measurement provides a result much faster and with a higher spectral resolution as determined by the tuning resolution of the seed laser and the spectrometer s resolution .
figures [ jsi ] ( a - c ) show experimentally reconstructed jsis for different pump bandwidths ; using second - harmonic generation lbo crystals with lengths of 3 , 7 , and 15 mm , we obtained spectral fwhm bandwidths of 4.1 , 2.1 , and 1.0 nm , respectively . the spectral resolution in the jsi measurement
is 0.1 nm for both spdc modes .
we find that while signal - mode bandwidth varies according to the pump bandwidth , that of the idler mode remains approximately unchanged .
this indicates that signal and pump group velocities are well matched , and consequently the phase - matching function has @xmath50 .
we estimate the heralded single - photon purity under the simplifying assumption that all spectral components are in phase , i.e. , @xmath66 . computing schmidt decompositions
, we obtain purities of 60.0% , 79.9% , and 90.6% for pump bandwidths of 4.1 , 2.1 , and 1.0 nm , respectively .
these estimated purities are in excellent agreement with the theoretical predictions ( 62% , 83% , and 91% for the same pump bandwidths ) shown in fig .
[ jsi ] ( d - f , g ) .
although our observed purity does not reach the attainable value ( 97% , predicted in fig .
[ predictions ] ( b - d ) ) for @xmath36 because of slightly mismatched group velocities of the pump and signal mode ( @xmath56 ) , it is much higher than other sources using nondegenerate spdc . the slightly shifted signal central wavelengths for the different pump bandwidths were due to a temperature change in the lbo crystals ( with the tuning sensitivity of @xmath10.7 nm / c@xmath67 around 173 c@xmath67 ) .
we also found that the spdc central wavelengths are insensitive to the temperature change in the ppktp crystal ( @xmath68 nm / c@xmath67 for the signal mode from 30 to 100 c@xmath67 ) ; the refractive - index changes of the three spdc modes are well cancelled to maintain the phase - matching condition .
the frequency - resolved opa can reveal very precise spectral intensity correlations .
however , it can not extract `` amplitude '' correlations including spectral phase , because the opa is a phase - insensitive amplifier for the vacuum field of the signal mode . in order to more directly estimate the purity without the spectral - phase assumption , we observed the second - order auto - correlation function at zero time delay @xmath69 for each spdc mode .
as described in @xcite , for an spdc source , @xmath70 , where @xmath71 is the purity of a heralded single photon state .
for example , @xmath72 for single - mode spdc and pure heralded single photons , while @xmath73 indicates a poissonian distribution , i.e. , multimode spdc and heralded single photons in completely mixed states .
in this experiment , using a fiber - based hanbury brown - twiss setup @xcite ( see fig .
[ setup ] ( c ) ) , we measured two apds single count rates @xmath74 and @xmath75 and coincidence count rate @xmath76 ; @xmath69 is then estimated by @xmath77 where @xmath78 100 mhz is the pump repetition rate . for our spdc source with an optimized ( 0.7-nm ) pump bandwidth ,
we observed @xmath79 @xmath80 for the idler ( signal ) mode , corresponding to a heralded single - photon purity of @xmath81% ( @xmath82% ) .
these purities are consistent with those estimated from the jsi measurement , indicating that the spectral phase of our spdc source does not degrade the purity . finally , we measured polarization - mode hong - ou - mandel interference ( homi ) @xcite of heralded single photons generated by sequential pump pulses .
the homi visibility is a direct measure of total state purity ( including the effect of the joint spectral phase ) and is crucial for many quantum information applications .
our setup is shown in fig .
[ setup ] ( d ) .
sequential detections of signal photons herald two idler photons separated by 10 ns , which corresponds to the time interval between two subsequent pump pulses .
a fiber polarization controller ( fpc ) adjusts the idler photon s polarization to be a superposition of horizontal ( @xmath83 ) and vertical ( @xmath84 ) polarization . by using a polarization - dependent optical delay line that delays only @xmath83-polarized photons for 10 ns , an orthogonally polarized
, temporally overlapped two - photon state @xmath85 is made probabilistically ( with 25% success probability ) for the case that the early - born photon with @xmath83-polarization passes the delay line , and the late - born photon with @xmath84-polarization is reflected by a polarizing beam - splitter ( pbs ) .
the two - photon st ate @xmath85 is projected onto the diagonal / anti - diagonal ( @xmath86 ) polarization basis ( @xmath87 , @xmath88 ) , and is then subjected to coincidence detections .
the results of our homi experiment for idler photons are shown in fig .
[ homi ] ( a ) .
a clearly - observed dip in coincidence counts indicates that the heralded idler photons are in highly pure states .
the solid curve is the best fit of the experimental data after subtracting background counts ( 2.4 counts for each data point ) ; we obtained an interference visibility of @xmath58% and a dip width of @xmath89 ps .
as a fitting function , we used a convolution of a gaussian function and the square of a triangular function , as predicted from the theoretical two - photon joint - spectral amplitude ( see appendix a ) . using a similar setup , we also observed the homi of signal photons heralded by sequential detections of idler photons ( see fig .
[ homi ] ( b ) ) .
a longer time of data accumulation was used for the homi measurement involving two signal photons because of less efficient smf coupling after the delay line . by using a gaussian fit ( see appendix a ) , the interference visibility and the dip width are estimated to be @xmath90% ( after subtracting 1.7 background counts ) and @xmath91 ps , respectively . although high and similar visibilities are observed for both measurements , as expected from the reconstructed jsi ( see fig .
[ jsi ] ( c ) ) , the widths of the two dips differ by a factor greater than 10 .
these observed visibilities are in good agreement with the purities estimated by the jsi and @xmath72 measurements , as well as with our theoretical predictions .
it is worthwhile discussing further possible improvements to our hsps .
our observed purity is essentially limited by the slightly mismatched group velocities of pump and signal modes . as shown in fig .
[ predictions ] , for example , perfect gvm is achieved by the pump , signal , and idler wavelengths of @xmath92 nm , 800 nm , and 1588 nm . however , in practice , preparation of the necessary somewhat high - power and broadband optical pulse at @xmath92 nm would require a bulky pump laser source , e.g. , a frequency - doubled optical parametric oscillator pumped by a ti - sapphire laser . with our current pump wavelength ( 521 nm ) ,
perfect gvm is satisfied by producing signal and idler photons at 762 and 1647 nm , respectively .
however , it is challenging to characterize photons with wavelengths greater than 1600 nm , for which standard apds are not sensitive ; however , superconducting single - photon detectors tuned for this wavelength may be a viable option . through numerical simulations , we found that eliminating the peripheral lobes of the phase - matching function for our current spdc source can enhance the purity up to 94% ; modulating poling periods in a ktp crystal can lead to a phasematching function without peripheral sidelobes , e.g. a gaussian function , instead of the typical sinc - function dependence which does present sidelobes @xcite .
furthermore , in our spdc source the poling modulation makes it possible to independently control signal and idler spectra : the signal - mode spectrum is shaped by the pump pulse spectrum , while the idler one can be controlled by the crystal length and poling modulation .
finally , we note that because our spdc source employs collinear phase - matching , it would be compatible with using waveguide structures so as to enhance the photon - pair generation efficiency .
we have demonstrated the generation of intrinsically pure heralded single photons using a highly nondegenerate , spectrally - factorable spdc source based on a ppktp crystal .
the signal photon at 775 nm permits the use of a si - apd to herald the idler photon , while the heralded idler photon at 1590 nm is in the telecom l - band , a spectral region where optical fibers are highly transmissive .
a broadband pump source together with matched group velocities of the pump and signal modes produces spectrally factorable photon pairs and pure heralded single photons without the need for spectral filtering . with an optimized pump bandwidth , a @xmath190% single - photon purity was observed in our jsi , @xmath69 , and homi measurements .
we also observed that our source can achieve a high heralding efficiency using appropriate collection optics .
we anticipate that incorporating the spdc source with temporal multiplexing techniques @xcite will enable near - deterministic single - photon generation to scale up systems for optical quantum information processing .
funding for this work has been provided by nsf grant no .
phy 12 - 12439 and phy 15 - 20991 and us army aro durip grant
w911nf-12 - 1 - 0562 .
@xmath59 in earlier version of the source , we observed slightly asymmetric collection efficiencies for the signal and idler modes .
we believe this occurred due to non - optimal crystal coatings : we used a dielectric coating on the output facet of the ppktp crystal so that signal and idler photons escape from the crystal with 99% efficiency , while 98% of pump power is reflected on the facet to reduce the power transmitted into the spdc collection beam paths .
the input facet was anti - reflection coated ( with 99% transmission ) for the pump and idler wavelengths , but not for the signal wavelength . with these coatings , however ,
the reflected pump beam at the output facet can also generate unwanted spdc in the backward direction .
the unwanted signal photon is then partially reflected by the non - ideally coated ( with @xmath190% transmission ) input facet of the ppktp crystal , and as a result , is coupled to a collection smf with @xmath12% coupling efficiency of the main spdc signal photon .
this also caused an asymmetry in the second - order auto - correlation function measurement ( @xmath93 and@xmath94 for the signal and idler modes ) , because the unwanted signal photon has a delayed temporal mode ( corresponding to the round - trip time inside the ppktp crystal ) relative to the one from the main spdc , so that the signal mode then has two temporal modes .
here we describe theoretical curves of our observed homi dips . a schematic of the experiment
is shown in fig .
[ setup ] ( a , d ) .
we assume that sequentially generated photon pairs have identical jsa @xmath95 , where @xmath96 and @xmath97 are respectively frequency detunings of the signal and idler photons . for a given sequentially generated four - photon state @xmath98 , the probability of sequentially generated heralded idler photons coincidence detection , is given by @xcite @xmath99 @xmath100 where @xmath101 is the time delay through the optical delay line .
@xmath102 denotes the dip function of homi . since our spdc source has the dispersion parameter @xmath103 , the jsa ( without normalization factor ) is well approximated by @xmath104 \mathrm{sinc } \left[\frac{(k'_p - k'_i)\omega_i l}{2 } \right ] .
\label{s}\end{aligned}\ ] ] substituting eq . into eq .
, we obtain @xmath105 t(t , ( k'_p - k'_i)l)^2 , \label{dip_idler}\end{aligned}\ ] ] where @xmath106 is a triangular function : @xmath107 we used eq . to find the best fit of our observed homi dip shown in fig .
[ homi ] ( a ) . for our spdc source , eq .
more closely follows @xmath108 which is reflected by the sinc phase - matching function , because the temporal width of the convoluted gaussian function @xmath109 is much narrower than that of the triangular function @xmath110 . on the other hand , using similar computation , we obtain a dip for the signal photons interference for @xmath111 : @xmath112 .
\end{aligned}\ ] ] this gaussian dip is obtained due to the pump gaussian spectral envelope function and the group velocity matching of the pump and signal modes .
thus , we used a gaussian function for fitting our observed homi of signal photons shown in fig .
[ homi ] ( b ) .
following a standard perturbative approach , the quantum state produced by spontaneous spdc may be expressed as @xmath113 here , @xmath114 represents the joint amplitude , and @xmath115 is a constant related to the conversion efficiency .
the joint amplitude can in turn be expressed as @xmath116 and of the pump envelope function which , under the assumption of a gaussian spectral shape , may be written as in eq .
( [ pumpenvelope ] ) . in the above equations , @xmath118 is the pump beam waist , @xmath2 is the crystal length , @xmath119 is the position of the pump beam waist relative to the crystal central plane , and @xmath3 is the poling period .
note that the phase in eq .
( [ phasematchfuct ] ) is dependent on the choice of coordinate system ; we have placed the origin on the crystal s second face . here
, we have made the following definitions ( an underlined symbol indicates a two - dimensional transverse vector ) in terms of the cartesian components of the signal and idler wavevectors in writing eq .
( [ phasematchfuct ] ) we have assumed that poynting vector walkoff for the three interacting fields may be neglected , an assumption which is warranted for a crystal cut parallel to the crystallographic axes , as is the case in our experiment ( see above ) . in eq .
( [ colectmode ] ) @xmath127 , for the signal(s ) and idler ( i ) , @xmath128 is a normalization factor , @xmath129 represents the beam radius at the beam waist assumed to coincide with the crystal - air interface , @xmath130 represents the angular orientation of the collection modes , and @xmath131 represents the height of the collection mode on the crystal - air interface .
note that for collinear spdc , as used in our experiments ( see above ) , @xmath132 and @xmath133 .
here @xmath135 incorporates various source parameters .
similarly , the single - photon detection rate ( written below for the signal photon ; a similar expression exists for the idler photon ) can be expressed as for a heralded single - photon source a relevant quantity is the heralding efficiency .
the idler - mode heralding efficiency @xmath139 is the probability that a single photon is present in the idler - mode fiber , conditioned on the presence of a single photon in the signal - mode fiber , and is defined by where a corresponding equation may be written for the signal - mode heralding efficiency . in calculating the ratio in eq .
( [ eta_i ] ) , the relationship @xmath141 is useful ( with a similar relationship for @xmath142 ) ; the results of this calculation are shown in fig .
[ predictions ] ( e ) .
|
we report on the generation of an indistinguishable heralded single - photon state , using highly nondegenerate spontaneous parametric downconversion ( spdc ) .
spectrally factorable photon pairs can be generated by incorporating a broadband pump pulse and a group - velocity matching ( gvm ) condition in a periodically - poled potassium titanyl phosphate ( ppktp ) crystal .
the heralding photon is in the near ir , close to the peak detection efficiency of off - the - shelf si single - photon detectors ; meanwhile , the heralded photon is in the telecom l - band where fiber losses are at a minimum .
we observe spectral factorability of the spdc source and consequently high purity ( 90% ) of the produced heralded single photons by several different techniques . because this source can also realize a high heralding efficiency ( @xmath0% )
, it would be suitable for time - multiplexing techniques , enabling a pseudo - deterministic single - photon source , a critical resource for optical quantum information and communication technology .
|
a nova eruption occurs when a white dwarf has accreted enough material from a companion star to trigger thermonuclear runaway on its surface ( e.g. , starrfield et al .
observations with modern telescopes have revealed the presence of strong shocks in nova systems . in particular , the discovery of gev @xmath2-rays from v407 cyg indicated that in at least some systems , the shocks were strong enough to accelerate particles to very high energies ( abdo et al . 2010 ) . in the case of v407 cyg ,
the shocks were the result of the nova ejecta colliding with a dense stellar wind from a mira variable companion ( abdo et al .
much to the surprise of the nova community , further observations with the _ fermi gamma - ray space telescope _ discovered @xmath2-ray emission from systems with main sequence companions ( e.g. , ackermann et al . 2014 ) . while observations across the electromagnetic spectrum have contributed to the discovery of shocks in novae ,
radio observations have proven to be particularly valuable . resolved radio images
can reveal non - spherical structure and provide evidence for multiple outflows , as in the case of v959 mon ( chomiuk et al . 2014 ) .
early monitoring of the radio emission , even when the ejecta are too small to resolve , has also prove to be an effective probe of shocks in novae .
the shocks can manifest as synchrotron emission with non - thermal brightness temperatures in spectacular ways as in v1723 aql ( weston et al .
2016a ) , or more subtle ways as in v5589 sgr ( weston et al .
when the radio synchrotron emission is bright enough , the shocks can be directly imaged with very long baseline interferometry , as in v959 mon ( chomiuk et al . 2014 ) and rs oph ( rupen et al . 2008 ; sokoloski et al . 2008 ) .
observations with modern x - ray telescopes have also proven to be useful in characterizing shocks in novae .
those novae embedded within the wind of a giant companion typically emitting bright , hard x - rays early in their evolution , such as rs oph ( bode et al .
2006 ; nelson et al . 2008 ) and v745 sco ( page et al . 2015 ; orio et al .
novae in systems with main sequence companions are also capable of producing strong x - rays , but typically later in their evolution such was the case in both v382 vel ( mukai & ishida 2001 ) and v5589 sgr ( weston et al .
2016b ) .
on 2015 february 11 , the nova v1535 sco was discovered in the constellation scorpius .
it was reported in vsnet - alert 18276 by p. schmeer and in cbet 4078 by t. kojina .
early optical and near infrared follow - up indicated the nova was fading rapidly ( walter 2015 ) .
the combination of rapid fading and a possible bright near - infrared counterpart indicated that the companion star could be an m giant , in which case the nova could be embedded in the wind from a giant companion ( walter 2015 ) .
srivastava et al .
( 2015 ) later argued that the companion could be a k giant based on pre - eruption 2mass photometry of a likely counterpart ( 2mass j17032617 - 3504178 ) .
spectroscopic observations indicated it was a he / n type nova ( walter 2015 ; srivastava et al .
early x - ray and radio observations ( nelson et al . 2015 ) indicated the presence of strong shocks in the ejecta , which was further evidence that the ejecta were expanding into a dense medium .
the nova was originally given the designation pnv j1703260 - 3504140 , then known as nova sco 2015 , and eventually given the official designation v1535 sco .
srivastava et al .
( 2015 ) also noted that narrowing of the h - line profiles in the infrared indicated the presence of a deceleration shock , and estimated the total mass ejected to be between @xmath3 and @xmath4 @xmath5 m@xmath6 ( where @xmath5 is the filling factor of the ejecta ) . during the 2015 outburst , no _ fermi gamma - ray space telescope _
staring mode observations were scheduled and the nova was not reported as a detection from the survey mode data . however , it must be noted that _
fermi _ is generally less sensitive to transients at the galactic center due to a combination of the high background , and the north - south `` rocking '' profile of the survey mode results in the galactic center having minimal exposure ( acero et al .
it is therefore possible that v1535 sco produced @xmath2-rays like other novae ( e.g. , v407 cyg and v745 sco ; abdo et al .
2010 ; cheung et al .
2014 ) , but none were detectable due to these limiting factors .
recent work by morris et al .
( 2017 ) argues that all novae produce @xmath2-ray emission , but we mainly detect the nearby ones with _ fermi_. in this paper , we present our multi - wavelength observations of v1535 sco made during 2015 . these observations were made with the karl j. jansky very large array ( vla ) , the very long baseline array ( vlba ) , the _ swift _ x - ray telescope ( xrt ) , and the small & moderate aperture research telescope system ( smarts ) .
we present our observations and data reduction methods in section 2 .
our knowledge about the distance to v1535 sco is discussed in section 3 . in section 4 ,
we present results from the multi - wavelength observations .
we discuss our findings in section 5 .
final conclusions are summarized in section 6 . throughout this paper
, we use the initial optical detection of the nova on 2015 february 11.837 ( mjd 57064.837 ) as day 0.0 .
we also use the term `` embedded nova '' to refer to any nova embedded in the wind material from its companion star ( e.g. chomiuk et al . 2012 ; mukai et al .
we began monitoring v1535 sco with the vla within 64 hours of its discovery in the optical .
the first epoch only included observations at c - band ( 4.0 - 8.0 ghz ) and ka - band ( 26.5 - 40.0 ghz ) .
all following epochs also included observations at l - band ( 1.0 - 2.0 ghz ) and ku - band ( 12.0 - 18.0 ghz ) .
to maximize spectral coverage , all bands were split into upper and lower sidebands . for the first 8 epochs
, we used 8-bit sampling for all frequencies .
this gave us a total bandwidth of 1.0 ghz at l - band , and 2.0 ghz at all other bands .
for the ninth epoch ( day 93.463 ) we used 3-bit sampling for ku - band and ka - band , resulting in total bandwidths of 6.0 ghz and 8.0 ghz , respectively .
for the final epoch ( day 122.363 ) , we used 3-bit sampling for c - band , ku - band , and ka - band , giving total bandwidths of 4.0 ghz , 6.0 ghz , and 8.0 ghz , respectively .
10 epochs were observed under the program vla/13b-057 , and one epoch ( day 93.463 ) was observed under the program s61420 .
the nova was detected during the first epoch ( day 2.7 ) and remained detectable at multiple frequencies for nearly 100 days .
we ceased observations once the nova had faded to the point where it was no longer detected at all frequencies .
ccccccc 2015 - 02 - 14 & 2.663 & b & 4.55 & 13.6 & 4.13 & 0.13 + & & & 7.38 & 13.6 & 2.786 & 0.085 + & & & 28.2 & 11.3 & 0.819 & 0.071 + & & & 36.5 & 11.3 & 0.675 & 0.090 + 2015 - 02 - 18 & 6.663 & b & 13.5 & 11.9 & 0.416 & 0.021 + & & & 17.4 & 11.9 & 0.344 & 0.024 + & & & 28.2 & 11.8 & 0.295 & 0.053 + & & & 36.5 & 11.8 & 0.376 & 0.074 + 2015 - 02 - 19 & 7.663 & b & 1.26 & 15.6 & 1.57 & 0.090 + & & & 1.74 & 15.6 & 1.21 & 0.081 + & & & 4.55 & 13.6 & 0.650 & 0.026 + & & & 7.38 & 13.6 & 0.439 & 0.018 + 2015 - 02 - 24 & 12.763 & b & 13.5 & 12.5 & 0.385 & 0.020 + & & & 17.4 & 12.5 & 0.456 & 0.024 + & & & 28.2 & 12.4 & 0.785 & 0.062 + & & & 36.5 & 12.4 & 0.845 & 0.087 + 2015 - 03 - 01 & 17.663 & b & 1.26 & 15.6 & < 0.317 & 0.10 + & & & 1.74 & 15.6 & < 0.317 & 0.066 + & & & 4.55 & 13.6 & 0.221 & 0.019 + & & & 7.38 & 13.6 & 0.192 & 0.017 + 2015 - 03 - 07 & 23.663 & b & 13.5 & 12.0 & 0.585 & 0.024 + & & & 17.4 & 12.0 & 0.755 & 0.031 + & & & 28.2 & 11.9 & 1.15 & 0.074 + & & & 36.5 & 11.9 & 1.68 & 0.11 + 2015 - 03 - 10 & 26.563 & b & 1.26 & 15.6 & < 0.374 & 0.074 + & & & 1.74 & 15.6 & < 0.271 & 0.054 + & & & 4.55 & 13.5 & 0.455 & 0.023 + & & & 7.38 & 13.5 & 0.536 & 0.021 + 2015 - 03 - 25 & 41.583 & b & 13.5 & 12.0 & 0.267 & 0.015 + & & & 17.4 & 12.0 & 0.299 & 0.017 + & & & 28.2 & 12.0 & 0.348 & 0.044 + & & & 36.5 & 12.0 & < 0.351 & 0.055 + 2015 - 03 - 25 & 41.633 & b & 1.26 & 15.6 & < 0.609 & 0.112 + & & & 1.74 & 15.6 & < 0.271 & 0.060 + & & & 4.55 & 13.8 & 0.222 & 0.017 + & & & 7.38 & 13.8 & 0.253 & 0.014 + 2015 - 04 - 07 & 54.563 & b & 13.5 & 12.5 & 0.457 & 0.028 + & & & 17.4 & 12.5 & 0.399 & 0.022 + & & & 28.2 & 12.4 & 0.323 & 0.047 + & & & 36.5 & 12.4 & 0.276 & 0.062 + 2015 - 04 - 08 & 55.733 & b & 1.26 & 15.7 & 0.87 & 0.10 + & & & 1.74 & 15.7 & 0.661 & 0.060 + & & & 4.55 & 13.7 & 0.483 & 0.024 + & & & 7.38 & 13.7 & 0.374 & 0.018 + 2015 - 04 - 18 & 65.513 & b & 1.26 & 16.6 & 0.444 & 0.082 + & & & 1.74 & 16.6 & 0.258 & 0.057 + & & & 4.55 & 14.6 & 0.256 & 0.019 + & & & 7.38 & 14.6 & 0.228 & 0.014 + 2015 - 04 - 19 & 66.493 & b & 13.5 & 12.0 & 0.209 & 0.017 + & & & 17.4 & 12.0 & 0.167 & 0.023 + & & & 28.2 & 12.0 & 0.173 & 0.049 + & & & 36.5 & 12.0 & < 0.289 & 0.083 + 2015 - 05 - 01 & 78.463 & b & 1.26 & 15.6 & < 0.32 & 0.11 + & & & 1.74 & 15.6 & < 0.378 & 0.071 + & & & 4.55 & 13.6 & 0.170 & 0.019 + & & & 7.38 & 13.6 & 0.139 & 0.015 + 2015 - 05 - 01 & 78.503 & b & 13.5 & 12.1 & 0.129 & 0.017 + & & & 17.4 & 12.1 & 0.164 & 0.022 + & & & 28.2 & 12.0 & 0.174 & 0.053 + & & & 36.5 & 12.0 & < 0.241 & 0.075 + 2015 - 05 - 16 & 93.463 & b@xmath7bna & 1.26 & 8.0 & < 0.531 & 0.177 + & & & 1.74 & 8.0 & < 0.336 & 0.078 + & & & 4.55 & 8.0 & 0.0727 & 0.023 + & & & 7.38 & 8.0 & 0.0708 & 0.017 + & & & 13.5 & 8.0 & 0.0518 & 0.018 + & & & 16.5 & 8.0 & < 0.0859 & 0.021 + & & & 29.5 & 9.0 & < 0.149 & 0.036 + & & & 35.0 & 9.0 & < 0.135 & 0.045 + 2015 - 06 - 14 & 122.363 & bna@xmath7a & 1.26 & 8.0 & < 0.78 & 0.12 + & & & 1.74 & 8.0 & < 0.233 & 0.078 + & & & 5.0 & 8.0 & < 0.0806 & 0.020 + & & & 7.0 & 8.0 & < 0.111 & 0.022 + & & & 13.5 & 8.0 & <
0.106 & 0.016 + & & & 16.5 & 8.0 & < 0.120 & 0.021 + & & & 29.5 & 9.0 & < 0.264 & 0.045 + & & & 35.0 & 9.0 & < 0.176 & 0.059 for all epochs , we used the absolute flux calibrator 3c286 .
complex gains calibrators were j1626 - 2951 for both l - band and c - band , and j1650 - 2943 for both ku - band and ka - band .
all vla data were calibrated with the nrao vla calibration pipeline , version 1.3.1 , which uses casa version 4.2.2 .
the pipeline script was executed on either a dedicated desktop or an nrao lustre node .
once calibrated , the data were exported to aips for additional flagging .
the fully - flagged data were imaged with difmap ( sheperd 1997 ) .
the images were then imported back into aips and flux densities were measured using the task ` jmfit ` .
the l - band and c - band data from 2015 april 18 were also calibrated in aips in order to look for polarization . also , the ku - band and ka - band data from 2015 march 25 were re - calibrated in aips to check the results of the casa pipeline .
our vla results are presented in table [ radiotab ] . for non - detections , we calculated the upper limit as the flux density at the nova location plus 3 times the image root mean square ( rms ) .
our uncertainties are calculated by adding the image rms and an absolute flux density uncertainty in quadrature .
perley & butler ( 2013 ) report that the vla absolute flux density calibration is stable to within @xmath8 for 1 to 20 ghz , and within @xmath9 for 20 to 50 ghz .
because v1535 sco is far from 3c286 , our uncertainty in the flux density calibration will be higher .
we adopt an absolute flux density uncertainty of @xmath9 for 1 to 20 ghz , and @xmath10 for above 20 ghz .
cccccc 2015 - 02 - 19 & 7.663 & 4.87 & 177.3 & 0.477 & 0.056 + 2015 - 02 - 24 & 12.763 & 4.87 & 177.3 & < 0.278 & 0.041 + ccccccc 00033634002 & 4.16 & 4084 & 591 & 0.0062 & 0.1578 & 0.1640 + 00033634003 & 11.04 & 3485 & 198 & 0.0555 & 0.0878 & 0.1428 + 00033634004 & 13.94 & 3407 & 431 & 0.0886 & 0.0553 & 0.1429 + 00033634005 & 17.93 & 4669 & 676 & 0.1342 & 0.0329 & 0.1666 + 00033634006 & 24.75 & 2959 & 252 & 0.0925 & 0.0174 & 0.1091 + 00033634007 & 31.51 & 4378 & 46 & 0.0102 & 0.0067 & 0.0161 + 00033634008 & 38.77 & 4568 & 27 & 0.0021 & 0.0054 & 0.0075 + 00033634009 & 45.55 & 2138 & & < 0.0011 & < 0.0011 & < 0.0082 + 00033634010 & 49.08 & 2575 & 5 & 0.0008 & 0.0021 & 0.0024 + 00033634011 & 53.07 & 4714 & 12 & 0.0016 & 0.0025 & 0.0041 + 00033634012 & 60.03 & 4845 & 13 & 0.0013 & 0.0015 & 0.0027 + 00033634013 & 66.71 & 2979 & 7 & 0.0007 & 0.0026 & 0.0028 + 00033634014 & 73.86 & 4934 & & < 0.0005 & < 0.0002 & < 0.0036 + 00033634015 & 80.32 & 4526 & & < 0.0002 & <
0.0001 & < 0.0024 + we had 2 epochs of vlba observations under the program vlba/15a-269 , both in c - band with a central frequency of 4.87 ghz and a total bandwidth of 256 mhz .
for both epochs , we used j1709 - 3525 as the phase reference source .
the bright sources j1656 - 3302 and j1713 - 3418 were also observed as a means of gauging the successful calibration of the nova .
for the first epoch , 3 antennas were not usable : hancock , north liberty , and mauna kea .
for the second epoch , hancock and mauna kea could again not be used .
the total time on source for each epoch was approximately 177.3 minutes ( or 2.96 hours ) .
both epochs were calibrated using standard routines in aips , including the new bandpass correction routine described by walker ( 2014 ) .
images were made in difmap .
the nova was detected in the first epoch , but not in the second epoch ( see table [ vlbatab ] ) .
schedules for further epochs were submitted , but conflict with a top - priority vlba program prevented them from being observed .
the position of the compact source detected in the first vlba epoch was ra 17@xmath1103@[email protected] @[email protected] , dec [email protected] .
we obtained 14 epochs of _ swift _ xrt observations .
the exposure times ranged from 2.5 to 4.9 ks .
the nova was detected by the xrt in 11 of the 14 observations .
details of the observations are given in table [ swiftab1 ] .
we divided the _
swift _ xrt detections into soft and hard bands based on the photon energy , @xmath15 .
we chose @xmath15<1.0 kev to be soft and @xmath15>1.0 kev to be hard .
our decision to use these designations is based on previous observations of novae which indicate which indicate that the @xmath15<1.0 kev band can be dominated by the super - soft x - rays , while the @xmath15>1.0 kev x - rays are likely to be from shocks ( e.g. , mukai et al . 2008 ; mukai et al .
the x - ray emission was initially hard .
the soft emission increased over the first @xmath020 days while the hard emission decreased over the same time period .
we had no detections with the xrt after day 67 .
our _ swift _ xrt results are plotted in figure [ swift_all ] .
( 6563 ).,width=336 ] line over the first 16.5 days of nova eruption .
note the rapid narrowing of the line as the fastest material becomes optically thin.,width=307 ] smarts monitors novae in the optical and near infrared and provides the results to the public via the stony brook / smarts spectroscopic atlas of ( mostly ) southern novae .
photometric observations of v1535 sco began on 2015 february 13.35 using the andicam dual - channel imager on the smarts 1.3 m telescope .
spectroscopic measurements of the ejecta began on 2015 february 13.5 as part of the smarts observation campaign using the chiron echelle spectrometer on the 1.5 m smarts telescope ( tokovinin et al .
2013 ) . as an example
, we show the spectrum from day 1.565 in figure [ smarts_exspec ] .
there is p cyg absorption present on the h@xmath16 ( 4861 ) , h@xmath2 ( 4340 ) , ( 4471 , 4921 , & 5879 ) , ( 5169 ) , and ( 7774 & 8446 ) lines .
this is unusual for a he / n nova , as they typically do not have much mass loss .
the h@xmath17 line width measurements are presented in table [ smarts_hatab ] , and we show the evolution of the h@xmath17 line from day 1.565 to day 16.548 in figure [ smarts_ha ] . because we are concerned with the outermost edge of the main component of the ejecta , we estimate the maximum velocity as the full width at 3@xmath18 ( fw3@xmath18 ) .
we convert the h@xmath17 fwhm velocities to fw3@xmath18 velocities by assuming the core of the spectral line has a gaussian shape , then we multiply by a factor of @xmath19 in order to get the @xmath20 velocities . our velocity evolution is shown in figure [ sco15_vnr ] . while the measured velocity of the ejecta appears to decrease with time , we must emphasize that what is actually being measured here is the velocity of the h@xmath17 emitting region , which does not necessarily correspond to the outer edge of the ejecta .
outlier over the first 16.5 days of the nova eruption .
the vertical line is at -4850 km s@xmath21 .
note that by day 6.5 the lines at 6455.8 and 6456.4 dominate this region of the spectrum.,width=307 ] the h@xmath17 ( 6563 ) line shows high velocity outliers ( both red and blue shifted ; see table [ smarts_hawings ] ) .
these outliers do not appear to decelerate as long as they are present .
in fact , the high velocity outliers appear to accelerate throughout the first week of the nova s evolution , although this could be the result of blending with lines from ( 6678 ) and ( 6455.8 & 6456.4 ) which are not detectable until later in the nova s evolution .
we show the evolution of the high velocity outliers in figures [ smarts_hab ] ( blue - shifted ) and [ smarts_har ] ( red - shifted ) .
these outliers could be explained by a shell of high velocity material .
however , the narrowness of the outlier lines would require that the shell had a very small thickness , and such a fast - moving thin shell should become diffuse ( and thus undetectable ) much faster than we see here .
another explanation is a bipolar outflow , where the narrowness of these high velocity outliers would point to a relatively well - collimated outflow .
it would also likely have a fairly small inclination angle for us to detect such large velocities .
such well - collimated bipolar outflows have been seen in several other novae .
examples of embedded novae with bipolar outflows are u sco ( lpine et al .
1999 ) and rs oph ( rupen et al . 2008 ; sokoloski et al .
non - embedded novae are also known to have bipolar outflows , such as v1494 aql ( iijima & esenoglu 2003 ) , v475 sct ( kawabata et al .
2006 ) , v445 pup ( woudt et al . 2009 ) , v5668 sgr ( bannerjee et al .
2016 ) , and possibly t pyx ( chesneau et al .
2011 ) .
outlier over the first 16.5 days of the nova eruption .
the vertical line is at 4550 km s@xmath21 .
note that the peak of the outlier clearly moves to higher velocities through day 6.5 .
also , note that the line at 6678 appears on day 4.5 and dominates this region of the spectrum by day 7.5.,width=307 ] the high velocity outliers in figures [ smarts_exspec ] , [ smarts_hab ] , and [ smarts_har ] are likely to be h@xmath17 .
we see marginal evidence for similar outliers with identical velocities around the h@xmath16 line . were these the [ ] lines , the 6482 line would be blueshifted by approximately 1150 km s@xmath21 , while the 6668 line would be redshifted by only about 300 km s@xmath21 .
low velocity and lines appear in these regions around day 5 and strengthen as the ejecta evolve .
lacking other plausible identifications , we identify the outliers as h@xmath17 ejected at high velocities . using the h@xmath17 line width as the approximate velocity evolution for the optically - thick component of the ejecta
, we can use @xmath22 to determine the approximate radial size of the ejecta at any given time @xmath23 .
we perform this integration numerically using the trapezoidal approximation . first , we interpolate our velocity measurements to give a smoother function to integrate over .
we assume that the velocity is constant between day 0 and our first spectroscopic measurement .
this is likely an oversimplification , although it may not be drastically far from the average velocity over this time , as the ejecta must first be accelerated to some ( unknown ) maximum velocity , then decelerate to the velocity we measured with our first spectroscopic observation .
our resulting radial size of the h@xmath17 emitting region is shown in the bottom panel of figure [ sco15_vnr ] .
the radial size of the h@xmath17 emitting region can serve as a minimum size of the ejecta .
a more realistic approximation of the radial size of the ejecta is to assume that the ejecta expand at roughly the velocity from the first observation : @xmath01659 km s@xmath21 .
however , neither of these simple models account for the high velocity outliers seen in the early spectra .
ccccc 1.565 & 47.5 & 1302 & 60.5 & 1659 + 2.491 & 37.3 & 1023 & 47.5 & 1304 + 4.493 & 36.0 & 987 & 45.8 & 1257 + 5.495 & 36.7 & 1006 & 46.7 & 1282 + 6.495 & 34.1 & 936 & 43.5 & 1193 + 7.484 & 32.8 & 899 & 41.8 & 1145 + 10.478 & 26.0 & 713 & 33.1 & 908 + 11.505 & 20.9 & 574 & 26.7 & 731 + 13.528 & 14.5 & 397 & 18.4 & 505 + 14.505 & 13.1 & 359 & 16.7 & 458 + 15.483 & 11.3 & 311 & 14.4 & 397 + 16.548 & 10.5 & 288 & 13.4 & 366 + 17.482 & 9.2 & 252 & 11.7 & 322 + 18.456 & 8.6 & 237 & 11.0 & 302 + 19.525 & 8.2 & 225 & 10.5 & 287 + 20.53 & 7.7 & 212 & 9.9 & 271 + 21.498 & 7.3 & 200 & 9.3 & 255 + 22.472 & 6.8 & 186 & 8.6 & 237 + 23.451 & 6.2 & 170 & 7.9 & 217 + 25.477 & 5.6 & 154 & 7.2 & 196 + 26.445 & 5.2 & 141 & 6.6 & 180 + 27.507 & 4.7 & 129 & 6.0 & 164 + 28.462 & 4.4 & 121 & 5.6 & 154 + 29.435 & 4.1 & 112 & 5.2 & 143 + 31.469 & 3.7 & 101 & 4.7 & 129 + 32.432 & 3.5 & 96 & 4.5 & 122 + 33.416 & 3.3 & 91 & 4.2 & 116 + 34.432 & 3.1 & 85 & 3.9 & 108 + 35.44 & 3.1 & 86 & 4.0 & 110 + 36.42 & 3.1 & 84 & 3.9 & 107 + 37.432 & 3.0 & 83 & 3.9 & 106 + 38.413 & 3.0 & 82 & 3.8 & 105 + 39.398 & 2.9 & 79 & 3.7 & 101 + 50.435 & 2.7 & 75 & 3.5 & 96 + 51.468 & 2.7 & 74 & 3.5 & 95 + 52.412 & 2.7 & 74 & 3.5 & 95 + 53.364 & 2.7 & 74 & 3.4 & 94 + 54.463 & 2.7 & 73 & 3.4 & 93 + 54.505 & 2.6 & 72 & 3.4 & 92 + 55.389 & 2.6 & 72 & 3.4 & 92 + 56.447 & 2.7 & 73 & 3.4 & 93 + 57.438 & 2.7 & 73 & 3.4 & 93 + 58.371 & 2.7 & 74 & 3.4 & 94 + 59.398 & 2.6 & 72 & 3.3 & 91 + 60.343 & 2.5 & 69 & 3.2 & 88 + 61.382 & 2.5 & 69 & 3.2 & 88 + 62.381 & 2.5 & 69 & 3.2 & 88 + 64.495 & 2.5 & 69 & 3.2 & 88 + 65.371 & 2.5 & 69 & 3.2 & 87 + 66.37 & 2.5 & 67 & 3.1 & 86 + 67.413 & 2.4 & 67 & 3.1 & 85 + 68.407 & 2.4 & 66 & 3.1 & 84 + 69.473 & 2.4 & 64 & 3.0 & 82 + 70.434 & 2.3 & 64 & 3.0 & 82 + 71.4 & 2.3 & 64 & 3.0 & 81 + 72.422 & 2.3 & 64 & 3.0 & 81 + 74.341 & 2.3 & 63 & 2.9 & 80 + 76.393 & 2.3 & 63 & 2.9 & 80 + 78.453 & 2.4 & 66 & 3.0 & 84 + 80.477 & 2.2 & 61 & 2.8 & 78 + 82.36 & 2.3 & 62 & 2.9 & 79 + 86.377 & 2.3 & 63 & 2.9 & 80 + 88.287 & 2.3 & 64 & 3.0 & 81 + 91.302 & 2.3 & 63 & 2.9 & 80 + 93.322 & 2.3 & 62 & 2.9 & 79 + 98.299 & 2.5 & 68 & 3.2 & 87 + 100.274 & 2.4 & 66 & 3.1 & 84 + 103.232 & 2.4 & 66 & 3.0 & 84 + 105.25 & 2.3 & 64 & 3.0 & 82 + 109.298 & 2.4 & 66 & 3.0 & 84 + 111.376 & 2.4 & 66 & 3.1 & 84 + 116.206 & 2.4 & 66 & 3.1 & 84 + 119.308 & 2.4 & 66 & 3.0 & 84 + 120.457 & 2.4 & 65 & 3.0 & 83 + line width of v1535 sco .
the dashed blue line shows the interpolation of the data used to estimate the radius of the h@xmath17 emitting region . *
bottom panel : * integrated radial size of the h@xmath17 emitting region.,width=336 ] ccccc 1.565 & 6458.25 & 4778 & 6658.67 & 4377 + 2.491 & 6458.09 & 4785 & 6660.00 & 4414 + 4.493 & 6456.03 & 4880 & 6665.90 & 4707 + 5.495 & 6455.45 & 4906 & 6668.08 & 4807 + 6.495 & 6455.5 & 4904 & 6669.98 & 4893 + 7.484 & 6456.3 & 4867 & 6670.75 & 4929 + 10.478 & 6456.52 & 4857 & 6671.3 & 4954 + 11.505 & 6456.91 & 4839 & & +
the distance to v1535 sco is highly uncertain .
srivastava et al .
( 2015 ) apply the maximum magnitude rate of decline ( mmrd ) relations from della valle & livio ( 1995 ) and downes & duerbeck ( 2000 ) to get distance estimates of @xmath24 and @xmath25 kpc , respectively . however , munari et al .
( 2017 ) reported a distance to v1535 sco of approximately 9.7 kpc , despite using the same mmrd method as srivastava et al .
this indicates the mmrd distance may be highly dependent on the data used , as munari et al .
( 2017 ) performed all their observations on a single telescope and srivastava et al .
( 2015 ) obtained their data from multiple telescopes via the american association of variable star observers ( aavso ) . furthermore , substantial uncertainty has been shed on the mmrd method in general by the work of kasliwal et al .
( 2011 ) and cao et al .
the srivastava et al .
mmrd distances seem unlikely as they place the nova on the opposite side of the galactic bulge , which should make it extremely reddened and obscured due to the intervening material .
we were able to constrain the distance to v1535 sco by using pre - outburst photometry to determine a spectral type for the companion .
the source 2mass1703261 - 350417 is within @xmath26 " of v1535 sco , and has @xmath27 , @xmath28 , and @xmath29 . without correcting for reddening
, we find near - ir colors of @xmath30 and @xmath31 . using the derived reddening from srivastava et al .
( 2015 ) of @xmath32 we can compute the necessary color corrections .
the reddening conversion functions of schlafly & finkbeiner ( 2011 ) give @xmath33 , @xmath34 and @xmath35 . using these
, we find the dereddened colors to be @xmath36 , @xmath37 , and @xmath38 , where the prime signifies that it is dereddened . using the spectral class system of covey et al.(2007 ) , the colors are consistent with a spectral class from k3 iii to m0 iii which
, in turn , give an absolute @xmath39 magnitude between @xmath40 to @xmath41 .
this gives a distance modulus of @xmath42 , or a distance between 6.2 and 19.6 kpc .
if we instead use @xmath43 , as reported in munari et al .
( 2017 ) , our distance range is 5.8 to 14.0 kpc . note that , although we can not definitively rule out the companion being a main sequence star using this method , if it were a main sequence progenitor system the distance modulus would be @xmath44 mags . since the peak brightness of the nova was @xmath45 , the absolute magnitude at peak would be @xmath46 , which is far too dim for a nova .
therefore , we can reasonably rule out a main sequence companion .
we therefore find it more likely that the companion star is an evolved star in the range between subgiant and giant .
because the progenitor distance is consistent with being located at the galactic center , and that is also where we expect most novae to occur , we will assume a distance of @xmath47 kpc for v1535 sco .
this distance agrees well with estimates in munari et al .
( 2017 ) based on both mmrd and the buscombe & de vaucouleurs ( 1955 ) method of estimating the absolute magnitude 15 days after optical maximum .
the full vla light curves for v1535 sco are presented in figure [ vlalc ] .
the nova was initially detected at relatively high flux densities , faded substantially , and then rose again to secondary peak around day 25 .
this secondary peak is unusual for novae , but not unprecedented ( e.g. , taylor et al .
1987 ; krauss et al . 2011 ; eyres et al . 2009 ; weston et al .
2016 ; finzell et al .
after day 27 , the source faded again until about day 56 when there is a tertiary peak .
this is highly unusual .
after the tertiary peak , the nova fades until it is no longer detectable .
the radio light curve for v1535 sco is significantly different from the radio light curve of a non - embedded nova . in a `` typical '' non - embedded nova ,
the radio emission is delayed from the optical emission by up to @xmath02 weeks , and then slowly rises to a maximum over several months .
this is due to the fact that non - embedded novae are typically _ thermal _ sources at radio wavelengths .
therefore , their flux density is directly related to the size of the ejecta . as the ejecta expand , the size of the emitting surface area increases and so does the flux density ( e.g. , hjellming 1996 ; seaquist & bode 2008 ) .
in contrast , the radio light curves for embedded novae are dominated by synchrotron radiation in the first @xmath0 weeks , and therefore have high flux densities early and fade as the ejecta cool ( e.g. , obrien et al . 2006 ) . .
non - detections are indicated with a downward - pointing triangle.,width=336 ] we measured the spectral index @xmath17 ( using @xmath48 ) for each of our vla observations .
we fit the flux densities to a power law using the non - linear least squares ` curve_fit ` function in the _ scipy _ package of python , weighted by the 1@xmath18 uncertainties .
we ignored upper limit values for this fit , using only solid detections .
if two observations were separated by less than 24 hours , they were combined to make a single spectral index measurement .
our calculated spectral indices are presented in table [ vlasitab ] , including the associated 1@xmath18 uncertainties from the variance - covariance matrix .
the evolution of the spectral index is shown graphically in figure [ vlasi ] . during the first week of its evolution ,
the nova s radio spectral index rises toward lower frequencies , an indicator of optically thin synchrotron emission .
the spectral index then switches to rising toward higher frequencies .
this is typical of sources emitting via optically thick thermal bremsstrahlung . to our knowledge ,
such a dramatic switch between a synchrotron - like @xmath17 to a thermal - like @xmath17 has never been seen before . however , the spectral index of v1324 sco showed evidence of being negative at early times followed by a positive slope later ( finzell et al .
2017 ) , and eyres et al .
( 2009 ) reported the presence of both positive and negative spectral indices present simultaneously for rs oph . at the tertiary peak around day 56
, the spectral index again rises toward lower frequencies and appears to be optically thin synchrotron . this apparent switch from a thermal - like to synchrotron - like spectral index is unprecedented . as the nova fades , its spectral index flattens as expected for an optically thin thermal bremsstrahlung source .
the evolution of v1535 sco s radio spectral index is very different from the majority of classical novae .
typically , a the radio emission begins with a positive ( rising toward higher frequencies ) spectral index while the ejecta is optically thick .
as the ejecta becomes optically thin , the radio spectral index flattens to value around -0.1 .
this transition from strongly positive to flat begins with the highest frequencies first and progresses through the lower frequencies as the ejecta dims overall . even in novae where shocks are detected such as v959 mon ,
the spectral index at early times is still usually positive ( chomiuk et al . 2014 ) .
a negative spectral index , especially during the first few weeks of a nova s evolution , usually occurs in systems with evolved companions producing a strong wind for the nova ejecta to shock against , as in rs oph ( e.g. , eyres et al . 2009 ) .
cccc 2015 - 02 - 14 & 2.663 & 4.536.5 & [email protected] + 2015 - 02 - 18 & 6.663 & 13.536.5 & [email protected] + 2015 - 02 - 19 & 7.663 & 1.267.38 & [email protected] + 2015 - 02 - 24 & 12.762 & 13.536.5 & [email protected] + 2015 - 03 - 01 & 17.663 & 4.557.38 & [email protected] + 2015 - 03 - 07 & 23.663 & 13.536.5 & [email protected] + 2015 - 03 - 10 & 26.563 & 4.557.38 & [email protected] + 2015 - 03 - 25 & 41.583 & 41.633 & 4.5528.2 & [email protected] + 2015 - 04 - 07 & 54.563 & 13.536.5 &
[email protected] + 2015 - 04 - 08 & 55.733 & 1.267.38 & [email protected] + 2015 - 04 - 18 & 19 & 65.513 & 66.493 & 1.2628.2 & [email protected] + 2015 - 05 - 01 & 78.463 & 78.503 & 4.5528.2 & [email protected] + 2015 - 05 - 16 & 93.463 & 4.5513.5 & [email protected] + our vlba detection on 2015 february 19 ( figure [ vlba1 ] ) was simultaneous with vla l - band and c - band observations .
using difmap , we modelled the vlba detection as a circular gaussian in the _ uv_-plane .
the resulting total flux density is 0.477 mjy , with an off - source image rms of 0.044 mjy / beam , and a diameter of 6.52 mas .
the resolving beam for this observation was 2.5 x 12.4 mas , so this could indicate a resolved source .
the brightness temperature ( t@xmath49 ) in k is given by the rayleigh - jeans relation ( e.g. , rohlfs & wilson 2006 ) @xmath50 where @xmath51 is the flux density in erg s@xmath21 @xmath52 hz@xmath21 at frequency @xmath53 , @xmath54 is the observing wavelength in cm , @xmath55 is the boltzmann constant in erg k@xmath21 , and @xmath56 is the solid angle of the source . for a resolved source with a circular gaussian shape , @xmath57 , where @xmath58 is the angular diameter of the source in radians .
therefore , the brightness temperature is calculated using @xmath59 for our source .
this gives a brightness temperature of 5.8@xmath6010@xmath61 k. this is significantly lower than the 10@xmath62 k @xmath63 of the synchrotron components in rs oph ( rupen et al .
2008 ) . however , we emphasize that our observation was made with lower resolution due to the drop - out of the mauna kea antenna .
it is possible that our single structure is actually composed of 2 or more components that are too compact to resolve with our configuration .
modelling the source as two point sources results in a similar total flux density ( 0.445 mjy ) and image rms ( 0.04 mjy / beam ) with the same reduced @xmath64 ( 1.09 ) as the circular gaussian model .
if we make the assumption that the angular size of the components is equal to the minor axis of the restoring beam ( 2.5 mas ) , the resulting @xmath63 estimates are 1.8@xmath6010@xmath65 k and 1.9@xmath6010@xmath65 k. these are still lower than the @xmath63 for rs oph , but are consistent with the @xmath63 for the compact components in v959 mon ( chomiuk et al .
2014 ) .
the simultaneous vla flux density at 4.55 ghz was 0.650 mjy . because the vlba is only sensitive to high brightness temperature ( i.e. , compact ) emission , and the spectral index from the vla observations of the nova ejecta was negative ( i.e. , larger flux density at lower frequencies ) ,
we conclude that the vlba emission is most likely from synchrotron radiation .
the compact emission on this day dominated the total emission at the 73% level .
the excess vla radio flux on this day may be due to thermal bremsstrahlung emission from the shocked plasma or the pre - shocked , cooler outflow . using the vlba flux density of the single circular gaussian model component , we estimate the magnetic field strength using the revised equipartition formula from beck & krause ( 2005 ) .
this formula requires knowledge of the path length through which the radiation propagates @xmath66 , the filling factor of the emitting material @xmath67 , the inclination angle of the field @xmath68 , and the proton - to - electron number density ratio , @xmath69 : @xmath70 where @xmath17 is the radio spectral index ( @xmath48 ) . to make our estimate
, we will assume a filling factor of 1 , and an inclination angle of 0 ( i.e. , the field is face - on ) .
for @xmath66 , we assume the emitting region was relatively thin , @xmath71 of the radius of the h@xmath17 emitting region ( see figure [ sco15_vnr ] ) at the time of the observation , which gives us @xmath72 cm . we used the typical values for @xmath69 of 40 and 100 .
the resulting magnetic field strengths were 0.14 g for @xmath73 and 0.18 g for @xmath74 .
if we assume that the path length is @xmath71 of the radius of an ejecta shell expanding at a constant velocity of 1659 km s@xmath21 ( the maximum velocity measured from the h@xmath17 linewidth ) , we get @xmath75 cm and a magnetic field of 0.13 g for @xmath73 and 0.17 g for @xmath74 . if we assume that the path length is @xmath71 of the radius of an eject expanding at a constant velocity of 4782 km s@xmath21 ( the median velocity from the high velocity outliers in the h@xmath17 spectra over the first 5.5 days ) , we get @xmath76 cm and a magnetic field of 0.10 g for @xmath73 and 0.13 g for @xmath74 .
these values are all larger than , but still comparable to , the magnetic field strengths estimated for rs oph of @xmath77 ( rupen et al .
assuming the same values of @xmath69 as we used , the rs oph magnetic field strength was between 0.087 g and 0.11 g. it should also be noted that rupen et al .
( 2008 ) estimated the magnetic field for rs oph over 20 days after the eruption , whereas we are estimating the field for v1535 sco only 7.7 days after eruption .
it is possible that the magnetic field declines over time as the ejecta expand and the shocks dissipate .
the optical and near - infrared light curve for v1535 sco is not consistent with light curves for other novae known to have red giant companions ( see figure [ smarts_comp ] ) . in particular , when one considers the difference between the peak magnitude and the current magnitude ( right hand side of figure [ smarts_comp ] ) , v1535 sco is clearly an outlier . in other novae with red giant companions ,
the companion star begins to dominate the difference light curve at longer wavelengths relatively early , while the shorter wavelengths are still dominated by the nova ejecta .
v1535 sco , on the other hand , does not follow this behavior .
instead , all wavelengths fade together for the duration of the light curve .
this indicates that the companion star is unlikely to be a red giant .
however , the presence of synchrotron emission in the radio ( sections 4.1 and 4.2 ) , the hard x - ray emission , and the rapid fading of the optical light curve all indicate the nova ejecta were shocking against an external medium . and @xmath78 fluxes.,width=312 ]
one possible explanation is that the companion star is not an m giant , but a k giant .
we tested this hypothesis by fitting our optical and near infrared fluxes with a model that combined a hot accretion disk and a stellar spectral energy distribution for a k giant star .
the quiescent unreddened optical colors of v1535 sco are close to 0 .
we modelled the optical - infrared ( oir ) spectral energy distribution ( sed ) as the sum of a mid - k giant plus an active accretion disk .
the accretion disk is constructed following the bertout , basri , & bouvier ( 1988 ) formalism .
the optically thick disk is constructed of a series of annuli , each of which emits as a black body at a temperature set by its distance from the star .
we set the inner edge of the accretion disk to the radius of a white dwarf ; exact details are not important since the boundary layer and the inner edge of the disk emit in the ultraviolet . on the red side of the peak ,
the accretion disk follows a power law f@xmath79 . with 7 oir fluxes ,
the model is under - constrained .
we fix the extinction a@xmath80 to be 2.16 mag and set the distance to 8.5 kpc .
we constrain the white dwarf to be hot .
we constrain the donor star to be k5 iii , with vjhk colors taken from koornneef ( 1983 ) .
free parameters in the fit are the magnitude of the donor star and the mass accretion rate .
an example of a fit is shown in figure [ smarts_sedfit ] .
we fit the ten seds obtained 2016 april 25 through 2017 april 02 . over that year
the donor star @xmath81 faded by [email protected] mag .
this fading could be a cooling after irradiation by the nova , or it could be part of a longer term variation .
the mean fit @xmath81 of 12.14 is consistent with the 2mass @xmath81 magnitude .
meanwhile the mass accretion rate has dropped linearly by a factor of 6 , from 12 to 2@xmath6010@xmath82m@xmath83 yr@xmath21 .
numbers should be taken with a grain of salt .
the mass accretion rate and the assumed distance are strongly correlated where the inferred mass accretion rate scales as the cube of the inferred distance . using our fits and assuming a distance of 8.5 kpc , we find that the absolute @xmath39 magnitude for v1535 sco in quiescence should be -1.2 .
covey et al .
( 2007 ) gives the absolute @xmath39 magnitude for a k3 iii star as -1.13 , which is remarkably close for such a simple model .
munari et al . (
2017 ) also conclude that the companion in v1535 sco is consistent with a k3 - 4 iii giant .
region of the spectrum as a function of time .
darker regions indicate more emission , brighter regions indicate absorption.,width=336 ] srivastava et al .
( 2015 ) presented measurements of the ejecta velocity for v1535 sco based on the pa@xmath16 emission line at 1.2818 @xmath84 m starting 7 days after the detection and ending 40 days after the detection .
their velocities are given as full width at half maximum ( fwhm ) of the line .
as with our h@xmath17 measurements , we approximate the fw3@xmath18 velocities by applying the same calculation as described in section 2.4 .
we then fit the line - width - vs - time data with a power law using the non - linear least squares ` curve_fit ` technique in the _ scipy _ package of python .
our resulting fit is somewhat different from that given in srivastava et al .
they report a power law of @xmath85 whereas our fit gives @xmath86 .
however , we should note that we are uncertain as to the methods used by srivastava et al .
( 2015 ) because they do not describe their fitting procedure in detail . to compare the pa@xmath16 results with our more
densely sampled h@xmath17 data , we fit our h@xmath17 fw3@xmath18 measurements from day 7.48 to day 39.4 using the same _ scipy _ method and find a power law of @xmath87 , which nearly agrees within uncertainties with the srivastava et al .
( 2015 ) value . of course
, this simple power - law fit completely ignores the complicated behavior observed during the first week following the explosion ( see figure [ sco15_vnr ] ) .
the smarts optical spectroscopic monitoring of v1535 sco also revealed blue - shifted h@xmath17 absorption features .
these features become noticeable around day 15 , after the h@xmath17 line has narrowed significantly ( see figure [ walterha ] ) .
the blue - shifted features have an initial velocity of @xmath88 km s@xmath21 , but slow to @xmath89 km s@xmath21 by day 247 .
these absorption features indicate the presence of a large amount of cool , neutral material ahead of the h@xmath17 emitting region .
it is possible that this neutral material was pre - shocked wind material being swept up by the nova ejecta in a `` snowplow '' fashion .
we fit the _ swift _ data using the ` xspec ` version 12.8.2 package .
the spectra were best fit using a combination of a black body ( bb ) and a thermal plasma ( apec ) .
the apec model includes contributions from free - free continuum and lines ( smith et al . 2001 ) .
only the first 5 detections had high enough signal - to - noise to successfully model the emission .
the results are given in table [ swiftab2 ] .
the hydrogen column density ( n@xmath1 ) appears to be fairly constant throughout the first 18 days , with a deviation on day 11 which is most likely the fitting software getting stuck in a local minimum .
this constant n@xmath1 is quite different from the sharply decreasing n@xmath1 observed in the embedded novae rs oph ( bode et al .
2006 ) and v745 sco ( page et al . 2015 ) .
the black body temperature appears to decrease by only a factor of 2 over the first 25 days , with a minor fluctuation around day 17.9 .
the hot plasma temperature , on the other hand , dramatically cools by at least 2 orders of magnitude during the same time period .
this is likely due to the shocks expanding and becoming radiatively efficient .
the @xmath90 of 0.96 reported by munari et al . ( 2017 ) implies n@xmath91 @xmath52 ( predehl & schmitt 1995 ) .
our values of n@xmath1 are nearly twice this , indicating a local enhancement from a possible stellar wind .
interestingly , our n@xmath1 values are quite similar to those measured for 2 recurrent novae known to have red giant companions : v745 sco and rs oph . in rs oph , the values for n@xmath1 were @xmath92 @xmath52 just 0.16 days after eruption , but fell to @xmath93 @xmath52 by day 3.23 ( page et al .
2015 ) and were @xmath94 @xmath52 by day 10 ( orio et al .
2015 ) . in rs oph , nelson et al .
( 2008 ) reported n@xmath95 @xmath52 14 days after eruption .
we stress that this does not directly indicate the presence or absence of a red giant , especially as the kt values for v1535 sco are similar to those that mukai & ishida ( 2001 ) found for v382 vel , which is known to have a main sequence companion .
however , it is important to note that v1535 sco was detected by _
swift _ only 5 days after its discovery in the optical , while v382 vel was not detected in x - rays until 20 days after optical discovery .
the delay in x - rays in v382 vel is often seen in non - embedded novae , and is thought to be due to the fact that it takes time for multiple outflows to be ejected and then collide
. it may therefore be more appropriate to compare the x - ray emission from v382 vel to the later ( > day 49 ) x - rays from v1535 sco .
overall , the x - ray emissions are somewhat puzzling . on the one hand ,
the flat n@xmath1 for several days is different from the behavior in known embedded novae . on the other hand , there appears to be local absorption in the system on levels similar to known embedded novae .
there are two other major points to consider .
first , the x - ray emission in v1535 sco was hard early in its evolution , which is much more typical of embedded novae .
second , those early x - rays were very bright . using our assumed distance of 8.5 kpc , the early kt of 40 kev , and n@xmath96 @xmath52 , we find an x - ray luminosity of @xmath97 erg s@xmath21 , which is much higher than any non - embedded nova we are aware of . for comparison ,
the x - ray luminosity for v382 vel was @xmath98 erg s@xmath21 ( mukai & ishida 2001 ) , while rs oph as over @xmath99 erg s@xmath21 ( mukai et al .
2008 ) . while the x - ray data do not provide proof that the nova was embedded in wind material from the companion star , it is at least a plausible explanation for the interesting behavior
bringing together all of our various observations and examining them together leads to some interesting conclusions .
figure [ rxofig ] shows information from all of our observing wavelengths with important times ( e.g. , the vlba detection and the second radio synchrotron emission event ) indicated . the initial negative radio spectral index combined with hard x - rays indicates synchrotron radiation produced by strong shocks .
this is expected for a nova exploding into a thick wind from a giant companion .
the transition from a negative to positive radio spectral index corresponds with both a softening of the x - rays while the total x - ray count rate remains relatively constant .
this indicates the shocks are weakening and the plasma is cooling . at the same time , the radio thermal photosphere is expanding and begins to dominate the radio emission .
cccccccc 2 & 4.16 & [email protected] & 67@xmath1411 & 349@xmath100 & > 41 & [email protected] & 373.39/377 + 3 & 11.04 & 3@xmath142 & 45@xmath101 & 0.0037@xmath102 & 3.8@xmath103 & 0.0027@xmath104 & 214.39/121 + 4 & 13.94 & [email protected] & 35@xmath144 & 5.9@xmath105 & [email protected] & 0.0037@xmath106 & 151.98/159 + 5 & 17.93 & [email protected] & 45@xmath107 & 0.74@xmath108 & 1.013@xmath109 & 0.0018@xmath110 & 203.73/138 + 6 & 24.75 & 18.4@xmath111 & 31@xmath112 & 4264@xmath113 & 0.27@xmath114 & 0.026@xmath115 & 80.10/88 + the ejecta cool and fade for several weeks until around day 49 when the x - ray count rate appears to increase slightly ( going from a non - detection to detection again ) and the radio spectral index again shows evidence of synchrotron emission .
there are two possible explanations for this second episode of synchrotron emission .
first , the ejecta encountered another dense medium to shock against . using the fit to the velocity from h@xmath17 emission lines and integrating over time
, the h@xmath17 emitting region had a radius of approximately 13 au by this time .
if we assume that the ejecta that were emitting synchrotron radiation at this time was instead travelling at a velocity of 1659 km s@xmath21 for 49 days , that would be at a radius of approximately 47 au .
if we assume that the synchrotron - emitting material was from the ejecta producing the high velocity outliers in the h@xmath17 spectra and travelled with a constant velocity of 4782 km @xmath21 , it would have travelled 135 au by day 49 .
it is possible that the ejecta caught up to material ejected during a previous nova eruption .
if we assume that the slowest material from a previous ejecta were moving at 50 km @xmath21 , the previous eruption would have occurred 1.2 years ago for a distance of 13 au , 4.5 years ago for a distance of 47 au , or 12.8 years ago for a distance of 137 au .
of these three , the 12.8 years seems the most likely , as it is hard to imagine the astronomy community could have missed multiple outbursts of this nova in the past 10 years .
however , a recurrence time of 12.8 years would still require that this system is a very young recurrent nova undergoing only its second or third outburst , or that we have missed a very large number of its previous eruptions .
the second possibility is that the second synchrotron event is the result of shocks within the ejecta itself .
we already have evidence for two different outflows from the width of the h@xmath17 line and the high velocity outliers ( see section 2.4 ) .
it is not hard to imagine an interaction between these two outflows , or possibly even a third outflow such as a hot , fast wind driven by thermonuclear burning on the surface of the white dwarf but hidden below the optical photosphere of the bulk of the ejecta . using the radial size
, we can estimate what we expect the brightness temperature of the ejecta to be at the time of the vlba observation using @xmath116 where @xmath117 is the radius of the spherically expanding ejecta and @xmath118 is the distance to the source ( e.g. , seaquist & bode 2008 ) .
we will start by assuming the size of the radio emitting region is the same as the h@xmath17 emitting region . from our fit to the h@xmath17 line widths presented in section 2.4 , we derive a radius for the h@xmath17 emitting region of @xmath119 cm on day 7.663 . using only the vlba flux of 0.477 mjy , and assuming a distance of 8.5 kpc
, this gives @xmath120 k. alternatively , the radius could be the distance travelled by ejecta moving with a velocity of 1659 km s@xmath21 , which gives is @xmath121 cm . using the vlba flux and a distance of 8.5 kpc ,
this material would have a @xmath122 k. both of these are significantly higher than the estimate of @xmath123 k using the distance - free formula for @xmath63 in section 4.2 . to get the largest possible radius
, we could assume the ejecta were travelling at 4782 km s@xmath21 for 7.663 days , giving a radius of @xmath124 cm and @xmath125 k. the discrepancy between the observational and expected @xmath63 can indicate several possibilities : first , the ejecta were unresolved by the vlba ; second , the nova is much closer than 8.5 kpc ; third , the ejecta shape is highly non - spherical ; or some combination of these reasons .
we should note that the expected @xmath63 using the largest possible radius agrees well with the observational @xmath63 assuming 2 point sources .
however , using the largest possible radius would only give us a single sphere with an angular size of 2.5 mas at 8.5 kpc ( which just happens to be the minor axis of the vlba restoring beam for our observation ) , not the 2 point sources modelled in section 4.2 , so we can not reconcile the measurement with the observation quite so easily . from section 3 ,
recall that the closest the nova can be is 5.8 kpc . at this distance
, the brightness temperature would be @xmath126 k for a radius of @xmath119 cm , @xmath127 k for a radius of @xmath121 cm , or @xmath128 k for a radius of @xmath124 cm .
the first two are still larger than our value of @xmath129 k we calculated for the single - component model in section 4.2 , although they are comparable to the two - component model .
the final one , using our maximum possible radius , agrees well with the single - component model @xmath63 in section 4.2 .
however , recall that we find it more likely that the high velocity outliers in the h@xmath17 spectra represent a bipolar outflow , not an expanding spherical shell . note that this approach to estimating what the brightness temperature should be at the time of the vlba detection assumes that the entire surface of the ejecta shell is emitting .
if the emission is coming from compact , jet - like structures as in rs oph ( rupen et al .
2008 ; sokoloski et al .
2008 ) , the value for @xmath117 would be smaller and the estimate for @xmath63 would be even larger . placing
the nova at further distances also increases the estimate for @xmath63 . because placing the nova closer does not make the values of @xmath63 match
, it is more likely that the vlba did not fully resolve the compact emitting region on day 7.663 . because of the highly - elliptical shape of the restoring beam , it is especially likely that there is unresolved structure in the north - south direction .
however , the image does look like there is some resolved structure in the east - west direction .
we find it likely that the compact component is non - spherical , possibly bi - lobed , with more extension in the east - west direction .
we used the emission measures for both radio and x - ray to investigate whether the secondary maximum in the radio light curve that occurs between day 23.663 and 26.563 can be explained by thermal radio emission from the x - ray emitting plasma .
for the x - rays , the emission measure depends on the volume of the emitting region : @xmath130 where @xmath131 and @xmath132 are the number densities of ions and electrons , respectively , and @xmath133 is the volume of emitting material . generally it is assumed that @xmath134 so that it simplifies to @xmath135 . to make a further simplification ,
the emitting volume is assumed to be thin compared to the total extent of the ejecta so that it can be approximated as a uniform slab : @xmath136 , where @xmath137 is the radius of the ejecta ( we are assuming the emitting region can be approximated as a cylinder ) and @xmath138 is the thickness of the emitting region . the x - ray emission measure can be determined from the hot plasma parameters in the fits to the _ swift _ data : @xmath139 where @xmath118 is the distance to the nova . assuming a distance of 8.5 kpc ( placing the nova approximately at the galactic center ) and using @xmath140 from day 24.75 ( see table [ swiftab2 ] )
gives @xmath141 @xmath142 .
assuming the ejecta travel with a constant velocity of 1659 km s@xmath21 , we can determine the approximate radius for the shell on day 24.75 was @xmath143 cm .
if we assume that the thickness of the emitting shell is @xmath71 of its radius , we can solve for the number density of the emitting electrons : @xmath144 @xmath142 emitting at a temperature of approximately @xmath145 k ( from kt@xmath146 in table [ swiftab2 ] ) .
note that if we assume a larger distance to the nova , both @xmath147 and @xmath132 increase .
the radio ( path length ) emission measure is given by : @xmath148 plugging in the same @xmath138 ( but this time in pc ) used for @xmath147 and using @xmath132 derived from the x - ray data gives @xmath149 @xmath150 pc .
the radio emission measure is also part of the equation for optical depth , @xmath151 ( rohlfs & wilson 2006 ) : @xmath152 therefore , another way to calculate the radio emission measure is given by : @xmath153 where @xmath154 is the electron temperature in k , @xmath53 is the observing frequency in ghz , @xmath151 is the optical depth at the observing frequency , and @xmath155 is a correction term usually assumed to be 1 . in order for the emission to be optically thick at the observed frequency , @xmath156 .
because we are trying to determine if the radio flux can be explained by the x - ray plasma , we set @xmath157 k from the kt@xmath146 value from the fits to the swift data on day 24.75 ( see table [ swiftab2 ] ) .
we use our highest observing frequency of 36.5 ghz because it provides the tightest constraint .
this gives us @xmath158 @xmath150 pc , which is more than two orders of magnitude larger than expected from the x - ray electron number density used above .
therefore , under the reasonable assumptions we have made , the x - ray emitting plasma can not account for the thermal radio emission detected around day 24 .
it is more likely that there is also a warm ( @xmath159k ) ionized ejecta that begins to dominate the radio emission at this time .
unfortunately , the uncertainty on @xmath140 from day 24.75 is very high .
if we use the @xmath140 value from day 17.93 with its much lower uncertainty and assume that the number density remains mostly constant until day 24 , we get @xmath160 @xmath142 , which would imply @xmath161 @xmath150 pc .
this makes the discrepancy between the two methods for determining the radio emission measure even greater , further strengthening our argument that the x - ray emitting plasma can not account for the radio flux density observed during the second radio maximum .
v1535 sco showed peculiar behavior at nearly every wavelength we observed .
the radio emission started bright , faded quickly , and then had two re - brightening events .
the spectral index for early radio emission and the second radio re - brightening event were consistent with optically thin synchrotron emission .
the x - ray emission began promptly with a hard spectrum .
there was a re - brightening in the x - rays which appeared to correspond to the second re - brightening event in the radio .
the lack of connection between the first radio re - brightening event and the x - ray emission implies that are are at least two emitting components in the ejecta : one shock - heated plasma and one thermal bremsstrahlung .
the optical observations indicate that the nova was discovered post - peak , and faded very fast .
optical spectroscopy also indicated the presence of two outflows : 1 ) a relatively slow ( @xmath01659 km s@xmath21 ) outflow ; and 2 ) a fast ( @xmath04782 km s@xmath21 ) , possibly bipolar outflow .
spectral monitoring of the h@xmath17 line at late times also indicated the presence of a neutral , dense surrounding cloud of emitting material .
the early hard x - rays combined with the radio synchrotron emission and the detection with the vlba strongly support the existence of strong shocks very early in the evolution of the nova .
such early strong shocks are most easily explained by the presence of a dense wind from the companion star .
the estimated magnetic field from the vlba detection was between 0.10 g and 0.18 g , which is larger than but comparable to the magnetic field strength found by rupen et al .
( 2008 ) for rs oph , a system known to have a red giant companion .
there was evidence for a second shock around day 50 .
the radio spectral index at this time changed from being consistent with optically thick thermal bremsstrahlung , to being synchrotron - like .
the x - ray emission also showed an increase at this time .
we posit that this second shock was the result of collisions between multiple outflows within the ejecta , but the presence of a shell of dense material ( possibly from a previous eruption ) has not been completely ruled out .
the nova had strong hard x - ray emission early in its evolution , but no detectable @xmath2-rays .
however , it should be noted that v1324 sco is the only nova _ fermi _ has detected which has distance comparable to v1535 sco ( finzell et al .
we also note that v745 sco , with its red giant companion and presumably denser wind , was only marginally detected by _
( cheung et al . 2014 ) .
it is very likely that v1535 sco produced at least some @xmath2-ray emission , but it was simply too far away for _ fermi _ to detect . to date
, only a handful of galactic novae with red giant companions are known , including the recurrent novae rs oph and v745 sco .
however , recent studies on the nova population in m31 indicate that there may be many novae with red giants that are not detected , and they may constitute @xmath030% of the nova eruptions in the milky way ( williams et al .
there is particular interest in discovering more of these novae because they are possible progenitors to type ia supernovae ( e.g. , starrfield et al .
in fact , dilday et al .
( 2012 ) claim that the supernova ptf 11kx originated from such a system .
novae that occur in a system with a red giant companion are often referred to as `` symbiotic novae '' .
however , the term `` symbiotic novae '' is also used to describe a particularly slow and long - lasting class of novae where nuclear burning on the surface of the white dwarf is sustained for several years , and sometimes decades such as pu vul , hm sge , and ag peg ( e.g. , iben jr . & fujimoto 2008 ) . in order to avoid confusion and better
describe the general class of thermonuclear novae with evolved companions , our collaboration has adopted the term `` embedded nova '' to refer to any nova that is embedded in the dense wind of its post - main sequence companion star , regardless of the duration of its optical maximum ( e.g. chomiuk et al .
2012 ; mukai et al .
2014 ) .
we find some evidence that v1535 sco is an embedded nova .
the early strong , hard x - rays and non - thermal radio emission argue for the presence of a pre - existing dense material to shock against .
also , the x - ray emission showed signs of absorption beyond what is expected for the measured @xmath90 , which also points to the nova being embedded in some pre - existing cloud .
the fast optical decline is also consistent with an embedded nova .
however , the long - term behavior of the optical and near - infrared light curves were unlike other novae known to have red giant companions .
we also did not observe the rapid decrease in n@xmath1 that previous x - ray observations of novae with known red giant companions have reported .
we produced a model of the quiescent system as a k giant with an accretion disk which is in good agreement with the observed optical and infrared magnitudes .
we therefore find that the companion star in this system is producing a significant stellar wind and is most likely a k giant , specifically a k3 - 5 iii , which is in agreement with munari et al .
several symbiotic binary systems with k giant stars are known , and they are often referred to as yellow symbiotics " ( e.g. , baella et al .
2016 ) .
we thank the anonymous referee for their thoughtful criticism of this manuscript which led to significant improvements .
we are grateful to c. c. cheung for his insight into the operations of the _ fermi gamma - ray space telescope_. we also appreciate the useful conversations about novae , shocks , and accretion at the stellar remnants at the junction meeting and the conference on shocks and particle acceleration in novae and supernovae .
we thank the nrao for the generous allocation of vla and vlba time for our observations .
we also thank neil gehrels and the _ swift _ mission for the generous allocation of target - of - opportunity time to observe v1535 sco .
we acknowledge support from nasa award nnx14aq36 g .
t. n. was supported in part by nasa award nnx13a091 g .
j. l. s. and j. h. s. w. were funded in part by nsf award ast-1211778 .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
access to smarts has been made possible by generous support from the office of the provost of stony brook university .
this research made use of _ scipy _ , an open - source library of numerical routines for scientific computing available at http://www.scipy.org ; _ aplpy _ , an open - source plotting package for python hosted at http://aplpy.github.com ; and _ astropy _ , a community - developed core python package for astronomy ( astropy collaboration , 2013 ) available at http://www.astropy.org .
many of the figures in this manuscript were made using the matplotlib 2d graphics package for python .
+ abdo , a. a. , ackermann , m. , ajello , m. , et al .
2010 , science , 329 , 817 acero , f. , ackermann , m. , ajello , m. et al .
2015 , , 218 , 23 ackermann , m. , ajello , m. , albert , a. , et al .
2014 , science , 345 , 554 baella , n. o. , pereira , c. b. , miranda , l. f. , & alvarez - candal , a. 2016 , , 151 , 100 banerjee , d. p. k. , srivastava , m. k. , ashok , n. m. , & venkataraman , v. 2016 , , 455 , l109 beck , r. , & krause , m. 2005 , an , 326 , 414 bertout , c. , basri , g. , & bouvier , j. 1988 , , 330 , 350 bode , m. f. , obrien , t. j. , osborne , j. p. , et al .
2006 , , 652 , 629 buscombe , w. , & de vaucouleurs , g. 1955 , obs , 75 , 170 cao , y. , kasliwal , m. m. , neill , j. d. , et al .
2012 , , 752 , 133 chesneau , o. , meilland , a. , banerjee , d. p. k. , et al .
2011 , , 534 , l11 cheung , c. c. , jean , p. , shore , s. n. , et al .
2014 , the astronomer s telegram , 5879 chomiuk , l. , krauss , m. i. , rupen , m. p. , et al .
2012 , , 761 , 173 chomiuk , l. , linford , j. d. , yang , j. , et al .
2014 , nature , 514 , 339 covey , k. r. , ivezi , . ,
schlegel , d. , et al .
2007 , , 134 , 2398 della valle , m. , & livio , m. 1995 , , 452 , 704 dilday , b. , howell , d. a. , cenko , s. b. , et al .
2012 , science , 337 , 942 downes , r. a. , & duerbeck , h. w. 2000 , , 120 , 2007 eyres , s. p. s. , o - brien , t. j. o. , beswick , r. , et al .
2009 , , 395 , 1533 finzell , t. , chomiuk , l. , munari , u. , & walter , f. m. 2015 , , 809 , 160 finzell , t. , chomiuk , l. , metzger , b. d. , et al . 2017 , ( submitted ) , arxiv:1701.03094 hjellming , r. a. 1996 , in asp conf .
93 , radio emission from the stars and the sun , ed .
a. r. taylor & j. m. paredes ( san francisco , ca : asp ) , 174 iben jr . ,
i. , & fujimoto , m. y. 2008 , in classical novae , ed . m. f. bode & a. evans ( 2nd ed . ; cambridge : cambridge univ . press ) , 34 iijima , t. , & esenoglu , h. h. 2003 , , 404 , 997 kasliwal , m. m. , cenko , s. b. , kulkarni , s. r. , ofek , e. o. , quimby , r. , & rau , a. 2011 , , 735 , 94 kawabata , k. s. , ohyama , y. , ebizuka , n. , et al .
2006 , , 132 , 433 koornneef , j. 1983 , , 128 , 84 krauss , m. i. , chomiuk , l. , sokoloski , j. l. , et al .
2011 , , 739 , 6 lpine , s. , shara , m. , livio , m. , & zurek , d. 1999 , , 522 , l121 linford , j. , nelson , t. , chomiuk , l. , et al .
2015 , the astronomer s telegram , 7194 morris , p. j. , cotter , g. , brown , a. m. , & chadwick , p. m. 2017 , mnras , 465 , 1218 mukai , k. , & ishida , m. 2001 , , 551 , 1024 mukai , k. , orio , m. , & della valle , m. 2008 , , 677 , 1248 mukai , k. , nelson , t. , chomiuk , l. , et al .
2014 , in asp con .
490 stella novae : past and future decades , ed .
a. woudt & v. a. r. m. ribeiro , 327 munari , u. , hambsch , f .- j .
, & frigo , a. 2017 , mnras ( submitted ) , arxiv:1703.09017 nelson , t. , orio , m. , cassinelli , j. p. , still , m. , leibowitz , e. , & mucciarelli , p. 2008
, , 673 , 1067 nelson , t. , linford , j. , chomiuk , l. , et al .
2015 , the astronomer s telegram , 7085 obrien , t. j. , bode , m. f. , porcas , r. w. , et al .
2006 , in proceedings of the 8th european vlbi network symposium , ed .
baan , w. , bachiller , r. , booth , r. , et al .
, 52 orio , m. , rana , v. , page , k. l. , sokoloski , j. , & harrison , f. 2015 , , 448 , l35 page , k.l . , osborne , j. p. , kuin , n. p. m. , et al .
2015 , , 454 , 3108 perley , r. a. , & butler , b. j. 2013 , , 204 , 19 predehl , p. , & schmitt , j. h. m. m. 1995 , , 293 , 889 rohlfs , k. , & wilson , t. l. 2006 , tools of radio astronomy , 4th revised and enlarged edition , berlin : springer rupen , m. p. , mioduszewski , a. j. , & sokoloski , j. l. 2008 , , 688 , 559 schlafly , e. f. , & finkbeiner , d. p. 2011
, , 737 , 103 seaquist , e. r. , & bode , m. f. 2008 in classical novae , 2nd edition .
cambridge astrophysics series , no .
43 , cambridge : cambridge university press , ed . m. f. bode & a. evans , 141 sheperd , m. c. 1997 , in asp conf .
125 , astronomical data analysis software and systems vi , ed .
g. hunt & h. e. payne ( san francisco , ca : asp ) , 77 smith , r. k. , brickhouse , n. s. , liedahl , d. a. , & raymond , j. c. 2001 , , 556 , l91 sokoloski , j. l. , rupen , m. p. , & mioduszewski , a. j. 2008 , , 685 , l137 srivastava , m. k. , ashok , n. m. , banerjee , d. p. k. , & sand , d. 2015 , , 454 , 1297 starrfield , s. , truran , j. w. , sparks , w. m. , & kutter , g. s. 1972 , , 176 , 169 starrfield , s. , iliadis , c. , timmes , f. x. , et al .
2012 , basi , 40 , 419 taylor , a. r. , pottasch , s. r. , seaquist , e. r. , & holli , j. m. 1987 , , 183 , 38 tokovinin , a. , fischer , d. a. , bonati , m. , et al .
2013 , , 125 , 1336 walker , r. c. 2014 , vlba scientific memo # 37 walter , f. 2015 , the astronomer s telegram , 7060 weston , j. h. s. , sokoloski , j. l. , metzger , b. d. , et al .
2016a , , 457 , 887 weston , j. h. s. , sokoloski , j. l. , chomiuk , l. , et al . 2016b , , 460 , 2687 williams , s. c. , darnley , m. j. , bode , m. f. , & shafter , a. w. 2016 , , 817 , 143 woudt , p. a. , steeghs , d. , karavska , m. , et al .
2009 , , 706 , 738
|
we present multi - wavelength observations of the unusual nova v1535 sco throughout its outburst in 2015 .
early radio observations were consistent with synchrotron emission , and early x - ray observations revealed the presence of high - energy ( > 1 kev ) photons .
these indicated that strong shocks were present during the first @xmath02 weeks of the nova s evolution .
the radio spectral energy distribution was consistent with thermal emission from week 2 to week 6 .
starting in week 7 , the radio emission again showed evidence of synchrotron emission and there was an increase in x - ray emission , indicating a second shock event .
the optical spectra show evidence for at least two separate outflows , with the faster outflow possibly having a bipolar morphology .
the optical and near infrared light curves and the x - ray n@xmath1 measurements indicated that the companion star is likely a k giant .
|
neutrino experiments have firmly demonstrated neutrinos are massive and have non - vanishing mixing angles among the different generations @xcite . since in the standard model neutrinos are massless this experimental results are a clear evidence of beyond standard model physics . from a general point of view majorana neutrino masses can be generated by adding to the standard model lagrangian the non - renormalizable effective lagrangian @xcite @xmath1 where @xmath2 and @xmath3 are the lepton and higgs electroweak doublets .
the presence of the effective lagrangian in ( [ eq : eff - op ] ) guarantees that after electroweak symmetry breaking neutrinos acquire majorana masses .
extensions of the standard model in which neutrinos have majorana masses realize the effective operator @xmath4 in different ways . however , among all these possibilities there are subsets that have a common feature , namely the order in perturbation theory at which the operator is realized . as will be discussed in sec .
[ sec : o5realizations ] a rather general classification of neutrino mass models can be done by gathering together these different subsets . making an exhaustive list of possible realizations of @xmath4 and their corresponding phenomenology is far out of the scope of the present discussion .
thus , instead of using this approach we will discuss two classes of generic models : models in which neutrino masses are generated radiatively at the 2-loop level and supersymmetric models with broken r - parity . in the former case we will stick to the babu - zee model @xcite , while for the latter we will discuss bilinear r - parity breaking models with a neutralino lsp @xcite .
at @xmath5 the heavy physical degrees of freedom , that when integrated out yield the dimension five ( @xmath6 ) effective operator @xmath4 , are no longer decoupled .
once specified they define a particular model for majorana neutrino mass generation , so different models lead to different realizations of @xmath4 .
as pointed out in the introduction these realizations can be classified according to the order in perturbation theory at which @xmath4 is generated , namely tree - level ( @xmath7 ) , one - loop ( @xmath8 ) , two - loops ( @xmath9 ) and three - loops ( @xmath10 ) , where @xmath11 denotes the number of loops in each case .
three - loop level realizations require order one yukawa couplings implying that models based on @xmath12 are , in general , not consistent with neutrino data once the requirement of perturbativity of the corresponding couplings is imposed .
a general and model independent classification of the tree - level @xmath7 and one - loop level @xmath13 realizations of @xmath4 has been carried out in ref .
@xcite and such an analysis could be , in principle , extended to the two and three loop order realizations .
regardless of the order at which the operator arises the procedure is based on the determination of all the possible gauge invariant renormalizable vertices within the loop for all the different possible topologies . in this procedure
the only gauge quantum numbers that are fixed are those of the external legs ( @xmath0 ) while those of the physical degrees of freedom flowing in the loop are free and should be fixed by the requirement of gauge invariance .
another approach for the same sort of classification is by using all the possible @xmath14 effective operators up to certain @xmath15 .
this method has been implemented in ref .
@xcite for effective operators up to @xmath16 .
the difference between these two approaches is that while the former focus exclusively on the @xmath4 operator the later covers this operator but also higher dimensional operators as e.g. @xmath17 .
note that these higher dimensional effective operators ( @xmath18 ) can give a dominant contribution to neutrino masses only if the leading effective operator @xmath4 is forbiden due to a new symmetry .
these cases have been throughout analysed in ref .
@xcite and we have nothing more to add here . the tree - level realizations of @xmath4 correspond to the different type of seesaw models ( type - i @xcite , type - ii @xcite and type - iii @xcite ) whereas for the 1-loop cases the number of possibilities ( models ) is much more larger ( is determined by the different @xmath19 assignments of the internal degrees of freedom flowing in the loop ) .
examples of @xmath13 include the zee model @xcite , models with scalar leptoquarks @xcite and models with extra scalars and fermions with nontrivial color charges @xcite .
examples of two - loop realizations include extended scalar sectors as in the case of the the babu - zee model @xcite and models with scalar leptoquarks @xcite .
up to our knowledge three - loop level realizations rely on extensions of both the scalar and fermion sectors , examples can be found in references @xcite . in models in which the operator @xmath4 is generated beyond the tree - level the lepton number breaking scale can readily be at or around the electroweak scale .
thus , models embedded in such realizations usually lead to testable predictions in either high - energy or high - intensity experiments .
in what follows we will discuss two main cases , the babu - zee model and supersymmetric bilinear r - parity violating models , paying special attention to some of their phenomenological implications .
in this model the standard model scalar sector is extended by the addition of two new scalars , @xmath20 and @xmath21 , both singlets under @xmath22 .
their couplings to standard model leptons is given by @xmath23 here , @xmath24 are the standard model ( left - handed ) lepton doublets , @xmath25 the charged lepton singlets , @xmath26 are generation indices and @xmath27 is the completely antisymmetric tensor . note that @xmath28 is antisymmetric , while @xmath29 is symmetric .
assigning @xmath30 to @xmath31 and @xmath21 , eq .
( [ yuks ] ) conserves lepton number .
lepton number violation in the model resides only in the following term in the scalar potential @xmath32 here , @xmath33 is a parameter with dimension of mass .
the setup of eq .
( [ yuks ] ) and eq .
( [ scalar ] ) generates majorana neutrino masses via the two - loop diagram shown in fig .
( [ fig : mnu ] ) .
the resulting neutrino mass matrix can be expressed as @xmath34 with summation over @xmath35 implied .
the parameters @xmath36 are defined as @xmath37 , with @xmath38 the mass of the charged lepton @xmath39 .
following @xcite we have rewritten @xmath40 and @xmath41 .
@xmath42 finally is a dimensionless two - loop integral given by @xmath43 for non - zero values of @xmath44 , @xmath42 can be solved only numerically .
we note that for the range of interest , say @xmath45 , @xmath42 varies quite smoothly between ( roughly ) @xmath46 .
phenomenological tests of this model have been studied in @xcite . among all of them
those involving @xmath47 can be regarded as the most stringent ones . in ref .
@xcite it has been shown that the corresponding decay branching ratio for this process can be written as @xmath48 with @xmath49 and @xmath50 the mass of the singly charged scalar .
figure [ muegamvmh ] shows the resulting lower limit on @xmath51 ) as a function of @xmath50 for the case of normal and inverted hierarchies .
note that the horizontal solid line indicates the upper limit set by the mega experiment @xcite and not the new one placed by the meg experiment , @xmath52 at 90% c.l .
@xcite . using the updated limits the constraints on the singly charged scalar mass would be even more stringent that the ones quoted here . in summary , in this model @xmath53 is guaranteed for singly charged scalar masses smaller than 590 gev ( 5.04 tev ) for normal ( inverse ) hierarchical neutrino masses , and larger or even much larger branching ratios are expected in general .
thus , a non - observation of this process in the next few years , at least for the case of inverse hierarchy , would certainly remove most of the motivation to study this model . ) as a function of the charged scalar mass @xmath50 for normal hierarchy ( left plot ) and inverted hierarachy ( right plot ) .
the three lines are for the current solar angle @xmath54 best fit value ( full line ) and 3 @xmath55 lower ( dashed line ) and upper ( dot - dashed line ) bounds .
other parameters fixed at @xmath56 , @xmath57 and @xmath58 ev@xmath59.,title="fig:",width=211,height=188 ] ) as a function of the charged scalar mass @xmath50 for normal hierarchy ( left plot ) and inverted hierarachy ( right plot ) .
the three lines are for the current solar angle @xmath54 best fit value ( full line ) and 3 @xmath55 lower ( dashed line ) and upper ( dot - dashed line ) bounds .
other parameters fixed at @xmath56 , @xmath57 and @xmath58 ev@xmath59.,title="fig:",width=211,height=188 ]
bilinear r - parity violation ( brpv ) provides an intrinsically supersymmetric framework for majorana neutrino masses ( for a review see @xcite ) . in these models
the superpotential includes , in addition to the mssm terms , also the term @xmath60 this term breaks not only r - parity but also lepton number in all three generations . in order to have a consistent model
a soft susy breaking term has to be added to the scalar potential , namely @xmath61 the presence of these terms induce a non - vanishing vacuum expectation value for the sneutrinos ( @xmath62 ) that give rise to a mixing among neutral gauginos and higgsinos with neutrinos . due to this mixing one of
the neutrinos acquires mass .
the effective neutrino mass matrix reads @xmath63 where @xmath64 ( with @xmath33 the higgsino mass term and @xmath65 ) and @xmath66 is the neutralino mass matrix .
the other two neutrinos acquire mass from one - loop corrections involving @xmath67 , @xmath68 and scalar loops , being the bottom - sbottom and tau - stau loops the most important contributions @xcite .
thus , the brpv model is an example of a model in which neutrino masses arise from different realizations of the @xmath69 operator .
the atmospheric and reactor angles are approximately given by the rotation angles that diagonalize the tree level mass matrix @xmath70 : @xmath71 the solar angle instead is obtained once the one - loop corrections are taken into account .
a remarkable feature of the bilinear r - parity breaking model is that the same parameters that determine neutrino physics also control the decay patterns of the lsp , thus it is always possible to establish a set of correlations between the lsp decay branching ratios and neutrino observables .
these correlations can be used as a tool to know whether brpv is responsible for the origin of neutrino masses .
once r - parity is broken the lsp is unstable and decays to standard model fermions .
thus astrophysical contraints on its nature do not hold any more and in principle any supersymmetric particle can be the lsp .
the decay patterns of all possible lsps and their relations with neutrino observables in the context of bilinear r - parity breaking models have been analysed in @xcite .
additional studies in more generic models including trilinear r - parity violating couplings have been carried out for slepton and sneutrino lsps @xcite . here
we will highlight some of the main features of this `` program '' in bilinear r - parity breaking models assuming the lightest neutralino to be the lsp .
the following discussion is based on ref .
@xcite .
the presence of the bilinear r - parity violating parameters induce not only a mixing between neutralinos and neutrinos but also a mixing between charginos and charged leptons , charged higgses and sleptons , cp - even ( cp - odd ) components of the neutral higgses and the corresponding cp - even ( cp - odd ) components of the sneutrinos @xcite .
all together these mixings determine the three - body leptonic , semi - leptonic and invisible neutralino decays : @xmath72 , @xmath73 , @xmath74 and @xmath75 .
given the above discussion it should now become clear that once the bilinear r - parity breaking parameters are constrained by neutrino experimental data the lightetest neutralino decays are expected to be constrained as well .
in fact , it turns out that the constraints imposed by neutrino physics assure : ( @xmath76 ) the neutralino always decay inside the detector ; ( @xmath77 ) the decay branching ratio for neutralino invisible decays never exceeds 10% ; ( @xmath78 ) different ratios of decay branching ratios are strongly correlated with neutrino mixing angles .
figure [ fig : neutralino - decaybr ] shows the corresponding correlations for the semi - leptonic final states @xmath79 , @xmath80 and @xmath81 . from these results and the measured neutrino mixing angles it can be established that if brpv is the mechanism responsible for the origin of neutrino masses and the lightest neutralino turn out the be the lsp the following measurements should be expected at lhc : @xmath82 in summary , in bilinear r - parity breaking models the decay patterns of the lsp are strongly correlated with neutrino mixing angles .
these correlations allow to set constraints on the different decay branching ratios of the lsp that in turn can be used to prove whether these models are responsible for the origin of neutrino masses and mixings .
from a general perspective majorana neutrino masses can be accounted for by the effective dimension five operator @xmath4 .
we have argued that the different `` incarnations '' of this operator can be classified according to the order in perturbation theory at which the operator is realized . in principle by using group theoretical arguments one could make , in a model independent way , an exhaustive list of all the possibilities at each order and up to the three - loop level .
models in which @xmath4 is realized radiatively ( @xmath83 , where @xmath11 denotes the numer of loops ) rely on tev scale physics .
thus , an obvious question is whether this new physics , and therefore the origin of neutrino masses , can be proved at e.g. the lhc . as illustrative examples
we have discussed what we consider two benchmark models : the babu - zee model and supersymmetry with bilinear broken r - parity .
[ sec : acnowledgements ] i would like to thank sergey kovalenko , werner porod and diego restrepo for the enjoyable collaboration on some of the subjects discussed here .
i want to especially thank martin hirsch for the collaboration that led to some of the papers quoted here and for the always enlightening discussions .
p. minkowski , _ phys .
_ b * 67 * 421 ( 1977 ) ; t. yanagida , in _ proc . of workshop on unified theory and baryon number in the universe _
, eds . o. sawada and a. sugamoto , kek , tsukuba , ( 1979 ) p.95 ; m. gell - mann , p. ramond and r. slansky , in _ supergravity _ , eds p. van niewenhuizen and d. z. freedman ( north holland , amsterdam 1980 ) p.315 ; p. ramond , _
sanibel talk _
, retroprinted as hep - ph/9809459 ; s. l. glashow , in _ quarks and leptons _ , cargse lectures , eds m. lvy , ( plenum , 1980 , new york ) p. 707 ; r. n. mohapatra and g. senjanovi , _ phys .
* 44 * , 912 ( 1980 ) , j. schechter and j. w. f. valle , phys . rev .
d * 22 * , 2227 ( 1980 ) . j. schechter and j. w. f. valle , phys .
d * 22 * , 2227 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , nucl .
b * 181 * , 287 ( 1981 ) ; r. n. mohapatra and g. senjanovic , phys .
d * 23 * , 165 ( 1981 ) ; c. wetterich , nucl .
b * 187 * , 343 ( 1981 ) ; r. foot , h. lew , x. g. he and g. c. joshi , z. phys .
c * 44 * , 441 ( 1989 ) .
a. zee , phys .
* b93 * , 389 ( 1980 ) .
|
models for majorana neutrino masses can be classified according to the level in perturbation theory at which the effective dimension five operator @xmath0 is realized .
the possibilities range from the tree - level up to the three - loop level realizations .
we discuss some general aspects of this approach and speculate about a model independent classification of the possible cases . among all the realizations , those in which the effective operator is induced by radiative corrections open the possibility for lepton number violation near -or at- the electroweak scale .
we discuss some phenomenological aspects of two generic radiative realizations : the babu - zee model and supersymmetric models with bilinear r - parity violation .
|
it is well established that a sizable fraction of the integrated light of stellar populations comes from the thermally pulsing asymptotic giant branch ( tp - agb ) phase @xcite . however , the size of this fraction has been subject of much discussion in the recent literature , with evolutionary population synthesis ( eps ) models of galaxies favouring either `` heavy '' @xcite or `` light '' @xcite flux contributions from tp - agb stars .
efforts to account for the tp - agb contribution to the integrated light of galaxies are based on two different techniques : \1 ) eps models based on the isochrone method @xcite start by adopting the best - available sets of evolutionary tracks and isochrones including the tp - agb phase . since this phase is notoriously challenging to model due to difficulties and uncertainties in the description of mixing , energy transport by convection , mass loss , and numerical aspects at present
there is no set of widely - accepted tp - agb model grids in the literature .
therefore , in general , the choice falls on models which at least try to reproduce basic observables of tp - agb stars in the magellanic clouds ( mcs ; e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this approach has been adopted in the popular @xcite and @xcite models , with some subtle technical differences and a posteriori corrections to consider e.g. circumstellar dust @xcite or shifts in the @xmath3 and @xmath4 of tp - agb models @xcite .
\2 ) eps models by @xcite , based on the fuel consumption theorem ( cf . * ? ? ?
* hereinafter fct ) , bypass the use of tp - agb evolutionary tracks and stellar isochrones , since they are rooted in the belief that detailed modelling of the tp - agb evolution is hopelessly uncertain to be useful in this context .
each single - burst stellar population is simply assigned , as a function of age , the tp - agb `` fuel '' that appears to best reproduce the tp - agb luminosity contribution measured in mc clusters .
besides the integrated fuel as a function of stellar age , the fct method employs other prescriptions , such as the fraction of fuel burnt by c- and m - type stars , and some rough dependence on metallicity .
it is somewhat puzzling and disturbing that , despite the efforts of calibrating these eps models on tp - agb data in the mc clusters ( albeit with different approaches ) , when taken without their `` corrections '' , the same eps models tend to overestimate the contribution of the tp - agb to the spectral energy distributions of galaxies .
for instance , @xcite models , based on the fct recipe , show an excess of tp - agb flux which is not observed in post - starburst galaxies @xcite .
likewise , @xcite evolutionary tracks predict @xmath5 more agb stars than observed in a sample of nearby galaxies observed with hst , which translates to a factor @xmath0 excess in their integrated near - infrared flux @xcite .
in this study we analyse a specific evolutionary effect which we identify as the likely reason for this unexpected incongruity .
we start describing in detail how the tp - agb integrated luminosity is predicted to vary with age ( sect .
[ models ] ) .
we find a significant boosting of the tp - agb contribution not predicted by the fct at ages @xmath6 gyr , that coincides with the age interval populated by some of the most massive clusters in the mcs ( sect .
[ controversy ] ) .
the far - reaching consequences and implications in the context of the tp - agb calibration are discussed in sect .
[ conclu ] .
to explore the detailed contribution of tp - agb stars to the integrated light of a ssp , we have computed a dense grid of stellar evolutionary tracks , assuming an initial composition with metallicity @xmath7 , helium abundance @xmath8 , and a scaled - solar distribution of metals ( cf . * ? ? ?
this corresponds to the case @xmath9}}={\mbox{\rm [ { \rm m}/{\rm h}]}}=-0.40 $ ] , which is suitable to describe intermediate - age clusters in the lmc .
stellar evolutionary tracks were computed with the parsec code @xcite until the first thermal pulse on the tp - agb , and then followed with the colibri code @xcite until the complete ejection of the stellar envelope .
tracks are closely distributed in initial mass @xmath10 , with a spacing @xmath11 in the proximity of the limiting maximum mass , @xmath12 , for a star to develop an electron - degenerate he - core after the main sequence ( ms ; with @xmath13 in this case ) . for low - mass stars ( those with @xmath14 ) , the evolution from the he - flash to the initial stage of quiescent he core - burning ( cheb ) is skipped by means of a standard and well - tested algorithm that preserves the stellar core mass and chemical profile , while converting part ot the helium in the core into carbon to take into account the nuclear energy necessary to lift the electron degeneracy .
convective core overshooting is adopted with an efficiency as detailed in @xcite .
in addition , we completely suppress breathing pulses towards the end of cheb phase , so as to avoid erratic track - to - track fluctuations in the cheb lifetimes .
evolutionary tracks are shown in the initial mass versus age ( vs. @xmath15 ) plots of fig .
[ fig_taum ] , color - coded as a function of the evolutionary phase ( left panel ) or luminosity ( right panel ) .
the inverse power - law relation between the initial mass and ms lifetime is clearly recognizable in the plot , as is the presence of a long - lived red giant branch ( rgb ) for masses @xmath16 .
also evident is that the cheb lifetime suddenly gets longer at @xmath17 , passing from the @xmath18 yr typical of low - mass stars to at least twice this value at @xmath19 .
this abrupt growth reflects the onset of cheb at significantly lower core masses and luminosities , compared to the case of low - mass stars .
these features are well - established and their consequences to the cheb phase are thoroughly discussed in @xcite .
moreover , we notice how brief the typical lifetime of the tp - agb phase ( a few @xmath20 yr ) is compared to the cheb lifetime , and its relatively narrow dynamical range over the relevant mass interval . in fig .
[ fig_taum ] a few horizontal lines are drawn to highlight the evolutionary phases intersected by the stellar _ isochrones _ at some selected ages : * the youngest isochrone , of @xmath21 gyr , crosses the ms , a very short section of `` rgb '' ( actually , this is just the quick core contraction phase that precedes he - ignition in a non - degenerate core ) , a long section of cheb , the early - agb ( e - agb ) , and the tp - agb . * similarly , the same phases are crossed by the 1.68-gyr isochrone , with the difference that its rgb is somewhat longer ( in terms of the spanned interval ) and more luminous , while the cheb is shorter . * at 1.55 gyr , the isochrone intercepts both a section of low - mass and bright cheb , and a section of higher - mass ( and fainter ) cheb .
this is the age range in which the most rapid changes in cheb morphology take place .
* at 1.63 gyr , the isochrone crosses three distinct tp - agb phases ( as well as three e - agbs ) , the first at @xmath22 , the second ( somewhat shorter ) at @xmath23 , and the third at @xmath24 .
the second tp - agb section raises quickly from below to above the 1.63 gyr line , so it is likely less populated than the other two .
it is somewhat obvious from fig .
[ fig_taum ] that the numbers of tp - agb stars will be `` boosted '' in isochrones with ages between 1.57 and 1.66 gyr , since they will all contain at least two well populated tp - agbs .
since the two branches have similar ( and high ) luminosities , their total integrated fluxes will be boosted as well .
a rough estimate of the boosting factor comes from the integral of @xmath10 , weighted by the imf , along the sections of the isochrone lines that correspond to the tp - agbs . given the limited range of masses involved
, we can naively expect a boosting factor of @xmath0 to be typical across this 0.1 gyr - wide age interval . ) , as a function of .
like in the previous fig .
[ fig_taum ] , different evolutionary stages are marked with different colors , and the vertical bars at the bottom mark the of the tracks which were actually calculated , rather than derived from interpolation . ]
let us now look at this problem in terms of the luminosity contribution of tp - agb stars , @xmath25 , to the integrated bolometric light of a single - burst stellar population ( ssp ) as a function of its age @xmath15 . as a first step ,
let us consider the integrated emitted energy of stellar tracks from their birth line on the pre - ms , up to any given evolutionary stage @xmath26 : @xmath27 figure [ fig_fuelm ] shows this quantity as a function of the initial stellar mass , as obtained from our set of evolutionary tracks , and it is color - coded as a function of the main evolutionary stages .
we can appreciate that the total emitted energy is a quite well - behaved function of .
the most notable feature in this plot is the presence of a significant contribution from the rgb at masses @xmath16 , which is counterbalanced by an increase in the cheb and tp - agb contributions at @xmath17 . except for these aspects ,
the run of the emitted energy with mass is quite smooth and covers a remarkably limited range .
for this particular set of tracks , the total post - ms emitted energy amounts to @xmath28 gyr , with the tp - agb contribution ranging from @xmath29 to @xmath30 gyr , depending on the stellar mass .
let us now look at the ssp integrated luminosity as derived from the isochrones , @xmath31 where @xmath32 initial mass function ( in this case
the * ? ? ?
* with a low - mass cut at @xmath33 ) . this quantity is shown in the left panel of fig .
[ fig_fuelt ] , as a function of age , and again separating the contribution from different evolutionary stages . notice the significant increase of the total tp - agb luminosity contribution at ages between @xmath34 and 1.66 gyr , the unequivocal evidence of the `` agb boosting '' that originates from the multiple tp - agbs at those ages .
moreover , notice that there is a period between 1.40 gyr and 1.55 gyr during which the increase in luminosity due to the appearance of the rgb is not compensated by a decrease in the tp - agb luminosity , so that also in this interval there is a temporary but more modest increase in the total emitted luminosity . only for
ages older than 1.66 gyr the evolution of the total luminosity seems to recover again the smooth trend defined by ages younger than 1.40 gyr .
the figure also contains some mild irregularities , especially at the end of the tp - agb phase , which derive either from the small variations in the rate at which agb stars are produced after the cheb phase , or from the variable numbers of thermal pulses occurring in the single tp - agb evolutionary tracks .
anyway , these fluctuations are small and not of concern here .
the most significant point , in the context of this study , is the presence of the agb - boosting period between @xmath34 and 1.66 gyr .
as already mentioned , the agb - boosting period derives from the sudden increase of the cheb lifetime at @xmath35 , that causes single isochrones to cross multiple sections of the tp - agb phase experienced by stars with slightly different initial masses .
approximations which do not take into account these different cheb lifetimes and how they reflect into isochrones , are not expected to show this feature . in particular , the fct ( cf . * ? ? ?
* ; * ? ? ?
* ) approximates the evolution of @xmath36 with the following equation : @xmath37 where @xmath38 is the integrated luminosity of the ms .
all the subsequent evolutionary stages are described by a single track of mass equal to the turn - off one , @xmath39 , uniformly weighted by the evolutionary rate at which stars with @xmath40 leave the ms , @xmath41 .
the right panel of fig .
[ fig_fuelt ] shows the result of applying equation [ ltpagb_fct ] , where the functions @xmath42 ( equivalent to `` tables of fuel '' ) are obtained from the evolutionary tracks , and are shown in our previous fig .
[ fig_fuelm ] .
we adopt the same layout as in the left panel , showing the contribution of every evolutionary phase to the integrated light with a different color .
evidently there is no sign of an agb - boosting period in this case .
the reason of such a remarkable difference in the predictions between the isochrone and fct methods can be easily caught considering that , to describe the post - ms phases of a ssp , the fct assumes a perfectly one - to - one correspondence between age - index and mass - index ( ms lifetime of a star with mass @xmath43 ) , while the isochrone method populates each phase by a finite range of initial stellar masses .
it follows that the fct approximation constrains any ssp to cross a given post - ms phase not more than once ( exemplified by the vertical lines in fig .
[ fig_taum ] ) , while multiple crossings may , in principle , take place with the isochrone method ( represented by e.g. the horizontal lines in fig .
[ fig_taum ] ) . of the two approaches shown in fig .
[ fig_fuelt ] , the one based on isochrones ( see also * ? ? ?
* ) is certainly best suited to represent the time evolution of single - burst stellar populations defined as generations of stars born at the same time and sharing the same age @xmath15 alike small - mass star clusters .
it is also the one that turns out to capture most details of the ssp luminosity evolution .
this is not the first time that the agb boosting at ages close to 1.6 gyr has been noticed . using isochrones derived from evolutionary tracks computed with a coarser mass resolution
, @xcite already noticed a temporary increase in the tp - agb production rate due to the flattening of the @xmath44 versus relation ( where @xmath44 is the age at the end of the cheb phase ) close to .
the increase was in proportion to @xmath45 , and appeared as a sharp , short - lived peak in the time evolution of integrated colors such as @xmath46 .
this feature has been present in many sets of the padova isochrones distributed since 2000 @xcite , and often is considered to be a bug by external users . by conveniently sampling a grid of isochrone ages one could completely eliminate it from the data .
the feature we identify and discuss in this work is essentially the same one , but this time it has been derived from a much more detailed and denser grid of stellar evolutionary tracks .
what was seen as a sharp peak in the production of agb stars appearing at a precise age by @xcite , it now appears as a triple tp - agb over a relatively wide age interval ( fig .
[ fig_fuelt ] ) .
let us give a closer look at this point .
[ fig_resolu ] illustrates essentially the same features as in the previous figs .
[ fig_taum ] and [ fig_fuelt ] , but now using grids of evolutionary tracks of progressively worse mass resolution .
these degraded grids are obtained from the high - resolution grid already described , by progressively eliminating tracks in the vicinity of the 1.76 one .
interpolated tracks and isochrones are then obtained using exactly the same algorithms as before .
four different resolutions are presented , namely @xmath11 , 0.02 , 0.04 and 0.08 , the first one being the original case already presented in previous figs .
[ fig_taum ] to [ fig_fuelt ] .
the presence of the 1.76 track ensures that , in all cases , we sample the maximum of he - burning lifetimes that occurs slightly above . for the @xmath47 and @xmath48 cases ,
the mass resolution is still good enough for a triple tp - agb to appear in the isochrones ( top - middle panels ) , although for a reduced age interval . as a consequence
, the agb - boosting period continues to appear in the integrated light ( bottom - middle panels ) , but confined to a smaller age range , _ and with a larger amplitude_. the larger amplitude is simply caused by the fact that tp - agb stars from a wider range of initial masses now appear in a narrower range of isochrone ages . in the case of the @xmath47 resolution , the effect is not dramatic , and the evolution of integrated light resembles very much the one obtained for the @xmath11 case . for @xmath48 , however , the agb - boosting appears over an age range of just @xmath49 gyr , and with a amplitude about 2.5 times larger than the one seen at @xmath11 . for the @xmath50 case , instead
, the situation is apparently very different .
there is no longer any single isochrone crossing three different tp - agb sections ( as seen in the top - right panel ) .
but anyway , an intense agb - boosting period still appears , and at slightly later ages than before .
this happens because there is an age interval ( between 1.66 and 1.73 gyr ) in which the isochrones cross a larger interval of while on the tp - agb phase .
this is actually the situation which was found by @xcite , and explained by means of the higher production rate of agb stars ( proportional to @xmath45 ) occurring in this age range .
it is obvious that , in addition to the four cases illustrated in fig .
[ fig_resolu ] , all sorts of intermediate situations may be created , depending on how well the mass interval around is sampled by the stellar evolutionary tracks that are used to build isochrones .
it may also be that different sets of evolutionary tracks , when produced at a @xmath51 resolution as our own , will reveal a more gradual change in the evolutionary features across .
this does not imply that such tracks will avoid the agb - boosting phase , since as demonstrated in the right panel of fig .
[ fig_resolu ] , the agb - boosting period can be generated even if the cheb burning lifetimes change over a mass interval as wide as @xmath50 .
more generally , we can affirm that _ as long as the cheb burning lifetime increases with mass above , in a way that largely compensates for the decrease in main - sequence and rgb lifetimes over the same interval , the agb - boosting period has to occur_. fig .
[ fig_resolu ] , and the simpler arguments in @xcite , show this clearly .
this effect can be eliminated only if we get rid of the increase in cheb burning lifetimes occurring above , or , alternatively , if incomplete descriptions of the light evolution of ssps ( as the one provided by the fct ) are adopted , as shown in our previous fig .
[ fig_fuelt ] . from the previous discussion ,
it is clear that the present results are a consequence of the abrupt changes in the evolutionary features at @xmath35 , as the electron degeneracy no longer develops in the core before he ignition : over a small interval of , the shortening of the rgb lifetime is followed by an abrupt reduction in the luminosity at which the he - burning phase takes place , that in turns causes a sharp increase of its lifetime .
these changes in the cheb lifetime and luminosity have been found in many other sets of evolutionary tracks since @xcite , as in e.g. @xcite , @xcite , @xcite , @xcite , @xcite and @xcite .
the appendix provides a more extensive comparison between different sets of tracks , revealing that the large increase on the cheb burning lifetimes above is not only a common feature in the tracks where the he - burning evolution is computed , but is also expected from the conditions at the stage of he - ignition , in nearly all sets of tracks computed up to that stage .
indeed , it seems well accepted @xcite that the longer cheb burning lifetimes slightly above derive from the smaller core masses ( and hence smaller initial he - burning luminosities ) at he - ignition , in these stars . in addition
, we recall that the same changes in core masses and he - burning luminosities at are at the origin of the appearance of secondary and dual red clumps , that are nowadays well observed in the milky way and other local group galaxies ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , and even in a few populous mc star clusters @xcite .
therefore , the agb - boosting effect is not a feature that pertains to a single , isolated set of evolutionary tracks .
moreover , our estimates of the emitted light ( or fuel ) _ as a function of the stellar initial mass _ are well in line with the behaviour suggested by @xcite , @xcite and @xcite .
in fact , the agb - boosting feature does not take place when we consider the behaviour of the evolutionary tracks as a function of the initial stellar mass ( fig .
[ fig_fuelm ] ) .
it only shows up if we look at the emitted light _ as a function of the age _ ,
provided we can count on stellar isochrones with a high age ( and mass ) resolution ( as in the left panel of fig .
[ fig_fuelt ] , and in the bottom panels of fig .
[ fig_resolu ] ) .
the previous analysis has demonstrated the expected appearance of triple tp - agb branches at ages @xmath2 gyr .
we designate this effect with the generic name of `` agb boosting '' instead of simply `` agb triplication '' , because details about its duration and `` boosting factor '' may well change as more detailed evolutionary tracks are computed .
but we emphasize that as demonstrated in the previous section as long as the cheb lifetimes sharply increase above , an agb - boosting period _ has _ to occur .
the agb - boosting phase is a temporary feature , preceded and followed by a smoothly - varying light contribution from tp - agb stars .
it can not last much longer than 0.1 gyr ( again depending on details of evolutionary tracks ) , as we can infer from figs .
[ fig_taum ] and [ fig_resolu ] . in the wider context of galactic evolution , this is a short - lived feature , and one may wonder whether it could be affecting the conclusions of eps models in any significant way .
should not this feature be completely smeared out when convolved with the continuous star - formation histories typical of star - forming galaxies ? and indeed , we believe this feature _ is _ smeared out in most nearby galaxies .
its critical importance resides in the fact that it affects , in a significant way , the mc clusters which are _ the classical calibrators _ of the contribution of tp - agb stars to eps models . vs. age plane ,
considering just the stars above the tip of the rgb at @xmath52 .
circles are used for stars in smc clusters , and triangles for the lmc .
stars marked in red belong to the clusters ngc 419 , 1806 , 1846 , 1751 and 1783 , for which the error bars indicate the approximate age range of their multiple populations .
the * middle panel * shows the age histogram of these tp - agb stars , whereas the * upper panel * shows the same but for the number of agb stars normalized to the clusters @xmath53-band integrated luminosities . in all panels ,
the light grey area indicates the narrow age interval for which the tp - agb population is expected to be boosted ( cf .
[ models ] ) . ]
let us consider fig .
[ fig_clusters ] , which shows the tp - agb stars in mc clusters from the classical compilation by @xcite .
the plot shows all the stars above the tip of the rgb at @xmath52 , for which we could attribute ages in the range from 0.1 to 3.3 gyr .
this exclude clusters which , despite containing a few candidate tp - agb stars , are too young ( like ngc 1850 , 1854 with @xmath54 gyr , cf .
* ) or too old ( like ngc 121 , 339 , 361 , 416 , 1841 and kron 3 ; cf .
* ; * ? ? ?
* ) to be of interest here .
cluster ages were in general taken from the recent compilation from @xcite which mostly includes age estimates derived from data that reaches the cluster turn - offs but for the following cases : * for the star clusters ngc 411 , 419 , 1806 , 1846 , 1751 and 1783 , which present clear signs of multiple populations , we adopt the mean ages and their dispersions found in the more recent and detailed cmd analyses based on hst data .
more specifically , for ngc 419 we adopt the @xmath55 gyr age range that encompasses the bulk of its star formation @xcite .
the same range is attributed to ngc 411 , given the great similarity of its color magnitude diagram with the ngc 419 one @xcite .
for ngc 1751 , 1783 and 1846 we adopt the full age ranges found in their centres by @xcite . for the latter clusters ,
similar age ranges are also obtained by @xcite , using essentially the same data but different methods and various sets of stellar models .
the ages for ngc 1806 are also taken from @xcite .
notice that for the all these clusters , the mean ages are similar ( to within 0.15 gyr ) to those found in the @xcite compilation , but for ngc 419 the cluster with the most tp - agb stars to which they assign a single age of 1.2 gyr .
the latter is actually the age of the youngest stellar population in ngc 419 as determined by @xcite , and not its mean age . * for ngc 152
, we take the age from @xcite , derived from hst wfpc2 photometry . * finally , for a few other clusters , like ngc 299 , 306 , 1652 , no reliable ages were found .
they contain just 9 candidate tp - agb stars , which certainly do not affect our discussion .
this sample is expected to be more or less unbiased in terms of its age distribution , in the sense that the original selection of clusters by @xcite was not aiming to sample any particular age range , nor any particular range of age - sensitive properties ( like the integrated colors ) . the bottom panel of fig .
[ fig_clusters ] shows the location of the tp - agb stars in these mc clusters in the @xmath56 vs. age plane , while the two upper panels show the histogram of stellar ages , using either straight star counts or counts normalized to the cluster @xmath53-band integrated luminosity which represents a rough measure of present cluster masses .
all these plots make evident a strong concentration of tp - agb stars close to the 1.5-gyr age range .
three factors are at play here : 1 .
this age range happens to contain some of the biggest intermediate - age lmc clusters .
classical examples are ngc 419 , 1806 , 1846 , 1751 and 1783 , with present masses as high as @xmath57 @xcite , and which contain 69 of the 129 tp - agb stars plotted in fig . [ fig_clusters ]
. curiously enough , these are clusters that present multiple turn - offs ( probably reflecting their large total masses ) and dual red clumps @xcite .
these observed properties are clearly indicating that their turn - off masses are effectively close to , and hence that their assigned mean ages should be reasonably accurate .
[ item_clust ] 2 .
the tp - agb numbers may indeed be boosted in these @xmath58 gyr clusters , due to the effect discussed in this paper .
all the big clusters with multiple turn - offs , indeed , present a substantial superposition between their age ranges and the 1.571.66 gyr boosting period .
[ item_boost ] 3 .
the tp - agb lifetime is likely to peak at masses of @xmath59 , which broadly corresponds to turn - off ages between @xmath60 and @xmath61 gyr .
several different sets of evolutionary models more or less independently indicate this ( e.g * ? ? ?
* ; * ? ? ?
[ item_lifetime ] it is the coincidence of these three factors to be particularly insidious . were mc clusters and their total masses uniformly distributed in age ( contrary to item [ item_clust ] ) , grouping several clusters in wide age bins would be enough to smear our any short - living evolutionary effect ( as the one mentioned in item [ item_boost ] ) ; but this is apparently not feasible , because the biggest mc clusters are preferentially found in the `` worst possible '' age interval , where the agb - boosting can fully have its impact . to complicate things ,
the tp - agb lifetimes are far from negligible in this mass / age range ( cf .
item [ item_lifetime ] ) . given the concentration of big mc star clusters in the agb boosting period , estimates of the tp - agb contribution to intermediate - age stellar populations
either involving numbers / lifetimes as in @xcite , integrated luminosities as in @xcite , or integrated colours as in @xcite and @xcite may be biased to too large values .
the problem is that these very populous intermediate - age mc clusters are presently not recognized as having a boosted agb population .
their tp - agb numbers and integrated luminosities are taken as representative of age intervals much wider than the 0.1-gyr interval in which the boosting occurs .
for instance , in @xcite the clusters likely affected by the boosting period are spread into two age bins spanning the complete age interval from 0.9 to 2 gyr ( which , indeed , happen to present the reddest mean integrated colours ) .
the problem is even worse in previous works , in which the cluster age determinations of intermediate - age mc clusters were typically much coarser , so that even wider age intervals could be affected .
therefore , the agb boosting period very likely causes an overestimation of the tp - agb contribution , which later propagates into eps models of galaxies . in all this matter ,
the exact age distribution of the stars in the mc clusters ngc 419 , 1751 , 1806 , 1846 and 1783 is really critical
. it would be enough to shift the age determination of these clusters by just 10 percent , to greatly increase / decrease the level of coincidence between their tp - agb ages , and the agb - boosting period .
on the other hand , removing these clusters from the calibration sample used to constrain tp - agb models would be detrimental , since this would dramatically reduce the numbers of observed tp - agb stars at our disposal .
moreover , the remaining clusters in the @xcite catalog ( see also * ? ? ?
* ) typically contain very few tp - agb stars per cluster ( less than 5 , except for ngc 1978 which has 12 , at an age of 2 gyr ) , and hence are strongly affected by stochastic fluctuations in their star counts and integrated colors .
the broad implication of our findings is that present models calibrated on mc clusters may be significantly overestimating the tp - agb flux contribution to models of distant galaxies .
precise numerical estimates of the excess factors are beyond the scope of this paper .
anyway , the way for improving this situation is quite clear to us : \(1 ) the data for tp - agb stars in intermediate - age star clusters in the magellanic clouds should be carefully revised , in view of re - deriving the lifetimes and integrated flux of their tp - agb stars . simply making plots of
agb star counts normalized to the @xmath53-band cluster luminosity @xcite and integrated colours @xcite versus a rough estimate of cluster age , with clusters binned into wide age bins , is not enough .
it is necessary to locate the exact position of each cluster with respect to the agb - boosting period .
approximations based on the fuel - consumption theorem are also to be avoided . describing the post - ms evolution of a given age with a single track with initial mass @xmath43 , the fct is not able , by construction , to cross any evolutionary phase more than once nor , more generally , to take into account any temporary increase in the rate of production of stars in post - ms stages due to changes in the post - ms lifetimes ( as in the effect described by * ? ? ?
in order to correctly take into account the agb - boosting effect , it is absolutely necessary ( a ) to model every cluster cmd using detailed isochrones calculated from tracks with a very fine mass resolution , ( b ) to precisely identify the actual turn - off mass , and its distribution in the clusters with multiple turn - offs .
this may not be an easy task .
hst - quality photometry is needed , as well as a careful consideration of effects such as incompleteness , crowding , mass segregation , and the field contamination .
however , @xcite demonstrate that such an accurate work of model fitting is not beyond reach .
\(2 ) in the absence of a representative sample of star clusters well distributed in age and containing large numbers of tp - agb stars , the use of galaxy fields to estimate the tp - agb contribution appears legitimate at least , their more continuous star formation history ( sfh ) would ensure that they are less affected by the agb - boosting effect than the present samples of mc clusters .
the basic requirement for a quantitative work is having reliable estimates of the galaxy sfhs .
this is the case of post - starburst galaxies @xcite , and of many nearby galaxies presently sampled with deep ground - based and hst photometry ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , it is also clear that when looking at entire galaxies , one will inevitably miss many of the details of the tp - agb evolution that we would like to constrain .
needless to say , these are all steps we are going to pursue in forthcoming papers .
we warmly thank julianne dalcanton for her great encouragement to this work , stephane charlot for the useful comments , and stefano rubele and leandro kerber for showing us that a better approach to studying magellanic cloud clusters is feasible .
many thanks also to the users of our isochrones who pointed out to their `` weird features '' with genuine curiosity rather than with distrust .
we acknowledge financial support from contract asi - inaf n. i/009/10/0 , from progetto di ateneo 2012 , university of padova , i d : cpda125588/12 , and from prin 2009 .
54 natexlab#1#1 , a. , marigo , p. , girardi , l. , salasnich , b. , dal cero , c. , rubele , s. , & nanni , a. 2012 , , 427 , 127 , g. , & charlot , s. 1993 , , 405 , 538 . 2003 , , 344 , 1000 , e. , ludwig , h .-
g . , steffen , m. , freytag , b. , & bonifacio , p. 2011
, , 268 , 255 , v. , deglinnocenti , s. , girardi , l. , marconi , m. , prada moroni , p. g. , & weiss , a. 2000 , , 354 , 150 , s. , & bruzual , a. g. 1991 , , 367 , 126 , c. , & gunn , j. e. 2010 , , 712 , 833 , c. , gunn , j. e. , & white , m. 2009 , , 699 , 486 , j. j. , et al .
2012 , , 200 , 18 , p. , woo , j .- h . , kim , y .- c . , &
yi , s. k. 2004 , , 155 , 667 , i. , chieffi , a. , limongi , m. , & straniero , o. 1999 , , 524 , 226 , a. , chaboyer , b. , jevremovi , d. , kostov , v. , baron , e. , & ferguson , j. w. 2008 , , 178 , 89 , j. a. , mould , j. , & blanco , v. m. 1990 , , 352 , 96 , l. 1999 , , 308 , 818 , l. , & bertelli , g. 1998 , , 300 , 533 , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , , 141 , 371 , l. , groenewegen , m. a. t. , weiss , a. , & salaris , m. 1998 , , 301 , 149 , l. , & marigo , p. 2007
, , 462 , 237 , l. , rubele , s. , & kerber , l. 2009 , , 394 , l74 , l. , et al .
2010 , , 724 , 1030 . 2013 , , 431 , 3501 , k. , et al .
2009 , , 138 , 1403 , r. a. , bruzual - a . , g. , charlot , s. , ballesteros - paredes , j. , & loinard , l. 2010 , , 403 , 1213 , p. , puzia , t. h. , kozhurina - platais , v. , & chandar , r. 2009 , , 137 , 4988 . 2011 , , 737 , 3 , m. , held , e. v. , rizzi , l. , girardi , l. , marigo , p. , & momany , y. 2008 , , 388 , 1185 , b. d. , et al .
2013 , , 772 , 8 , m. , et al .
2010 , , 722 , l64 , c. 1998 , , 300 , 872 . 2005 , , 362 , 799 , c. , daddi , e. , renzini , a. , cimatti , a. , dickinson , m. , papovich , c. , pasquali , a. , & pirzkal , n. 2006 , , 652 , 85 , p. , bressan , a. , nanni , a. , girardi , l. , & pumo , m. l. 2013 , , 434 , 488 , p. , & girardi , l. 2007 , , 469 , 239 , p. , girardi , l. , bressan , a. , groenewegen , m. a. t. , silva , l. , & granato , g. l. 2008 , , 482 , 883 , j. , et al .
2012 , , 748 , 47 , n. e. d. , greggio , l. , renzini , a. , carollo , c. m. , & maraston , c. 2013 , , 772 , 58 , e. w. , suntzeff , n. b. , & mateo , m. 1996 , , 34 , 511 , a. , cassisi , s. , salaris , m. , & castelli , f. 2004 , , 612 , 168 , g. , udalski , a. , kubiak , m. , szymanski , m. , wozniak , p. , & zebrun , k. 1998 , acta astronomica , 48 , 175 , o. r. , schrder , k .-
p . , hurley , j. r. , tout , c. a. , & eggleton , p. p. 1998
, , 298 , 525 , a. , & buzzoni , a. 1986 , in astrophysics and space science library , vol .
122 , spectral evolution of galaxies , ed . c. chiosi & a. renzini , 195231 , r. m. , shara , m. , fall , s. m. , & zurek , d. 2000 , , 119 , 197 , s. , girardi , l. , kozhurina - platais , v. , goudfrooij , p. , & kerber , l. 2011 , , 414 , 2204 , s. , girardi , l. , kozhurina - platais , v. , kerber , l. , goudfrooij , p. , bressan , a. , & marigo , p. 2013
, , 430 , 2774 , s. , kerber , l. , & girardi , l. 2010 , , 403 , 1156 , e. e. 1955 , , 121 , 161 , d. , et al .
2013 , , 765 , l41 , a. v. , greggio , l. , & renzini , a. 1989 , , 69 , 911 . 1990 , , 364 , 527 , b. l. , et al .
2013 , , 554 , a33 , j. t. , marshall , j. r. , & zijlstra , a. a. 2005 , , 442 , 597 , a. , & ferguson , j. w. 2009 , , 508 , 1343 , d. r. , et al .
2011 , , 739 , 5 , s. , gallazzi , a. , charlot , s. , pierini , d. , & pasquali , a. 2013 , , 428 , 1479
at the request of the referee , we provide here additional details about our models , and comparisons with other authors .
they are mostly aimed at testing the reliability of our results , especially in regard to the sharpness of the transition in evolutionary features taking place at @xmath62 . before proceeding
, we should emphasize that our models provide _ by far _
the best mass resolution available in the literature .
any comparison with other authors is very much limited by this fact .
evolutionary tracks , as a function of initial mass .
the crosses represent the values for tracks actually computed , which are then connected by straight lines . *
bottom panel : * the total lifetimes in the rgb ( red ) and cheb ( blue ) phases as a function of mass . *
middle panel : * the maximum luminosities reached at the stage of he - ignition , denoted as @xmath63 ( red ) , and minimum luminosities reached at the stage of quiescent he - burning , denoted as @xmath64 ( blue ) .
in addition , the luminosities provided by the he - burning core at different stages of the cheb evolution are plotted with dotted lines ( for @xmath65 , 0.7 , 0.5 , 0.3 , and 0.1 , from bottom to top ) . *
top panel : * the mass of the h - exhausted core at the beginning of the cheb phase ( red ) .
the dotted lines denote the increased core mass at successive stages of the cheb evolution ( for @xmath66 decreasing from 0.95 to 0.1 from bottom to top , as before ) . ]
as already discussed , the agb - boosting owns its origin to the sharp increase in cheb burning lifetimes at masses close to .
this increase is further illustrated in the bottom panel of fig .
[ fig_ourlifetimes ] , which details how the relevant lifetimes change with the stellar initial mass .
more specifically , we plot the run of the `` rgb lifetime '' , @xmath67 , measuread as the time span between the h - exhaustion in the core ( @xmath68 ) and the maximum luminosity reached soon after he - ignition , and the run of the he - burning lifetime , @xmath69 , measured as the time span between the he - ignition and the he - exhaustion in the core ( @xmath70 ) .
it can be seen that both quantities gradually change as approaches the limit , but the change becomes rather abrupt as is crossed .
similar changes occur also in the maximum and minimum luminosities reached at the initial stages of he - burning , denoted as @xmath63 and @xmath64 .
it is remarkable that the quantities at the rgb tip , @xmath67 and @xmath63 ( both marked in red ) , are derived from stellar models computed on a consistent way from the pre - main sequence up to the he - ignition , therefore they are _ not affected _ by the artificial method performed to build quiescent he - burning models of masses @xmath16 .
the same applies to the core mass at he - ignition , which is measured at a stage in which it is still not altered by the he - burning .
nonetheless , all models indicate the same discontinuity at @xmath71 , namely a sharp decrease in both @xmath67 , @xmath63 , and @xmath72 . this sharp decrease is clearly indicating a discontinuity caused by the stellar physics , rather than by the numerical algorithms adopted to skip the he - flash in low - mass models . and
indeed , the diagram of central temperature vs. central density shown in fig . [ fig_bifu ] shows a clear bifurcation in the evolutionary paths , that sharply separates the models with @xmath74 from those with @xmath75 .
the former ignite helium in non - degenerate conditions , while the latter decisely cool at increasing density , delaying he - ignition to a stage further along in the rgb .
this bifurcation has been found in numerous sets of published stellar tracks in the past .
what is unusual is just to show it with a mass resolution as small as @xmath11 , as in our case .
after he - ignition , low - mass models have their evolution interrupted and continued from a zero - age - horizontal branch model built with a suitable mass and chemical composition ( see sect . [ models ] ) , while intermediate - mass models are continuously computed until the tp - agb phase .
this dichotomy in the way models are evolved could create additional differences between the two mass ranges ( in addition to those caused by the onset of electron degeneracy ) , because the convective regions developed in the core during he - ignition could have been treated differently in the two cases . especially worrying would be if the overshooting scheme adopted in the tracks were to extend the convective cores in the models just massive enough to survive the he - flash . however ,
inspection of the models shows that the convective cores of low - mass cheb models are larger than those of he - burning stars above the @xmath71 discontinuity ( top panel of fig .
[ fig_ourlifetimes ] ) .
this is simply due to the larger h - exhausted cores of the former models , which are a common feature , irrespective of the adopted efficiency of core overshoot and of the way they have been computed . in spite of that , in models with mass just above the discontinuity the he - burning lifetimes is about twice that of the @xmath14 models .
again this feature is common to all models irrespective of the efficiency of core overshooting and due to the lower he - burning luminosity which is needed to sustain a smaller he core , i.e. to the he - core mass luminosity relation .
the latter can be appreciated by looking at the luminosity coming from he - burning reactions , @xmath76 , at several stages of the cheb evolution ( middle panel of fig .
[ fig_ourlifetimes ] ) .
@xmath76 is indeed dramatically smaller for the cheb models with the core mass close to the @xmath77 minimum , soon after he - ignition ( as for the curves with @xmath78 and @xmath79 ) . at later stages of cheb ,
the core masses grow and the differences between the @xmath76 values between tracks of varying masses become smaller ( as seen in the top and middle panels of fig .
[ fig_ourlifetimes ] ) ,
but anyway the very different @xmath76 at the start of the cheb ensure the much longer lifetimes of the models slightly above @xmath17 .
thus , no matter the way the models have been computed , a discontinuity in the he core masses gives rise to a discontinuity in the he - burning lifetimes , with models possessing smaller cores having significantly prolonged @xmath69 .
; the core mass at he - ignition , @xmath72 ; the maximum luminosity soon after he - ignition , @xmath80 ; the total rgb plus subgiant branch lifetime , @xmath67 ; and the total core he - burning lifetime , @xmath81 .
the models are taken from ( * ? ? ?
* sgr 89 ) ( * ? ? ?
* pshte 98 ) ( * ? ? ?
* dcls 99 ) ( * ? ? ?
* basti ) ( * ? ? ?
* y@xmath82 ) ( * ? ? ?
* dsep ) ( * ? ? ?
* garstec ) , and from this work ( parsec ) . ]
the abrupt changes occurring in the evolutionary features and lifetimes in the vicinity of have been already discussed by many authors , using grids of stellar models computed with different input physics and various degrees of completion and mass resolution . some of these models , chosen among the most complete available in the literature , and with initial chemical composition similar to the lmc one , are compared to ours in the different panels of fig .
[ fig_otherlifetimes ] .
the comparison is performed as a function of @xmath83 , because convective overshooting has the effect of systematically increasing the core masses and hence of reducing the value of ( see e.g. * ? ? ?
in other words , we will be using @xmath83 as a proxy for the core mass developed after the main sequence , so that we can directly compare at least in the vicinity of models which naturally develop very different core masses .
since the definition of might slightly differ from author to author , we simply redefine for each set , as being the mass value that presents the fastest variation of the quantity @xmath63 with mass .
@xcite were the first to explore the issue of how fast is the transition in evolutionary features across , in the context of the long - sought `` rgb phase transition '' defined by @xcite .
they computed a series of canonical ( without overshooting ) models at a typical separation of @xmath84 , but only up to the he - ignition .
it can be seen that the behaviour of the present models is quite similar to those of @xcite , apart from modest offsets in the plotted luminosities and lifetimes , and for @xcite models suggesting a more gradual variation of the stellar properties for masses above @xmath12 .
the offsets in luminosities and lifetimes are no surprise , given the differences in the initial chemical composition , and the large changes in the physical input adopted by stellar models in the last two decades .
particularly interesting is the comparison presented at the top panel , which presents the maximum mass coordinate of the maximum temperature reached inside the star before the he - flash , @xmath85 .
a value larger than zero is usually interpreted as the signature of an electron - degenerate core , with its temperature inversion caused by the efficient conduction and by neutrino cooling .
the mass at which @xmath85 becomes null has sometimes been used as a definition for ( e.g. * ? ? ?
it is interesting to note that in our models @xmath85 becomes null already at @xmath86 , which is at least 0.03 smaller than the mass value for which the evolutionary features more rapidly change , .
such a distinction between limiting masses could not have been detected in @xcite models , given their @xmath84 resolution .
apart from this fine detail , the run of @xmath85 with mass is surprisingly similar between these two sets of models .
the behaviour of @xcite models seem to be confirmed by those from @xcite and @xcite , although the latter are computed with a slightly worse resolution in mass . on the other hand , @xcite and
@xcite compute the complete he - burning evolution for models below and above the @xmath12 limit , finding that the cheb lifetime more than doubles for stars slightly more massive than @xmath12 as also found in our models .
very similar behaviour is also presented by the garstec tracks from @xcite .
also suggestive is the comparison between our models and some recent ones such as the yale - yonsei @xcite , basti @xcite , and dsep @xcite .
these sets of models predict a quite similar evolution of the quantities at the he - ignition , although , in all these cases , the finest details of the transition at are missed because of the limited mass resolution ( of about @xmath84 in the best case ) . in the case of basti models
, the comparison can be extended to the he - burning luminosities and lifetimes , which compare quite well . in basti ,
the longest cheb burning lifetimes are found just a few hundredths of solar masses above although this trend is not very clear give the scarcity of tracks computed in the immediate vicinity of .
another interesting point is that the basti tracks present the same trend of presenting the smallest luminosities at he - ignition in the mass interval immediately above .
this behaviour is probably reflecting the minimum in the core mass at he - ignition , which is also found in the same interval .
similar trends are also observed in the dsep models , which behave in a way very similar to basti ones .
unfortunately , dsep does not contain he - burning models in the immediate viciniy of , and largely fail to sample the mass interval for which he - burning lifetimes are expected to exceed 100 myr .
overall , we find a good level of agreement between the behaviour of our models and others found in the literature , despite the great differences in parameters such as the overshooting efficiency , initial metallicity and helium content . since our models have by far the best mass resolution among these different sets , and are the most complete in the calculation of he - burning phases , it is natural that they provide the most detailed description of the @xmath87 interval .
as surprising as it can be , they also provide the sharpest transition in the stellar evolutionary features across this mass interval .
|
in the recent controversy about the role of tp - agb stars in evolutionary population synthesis ( eps ) models of galaxies , one particular aspect is puzzling : tp - agb models aimed at reproducing the lifetimes and integrated fluxes of the tp - agb phase in magellanic cloud ( mc ) clusters , when incorporated into eps models , are found to overestimate , to various extents , the tp - agb contribution in resolved star counts and integrated spectra of galaxies . in this paper , we call attention to a particular evolutionary aspect , linked to the physics of stellar interiors , that in all probability is the main cause of this conundrum . as soon as stellar populations intercept the ages at which rgb stars first appear , a sudden and abrupt change in the lifetime of the core he - burning phase causes a temporary `` boost '' in the production rate of subsequent evolutionary phases , including the tp - agb . for a timespan of about 0.1 gyr ,
triple tp - agb branches develop at slightly different initial masses , causing their frequency and contribution to the integrated luminosity of the stellar population to increase by a factor of @xmath0 .
the boost occurs for turn - off masses of @xmath1 , just in the proximity of the expected peak in the tp - agb lifetimes ( for mc metallicities ) , and for ages of @xmath2 gyr .
coincidently , this relatively narrow age interval happens to contain the few very massive mc clusters that host most of the tp - agb stars used to constrain stellar evolution and eps models .
this concomitance makes the agb - boosting particularly insidious in the context of present eps models .
as we discuss in this paper , the identification of this evolutionary effect brings about three main consequences .
first , we claim that present estimates of the tp - agb contribution to the integrated light of galaxies derived from mc clusters , are biased towards too large values .
second , the relative tp - agb contribution of single - burst populations falling in this critical age range can not be accurately derived by approximations such as the fuel consumption theorem , that ignore , by construction , the above evolutionary effect .
third , a careful revision of agb star populations in intermediate - age mc clusters is urgently demanded , promisingly with the aid of detailed sets of stellar isochrones .
|
transitional disks are protoplanetary disks with heavily depleted gaps or cavities in their inner regions ( e.g. * ? ? ?
* ; * ? ? ?
they are thought to represent a transitional state in disk evolution , in which the gas - rich primordial disk gradually disperses due to planet formation or other effects ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
one transitional disk target that has garnered particular attention in recent years is , a 35 myr old @xmath01 @xmath1 star located approximately 140 pc away @xcite in the taurus - auriga region .
a wide gap in the disk extending out to @xmath050au can be inferred from the infrared spectral energy distribution ( sed ; @xcite ) , but it has also been spatially resolved both in scattered light at near - infrared wavelengths @xcite and in thermal emission at mm / sub - mm wavelengths @xcite . in near - infrared images , the outer edge of the gap appears off - center from the star , suggesting an eccentric gap @xcite , although the mm / sub - mm images are consistent with a circular gap @xcite .
in addition , @xcite discovered a protoplanet candidate at @xmath01621au using sparse aperture masking ( sam ) . while some follow - up observations only provided upper limits @xcite ,
newest results from h@xmath2 differential imaging confirm the accreting protoplanet hypothesis ( k. follette , p.c . ) . here
we present new scattered - light images of the lkca 15 disk taken at visible wavelengths using polarimetric differential imaging ( pdi ) .
the observations provide better spatial resolution and a more robust determination of the distribution of reflected light from the disk than previous data based on infrared angular differential imaging ( adi ) , in which forward modeling is necessary to account for self - subtraction effects @xcite .
lkca 15 was observed on two nights during science verification time for the newly commissioned high - contrast imager sphere ( spectro - polarimetric high - contrast exoplanet research ; @xcite ) at the european southern observatory s very large telescope facility .
its visible - light instrument , zimpol ( zurich imaging polarimeter , @xcite ) , was used with the extreme adaptive optics system ( saxo ; @xcite ) to obtain deep imaging polarimetry in the ` very broad band ' filter spanning the @xmath3 and @xmath4 bands ( 590890 nm ) .
zimpol offers extreme polarimetric contrast @xcite at a pixel scale of [email protected] and a field of view of 35@xmath535 . the first observation , on 2015 - 02 - 02 , comprised seven polarimetric cycles of four half - wave plate positions ( stokes @xmath6 , @xmath7 , @xmath8 , @xmath9 ) , each exposed for 8 @xmath5 20 seconds at a time , for a total integration time of 75 minutes . the second observation was executed in the same way on 2015 - 02 - 12 .
the exposure time had been chosen to saturate the star s point - spread function ( psf ) to maximize sensitivity at larger separations . however , the performance of the adaptive optics was poor and highly variable throughout both observations , which rendered the star unsaturated in most frames and yielded a full width at half maximum ( fwhm ) of @xmath035mas ( @xmath10 ) .
this can be ascribed to lkca 15 s apparent magnitude of @xmath11mag at the faint limit of saxo s guide star specifications , combined with unfavorable weather conditions and high airmass ( @xmath01.5 ) .
furthermore , a `` comet - tail '' aberration of the psf core is observed in some frames , likely caused by malfunction of the atmospheric dispersion corrector ( adc ) . after correcting the raw zimpol data for cosmetics and field distortion , we applied the pdi data reduction method detailed in @xcite .
it includes the following steps for reducing instrumental polarization : ( 1 ) flux equalization of each simultaneous pair of raw images in order to remove the polarization of the stellar psf , ( 2 ) double - quotient calculation of stokes @xmath12 and @xmath13 polarized flux images for each polarimetric cycle , ( 3 ) empirical correction of polarimetric throughput differences in stokes @xmath12 and @xmath13 , and ( 4 ) empirical correction of angular misalignment of the half - wave plate .
lastly , we transform the cartesian coordinate system of the stokes formalism @xmath14 into a polar coordinate system @xmath15 in which positive @xmath16 polarization is defined as azimuthal with respect to the star @xcite . under the single - scattering assumption
, the scattering polarization should appear in the @xmath16 image ( typically as positive signal ) , whereas the @xmath17 image should contain no scattered - light signal and can be used to estimate the noise level in the observation @xcite . while this assumption does not strictly hold for inclined , optically thick disks , as is the case for lkca 15 , the resulting deviations are small and automatically accounted for in the error estimation @xcite .
since we find no significant differences between the two epochs , we combine both data sets into one final set of stokes @xmath18 images , which improves the quality of the results . since the data are sensitivity - limited , we apply no frame selection .
we convolve the polarization images with a circular aperture of 7 pixels (= 25mas ) in diameter to reduce shot noise .
and @xmath17 polarized flux images , respectively . assuming single scattering
, the disk appears as positive flux in the @xmath16 image , while the @xmath17 image provides a noise estimate .
all images are convolved with a 7-pixel (= 25mas ) diameter circular aperture to reduce shot noise .
the small insets show the shape of the psf core .
the location of the star is marked with black - and - white circles .
the position angle of the north direction is @xmath19 .
the black triangles indicate the zero level in the color scales . *
( a ) * field of view of 18 at a linear stretch with a maximum of @xmath0@xmath20 times the brightest pixel of the stellar psf ( which is lightly saturated in some of the exposures ) .
the known crescent of scattered light from lkca 15 s outer disk is visible , along with a newly discovered inner disk component close to the star , saturating the color scale .
the white corners delimit the area used for ( b ) . *
( b ) * the central 06 shown at 3x magnification at a linear stretch with a maximum of @xmath0@xmath21 times the brightest pixel of the stellar psf , showing an unsaturated view of the inner disk component .
the crosshairs indicate the reported @xmath22-band location of lkca 15 b in @xcite .
the dashed circle marks the noise - dominated central area . * ( c ) * signal - to - noise ( s / n ) maps at a stretch of @xmath23\sigma$ ] . the same noise map is used for both the @xmath16 and @xmath17 images .
both the outer disk and the new inner disk component are detected at high significance . *
( d ) * the 18 field of view with each pixel multiplied by the square of its de - projected distance from the star , assuming an inclination of 50@xmath24 . ]
our @xmath25-band imaging polarimetry of lkca 15 is shown in fig .
[ f : images ] . despite the bad psf quality
, the instantaneous nature of dual - beam polarimetry allowed us to remove the starlight effectively and resolve the lkca 15 disk at high contrast and unprecedented angular resolution .
the @xmath16 image reveals a crescent of polarized flux tracing the shape of the near - side edge of the outer disk seen in infrared angular adi observations @xcite , and a previously unknown bright area of polarized flux surrounding the star at small separations ( @xmath26 , see panel [ f : images]b ) .
both features appear consistently positive in the @xmath16 image and are significantly brighter than the residuals in the @xmath17 image at similar radii , which identifies them as scattered light from circumstellar material .
[ f : images]c provides signal - to - noise ( s / n ) maps produced by dividing the pixel values of star - centered annuli in the @xmath16 and @xmath17 images by the standard deviation of the @xmath17 image in the same annuli .
both features achieve s / n values above @xmath27 over a large number of resolution elements and are therefore detected at high confidence .
the inner disk is detected down to a separation of @xmath01fwhm ( 36 mas ) . to estimate the dust distribution from the scattered - light intensity , we compensate for the @xmath28 fall - off of the illumination of dust particles with physical distance @xmath29 from the star . we calculate @xmath29 for each pixel by deprojecting its apparent distance from the star , assuming its light originates from a plane inclined by 50@xmath24 with its line of nodes at a position angle of 60@xmath24 ( cf .
[ f : images]d shows the @xmath16 and @xmath17 images after multiplying each pixel value with @xmath30 .
the disk gap is visible as an abrupt step in brightness at almost all position angles , including the faint far side that had eluded all previous scattered - light observations .
the reduced brightness of the disk surface near the minor axis is likely due to the dependence of scattering polarization on the phase angle .
however , we note that phase angles should change by at most @xmath06@xmath24 along radial paths on the visible disk surface ; thus , the location of the gap edge should remain largely unaffected . the eastern ansa appears brighter than the western one on the 1@xmath31 level ( evaluated in the @xmath30-scaled image in 50-pixel diameter apertures ) . -scaled @xmath16 image as in fig .
[ f : images]d displayed at a harder stretch , whereas the light blue ellipse represents our fit to its gap ( see text ) .
the location of the star is marked with a star symbol . * ( a ) * visual confirmation of fit for the blue ellipse . * ( b ) * modified maximum merit fit to the best - fit model from the adi - based forward - modeling analysis in @xcite overplotted in yellow .
the ellipse is registered against the underlying image by its star position . * ( c ) * fit ellipse ( green ) to the gap edge in 7 mm emission as reported in fig . 1 of @xcite , registered by the assumed star position .
* ( d ) * the same ellipse as in ( c ) aligned with our blue fit ellipse , demonstrating that the shapes match well .
the small green circle marks the true position of the star relative to the green ellipse . ]
[ f : ellipses ] l|@r@(r@ , r@)r@(r@ , r@ ) + & & + semimajor axis @xmath32 & 338 & @xmath33 & @xmath34 & 47.4 & @xmath35 & @xmath36 + semiminor axis @xmath37 & 245 & @xmath38 & @xmath39 & 34.3 & @xmath40 & @xmath41 + major - axis offset @xmath42 & @xmath43 & @xmath33 & @xmath44 & @xmath45 & @xmath35 & @xmath46 + minor - axis offset @xmath47 & @xmath48 & @xmath33 & @xmath34 & @xmath49 & @xmath35 & @xmath36 + + & & + major axis ( east ) @xmath50 & 69 & @xmath51 & @xmath39 & 9.7 & @xmath52 & @xmath53@xmath54 + major axis ( west ) @xmath55 & 74 & @xmath33 & @xmath56 & 10.4 & @xmath57 & @xmath53@xmath58 + minor axis ( north ) @xmath59 & 51 & @xmath60 & @xmath61 & 7.2 & @xmath62 & @xmath63 + minor axis ( south ) @xmath64 & 38 & @xmath65 & @xmath66 & 5.3 & @xmath57 & @xmath67 + [ 1 mm ] adi observations of circumstellar disks suffer from flux loss and oversubtraction effects @xcite and therefore require forward modeling to interpret their results @xcite .
imaging polarimetry , on the other hand , yields morphologically sound representations of the disk s appearance in polarized light ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we characterize the sharp gap edge in the @xmath30-scaled @xmath16 image by fitting it with an ellipse using a modified version of the maximum merit method introduced in @xcite .
we generate a number of ellipses with semimajor and -minor axes @xmath32 , @xmath37 displaced from the star s position by offsets @xmath42 , @xmath47 along the two axes , respectively . for a parameter set @xmath68
, we consider the two concentric elliptical annuli bounded by the triplet of ellipses @xmath69 , @xmath68 , and @xmath70 with an annulus width of @xmath71 pixels (= 36 mas = 1 fwhm ) .
the merit function @xmath72 is defined as the mean pixel value in the outer annulus of the @xmath30-scaled @xmath16 image minus the mean value of the inner annulus .
the best fit is achieved when the merit function is maximized , i.e.,where the mean brightness increase from the inner to the outer annulus is greatest . as a measure of uncertainty , we define the `` well - fitting '' family of ellipses by a threshold of @xmath73 , where @xmath74 is the standard deviation of merits of ellipses in the @xmath17 image . as demonstrated in fig .
[ f : ellipses]a , the best - fit ellipse matches the visual locus of the gap edge convincingly .
its numerical parameters and uncertainties are provided in table [ t : num ] .
figure [ f : ellipses]b compares this fit to a second ellipse obtained by applying the maximum merit method to the best - fit model from our adi forward - modeling analysis in @xcite .
the two ellipses share similar sizes , aspect ratios , and significant minor - axis offsets @xmath47 .
however , while we found a marginal major - axis offset @xmath42 towards the west in @xcite , our present data are consistent with @xmath75 .
[ f : ellipses]c shows the ellipse fit to the reconstructed 7 mm emission images reported in @xcite superimposed on our fit .
their ellipse is centered on the star , which results in a significant mismatch with our off - centered ellipse . while the two ellipses can be brought to coincide well with each other by translating one of them by @xmath76mas ( fig .
[ f : ellipses]d ) , this offset is highly significant compared to the astrometric accuracy of the 7 mm interferometry ( 10mas , @xcite ) and of the zimpol data ( @xmath77mas ) and must therefore be physical .
image along the major axis averaged over a width of 21 pixels (= 76 mas ) centered on the star ( thick black curve ) .
the family of thin grey curves are obtained by adding similar profiles taken in the @xmath17 image at various position angles to the black curve , and therefore visualize the noise level in the data .
the contrast of a given pixel is calculated by the ratio of its value and the maximum pixel in the stellar full - intensity psf .
the normalization is approximate due to the stellar psf being saturated in some exposures .
the blue dashed lines mark the location of the light blue fit ellipse in fig .
[ f : ellipses ] .
the red lines are exponential fits to the inner scattering component .
positive separation values correspond to the western side of the disk .
the size of the psf s fwhm is given for comparison . * ( b ) * profile of the @xmath16 image along the minor axis averaged over a width of 81 pixels (= 292 mas ) centered on the star .
positive separation values correspond to the northern side of the disk . ]
[ f : profiles ] figures [ f : profiles]a and [ f : profiles]b plot the surface brightness profiles in the @xmath16 image ( without @xmath30 scaling ) along the major and the minor axis , respectively .
the plotted values are pixel values as a function of separation averaged over a width of 21 pixels (= 76mas ) and 81 pixels (= 292mas ) , respectively , and normalized by the brightest pixel value of the star s intensity psf .
since some exposures in the data are saturated , the normalization is only approximate .
the new inner disk component appears as a roughly exponential rise of surface brightness within a radius of @xmath0025 .
we fit each side of the structure in each of the two profiles as proportional to @xmath78 , with @xmath79 the angular separation from the star and @xmath80 a characteristic length scale , using least - squares fitting .
these fits are listed in table [ t : num ] and illustrated in fig .
[ f : profiles ] . to gauge the uncertainty of these results ,
we extract a number of surface brightness profiles from the noise - like @xmath17 image at various position angles , add each of them to the science profiles from the @xmath16 image , and recalculate the exponential fits .
the standard deviations of these `` disturbed '' @xmath80 values are then adopted as the error bars for the table values of @xmath80 .
the disturbed surface brightness profiles are shown in grey in fig .
[ f : profiles ] .
the polarized intensity is likely underestimated at the innermost separations due to overlapping psf wings in the @xmath12 and @xmath13 images @xcite .
the newly discovered inner disk component , detected from @xmath030au down to the inner limit of @xmath07au ( fig .
[ f : profiles ] ) , overlaps with the location of the protoplanet candidate at @xmath01621au derived from sam observations @xcite . as with the outer disk , we expect anisotropic scattering to produce a distinct full - intensity forward - scattering maximum on the near side , which coincides with the position angle of the protoplanet candidate
. taken together with the forward - scattering maximum of the outer disk edge @xcite , these disk features likely interfere with the reconstruction of sam data , as in the case of t cha @xcite .
however , given the recent h@xmath2 detection ( k. follette , p.c . ) , a protoplanet embedded in the inner disk similar to hd100546 b @xcite remains the most plausible interpretation .
the inner disk component also sheds light on the extent of the gap as inferred from previous non - resolved observations . proposed a three - component structure with an outer disk at @xmath050au consistent with the millimeter cavity , an inner disk starting at the dust sublimation radius to explain the near - infrared flux , and an optically thin component out to @xmath05au to explain the prominent silicate emission feature .
our detection of scattered light between 730au shows illuminated dust is present at these locations above the disk mid plane .
an extrapolation of the gap wall temperature ( @xmath81k@xmath82 ) yields a temperature of 150250k , consistent with the presence of the 10 and 20@xmath83 m silicate emission feature in the sed .
we hypothesize that this new scattering structure represents the optically thin part of the inner disk extending out to at least 30au .
a quick adaptation of the disk model from @xcite indicates that this can be reconciled with the sed , since the contribution of the optically thin part to the sed is minor .
this geometry ( narrow gap , depleted inner disk , millimeter cavity ) qualitatively matches the imprint of a single planet on a disk ( e.g. * ? ? ?
* ; * ? ? ?
* ) .
our @xmath25-band polarimetry confirms the prediction from @xmath22-band adi forward - modeling @xcite that the outer boundary of the gap in the lkca 15 disk appears off - centered in reflected light , with a highly significant offset of @xmath76mas ( a third of the gap s projected radius ) between the star and the gap s apparent center .
while disk gaps are known to vary in size as a function of observed wavelength in systems with planet - induced dust filtering @xcite , a lateral displacement of the gap is more difficult to justify and has not been predicted by numerical simulations . the outer disk s vertical structure can not account for this displacement since the gap wall profile is tapered ( ` round ' ) rather than vertical @xcite .
this is qualitatively confirmed by the wide , homogeneous distribution of scattered light in fig .
[ f : images]d .
another explanation is provided by shadowing .
originally proposed an inner disk coplanar with the outer disk , casting a shadow onto the outer disk s inner wall @xcite .
our forward modeling in @xcite achieved significantly better fits with an unshadowed gap wall , which led us to speculate that the inner disk and its shadow were tilted out of the outer disk s plane . on the other hand , this arrangement should result in dark lanes cutting across the outer disk where the shadow plane intersects it , as seen in hd 142527 @xcite .
no such lanes are apparent in the lkca 15 imagery .
this conundrum can be resolved if the inner disk has a slightly higher inclination than the outer disk ( fig .
[ f : sketch ] ) .
the near - side disk rim as observed from earth would then be fully exposed to stellar irradiation , whereas the far - side disk surface would be obscured at the rim and only become visible further out where the disk surface flares out of the shadow .
this would create an off - centered apparent gap in reflected light even with a symmetric physical disk gap .
for the disk architecture in @xcite , the requisite differential tilt would be of order 1@xmath24 , which could be induced by a planet on a tilted orbit ( cf .
@xcite ) . a similar effect could be caused by a shadowing structure vertically extending from an otherwise coplanar inner disk @xcite , though this would imply dramatic variability on the timescale from days to years .
while we can not exclude this possibility , we find no significant variability in our observations ( @xcite ; this work ) .
we thank a. isella and the anonymous referee for helpful correspondence .
this work has been carried out within the frame of the national centre for competence in research planets supported by the swiss national science foundation .
ct , spq , hms , mrm acknowledge support from snsf .
ct is supported by the european commission under marie curie ief grant no .
jo acknowledges support from the millennium nucleus rc130007 ( chilean ministry of economy ) .
jc was supported by nsf under award 1009203 .
|
we present the first optical ( 590890 nm ) imaging polarimetry observations of the pre - transitional protoplanetary disk around the young solar analog lkca 15 , addressing a number of open questions raised by previous studies .
we detect the previously unseen far side of the disk gap , confirm the highly off - centered scattered - light gap shape that was postulated from near - infrared imaging , at odds with the symmetric gap inferred from millimeter interferometry .
furthermore , we resolve the inner disk for the first time and trace it out to 30au .
this new source of scattered light may contribute to the near - infrared interferometric signal attributed to the protoplanet candidate lkca 15 b , which lies embedded in the outer regions of the inner disk .
finally , we present a new model for the system architecture of lkca 15 that ties these new findings together . these observations were taken during science verification of sphere zimpol and demonstrate this facility s performance for faint guide stars under adverse observing conditions .
= 1
|
in paper @xcite we have proposed a method for distinguishing the charges of leptoquarks collisions is considered . ] ( lq ) produced in the reaction @xmath1 the idea is based on the radiative amplitude zero ( raz ) effect , i.e. the absence of photon emission at some directions depending on the leptoquark electric charges . in this paper we give a detailed numerical analysis of this method applying it to hera and possible lep+lhc experiments . in ref .
@xcite we have derived the analytical formulas for the unpolarized cross sections of the hard subprocesses @xmath2 we have shown that they are proportional to a factor @xmath3 .
so , at some photon emission angles where this factor vanishes the cross sections have exact zeros whose positions are completely determined by the electron ( positron ) and quark charges , @xmath4 and @xmath5 , respectively . of course in reality the leptoquarks decay into two fermions .
considering the electron decay channel we have got the following clear signature : the hard photon , the electron ( or positron ) and a hard ( quark ) jet with complete kinematical reconstruction of the subprocess : @xmath6 in this paper we analyse numerically the manifestations of the raz effect for the case of this three - body final state subprocess .
we make some quantitative conclusions on a possibility to distinguish the charges of leptoquarks produced in @xmath0 collisions taking account of the proton structure functions .
our calculations were made with the help of the comphep package @xcite created for automatic calculation of cross sections at tree level in the standard model and beyond .
for the numerical integration over three - particle phase space and for event generation we have used the bases package @xcite .
the amplitude of subprocess @xmath7 is represented by the feynman diagrams displayed in figs . 1 , 2 . in fig .
1 the diagrams corresponding to the signal lq production process ( 2 ) are represented .
the diagrams of fig . 2 correspond to the sm contribution : deep inelastic scattering associated with hard photon emission . using the comphep package we have calculated analytically the unpolarized squared matrix elements for subprocesses ( [ eq : iiii ] ) .
but in contrast to our previous paper @xcite we have introduced finite widths into the lq propagators .
then we have evaluated numerically the contribution of separate ( quark ) constituents to the integrated cross sections by the convolution of the subprocess cross section with the corresponding parton distribution function : @xmath8 here @xmath9 is the element of 3-body phase space ; @xmath10 is the cross section of subprocess ( [ eq : iiii ] ) ; @xmath11 is the squared cms energy of the electron - proton system ; the squared cms energy of the hard subprocess equals @xmath12 ; the quark distribution function is denoted by @xmath13 and the 4-momentum transfer scale is taken to be @xmath14 .
the photon emission angle is denoted by @xmath15 .
the direction @xmath16 is along the proton beam .
as the lower bound @xmath17 we use the value @xmath18 to take into account the quasi - resonant peak , where @xmath19 is the lq mass .
the function @xmath20 corresponds to the kinematical cuts which we impose to get a realistic distribution and values of cross sections . in particular
this function has to include cuts on the photon energy and on the angles between the photon and all charged particles in ( 2 ) to remove infrared and collinear divergences .
it is also necessary to apply cuts for a reliable separation of the hard jet .
all the fermion masses are put equal to zero in our calculations .
the numerical analysis of the corresponding cross sections is given for two cases : * hera : @xmath21 , electron beam energy @xmath22 , integrated luminosity @xmath23 ; * lep+lhc : @xmath24 , @xmath25 gev , annual integrated luminosity @xmath26 .
one of the general restrictions on the lq - fermion interaction is the chirality of the fermion - lq coupling ( see @xcite and references therein ) .
therefore we consider only either couplings with left - handed leptons or with right - handed ones . in both cases we use in the numerical analysis
an `` electroweak '' value for this constant , @xmath27 . as to lq mass we analyse cross sections in the following ranges * hera : @xmath28 ; * lep+lhc : @xmath29 . for the parton densities we used the parametrizations steq2p @xcite and mrs - a @xcite , which take account of recent hera data .
both parametrizations gave the same results within calculation error .
let us describe in more detail the cuts which we apply in the laboratory frame . for a proper background analysis
we introduce kinematical cuts for this reaction in such a way as to suppress contributions from the standard model feynman diagrams ( see fig .
2 ) . recall that for the problem under consideration the lq mass is supposed to be already known from resonant lq production .
so we can introduce a narrow cut on the invariant mass of the outgoing electron ( or positron ) and the ( quark ) jet : @xmath30 where @xmath31 is the lq width .
we have calculated @xmath31 at tree level for each value of the lq mass .
we introduce also a cut on the angle between the outgoing electron ( positron ) and the quark : @xmath32 . in the case of lq decay
the products are moving back - to - back in the lq rest frame , so this cut suppresses the standard model contributions and does not change the lq signal .
then we have to introduce cuts on the energies and emission angles of the outgoing electron ( positron ) and quark to exclude the forward and backward cones and various soft contributions : @xmath33 we also apply cuts on the photon energy and emission angle : @xmath34 , @xmath35 .
finally we introduce cuts on the angles between the outgoing photon and the electron ( positron ) and quark : @xmath36 . these cuts remove the collinear divergences . for this set of cuts and with @xmath37
the cross sections of reaction ( [ eq : iiii ] ) calculated in the standard model are : [ cols= " < , < " , ]
* feynman diagrams with lq signal . * feynman diagrams for subprocesses @xmath38 in sm .
* distributions in @xmath39 for @xmath40 with the @xmath41 leptoquark contribution .
these distributions are obtained at @xmath42 for hera ( @xmath43 ) and at @xmath44 for lep+lhc ( @xmath45 ) with kinematical cuts applied . * cross sections in dependence on the lq mass at hera . * cross sections in dependence on the lq mass at lep+lhc .
* distributions in the photon emission angle @xmath15 for valence quarks ( @xmath46 for hera and @xmath47 for lep+lhc ) in the case of scalar lq .
|
we analyse numerically manifestations of the radiative amplitude zero ( raz ) effect in single leptoquark production associated with hard photon emission .
we present some quantitative conclusions on the possibility to distinguish leptoquark charges produced in @xmath0 collisions taking account of three - body final state subprocesses and of proton structure functions . applying this method to hera and possible lep+lhc experiments we show that the raz analysis can serve as a tool to determine the leptoquark electric charge up to large leptoquark masses .
-2 cm -1 cm inp msu preprint-95 - 12/376 + qmw - ph-95 - 13 + hep - ph / yymmnn + 17-may-1995 * * v. ilyin , a. pukhov , v. savrin , a. semenov , + _ institute of nuclear physics of moscow state university , 119899 moscow , russia _ + and + w. von schlippe + _ queen mary & westfield college , london , england _
( to be submitted to physics letters )
|
in the imaginary time formalism of finite temperature field theory , temporal correlators are related to several dynamical properties through a prescription of analytical continuation .
especially the imaginary time correlator can be described by the same spectral function as that of retarded and advanced green functions .
therefore the spectral functions extracted from the imaginary time correlators , which are able to be calculated in lattice qcd , enable us to investigate important physics at finite temperature , e.g. dilepton rates @xcite , light mesons @xcite , heavy quarkonia @xcite , glueballs @xcite , transport coefficients @xcite .
one of the most important quantities at finite temperature qcd is the spectral function of heavy quarkonium , which plays the key role for understanding the quark gluon plasma ( qgp ) formation in heavy ion collision experiments e.g. the rhic experiment at brookhaven national laboratory .
recent studies on the spectral function of charmonium above @xmath3 suggest that hadronic excitations corresponding to @xmath4 may survive in the deconfinement phase till relatively high temperature @xcite .
such results of strongly interacting qgp may affect the scenario of @xmath4 suppression @xcite . therefore the determination of the accurate dissociation temperature of @xmath4 is required for many phenomenological studies . as somewhat different from the `` direct '' @xmath4 suppression ,
recently an `` indirect '' @xmath4 suppression was proposed @xcite .
the total yield of @xmath4 is not only from the direct production , about 40% of the @xmath4 s that come from the higher states @xmath5 and @xmath1 @xcite . if the higher states dissociate at lower temperatures , a part of the @xmath4 suppression may be observed in experiments at a lower temperature than that of @xmath4 .
therefore the study of @xmath5 and @xmath1 states and their dissociation temperatures is a very interesting subject for the physics of qgp .
several groups have already investigated the @xmath1 states , and their results indicate that the lowest peak of the spectral function , corresponding to the @xmath1 state , may dissociate just after the transition temperature @xcite . on the other hand
there is no lattice qcd study on the @xmath5 states above @xmath3 , because it is difficult to extract information of non - lowest states such as @xmath5 from the correlator on the lattice at finite temperature . in such studies of temporal correlators at finite temperature
, one has to take account of a characteristic contribution caused by the finite temporal extent .
usually a meson correlator is interpreted by a diagram with quark and anti - quark propagators like it is sketched in fig .
[ fig : corr](a ) . however in the case of a system with a finite temporal extent , a wraparound ( scattering ) contribution , as shown in fig .
[ fig : corr](b ) , has to be included as well .
it is similar to the situation at zero temperature e.g. the study of two pion correlators @xcite , and pentaquark correlator @xcite . because these correlators include multi - hadron states as intermediate states , and each state is also allowed as a color singlet state , which can have a finite energy state in the confined phase . at zero temperature the calculation should be performed , in principle , with infinite temporal extent , therefore , the wraparound effect is not physical .
the contribution is usually removed using e.g. dirichlet boundary conditions @xcite or appropriate analyses @xcite . in finite temperature calculations using the imaginary time formalism , the temporal extent is determined by ( the inverse of ) the temperature , and the quark fields have to have anti - periodic boundary conditions . in the confinement phase , an expectation value of the diagram in fig.[fig : corr](b ) vanishes because it is related to a single quark propagation , which is exponentially suppressed with its infinite energy due to the confinement . on the other hand , in the deconfinement phase
, the contribution may yield a finite expectation value .
this is similar to the discussion of the polyakov loop expectation value .
for example in the free quark system , when one adopt a meson - like operator with degenerate quark masses and zero spatial momentum , the correlator of the diagram in fig.[fig : corr](a ) behaves like @xmath6 , where @xmath7 is the quark mass ( or the energy of a single quark state ) .
( here we note that equations are written with dimension - less variables throughout this paper , if there is no comment on the units . ) on the other hand , the diagram in fig.[fig : corr](b ) behaves like @xmath8 , where @xmath9 is the inverse of the temperature .
the latter diagram provides a constant contribution to the correlator , and the contribution may dominate in the large imaginary time region as a zero energy mode . due to the contribution , correlators of such meson - like operators , e.g. charmonia ,
may be drastically changed just after the deconfinement transition .
the latter type diagram , in general , yields t - dependent contributions which depends on a difference between single quark energies for each quark propagator . even in non - degenerate quark masses or non - zero spatial momentum cases ,
the diagram can yield the constant contribution when the energies are identical .
however we do not discuss the case in this paper .
the finite momentum case has been discussed in ref .
@xcite at high temperature limit .
this paper is organized as follows . in sect .
[ sec : freetheory ] we discuss the correlators of meson - like operators and its spectral functions in a system of free quarks with finite temporal extent .
some numerical results are also presented . in this section
we also discuss the physical interpretation of the constant contribution and some analysis methods to avoid this contribution . in sect .
[ sec : quench_qcd ] we show numerical simulations in quenched qcd on anisotropic lattices , and demonstrate effects of the constant contribution in the interacting case and the analyses to avoid the contribution . as a result of these analyses
we show that the @xmath1 state yields a small change except for the constant mode in its spectral function after the deconfinement transition like the @xmath4 state at least up to @xmath10 .
at last we discuss results on the constant contribution .
in order to investigate the constant contribution as mentioned in the introduction , we first consider the free quark case of qcd , in which the constant contribution can be easily calculated . here
we define the meson - like correlators with quark bilinear operators , @xmath11 , in the free quark case this gives for the correlator , @xmath12 where @xmath13 are appropriate @xmath14 matrices , i.e. @xmath15 , @xmath16 , @xmath17 , and @xmath18 for pseudoscalar(ps ) , vector(v ) , scalar(s ) , and axialvector(av ) channels respectively . in this section and
later we always adopt the anti - periodic boundary condition for quark fields in temporal direction to suppose a finite temperature case , and periodic boundary conditions in spatial directions . in this paper
we calculate correlators for degenerate quark masses , constructed from bilinear operators , with vanishing spatial momentum as the simplest case , which is aiming at studies of charmonium at finite temperature discussed in sect .
[ sec : quench_qcd ] .
the spectral function of the correlator is defined by @xmath19 and has been calculated in ref .
@xcite in the high temperature limit . in the spectral representation ,
the constant contribution provides a term proportional to @xmath0 , which corresponds to a constant mode in the meson - like correlators .
this term is an odd function in frequency @xmath20 , which is one of the basic properties of bosonic spectral function . from the results in @xcite one can find several characteristic properties of the constant contribution .
first there is no constant mode in the ps channel , however the other correlators have the constant mode , which comes from the scattering contribution .
the correlator in the s channel has also no constant mode at vanishing quark mass , where the chiral symmetry is restored . secondly , in the v channel ( and the av channel in the chiral limit )
, there is no constant contribution coming from the zero relative momentum process . on the other hand
, there is one in the av and s channels ( and it disappears in chiral limit ) and @xmath21 ( if quark mass is finite ) respectively . ] .
these properties are very important especially in a finite volume calculation .
when the minimum nonzero momentum due to the finite volume is large , the contribution of the zero relative momentum processes dominates in the correlators .
furthermore , a two quark state with zero relative momentum is forbidden in the p - wave state . therefore the relative contribution of the constant mode increases exponentially as the spatial volume decreases in the p - wave state at finite quark mass . as a result
, we can expect a very different behavior of correlators between s - wave and p - wave states in the case of finite quark mass and finite ( but relatively small ) volume .
thirdly , correlators for each channel coincide at the midpoint @xmath22 as mentioned in ref .
these characteristic properties will be confirmed also by the numerical results presented in the next section ( [ sec : free_lattice ] ) .
it is important to consider physical interpretations of this contribution .
the term @xmath0 remains in the continuum form and infinite volume results @xcite , while it disappears at zero temperature , i.e. infinite temporal extent .
therefore the contribution is caused by a kind of physical thermal effect .
the origin of the effect could be considered as a scattering with the thermal bath ( landau dumping ) which is related to some transport coefficients . from the kubo formula , for example , a derivative of the spectral function in the v channel , @xmath23 which is defined by eq .
( [ eq : spf ] ) using @xmath24 , is related to the electrical conductivity @xmath25 @xcite , @xmath26 let us also note on the dilepton rate .
the dilepton rate is extracted from the spectral functions in the vector channel of spatial ( @xmath27 with @xmath24 ) and temporal ( @xmath28 with @xmath29 ) components , @xmath30 @xmath31 is the charge susceptibility and it has only the @xmath0 term which is equivalent to that of @xmath27 with the opposite sign . therefore the constant contribution in the dilepton rate cancels at least for the free quark case . in this subsection we present numerical calculations of the meson - like correlators for free quarks on lattices .
the calculations are performed on isotropic @xmath32 lattices , and the volume dependence is discussed using results for @xmath33 lattice .
the free quark is described by the wilson quark action with a bare quark mass of @xmath34 .
the boundary conditions are anti - periodic in temporal direction and periodic in spatial directions . figure [ fig : free1 ] shows correlators ( eq .
( [ eq : corrdef ] ) ) for each channel .
one can clearly see a constant contribution in the s and av channels .
the qualitative difference between the ( ps , v ) and ( s , av ) channels and the coincidence of correlators at the mid point @xmath35 can be explained as mentioned in the previous section .
figure [ fig : free2 ] shows the effective masses of the correlators in fig . [ fig : free1 ] .
the effective mass is defined by correlator at successive time slices .
@xmath36 } } { \cosh{\left[m_{\mbox{eff}}(t)\left(\frac{n_t}{2}-t-1\right)\right]}}\ ] ] the effective masses of the meson - like correlators with free quarks should approach the energy of the two quark state without momentum for s - wave states ( ps and v channels ) and with a minimum momentum for p - wave states ( s and av channels ) except for the zero energy mode . in the wilson quark action ,
the free quark propagator satisfies the dispersion relation , @xmath37 where @xmath38 and @xmath39 is the energy of a quark with momentum @xmath40 . using eq .
( [ eq : free_disp ] ) , one can calculate the lowest energies of the two quark states in s - wave and p - wave states .
the values obtained for @xmath41 and @xmath34 are shown in fig .
[ fig : free2 ] . in the ps and
v channels , the zero relative momentum processes have no constant contribution and their effective masses approach the expected values . in the s and av channels , on the other hand , their effective masses approach the zero energy level rather than the energy of two quark states .
next i present the results for the different volume .
figure [ fig : free1b ] and [ fig : free2b ] are results of the meson - like correlators and their effective masses on a @xmath42 lattice with the same quark mass , @xmath34 , respectively . in the latter section ( [ sec : quench_qcd ] ) , we present simulations in quenched qcd at finite temperature , where the finite temperature lattices have the similar aspect ratio , @xmath43 , and the lowest state energy in lattice units to this free quark calculation with @xmath44 .
although we find the large volume dependence between @xmath41 and @xmath44 , almost no volume dependence is seen for @xmath45 . at a sufficiently large aspect ratio , we find no clear constant contribution , but the finite constant contribution exists when the quark mass is finite .
since the lattice qcd is formulated in euclidean space - time , we can isolate the lowest state contribution from correlators in the large euclidean time region , while studies of non - lowest states are not so easy .
when one want to know the information about nonzero energy states , the constant mode makes it difficult .
the lowest scattering process yields only a constant mode in the correlators considered in this paper , and does not affect the other dynamics at higher energy , @xmath46 , at least in the free quark case . even in the case of interacting qcd ,
if the width of the peak structure in the spectral function at @xmath47 is sufficiently small , e.g. @xmath48 , the contribution yields only a constant mode in the correlators @xcite . therefore , in order to investigate the nonzero energy states in the correlators , it is useful to remove the constant mode from the correlators in an appropriate analysis . in this subsection
we present two analysis methods to avoid the constant contribution from the correlators .
of course the cosh + constant fit enables us to do so , but here we present the methods without a fit analysis .
the constant mode in the correlators can be removed by the derivative of the correlator with respect to @xmath49 @xcite , where the differential correlator @xmath50 is defined by a derivative of @xmath51 , @xmath52 here @xmath50 is equivalent to the symmetric derivative at @xmath53 .
one can define the effective mass of the differential correlator , @xmath54 } } { \sinh{\left[m_{\mbox{eff}}^{\mbox{diff}}(t ) \left(\frac{n_t}{2}-t-1\right)\right]}}. \label{eq : effmassdiff}\end{aligned}\ ] ] since , however , the overlap of each state is modified by a factor of the energy of the state in the analysis , the plateau of the effective masses becomes narrow .
furthermore , in general , a derivative of an observable increases the statistical fluctuations , thus the analysis may be difficult to apply to actual numerical calculations . therefore we adopt a rather different approach to remove the constant contribution .
we define the midpoint subtracted correlator , @xmath55 one can also define the effective mass of the subtracted correlator , @xmath56 } } { \sinh^2{\left[\frac{1}{2}m_{\mbox{eff}}^{\mbox{sub}}(t ) \left(\frac{n_t}{2}-t-1\right)\right]}}. \label{eq : effmasssub}\end{aligned}\ ] ] figure [ fig : free4 ] ( left panel ) shows the effective mass @xmath57 , defined in eq .
( [ eq : effmasssub ] ) , from the free quark results for @xmath41 in fig .
[ fig : free1 ] .
the effective masses are equivalent to the usual effective mass shown in fig .
[ fig : free2 ] except for the effects of the constant mode .
the expected energies of the lowest two quark states for the s - wave and the p - wave states can be calculated from the free quark dispersion relation , eq .
( [ eq : free_disp ] ) .
the values are shown in fig .
[ fig : free4 ] as well as in fig .
[ fig : free2 ] .
in contrast to the case of the usual effective masses in fig . [
fig : free2 ] , the effective masses @xmath57 approach the expected values even in the p - wave states .
the analysis to avoid the constant contribution works well at least in the free quark case .
the method using the midpoint subtracted correlators can also be applied to studies of spectral functions @xmath58 by a modification of the kernel , @xmath59 by using the alternative kernel @xmath60 , it is possible to extract the spectral function without the contribution in @xmath61 from e.g. the maximum entropy method .
next we consider the second method to avoid the constant contribution .
the method utilizes the @xmath62 transformation properties of the correlators , where the @xmath62 transformation is performed on the lattice through the transformation of temporal link variables , @xmath63 where @xmath64 and 2 , for any @xmath65 , at a time slice ( we adopt @xmath35 ) .
first we consider the @xmath62 transformation properties for the diagrams in fig .
[ fig : corr](a ) and ( b ) . the contribution of fig .
[ fig : corr](a ) has a @xmath62 center symmetry , while that of fig .
[ fig : corr](b ) is no longer invariant under the @xmath62 transformation . in case of the static limit
, one can easily understand it by the analogy of polyakov loop and wilson loop .
even in the non - static case , the properties can be derived using a hopping parameter expansion of the quark propagator . in any order of the expansion
, one can show this property like in the static limit .
furthermore , for the @xmath62 variant contribution , the @xmath62 factor can be factored out .
therefore one can cancel the @xmath62 variant contribution by averaging the correlators over the three sectors which are characterized by the location of the polyakov loop in the complex plane . here
we define the averaged correlator @xmath66 from @xmath67 , @xmath68 and @xmath69 , which are correlators on the configurations with an argument of the polyakov loop in the complex plane fulfilling @xmath70 , @xmath71 and @xmath72 respectively , where @xmath73 is a value of the polyakov loop on the given configuration , @xmath74 each correlator can be calculated on the @xmath62 transformed gauge configurations .
the second method to avoid the constant contribution is in a way similar to removing the backward quark propagations by using a cancellation between results with periodic and anti - periodic boundary conditions in the @xmath49 direction .
however the second method is beyond a technical step to separate the constant mode from the others .
the effect of @xmath62 symmetrization is to remove from the partition function all states with non - integer baryon number ( whether in quenched or full qcd ) .
this has been discussed , e.g. , in ref .
@xcite . here
we should mention that the cancellation of the constant mode is not exact in the second method . because , e.g. , in the case of a diagram with a 3 times wrapping quark line , the contribution also has the @xmath62 center symmetry .
therefore it will not be canceled in the analysis .
( this situation is also similar to the cancellation of the backward quark propagation as mentioned above .
this cancellation is also not exact .
the remaining @xmath62 symmetric contribution can be removed by considering the zero - baryon partition function , as discussed as well in ref .
@xcite . ) however , the @xmath62 invariant constant contributions is of @xmath75 or less , while that for the leading ( canceled ) contribution is @xmath76 .
therefore the averaged correlator is useful to see a qualitative effect of the constant contribution in the correlators , when @xmath77 is not so small . from the averaged correlator in eq .
( [ eq : defavecorr ] ) , one can define the effective mass @xmath78 , @xmath79 } } { \cosh{\left[m^{\mbox{ave}}_{\mbox{eff}}(t ) \left(\frac{n_t}{2}-t-1\right)\right]}}.\ ] ] the effective masses @xmath78 for the free quark case are presented in the right panel of fig .
[ fig : free4 ] . similar to the effective masses from the midpoint subtracted correlators , @xmath78 approach the expected nonzero lowest energies even for the p - wave states .
however the method does not exactly remove the constant contribution , a part of the constant contribution remains .
the method using the midpoint subtracted correlators does not work for correlators with finite momentum and non - degenerate masses of valence quarks , e.g. heavy - light meson .
although these cases are out of the scope of this study , the averaged correlator method works even for such meson - like operators . furthermore provided the peak structure of spectral function at @xmath80 has a wide width @xmath81 due to interactions , the subtracted correlator method may not even work in the case of correlators with zero momentum and degenerate quarks , while the averaged correlator method can be applied even in this case .
in this section we demonstrate effect of the constant contribution and several analyses to avoid it in quenched qcd at finite temperature . in the simulation
we calculate the charmonium correlators , which are defined like in sect .
[ sec : freecorr ] at the charm quark mass .
the temporal correlators ( or the spectral functions ) of charmonia are important to discuss the @xmath4 suppression which is considered as one of the most important signals of qgp formation in heavy ion collision experiments .
several groups have investigated the charmonium correlators in finite temperature lattice qcd @xcite . in such studies
one has to take into account the constant contribution to the charmonium correlators ( or spectral functions ) . because this contribution may not be negligible in finite temperature lattice qcd .
in fact it is a genuine effect in finite temperature qcd as mentioned in sect .
[ sec : freecorr ] . in finite temperature lattice
qcd the temporal correlators are restricted within the inverse of the temperature . in order to keep sufficient number of data points in temporal direction at high temperature , we adopt an anisotropic lattice , on which the temporal lattice spacing @xmath82 is smaller than the spatial one @xmath83
. the temperature is varied by the temporal lattice sizes @xmath84 with a fixed temporal lattice spacing @xmath82 . such calculations with a fixed lattice spacing also have the advantage to investigate thermal effects in correlators , i.e. one can directly compare the correlators at different temperatures without need for information on the beta - function . in this
section physical quantities are expressed in @xmath85 units when there is no comment on the units .
the gauge configurations have been generated by an standard plaquette gauge action with a lattice gauge coupling constant , @xmath86 and a bare anisotropy parameter @xmath87 .
the definition of the action and parameters are the same as that adopted in ref .
lattice sizes are @xmath88 where @xmath89 at @xmath90 and @xmath91 , 26 and 20 at @xmath92 .
the lattice spacings are @xmath93 gev and @xmath94 gev which are determined from the hadronic radius @xmath95 fm .
the physical volume size is about @xmath96 and temperatures at @xmath91 , 26 and 20 are @xmath97 , @xmath98 and @xmath99 respectively .
the temperatures have been determined by the peak position of the polyakov loop susceptibility from which we found the critical temperature around @xmath100 .
the finite temperature lattices have been generated up to 300 configurations after 20,000 sweeps for equilibration , each configuration is separated by 500 sweeps to reduce the autocorrelation . at zero temperature , we have 60 configurations with the same condition . in the calculations ,
the error analysis is performed by the jackknife estimation .
for the quark fields , we adopt an @xmath101 improved wilson quark action with tadpole improved tree level clover coefficients .
although the definition of the quark action is the same as in ref .
@xcite , we adopt a different choice of the wilson parameter @xmath102 to suppress effects of lattice artifacts in higher states of charmonium @xcite
. then the parameters in the quark action are @xmath103 , @xmath104 , @xmath105 and @xmath106 .
figure [ fig : quench1 ] shows the usual effective masses calculated with the parameters listed above .
the results at zero temperature satisfy a relativistic meson dispersion relation with the same anisotropy as the renormalized gauge anisotropy @xmath107 .
the left and right panels in fig .
[ fig : quench1 ] show the results from the local and spatial extended operators .
the latter operators are defined by @xmath108 with a smearing function @xmath109 in coulomb gauge . in this calculation
the spatially extended operators are adopted only in the source operators , the sink operators are local in any cases .
the smearing function is the same as that in ref .
@xcite , i.e. @xmath110 where @xmath111 and @xmath73 are parameters determined by a matching with the charmonium wave function as @xmath112 and @xmath113 . from the zero temperature calculation with the spatially extended operators we obtain the charmonium masses in physical units , @xmath114 mev and @xmath115 mev , which are slightly heavier than their experimental values @xmath116 mev and @xmath117 mev @xcite .
the experimental values in lattice units are also shown in fig .
[ fig : quench1 ] .
the results for the p - wave states are rather noisy at our current statistics , and the actual values are not so important for this study .
therefore we do not show the numbers in mev , however their effective masses show a plateau at reasonable values . in this section
we discuss the result of the charmonium correlators at finite temperature .
the definitions for several types of effective masses can be taken over from that of the free quark calculations in sect .
[ sec : freetheory ] .
first we show the usual effective masses , @xmath118 at finite temperature in fig .
[ fig : quench2 ] .
since these are results obtained with local operators , there is no plateau region at any temperature or channel we chose .
it can be expected from zero temperature calculation and the effective masses show reasonable behavior below @xmath3 , at which the charmonium correlators does not show large difference from the zero temperature results @xcite .
we find rather large statistical fluctuations in the p - wave states , a reason for that is given later . just above @xmath3 , we see drastic changes in the p - wave state channels .
the changes may cause the dissolution of @xmath1 states just above @xmath3 as already reported in ref .
@xcite .
on the other hand , the s - wave states , i.e. the ps and v channels , show small changes up to @xmath2 .
this temperature dependences are shown in fig . [
fig : quench3 ] for each channel .
when we compare the effective masses at the same time slices , we find no large change up to @xmath2 in both channels as well .
the results are consistent with previous lattice studies of charmonium spectral functions @xcite , in which the spectral functions for the ps and v channels in the deconfinement phase ( but at not so high temperature ) show a peak structure around the @xmath119 and @xmath4 masses like the result at zero temperature . as discussed for free quarks in sect .
[ sec : freetheory ] , the constant contribution may dominate the meson correlators for some channels in the deconfinement phase . in order to see the contribution in the meson correlators we calculate the effective mass from the midpoint subtracted correlator , @xmath57 defined in eq .
( [ eq : effmasssub ] ) .
figure [ fig : quench4 ] shows the results of @xmath57 in quenched qcd at finite temperature . as we expected , the effective masses at each temperature in the s - wave states show small differences from that of zero temperature , when comparing them at the same time slices . in the p - wave states , however , the drastic changes of the usual effective masses @xmath118 are absent in the effective masses from the midpoint subtracted correlators @xmath57 .
furthermore the similar behaviors of the effective masses hold till @xmath2 as in the case of the s - wave states .
the results indicate that the drastic changes of the usual correlators ( and effective masses ) are caused only by the constant contribution to the correlators .
the situation is very similar to the free quark case discussed in sect .
[ sec : freetheory ] .
next we discuss the polyakov loop sector dependence of meson correlators as mentioned in sect .
[ sec : free_ploop ] .
first of all we show time histories of the polyakov loop on our finite temperature configurations in fig .
[ fig : quench5 ] . for all configurations above @xmath3
the polyakov loop lies in the real sector .
thus fig .
[ fig : quench2 ] gives the result on configurations with the polyakov loop being in the real sector .
the correlators for the imaginary sectors are calculated on configurations generated by @xmath62 transformations from that in the real sector . as an example for the imaginary sectors
we show the effective masses calculated from @xmath67 , @xmath68 , @xmath69 and @xmath66 , which are defined in sect .
[ sec : free_ploop ] , in fig.[fig : quench6 ] . as discussed in sect .
[ sec : free_ploop ] , the @xmath62 transformation changes a part of the constant contribution by a factor of @xmath62 , i.e. @xmath120 where @xmath64 and 2 . since our definition of the meson correlators is real , the factor acts like @xmath121 .
it means that the constant contributions in the imaginary sectors become negative when the value in the real sector is positive .
hence the correlators of the p - wave states approach a positive constant value like in the real sector , as shown in fig.[fig : free1 ] , while the correlators for the imaginary sectors approach a negative value . as a result
the effective masses of the p - wave states diverge at a time slice as shown in fig .
[ fig : quench6 ] . here
we comment on the large statistical fluctuations of correlators for the p - wave states just below @xmath3 as can be seen in fig .
[ fig : quench2 ] .
below @xmath3 , in principle , the constant contribution vanishes due to the infinite energy of a single quark state in the confinement phase .
however the contribution on each configuration may have a small value near @xmath3 even in confinement phase , like the polyakov loop just below @xmath3 in fig .
[ fig : quench5 ] . in this case
the effect of the constant contribution appears in the correlators .
since the polyakov loop sector frequently changes below @xmath3 , the effects appear as the large fluctuation in the correlators . when the averaged correlator defined in eq .
( [ eq : defavecorr ] ) is calculated , one can expect that most of the constant contributions are canceled like in the free quark case ( sect .
[ sec : free_ploop ] ) .
figure [ fig : quench7 ] shows the results of @xmath78 at each temperature .
we find almost the same behavior as in fig . [
fig : quench4 ] .
the results also suggest that the drastic changes in the p - wave states just above @xmath3 are coming from the constant contribution only
. finally we present results of meson correlators with spatially extended operators .
the spatially extended operators are constructed with a smearing function which yields a spatial distribution of quark ( and anti - quark ) source(s ) .
the distribution also changes the momentum distribution of the quark propagation .
for example , a point source function yields any ( lattice ) momenta with equal distribution , while a wall source function yields only quark propagation with zero momentum . in the case of our source
the ( smearing ) function provides quark propagation with any momenta but quark propagators with small momentum are enhanced .
therefore the correlators with spatially extended operators defined , as in sect .
[ sec : setup ] , yield ( relatively ) large overlap with the lower energy states .
figure [ fig : quench8 ] is the result for the usual effective masses with the spatially extended operators .
other conditions are the same as in fig .
[ fig : quench2 ] . below @xmath3 ,
in contrast to the case of local operators the effective masses reach a plateau due to the larger overlap with the lowest state . above @xmath3
the effective masses of the p - wave states change more than in the case of local operators . because the constant contribution is enhanced by the smearing function more than the other contributions in the case .
the effect of the constant contribution is also present in the v channel . at least in the free quark case , as we discussed in sect .
[ sec : freecorr ] , the constant contribution to the v channel is smaller than that to the p - wave states for finite quark masses and small volumes .
in fact , even in quenched qcd , it is difficult to see the effect in the v channel in local operators .
however we find small but finite changes even in the v channel above @xmath3 while no changes are seen in the ps channel as we expected .
we can say that the small changes in the v channel is caused by the constant contribution , because the analysis with the midpoint subtracted correlators shows almost no changes in the v channel above @xmath3 as it can be seen in fig .
[ fig : quench9 ] . similar to the results with local operators ,
the effective masses of the p - wave states also show small change at any temperatures .
figure [ fig : quench10 ] is the same figure as fig .
[ fig : quench7 ] , but with spatially extended operators .
we find similar behavior for both cases of operators .
let us discuss details on @xmath1 states using the results of the previous sections .
in this study we find that the drastic changes of the ( usual ) correlators in the p - wave states are caused by the constant contribution , while the correlators without the constant mode yield small change till , at least , @xmath2 even in the p - wave states .
the changes are of similar size as that of the s - wave states as one can see in fig .
[ fig : quench11 ] and fig .
[ fig : quench12 ] . in the case of the spatially extended operators ,
the averaged effective mass @xmath78 for the av channel gradually reduces as the temperature increases .
the change seems to be effected by the constant contributions with the @xmath62 center symmetry as discussed in sect .
[ sec : free_ploop ] .
in fact , we can not find any hints of such a change in @xmath57 in the left panel of fig .
[ fig : quench12 ] .
furthermore it is reasonable to find the effect only in case of the spatially extended operators , because the @xmath62 symmetric constant contribution is enhanced by the smearing function .
although the constant mode coming from the scattering process is important to investigate some transport phenomena , it does not affect the spectral function in the @xmath122 region for the charmonium correlators with zero spatial momentum .
therefore the dissolution of @xmath1 states just above @xmath3 as discussed in e.g. ref .
@xcite might be misleading . when there is the constant contribution , the hadronic state is no longer the lowest state .
the situation provides some difficulties in the analysis to reconstruct the spectral function in the higher energy part @xmath122 even if the mem is applied @xcite .
the difficulties may be understood by the fact that most studies of the charmonium spectral function can not reproduce the excited states correctly especially at finite temperature ( even below @xmath3 ) .
furthermore , in general , correlators for the p - wave states are noisier than that for the s - wave states as one can see in fig .
[ fig : quench1 ] . in principle
the mem analysis can extract the correct spectral function even if the constant contribution exists , however it is difficult to reproduce the non - lowest part of the spectral function correctly by the mem analysis at finite temperature with typical statistics . in order to investigate the properties of @xmath1 states above @xmath3
, a careful analysis taking the constant contribution into account is necessary .
a correct mem analysis of the correlator with and without the constant contribution should give the same results , up to the low energy region @xmath61 .
such a comparison can be easily performed by the mem analysis with the alternative kernel of eq .
( [ eq : subkernel ] ) for midpoint subtracted correlators . failure to do so signals unreliable mem results .
we also find a polyakov loop sector dependence of the constant contribution in quenched qcd .
to study transport coefficients from the behavior of spectral function at @xmath80 one has to pay attention to the polyakov loop sector especially in quenched qcd calculations . from a practical viewpoint , the dependence on the polyakov loop sector is useful to estimate the constant contribution as discussed in sect . [
sec : free_ploop ] and sect .
[ sec : quench_ploop ] .
the constant contribution may also explain some other results of previous studies of temporal correlators .
first , in the study of the spectral function at finite temperature by the mem analysis , one sometimes find a kind of divergence of the spectral function near @xmath80 , it may be caused by the constant contribution .
of course it may be a simple failure of the analysis , however the contribution near @xmath80 ought to exist except for the ps channel .
secondly , in the s - wave states of charmonium , several groups have reported that the correlator for the ps channel shows almost no change above the deconfinement transition , while that of the v channel shows a small but significant change @xcite .
this can also be seen from our calculation in sect .
[ sec : quench_qcd ] .
it can be explained by the constant contribution as we did in this paper .
thirdly , when one considers a spatial boundary condition dependence of the charmonium spectrum at finite temperature as in ref .
@xcite , the p - wave states ( in the free quark case ) have the lowest energy using a `` hybrid boundary '' , which is an anti - periodic boundary in one spatial direction and periodic in the other directions . on the other hand
, the p - wave state has a larger energy with periodic boundaries in any spatial directions .
however the ( usual ) effective mass of the p - wave state correlators for free quark calculations provides a smaller value in periodic boundary conditions than in the hybrid boundary case
. this unexpected behavior of the effective mass is also explained by the scattering processes because of suppression of the scattering contribution by the anti - periodic boundary in a spatial direction .
finally , the polyakov loop sector dependence of the chiral order parameter in quenched qcd @xcite may be also explained by the effect of ( @xmath62 variant ) wrapped contributions in the chiral condensate .
as we discussed in the last section , the constant contribution affects the temporal correlators at finite temperature in various cases .
therefore it is important to take this contribution into account for studies of temporal correlators in lattice qcd at finite temperature .
the author thanks f. karsch , s. sasaki , s. datta , t. yamazaki , t. doi , h. matusufuru , p. petreczky , s. aoki and c. schmidt for helpful discussions and comments .
this work has been authored under contract number de - ac02 - 98ch1 - 886 with the u.s .
department of energy .
the simulations have been performed on supercomputers ( nec sx-5 ) at the research center for nuclear physics ( rcnp ) at osaka university and ( nec sx-8 ) at the yukawa institute for theoretical physics ( yitp ) at kyoto university .
f. karsch , e. laermann , p. petreczky , s. stickan and i. wetzorke , phys .
b * 530 * , 147 ( 2002 ) [ arxiv : hep - lat/0110208 ] .
de forcrand _ et al . _
[ qcd - taro collaboration ] , phys .
d * 63 * , 054501 ( 2001 ) [ arxiv : hep - lat/0008005 ] .
t. umeda , r. katayama , o. miyamura and h. matsufuru , int .
j. mod .
a * 16 * , 2215 ( 2001 ) [ arxiv : hep - lat/0011085 ] . t. umeda , k. nomura and h. matsufuru , eur .
j. c * 39s1 * , 9 ( 2005 ) [ arxiv : hep - lat/0211003 ] .
s. datta , f. karsch , p. petreczky and i. wetzorke , phys . rev .
d * 69 * , 094507 ( 2004 ) [ arxiv : hep - lat/0312037 ] .
m. asakawa and t. hatsuda , phys .
* 92 * , 012001 ( 2004 ) [ arxiv : hep - lat/0308034 ] .
g. aarts , c. r. allton , r. morrin , a. p. o. cais , m. b. oktay , m. j. peardon and j. i. skullerud , pos * lat2006 * , 126 ( 2006 ) [ arxiv : hep - lat/0610065 ] .
a. jakovac , p. petreczky , k. petrov and a. velytsky , arxiv : hep - lat/0611017 . n. ishii , h. suganuma and h. matsufuru , phys . rev .
d * 66 * , 094506 ( 2002 ) [ arxiv : hep - lat/0206020 ] .
g. aarts and j. m. martinez resco , jhep * 0204 * , 053 ( 2002 ) [ arxiv : hep - ph/0203177 ] . s. gupta , phys
. lett .
b * 597 * , 57 ( 2004 ) [ arxiv : hep - lat/0301006 ] .
a. nakamura and s. sakai , phys .
lett . * 94 * , 072305 ( 2005 ) [ arxiv : hep - lat/0406009 ] .
t. hashimoto , k. hirose , t. kanki and o. miyamura , phys .
lett . * 57 * , 2123 ( 1986 ) .
t. matsui and h. satz , phys .
b * 178 * , 416 ( 1986 ) .
f. karsch , d. kharzeev and h. satz , phys .
b * 637 * , 75 ( 2006 ) [ arxiv : hep - ph/0512239 ] .
s. digal , p. petreczky and h. satz , arxiv : hep - ph/0110406 . l. antoniazzi _ et al . _
[ e705 collaboration ] , phys .
* 70 * , 383 ( 1993 ) . c. kim , nucl .
suppl . * 129 * , 197 ( 2004 ) [ arxiv : hep - lat/0311003 ]
. t. t. takahashi , t. umeda , t. onogi and t. kunihiro , phys .
d * 71 * , 114509 ( 2005 ) [ arxiv : hep - lat/0503019 ] .
g. aarts and j. m. martinez resco , nucl . phys .
b * 726 * , 93 ( 2005 ) [ arxiv : hep - lat/0507004 ] .
f. karsch , e. laermann , p. petreczky and s. stickan , phys .
d * 68 * , 014504 ( 2003 ) [ arxiv : hep - lat/0303017 ] .
p. petreczky and d. teaney , phys .
d * 73 * , 014508 ( 2006 ) [ arxiv : hep - ph/0507318 ] .
p. majumdar , y. koma and m. koma , nucl . phys .
b * 677 * , 273 ( 2004 ) [ arxiv : hep - lat/0309003 ] .
s. kratochvila and p. de forcrand , phys .
d * 73 * , 114512 ( 2006 ) [ arxiv : hep - lat/0602005 ] .
h. matsufuru , t. onogi and t. umeda , phys .
d * 64 * , 114503 ( 2001 ) [ arxiv : hep - lat/0107001 ]
. w. m. yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * ( 2006 ) 1 .
t. umeda and h. matsufuru , pos * lat2005 * , 154 ( 2006 ) [ arxiv : hep - lat/0510026 ] .
h. iida , t. doi , n. ishii , h. suganuma and k. tsumura , phys .
d * 74 * , 074502 ( 2006 ) [ arxiv : hep - lat/0602008 ] .
s. chandrasekharan , d. chen , n. h. christ , w. j. lee , r. mawhinney and p. m. vranas , phys .
* 82 * , 2463 ( 1999 ) [ arxiv : hep - lat/9807018 ] .
|
we discuss a constant contribution to meson correlators at finite temperature . in the deconfinement phase of qcd
, a colored single quark state is allowed as a finite energy state , which yields to a contribution of wraparound quark propagation to temporal meson correlators .
we investigate the effects in the free quark case and quenched qcd at finite temperature .
the `` scattering '' contribution causes a constant mode in meson correlators with zero spatial momentum and degenerate quark masses , which can dominate the correlators in the region of large imaginary times . in the free spectral function ,
the contribution yields a term proportional to @xmath0 . therefore this contribution is related to transport phenomena in the quark gluon plasma .
it is possible to distinguish the constant contribution from the other part using several analysis methods proposed in this paper . as a result of the analyses
, we find that drastic changes in charmonium correlators for @xmath1 states just above the deconfinement transition are due to the constant contribution .
the other differences in the @xmath1 states are small
. it may indicate the survival of @xmath1 states after the deconfinement transition until , at least , @xmath2 .
|
absorption line systems detected in the spectra of quasi stellar objects ( qso ) originate in intervening galaxies or protogalaxies . among the different classes of absorbers , the damped lyman @xmath0 systems ( dla ) are those characterized by n(hi )
@xmath1 10@xmath2 @xmath3 and by showing many low ionization species such as feii , siii , crii , znii and oi , which are in their dominant ionization stage for a hi region .
the high accuracy in the hydrogen column density determination derived from the damped wings and the absence of significant ionization corrections allow accurate absolute abundance determinations of the gas phase elements . since dla systems are observable up to the highest redshift they provide a unique tool for the study of the cosmic chemical history of the early universe .
such observations are complementary to the hubble deep field ones ( mobasher et al . 1996 ) which show that high redshift galaxies have a large variety of morphological types including many peculiar objects with traces of interaction and merging : interestingly , a large fraction are starburst galaxies , as suggested by their colors .
the relative elemental abundances rather than absolute ones are a valuable diagnostic of the first elemental buildup .
this is because absolute abundances are affected by all the model assumptions , whereas abundance ratios generally depend only on the assumed nucleosynthesis and stellar lifetimes .
the way of using the information contained in abundance ratios is to compare the observed ratios with predictions from chemical evolution models which take into account detailed stellar nucleosynthesis and lifetimes . as in the milky way
, we hope to distinguish halo - like abundances , produced by type ii sne and characterized by enhancements of @xmath0 elements with respect to the iron - peak elements , from disk - like abundances , produced by the cumulative effect of both type ia and type ii sne , where the abundance ratios become progressively solar .
the halo abundance pattern follows from the fact that fe arises mostly from type ia supernovae with some contribution from type ii , while the reverse is true for si and o. the longer lifetimes of the snia , which are believed to have progenitors with masses of 1 - 8 m@xmath4 , result in a delayed iron enrichment compared to the major snii products such as @xmath0-elements ( tinsley , 1980 ; greggio and renzini , 1983a ; matteucci and greggio , 1986 ) .
therefore , the behaviour of the [ @xmath0/fe ] ratios is mainly dependent on the relative lifetimes of type ia and ii supernovae .
another key element is nitrogen and the ratio n / o . this element , in fact , is thought to originate mainly from low and intermediate mass stars and to be a `` secondary '' element , in the sense that it is produced proportionally to the initial stellar metallicity . as a consequence nitrogen is restored into the interstellar medium with a large temporal delay relative to oxygen , which is produced in massive stars and
is a `` primary '' element , namely its production is independent of the initial stellar metallicity .
for this reason , the @xmath0/fe and n / o ratios can be used as cosmic clocks and represent a clue in understanding the nature of high redshift objects such as qso ( hamman and ferland , 1993 ; matteucci and padovani , 1993 ) and dla systems ( matteucci , 1995 ) .
the dla systems are generally believed to be the progenitors of the present - day spiral galaxies ( wolfe et al .
however , it has also been suggested that they could be dwarf galaxies , since they have a large amount of gas , low metallicities and low dust - to - gas ratios ( pettini et al . 1990;meyer and york 1992 ; steidel 1994 ) .
whether dla are proto - spirals or proto - dwarfs can be indicated by their observed abundance pattern . to this purpose in this paper
we will present models of chemical evolution for galaxies of different morphological type and compare them with the observed abundances in dla .
so far chemical abundances have been measured in a set of dla systems and with a variety of resolutions and accuracies by black et al .
( 1987 ) , meyer & york ( 1987 ) , chaffee et al .
( 1988 ) , meyer et al .
( 1989 ) , meyer & roth ( 1990 ) , rauch et al .
( 1990 ) , pettini et al .
( 1990 ) , meyer & york ( 1992 ) , pettini et al .
( 1994 ) , fan & tytler ( 1994 ) , wolfe et al .
( 1994 ) , pettini et al .
( 1995 ) , lu et al .
( 1995a ) , steidel et al .
( 1995 ) , lu et al ( 1995b ) . here
, we would like to focus on the abundances which could be used to compare with chemical evolution models and in particular with the few nitrogen detections so far reported in the literature .
a summary of the abundances for the systems for which nitrogen detection or upper limits are available are reported in table 1 and 2 .
all the original abundances have been renormalized to the solar abundances of anders & grevesse ( 1989 ) , with the exception of @xmath5=@xmath6 taken from hannaford et al .
( 1992 ) ( [ m / h ] @xmath7 log(m / h)-log(m / h)@xmath8 ) .
the abundances in the tables are derived from column density ratios of different atomic and ionic species in the gas phase , and can be translated to elemental abundances only if ionization effects and dust depletion are negligible or can be accounted for . in the following
we discuss shortly these two effects . [ cols="<,^,<,<,<,<,<,<,<",options="header " , ] @xmath9 values taken from lipman ( 1995 ) .
the n / o versus o / h distribution is shown in fig . 1 where the models with only one burst together with the models for the solar neighbourhood are shown .
the dot - dashed curve is from model n. 5 from kunth et al .
( 1995 ) ( model 3 of table 3 ) and it can reproduce the observations of izw 18 . incidentally , we note that this model can fit very well the n / o upper limit by pettini et al .
( 1995 ) but for a burst age shorter than 20 myr .
on the other hand , pettini s et al.s point , as well as the other points from lipman ( 1995 ) , could be marginally reproduced also by the evolution of a disk galaxy in the earliest phases of its evolution ; in particular during the halo phase , as it is shown in the figure .
in fact , the continuous curves in fig .
1 represent the predictions from the model of chemical evolution of the solar neighbourhood under different assumptions about the nucleosynthesis of n. in particular , the area delimited by those models goes from purely secondary n in stars of all masses ( the straight line at the right end ) to primary n in massive stars and secondary and primary n in low and intermediate mass stars . on the other hand , the n / o ratios observed by molaro et al.(1995 ) and green et al.(1995 ) are reproduced by models with a strong starburst and very strong galactic wind ( models 4 and 5 in table 2 ) , so that models for the solar neighbourhood seem to be completely ruled out in explaining these two dla systems .
thus we show that it is possible to have large n / o ratios even at low o abundances if the nitrogen produced by massive stars , restored on relatively short time scales , is primary and with a strong differential effect in the galactic wind . as shown in fig .
2 the same model produces solar ratios of alpha - elements ( such as s , o and si ) to iron peak elements ( such as fe and zn ) , as it is observed at least in few dla . in particular , in fig
2 we show the predictions of the models for o and si relative to iron to be compared with the data of table 4 . the x - axis does not extend to abundances larger than [ fe / h]=-2.0 since in one - burst models the metallicity does not increase any further . sulphur is not shown here but it should closely follow oxygen . on the other hand , the predicted [ si / fe ] is lower than [ o / fe ] and the reason for this resides in the fact that more si than o is produced in type ia sne ( nomoto et al .
1984 ; thielemann et al .
1993 ) and that we assume the same @xmath10 parameter for si and o. studies of abundances in halo stars do not allow us to discriminate clearly on this point ( franois , 1986 ; ryan et al . 1991 ; primas et al . 1994 ; mcwilliam et al . 1995 ) .
therefore , it does not seem safe to assume [ si / o]=0 in order to derive oxygen abundances in dla systems ( see lipman , 1995 )
. however , as indicated in figure 2 the observed [ o / fe ] is lower than [ si / fe ] for the dla observed by molaro et al .
a possible explanation for this is that , in the framework of the differential galactic wind , o , which originates mostly in type ii sne , should be lost from the galaxy in a larger percentage than silicon , which originates also from type ia sne . in other words ,
si should be treated as we do with c and fe .
however , numerical experiments show that even in this case is difficult to invert the situation and have [ si / fe ] higher than [ o / fe ] , as indicated by the data taken at face values ! in figure 3
we show the predicted n / o vs. o / h from models with 4 short bursts ( 20 myr each ) all occurring inside the first gyr from the start of star formation , so that they can be representative of high redshift objects . in particular , we show the predictions of model 6 and 7 which are similar to model 3 and 2 , respectively , but with four bursts of star formation.these models show that in principle the dla systems from green et al .
( 1995 ) and from molaro et al .
( 1995 ) could be explained by a galaxy like izw18 experiencing more than one burst of star formation and observed during the interburst period , without invoking an extremely large wind parameter .
in fact , during this phase the n / o ratio increases as due to the fact that oxygen is no more produced while n continues to be restored from low and intermediate mass stars .
this is a well known effect as shown by pilyugin ( 1992;1993 ) . in figs .
4 and 5 we show the [ o , si / fe ] ratios predicted by model 6 and 7 , respectively . here too we can see the oscillating behaviour due to the alternating burst and quiescent phases .
in particular , in both models the [ o , si / fe ] ratios decrease during the interburst phase and increase again during the bursts . the difference between the two models
resides in the efficiency of the wind in model 7 , which reflects in a stronger variation of the abundance ratios between the burst and the interburst phases .
the reason for this variation in model 7 is that the stronger wind acting during bursts is responsible for very low absolute abundances at the end of each burst , so the increase of fe during the interburst is stronger relatively to model 6 where the absolute abundances at the end of each burst are higher .
finally , it is worth noting that current models for elliptical galaxies ( see matteucci and padovani , 1993 ) would never reproduce such the high n / o ratio observed by molaro et al .
( 1995 ) for such low metallicities . in figure 1 , for example , the predictions of models for an elliptical galaxy of initial luminous mass @xmath11 would lie at the right side of the solar neighbourhood curve for only secondary nitrogen , and even in the case of primary n from massive stars would be completely outside of the metallicity region where the dla systems are observed .
the abundance pattern of non refractory elements in some primeval dla systems at high redshift seems not to follow the pattern observed in the halo of the milky way .
this suggests that these objects had a different chemical evolution than the milky way and the spirals in general , at variance with the general belief that dla are the progenitors of present day spirals .
we have shown that the most promising models to explain the observed abundances are those succesfully applied to the dwarf irregular galaxies such as izw18 ( marconi et al .
1994 , kunth et al . 1995 ) .
the conclusion about these dla systems being dwarf galaxies has also been independently suggested by meyer and york ( 1992 ) and by steidel et al .
( 1994 ) from the low abundances found in the few dla observed at low redshifts . - in order to explain the high n / o ratios observed in two dla systems by green et al .(1995 ) and molaro et al.(1995 ) one has to assume that these systems are dwarf irregular galaxies experiencing their first or one of their first bursts of star formation .
these galaxies should also experience strong enriched galactic winds carrying away mostly the products of sn ii explosions such as oxygen and other @xmath0-elements .
in particular , the high n / o abundance ratios could represent either the situation of an interburst phase where n increases and o does not , or the situation of a burst triggering an extremely strong and enriched galactic wind .
-nitrogen in massive stars should have a primary origin as already suggested by matteucci ( 1986 ) .
this seems to be possible but is strongly dependent on the assumed treatment of convection in stellar interiors .
the nucleosynthesis of n in stars of all masses , especially in low and intermediate mass stars , which are the main producers of this element , needs revision and a homogeneous set of calculations for stars of all masses is required .
new yields , but only for stars below 4 @xmath12 , have been recently computed by marigo et al .
unfortunately , the mass range above that mass limit is very important for nitrogen production and therefore firm conclusions can not yet be drawn . - some of the differences in the abundances observed among different dla systems could be due to the fact that some of them are proto - spirals ( see the systems observed by pettini et al .
1995 and lipman , 1995 ) and some are proto - dwarfs ( see the systems observed by molaro et al .
1995 and green et al .
1995 ) - abundance ratios between elements produced from stars at different rates such as n / o and @xmath0/fe represent a very useful tool either to date a galaxy or to understand the nature of high red - shift objects .
|
high redshift dla systems suggest that the relative abundances of elements might be roughly solar , although with absolute abundances of more than two orders of magnitude below solar .
the result comes from observations of the [ sii / znii ] ratio , which is a reliable diagnostic of the true abundance , and from dla absorbers with small dust depletion and negligible hii contamination .
in particular , in two dla systems nitrogen is detected and at remarkably high levels ( vladilo et al .
1995 , molaro et al .
1995 , green et al .
1995 , kulkarni et al .
1996 ) .
here we compare the predictions from chemical evolution models of galaxies of different morphological type with the abundances and abundance ratios derived for such systems .
we conclude that solar ratios and relatively high nitrogen abundances can be obtained in the framework of a chemical evolution model assuming short but intense bursts of star formation , which in turn trigger enriched galactic winds , and a primary origin for nitrogen in massive stars .
such a model is the most successful in describing the chemical abundances of dwarf irregular galaxies and in particular of the peculiar galaxy izw18 .
thus , solar ratios at very low absolute abundances , if confirmed , seem to favour dwarf galaxies rather than spirals as the progenitors of at least some of the dla systems .
|
this paper addresses the problem of a manifestly invariant renormalization procedure for supersymmetric ( susy ) gauge theories .
a peculiar feature of supersymmetric gauge theories is nonpolynomial structure of the action .
the only possibility to reduce the action to a polynomial form is to use the wess - zumino gauge .
however in this case the manifest supersymmetry is lost and the powerful superdiagram technique can not be used . in supersymmetric gauges an infinite number of primitively divergent diagrams is present , which makes the theory formally nonrenormalizable .
nevertheless it can be shown that supersymmetry and gauge invariance reduce the number of independent counterterms , and the usual charge and wave functions renormalizations are sufficient .
the corresponding procedure was firstly constructed for susy qed by one of the present authors @xcite and then applied to abelian and non - abelian models in papers @xcite .
however the proof given in these papers relies on the assumption of a possibility to use an intermediate regularization preserving the symmetry of the theory .
most popular gauge invariant regularizations like dimensional regularization @xcite , dimensional reduction @xcite or lattice regularizations break supersymmetry @xcite .
higher covariant derivative regularization @xcite in principle may be used , but the calculations are quite involved ( see however @xcite ) .
a possible alternative is presented by algebraic renormalization . in this method
one uses an arbitrary regularization or subtraction scheme , breaking the symmetry of the theory ( for example the momentum cut - off regularization or the differential renormalization @xcite ) .
the symmetry is restored at the second step by tuning finite counterterms to provide relevant generalized ward identities ( gwi ) for the renormalized green functions .
this method was applied successfully to susy gauge theories in the papers @xcite , where the invariant renormalizability of @xmath0 and @xmath1 non - abelian susy gauge models was proven in the framework of algebraic renormalization .
however a practical implementation of the algebraic renormalization is rather cumbersome as the procedure requires a tuning of a large number of ( noninvariant ) counterterms .
recently a new method of invariant renormalization was proposed @xcite , which provides automatically the renormalized green functions possessing a relevant symmetry for arbitrary intermediate regularization .
this method was formulated as a special subtraction procedure which incorporates gwi .
solving explicitly the corresponding gwi one reduces the problem of renormalization in an arbitrary regularization scheme to explicitly gauge invariant procedure .
so instead of the two step algebraic renormalization we have the one step procedure which guarantees the symmetry of the renormalized theory . in the present paper
we discuss how this method may be generalized to supersymmetric gauge theories .
renormalization of susy qed is described in details .
a corresponding procedure for non - abelian susy gauge models is under consideration .
the paper is organized as follows : in section [ section_susy_qed ] we introduce the notations and remind some information about supersymmetric quantum electrodynamics .
the universal invariant renomalization scheme for the model is constructed in the next section [ section_universal_renormalization ] and illustrated by an example of the one - loop renormalization in section [ section_one_loop ] .
the results are discussed in the conclusion .
[ section_susy_qed ] in the superspace @xmath0 supersymmetric electrodynamics is described by the following action : @xmath2 where @xmath3 and @xmath4 are chiral superfields , @xmath5 is a real superfield .
the superfield @xmath6 in the abelian case is defined as @xmath7,\ ] ] where @xmath8 is the usual supersymmetric covariant derivative @xmath9 the integration over the superspace in the equation ( [ sqed_action ] ) is defined as @xmath10 up to surface terms these expressions can be written in the explicitly supersymmetric form : @xmath11 here we use the following notations : @xmath12 action ( [ sqed_action ] ) is invariant under the gauge transformations @xmath13 where @xmath14 is an arbitrary chiral superfield .
this invariance allows to gauge away some components of @xmath15 , resulting in the following equation : @xmath16 where @xmath17 is a gauge field , @xmath18 is a maiorana spinor and @xmath8 is a real scalar auxiliary field . in this gauge ,
which is called wess - zumino gauge , residual gauge transformations depend only on a single parameter , while the action is polynomial .
however , this gauge breaks explicitly supersymmetry of the model .
quantization of model ( [ sqed_action ] ) is described in details in book @xcite and is not considered here .
we only note , that the gauge fixing is made by adding the term @xmath19 @xmath20 being a constant .
[ section_universal_renormalization ] in this section we consider a renormalization procedure , which may be used with arbitrary not necessary gauge invariant regularization providing renormalized correlators satisfying automatically susy gwi . in the language of counterterms
it means , that no noninvariant counterterms are needed and renormalization freedom is restricted to the choice of gauge invariant local terms , which are not fixed by gwi .
we assume , that a regularization used for calculations is manifestly supersymmetric allowing to use supergraph technique .
the construction presented below may be used for arbitrary susy gauge models formulated in terms of corresponding superfields .
however in this paper we concentrate on renormalization of susy qed and explicit equations will refer to this model .
we follow the approach , developed in @xcite , where the subtraction procedure incorporating automatically gwi was proposed . to avoid the appearance of spurious infrared divergences we shall work in the `` diagonal '' gauge , corresponding to @xmath21 in equation ( [ gauge_fixing ] ) .
consideration of a general gauge requires additional infrared regularization , which in the abelian case may be achieved by simply adding the mass term for gauge field and putting @xmath22 after calculation of integrals ( see @xcite ) .
it is well known @xcite , that the degree of divergency of a diagram with @xmath23 external lines of chiral and antichiral superfields and @xmath24 external lines of the gauge superfield is equal to @xmath25 where @xmath26 is a number of @xmath27 , @xmath28 , @xmath29 or @xmath30 propagators .
therefore , divergencies are present in the following green functions : will be absent for @xmath31 .
however , in the general case it is impossible to ignore their existence . ]
@xmath32 \equiv \qquad\vphantom{\bigg(}\nonumber\\ & & \equiv \int d^4x_1\ldots d^4x_n\,d^4y\,d^4z\ , \frac{\delta^{n+2}\gamma}{\delta v_{x_1}\ldots \delta v_{x_n}\ , \delta\phi_y\,\delta\phi^+_z}\bigg|_{v,\phi,\tilde\phi=0 } \times\\ & & \qquad\qquad\qquad\qquad \times \exp\big(i ( p_1)_\mu x_1^\mu + \ldots + i ( p_n)_\mu x_n^\mu + i q_\mu y^\mu + i k_\mu z^\mu\big).\nonumber\\ \nonumber\\ & & ( 2\pi)^4 \delta^4\big((p_1)_\mu + \ldots + ( p_n)_\mu\big)\ , \pi \big[(\theta_{x_1},p_1),\ldots ( \theta_{x_n},-p_1-\ldots - p_{n-1})\big ] \equiv \qquad\vphantom{\bigg(}\nonumber\\ & & \equiv \int d^4x_1\ldots d^4x_n\ , \frac{\delta^{n}\gamma}{\delta v_{x_1}\ldots \delta v_{x_n } } \bigg|_{\phi,\tilde\phi , v=0 } \exp\big(i ( p_1)_\mu x_1^\mu + \ldots + i ( p_n)_\mu x_n^\mu \big)\end{aligned}\ ] ] and in the functions @xmath33 , which are constructed similar to functions @xmath34 , but differentiation is performed over @xmath4-fields .
note , that functions @xmath35 for odd @xmath36 are equal to 0 , because contributions of diagrams with loops of @xmath3 are cancelled with contributions of corresponding diagrams with @xmath4-loops .
the functions @xmath34 , @xmath33 and @xmath35 satisfy ward identities @xcite , which in our notations can be written in the following form : @xmath37 = \vphantom{\frac{1}{2}}\nonumber\\ & & = 2 \bar d^2_{x_1 } \delta^4(\theta_y -\theta_{x_1})\times\nonumber\\ & & \qquad\qquad \gamma\big[(\theta_{x_2},p_2),\ldots,(\theta_{x_n},p_n);(\theta_{x_1},q+p_1 ) , ( \theta_z ,- q - p_1-\ldots - p_n)\big ] + \nonumber\\ & & + 2 d^2_{x_1 } \delta^4(\theta_{x_1}-\theta_z ) \times\nonumber\\ & & \qquad\qquad \times \gamma\big[(\theta_{x_2},p_2),\ldots,(\theta_{x_n},p_n);(\theta_y , q ) , ( \theta_{x_1},-q - p_2-\ldots - p_n)\big ] ; \\
\nonumber\\ \label{ward_identity3 } & & \big(d^2_x+\bar d^2_x\big)\bigg(\pi\big[(\theta_x , p),(\theta_y ,- p)\big ] - \frac{1}{2 e^2\xi}\,p^2 \delta^4(\theta_x-\theta_y)\bigg)=0;\\ \nonumber\\ & & \big(d^2_{x_1 } + \bar d^2_{x_1}\big)\,\pi \big[(\theta_{x_1},p_1 ) , \ldots , ( \theta_{x_{n-1}},p_{n-1 } ) , ( \theta_{x_n},-p_1-\ldots - p_{n-1})\big ] = 0 , \quad n>2 .
\nonumber\\ \label{ward_identity4}\end{aligned}\ ] ] the supersymmetric covariant derivative in momentum representation is written as @xmath38 where @xmath39 is the momentum , corresponding to @xmath40-coordinate .
ward identities for functions @xmath33 has the same structure . as the chiral projector @xmath41 can be presented as @xmath42 these identities express the green functions with at least one `` chiral '' gauge field component in terms of correlators with less number of chiral gauge components . equations ( [ ward_identity1 ] ) , ( [ ward_identity3 ] ) and ( [ ward_identity4 ] ) are written for susy qed .
however they have essentially the same structure in non - abelian susy models , differing by the rhs which includes in this case also correlators with faddeev - popov ghost lines .
our strategy may be formulated as follows : we firstly consider one - loop diagrams and renormalize in arbitrary ( infrared safe ) way the diagrams at the rhs of equations ( [ ward_identity1 ] ) ( [ ward_identity4 ] ) . then having in mind that in an anomaly free theory gwi for correlators calculated with arbitrary subtraction scheme may be violated only by local terms
, we define the renormalized green functions at the lhs of equations ( [ ward_identity1 ] ) ( [ ward_identity4 ] ) in such a way , that these identities are satisfied automatically .
this renormalization still may be incomplete , as gwi obviously allow to add to the vertex function at the lhs an arbitrary gauge invariant counterterm .
these counterterms as usual are free parameters which may be fixed by normalization conditions . in the case of susy
qed this procedure looks as follows : first of all it is necessary to renormalize one - loop two - point green function of matter superfields .
the terms , proportional to @xmath43 and not containing @xmath5 in the effective action , have the following structure @xmath44 hence , the two - point function of the matter field can be written as @xmath45 = \bar d^2_x\ , d^2_x \delta^4(\theta_x-\theta_y)\,\sigma(q).\ ] ] renormalized two - point function is defined by subtraction @xmath46 where @xmath47 is a normalization point .
after renormalization of two - point function it is necessary to construct renormalized vertex functions @xmath34 and @xmath48 .
due to supersymmetry the function @xmath34 can be presented in the form @xmath49 = \sum\limits_i b_i(\theta , p ) f_i(p),\ ] ] where @xmath39 and @xmath50 denote all set of arguments , @xmath51 are polynomials in @xmath39 and @xmath50 , which are some linear independent combinations of covariant derivatives , acting on the product of @xmath52 , and @xmath53 are scalar functions of external momenta .
renormalization is performed by subtracting polynomials @xmath54 from the functions @xmath53 .
we choose these polynomials in such a way , that the resulting function satisfies gwi ( [ ward_identity1 ] ) ( [ ward_identity4 ] ) where the rhs includes renormalized functions with less number of chiral external lines .
let us define `` partially renormalized '' function @xmath55 = \sum\limits_i b_i(\theta , p ) \big(f_i(p)-p_i(p)\big),\ ] ] where @xmath54 are some polynomials .
then the lhs of equation ( [ ward_identity1 ] ) may be written in the form @xmath56 = \sum\limits_i ( d^2_{x_1 } + \bar d^2_{x_1 } ) b_i(\theta , p ) \big(f_i(p)-p_i(p)\big).\ ] ] the combinations @xmath57 in the general case are not independent .
it is convenient to introduce linear independent polynomials @xmath58 , proportional to the chiral parts of @xmath59 : @xmath60 expanding the rhs of equation ( [ ward_identity1 ] ) over @xmath58 it is possible to rewrite the ward identity as @xmath61 taking into account , that @xmath58 are linear independent , this equation is equivalent to the following system of linear equations : @xmath62 which represents ward identity ( [ ward_identity1 ] ) expanded in terms of linear independent polynomials @xmath63 . if the polynomials @xmath54 satisfy system ( [ eq ] ) , the function @xmath64 defined by equation ( [ partially_renormalized_function ] ) satisfies susy gwi . indeed ,
@xmath65 = \sum\limits_i c_{ij } q_j(\theta , p ) \big(f_i(p)-p_i(p)\big ) = \sum\limits_j q_j(\theta ,
p ) r_j(p).\ ] ] note , that a choice of @xmath66 is not unique , because any solution of ward identity is defined up to a term @xmath67 , which can not be determined from the ward identity . however , this freedom is irrelevant for our procedure . to eliminate the remaining ultraviolet divergences it is sufficient to subtract from @xmath64 gauge invariant local counterterms @xmath68 @xmath69 = \gamma^r[p,\theta]-p_{gi}.\ ] ] these counterterms obviously can not be fixed by gwi .
so we succeeded to reduce the subtraction procedure in an arbitrary regularization scheme to subtraction of gauge invariant counterterms .
having obtained the function @xmath70 , it is necessary to substitute it into the rhs of renormalized ward identity ( [ ward_identity1 ] ) for the function @xmath71
. then the process is repeated .
so , we constructed a recurrent procedure , which defines all functions @xmath72 .
this procedure is illustrated in the next section by an example of one - loop renormalization of the function @xmath73 in the momentum cut - off regularization .
the functions @xmath74 are constructed in a similar manner . at the next step
it is necessary to renormalize green functions @xmath35 , corresponding to diagrams without external lines of matter superfields .
due to supersymmetry they can be written in the form , similar to equation ( [ gammabf ] ) : @xmath75 = \sum\limits_i b_i(\theta , p ) f_i(p).\ ] ] for example , the function @xmath35 for @xmath76 can be presented as @xmath77 = f_1(p)\,p^2\,\pi_{1/2}\,\delta^4(\theta_x-\theta_y ) + f_2(p)\,\delta^4(\theta_x-\theta_y),\qquad\end{aligned}\ ] ] where @xmath78 in this case @xmath79 as before , the renormalized green functions are obtained by subtracting from @xmath53 some polynomials chosen to provide gwi for the function @xmath80 @xmath81 = \sum\limits_i b_i(\theta , p ) \big(f_i(p ) - p_i(p)\big),\ ] ] for the two - point green function substitution of ( [ original_pi2 ] ) into equation ( [ ward_identity3 ] ) gives the following equation : @xmath82 therefore , @xmath83 while the function @xmath84 can not be defined from gwi and corresponds to a gauge invariant counterterm .
it is convenient to choose @xmath85 where @xmath86 is a normalization point , which can be different from @xmath47 .
then the renormalized two - point green function can be written as @xmath87 = \frac{1}{32 e^2 \xi}\big(d^2 \bar d^2 + \bar d^2
d^2\big ) \delta^4(\theta_x-\theta_y ) + \nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\qquad\qquad + \big(f_1(p ) - f_1(\mu_\pi)\big)\,p^2\,\pi_{1/2}\ , \delta^4(\theta_x-\theta_y),\qquad\end{aligned}\ ] ] this green function satisfies the equation @xmath88 - \frac{1}{2 e^2 \xi}\,p^2 \delta^4(\theta_x-\theta_y)\bigg)=0,\ ] ] which is a supersymmetric generalization of transversality condition in the usual quantum electrodynamics .
green functions , containing only external @xmath5-lines with @xmath89 , can be renormalized similarly .
it is sufficient to put in equations ( [ rewritten_gwi ] ) and ( [ eq ] ) @xmath90
. then equation ( [ partially_renormalized_function ] ) may be used to define the renormalized function @xmath80 .
so , the one - loop renormalization procedure is finished .
due to the locality of the subtraction terms in the limit , when regularization is removed , the present scheme is equivalent to adding the counterterms @xmath91 it is important to note , that due to the presence of @xmath92-functions in @xmath59 all these terms can be presented as integrals over a single @xmath50 . for the noninvariant regularization these counterterms
certainly can be not gauge invariant .
having constructed one - loop counterterms , one can calculate two - loop diagrams and perform similar renormalization at the two - loop level .
all combinatorics is given by the standard r - operation .
after the renormalization in the each loop renormalized green functions will automatically satisfy renormalized ward identities .
it means , that the present scheme provides gauge invariant result for the effective action even if a regularization is not gauge invariant .
[ section_one_loop ] to illustrate application of the scheme , described above , let us consider one - loop renormalization of @xmath0 supersymmetric qed , regularized by cutting - off loop momenta in the gauge with @xmath21 . the diagram , corresponding to the one - loop two - point green function of the matter field is presented at figure [ figure_anomalous_dimension_diagram ] . after simple calculations , using feinman rules , described in book @xcite , in the euclidean space its contribution to the effective action can be written as @xmath93 where integration over loop momentum @xmath94 is defined as @xmath95 and @xmath96 is an ultraviolet cut - off . therefore , according to equation ( [ sigma_definition ] ) the function @xmath97 can be written as @xmath98 two - point function for the matter superfields is renormalized by subtraction ( [ sigma_subtraction ] ) , which corresponds to @xmath99 the next diagram to be considered is the one - loop three - point vertex function , which is described by the diagrams , presented at figure [ figure_vertex_diagrams ] . having calculated them in the chosen regularization we obtained , that the corresponding three - point function can be written as @xmath100 = \nonumber\\ & & \qquad = \bigg\ { -\int\limits^m \frac{d^4k}{(2\pi)^4 } \bigg(\frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big ) } + \frac{e^2}{8 k^2 \big((k+q+p)^2 + m^2\big)}\bigg ) -\nonumber\\ & & \qquad - \int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2 ( k+q+p/2)_\mu}{4 k^2 \big((k+q)^2 + m^2\big ) \big((k+q+p)^2 + m^2\big)}\,\frac{1}{4}\bar d_x \gamma^\mu \gamma_5 d_x + \nonumber\\ & & \qquad + \int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big)\big((k+q+p)^2 + m^2\big)}\ , p^2 ( \pi_{1/2})_x\bigg\ } \times\nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\qquad\qquad \times \bigg(\bar d^2_x \delta^4(\theta_y-\theta_x)\ , d^2_x \delta^4(\theta_x-\theta_z)\bigg ) . \qquad\qquad
\vphantom{\bigg(}\end{aligned}\ ] ] the structure functions @xmath59 entering equation ( [ gammabf ] ) are @xmath101 renormalized vertex function is defined by subtracting the corresponding polynomials : @xmath102 = \nonumber\\ & & = \bigg\ { -\int\limits^m \frac{d^4k}{(2\pi)^4 } \bigg(\frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big ) } + \frac{e^2}{8 k^2 \big((k+q+p)^2 + m^2\big)}\bigg ) - p(p ) -\nonumber\\ & & -\int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2 ( k+q+p/2)_\mu}{4 k^2 \big((k+q)^2 + m^2\big ) \big((k+q+p)^2 + m^2\big ) } \,\frac{1}{4}\bar d_x \gamma^\mu \gamma_5 d_x + \nonumber\\ & & + \int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big)\big((k+q+p)^2 + m^2\big)}\ , p^2 ( \pi_{1/2})_x\bigg\ } \times\nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad \times \bigg(\bar d^2_x \delta^4(\theta_y-\theta_x)\ , d^2_x \delta^4(\theta_x-\theta_z)\bigg ) . \qquad
\vphantom{\bigg(}\end{aligned}\ ] ] note , that we do not subtract anything from the last two terms , because the corresponding integrals are convergent .
taking into account , that @xmath103 the left hand side of ward identity can be written as @xmath104 = \vphantom{\frac{1}{2}}\nonumber\\ & & = \bar d^2_{x } \delta^4(\theta_y -\theta_{x } ) \bar d^2_{x } d^2_{x } \delta^4 ( \theta_x-\theta_z ) \times\vphantom{\frac{1}{2}}\nonumber\\ & & \qquad\qquad\qquad \times \bigg(-\int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2}{8 k^2 \big((k+q+p)^2 + m^2\big ) } - p(p ) \bigg)+\qquad \nonumber\\ & & + d^2_{x } \delta^4(\theta_{x}-\theta_z ) d^2_{x } \bar d^2_{x } \delta^4(\theta_x-\theta_y ) \times\vphantom{\frac{1}{2}}\nonumber\\ & & \qquad\qquad\qquad \times \bigg(-\int\limits^m \frac{d^4k}{(2\pi)^4 } \frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big ) } - p(p ) \bigg)\end{aligned}\ ] ] while the right hand side of the ward identity is written as @xmath105 thus , in this case @xmath106 taking into account , that @xmath107 we obtain , that @xmath108 therefore , the renormalized three point function is equal to @xmath109 = \bigg\ { \sigma^r(q ) + \sigma^r(q+p ) -\nonumber\\ & & \qquad - \int \frac{d^4k}{(2\pi)^4 } \frac{e^2 ( k+q+p/2)_\mu}{4 k^2 \big((k+q)^2 + m^2\big ) \big((k+q+p)^2 + m^2\big)}\,\frac{1}{4}\bar d_x \gamma^\mu \gamma_5 d_x + \nonumber\\ & & \qquad + \int \frac{d^4k}{(2\pi)^4 } \frac{e^2}{8 k^2 \big((k+q)^2 + m^2\big)\big((k+q+p)^2 + m^2\big)}\ , p^2 ( \pi_{1/2})_x\bigg\ } \times\nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad \times \bigg(\bar d^2_x \delta^4(\theta_y-\theta_x)\ , d^2_x \delta^4(\theta_x-\theta_z)\bigg ) .
\qquad\qquad \vphantom{\bigg(}\end{aligned}\ ] ] it is evident , that our subtraction ( together with the corresponding subtraction for @xmath33 ) is equivalent to adding the following counterterms : @xmath110 where we substituted the explicit expression for @xmath111 , constructed above .
so , the renormalization of the three - point function is finished . at the next step it is necessary to consider four - point function with two external matter superfields and continue the recurrent process .
the one - loop contribution to the two - point green function of the gauge field is defined by diagrams , presented at figure [ figure_beta_diagrams ] and is equal to @xmath112 note , that due to using the regularization , which breaks gauge invariance in equation ( [ one - loop_beta_contribution ] ) noninvariant terms are present .
equation ( [ one - loop_beta_contribution ] ) corresponds to @xmath113 in equation ( [ original_pi2 ] ) .
( it is not difficult to find explicit expression for @xmath114 , but it is rather lengthy and we do not present it here . ) after subtraction ( [ pi2_subtraction ] ) renormalized contribution of diagrams with two external @xmath5-lines to the effective action can be presented as @xmath115 so all noninvariant terms disappear and the result becomes finite .
the @xmath116-function , corresponding to expression ( [ v2_gamma_contribution ] ) , is equal to @xmath117 and agrees with calculations , made by dimensional reduction @xcite .
it is easy to verify , that the above subtraction is equivalent to adding the following counterterms : @xmath118 which are not gauge invariant .
in this paper we presented a renormalization procedure for susy qed , which guarantees gauge invariance of the renormalized theory for any intermediate regularization .
susy ward identities are incorporated into subtractions , which allows to avoid the appearance of noninvariant counterterms .
the same procedure obviously may be applied to non - abelian supersymmetric models provided they are formulated in terms of corresponding superfields .
some technical problems related to more complicated structure of non - abelian theories will be considered elsewhere .
|
manifestly invariant renormalization scheme for supersymmetric gauge theories is proposed .
this scheme is applied to supersymmetric quantum electrodynamics . 6.1 in 8.70 in .03 in .03 in .03 in = -12pt
|
historically the first speculations about the existence of active galactic nuclei ( agn ) and super massive black holes ( smbhs ) came from the huge energy requirements implied by the discovery of distant powerful radio galaxies and quasars . today
, radio observations remain crucial to understanding the role of smbhs in astrophysics . only at radio wavelengths
is it possible to image the region immediately surrounding the smbh central engine and the relativistic jets which apparently originate with the smbh .
typical resolution obtained with very long baseline interferometer ( vlbi ) observations at centimeter wavelengths is of the order of 0.001 arccsecond ( 1 milliarcsec ) .
thus , for nearby sources such as those located in the virgo cluster , a linear resolution of 1 milliarcsec corresponds to only about 0.1 parsec or about 100 schwartzchild radii for the smbh located in the nucleus of m87 .
for @xmath1 the resolution for earth - based vlbi is @xmath2 10 kpc .
however , using an earth to space interferometry at @xmath3 1 cm or earth - based systems at millimeter wavelengths , the resolution is improved by more than another order of magnitude .
finally , we note that at radio wavelengths , there is no obscuration , even close to the smbh . however , not all radio sources are due to agn ; and not all agn and smbhs are radio sources . nearly all observed extragalactic radio emission is probably due to synchrotron radiation from ultra relativistic electron with energies @xmath3 10 gev moving in weak magnetic fields with @xmath4 to @xmath5 gauss .
the high energy electrons are thought to be accelerated in one of two ways ; either by a central engine associated with accretion onto a smbh in elliptical galaxies or in quasars , or by supernovae following massive star formation ( starbursts ) in the nucleus of early type galaxies .
unfortunately , both processes are often referred to as agn , and this has led to considerable confusion in the literature .
above @xmath6 mjy , the radio source number - flux density relation is dominated by sources driven by smbhs .
these more powerful sources are characterized by extended radio lobes , and by highly beamed relativistic jets extending from a few parsecs to hundreds of kiloparsecs from the smbh . at microjy levels there is an increasing contribution from star formation related activity rather than from a smbh .
however , there is an uncomfortably large spread in the observed microjy source count even among different observers using the same instrument , the vla , in the same field .
most likely these discrepancies are the result of systematic errors in the reported flux densities due to uncertainties in corrections for the effects of resolution @xcite .
a recent complexity comes from the arcade 2 balloon measurement of a 3 ghz sky brightness of @xmath7 k which is about 5 sigma above that expected from known radio sources , suggesting a possible population of previously unrecognized weak sub microjansky sources ( * ? ? ?
* ( fixen et al . 2011 ) ) . however , deep 3 ghz vla observations showed no evidence for any source population greater than about 30 nanojy .
any population of weaker sources that could produce the excess sky brightness would need to have a sky density greater than 6 x @xmath8 per ster , or 60 times greater than the density of the faintest ( mag 29 ) galaxies in the hubble ultra deep field ( * ? ? ? * ( condon et al .
2012 ) ) . thus if the excess background temperature observed by arcade
is real and due to discrete sources , these sources can not be associated with any known galaxy population .
it will be important to verify that the arcade 2 results were not contaminated by unrecognized galactic radio emission .
radio emission due to agn can usually be distinguished from that due to star formation in a variety of ways .
@xmath9 * morphology : * star formation sources may have dimensions of the order of a few tenths of an arcsecond or a few kiloparsecs at cosmological distances while smbh driven agn sources are typically very small , of the order of 0.001 arcsec ( 10 pc ) or less , and are coincident with the galaxy nucleus or qso .
quasars and agn powered by smbhs are often variable on time scales as short as days with corresponding changes in their morphology indicating highly collimated outflows with apparent superluminal velocity .
smbh driven agn may also contain extended lobes tens or hundreds of kiloparsecs distant from the compact nucleus , and sometimes show optical and radio jets joining the nucleus and radio lobes .
@xmath9 * radio spectra : * star forming sources and the extended radio lobes of agn generally have steep radio spectra . due to synchrotron self absorption ,
the compact sources generally have flat or even inverted spectra .
@xmath9 * brightness temperature : * star forming sources mostly have measured brightness temperature up to @xmath10 k while the compact flat spectrum sources associated with agn have brightness temperature @xmath11 k or more .
@xmath9 * radio luminosity : * star forming regions typically have a radio luminosity close to @xmath12 w / hz and follow the well known correlation between radio and fir luminosity @xcite .
radio galaxies and quasars driven by smbhs may be @xmath13 times more luminous so their radio luminosity greatly exceeds that expected from the radio / fir relation characteristic of star forming regions .
@xmath9 * x and @xmath14-ray emission : * star forming regions are only weak x - ray sources with typical luminosity @xmath15 ergs / sec while smbh driven agn can be strong x - ray , @xmath14-ray , and tev sources .
@xmath9 * host galaxies : * smbh driven radio sources are located in the nuclei of elliptical galaxies or are associated with quasars which themselves are thought to be the bright nuclei of elliptical galaxies that greatly outshine their host galaxy .
low ( optical ) luminosity agn are typically found in early type spiral ( often classified as seyfert ) galaxies .
radio emission from star forming regions is typically associated with spiral galaxies but may also be found in the host galaxies of radio quiet quasars ( see section 4 ) .
perhaps the first suggestions that the nuclei of galaxies may contain more than just stars came from sir james jeans in 1929 who remarked in his book on astronomy and cosmogony ( * ? ? ?
* ( jeans 1929 ) ) , + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the centres of the nebulae are of the nature of singular points at which matter is poured into our universe from some other and entirely spatial dimension so that to a denizen of our universe , they appear as points at which matter is being continuously created . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the modern understanding of the important role of galactic nuclei probably began with the famous paper by karl seyfert ( 1943 ) who reported on his study of broad strong emission lines in the nucleus of seven spiral nebulae .
interestingly , although seyfert s name ultimately became attached to the broad category of spiral galaxies with active nuclei , his 1943 paper received no citations until 1951 , and apparently went unnoticed until baade and minkowski ( * ? ? ?
* ( 1954 ) ) drew attention to the similarity of the cygnus a radio source spectrum with that of the galaxies studied by seyfert . not until the 1949 nature paper by bolton , stanley , and slee ( * ? ? ?
* ( 1949 ) ) did astronomers finally recognize the vast energy requirements of radio galaxies .
bolton et al .
had identified three of the strongest discrete radio sources with the crab nebula , m87 , and ngc 5128 , until that time the discrete radio sources were widely thought to be associated with galactic stars .
this was understandable , as karl jansky and grote reber had observed radio emission from the milky way . the milky way is composed of stars , so it was natural to assume that the discrete radio sources had a stellar origin .
bolton et al . understood the importance of their identification of the taurus a radio source with the crab nebula which was widely recognized as the remnant of the 1054 supernova reported by chinese observers .
bss correctly identified two other strong sources with m87 and ngc 5128 , but realizing that if they were extragalactic , their absolute radio luminosity would need to be a million times more luminous than that of the crab nebula , they argued that `` ngc 5128 and ngc 4486 ( m87 ) have not been resolved into stars , so there is little direct evidence that they are true galaxies . ''
so they concluded that they are within our own galaxy .
indeed their paper carried the title positions of three discrete radio sources of galactic radio frequency radiation . "
john bolton later argued that he really did understand that m87 and ngc 5128 were very luminous radio sources , but that he was concerned that that in view of their apparent extraordinary radio luminosity , nature might not publish their paper .
the following years saw the identification of more radio galaxies , and the changed paradigm which had previously considered all discrete radio sources to be stellar to one with most high latitude sources were assumed to be extragalactic .
the energy requirements were exacerbated in 1951 with the identification of cygnus a , the second strongest radio source in the sky with a magnitude 18 galaxy at what was then considered a high redshift of 0.056 and a corresponding radio luminosity about @xmath16 times more luminous than m87 and ngc 5128 ( * ? ? ? * ( baade and minkowski 1954 ) ) .
the total energy contained in relativistic particles and magnetic fields in the radio lobes of cygnus a and other powerful radio galaxies was estimated to be at least @xmath17 ergs @xcite .
hoyle , fowler , burbidge and burbidge ( * ? ? ?
* ( 1964 ) ) were apparently the first to call attention to gravitational collapse as a possible energy source to power radio galaxies . by the middle of 1960 , many radio sources had been identified with galaxies having red shifts up to 0.24 @xcite .
typically the optical counterpart of strong radio sources was identified with an elliptical galaxy that was the brightest member of a cluster . in 1960 , rudolph minkowski ( * ? ? ? * ( 1960 ) ) identified 3c 295 with a mag 20 galaxy at z=0.46 .
3c 295 is about ten times smaller than cygnus a and ten times more distant consistent with the idea that the smallest radio sources might be path finders to finding very distant galaxies .
but a few months later caltech radio astronomers identified the first of several very small sources with what appeared to be galactic stars , thus raising questions about the extragalactic nature of other small diameter radio sources .
while searching for ever more distant radio galaxies , caltech radio astronomers john bolton and tom matthews identified 3c 48 with an apparent stellar object . at the 107th
meting of the american astronomical society held in new york in december 1960 , allan sandage @xcite reported the discovery of the first true radio star . " before he left to return to australia , john bolton @xcite speculated that 3c 48 had a high redshift of 0.37 , but was apparently dissuaded by jesse greenstein and ira bowen on the grounds that there was a 3 or 4 angstrom discrepancy among the corresponding rest wavelengths . in a subsequent analysis of the complex emission line spectrum ,
jesse greenstein @xcite interpreted the 3c 48 spectrum in terms of emission lines from highly ionized states of rare earth elements .
he briefly considered a possible redshift of 0.37 , but quickly dismissed the possibility that 3c 48 was extragalactic .
nearly two years would pass , and other compact radio sources would be identified as galactic stars before a series of lunar occultations would lead to the identification of 3c 273 with a star like object at a redshift of 0.16 and the immediate realization that 3c 48 was also extragalactic with a redshift of 0.37 leading to the recognition of quasi stellar radio sources or `` quasars '' as the extremely bright nuclei of galaxies .
the apparent high radio as well as optical luminosity of quasars , coupled with their very small dimensions presented a further challenge to understanding the source of energy and how this energy is converted to relativistic particles and magnetic fields .
the following years led to the identification of more quasars at ever larger redshifts and the suggestion that quasars are powered by accretion onto super massive black holes ( smbh ) with masses up to @xmath18 solar masses or more @xcite . generally , the identified quasars had a significant uv excess compared with stars , so due to the redshift of their spectrum , they appeared blue on photographic plates facilitating their identification with radio sources with even modest position accuracy . in 1965 , sandage noted that the density of blue stellar objects on the sky was some thousand time greater than that of 3c radio sources .
sandage argued that what he called `` quasi stellar galaxies '' are related to quasars , except that they are not strong radio sources .
but , his paper was widely attacked , perhaps in part because of the perceived irregular treatment by the astrophsyical journal .
sandage s paper was received on may 15 , 1965 at the astrophysical journal , but s. chandrasekar , the apj editor was apparently so impressed by sandage s claim for a new constituent of universe " that he held up publication of the journal , and sandage s paper appeared in the may 15 issue
. tom kinman @xcite along with lynds and villere @xcite argued that most of sandage s blue stellar objects were only blue galactic stars , while fritz zwicky @xcite pointed out that he had previously called attention to this phenomena , and he later accused sandage of `` one of the most astounding feats of plagiarism '' @xcite .
as it turned out , most of sandage s blue stellar objects were just that , blue stellar objects , " and
only some ten percent of optically identified quasars are strong radio sources .
but , it has now been more than half a century since we have divided quasars into the two classes of radio loud and radio quiet quasars , and it has still not been clear if there are two distinct populations or rather whether the radio loud population is merely the extreme end of a continuous distribution of radio luminosity .
proponents of each interpretation claim that the other interpretation is due to selection effects . many of the previous investigations designed to distinguish between radio loud and radio quiet quasars were limited by contamination from low luminosity agn with absolute optical magnitudes greater than -23 , biased samples based on radio rather than optical selection criteria , and inadequate sensitivity to detect radio emission from most of the radio quiet population . in an attempt to overcome these limitations , kimball et al .
@xcite observed 179 quasars selected from the sdss .
all of the these quasars were within the redshift range 0.2 to 0.3 and were brighter than @xmath19 = -23 so were genuine quasars that presumably contained a smbh .
the observations were made with the jansky very large array at 6 ghz reaching an rms noise of 6 @xmath20jy .
all but about 6 quasars were detected as radio sources with an observed radio luminosity sharply peaked between @xmath21 and @xmath22 watts / hz characteristic of the radio luminosity typically observed from star forming galaxies .
about ten percent of the sdss sample are strong radio sources with radio luminosities ranging up to @xmath23 w / hz .
kimball et al . concluded that the radio emission from radio quiet quasars is due to star formation in the host galaxy .
similar conclusions were reached by padovani et al .
( 2011 , 2014 ) based on the identification and classification of the microjy radio sources found in a deep vla survey of the extended chandra deep field south .
based on radio , optical , ir , and x - ray data , padovani et al . concluded that the microjy radio emission from agn , like that of galaxies , is powered primarily by starbursts , and not the smbhs which powers the agn .
condon et al .
@xcite argue that these starbursts are fueled by the same gas that flows into the smbh that powers the quasar and thus accounts for the co - evolution of star formation and smbhs .
shortly after the recognition of quasars , radio source observations in both the soviet union @xcite and the u.s .
@xcite demonstrated variability on time scales of months or less .
this presented a problem .
causality arguments suggested linear dimensions , d @xmath24 c@xmath25 where c is the speed of light and @xmath25 the characteristic time scale of the observed variability . knowing the quasar redshift and corresponding distance puts a limit to the angular size which for many variable sources was only @xmath26 arcseconds and the corresponding lower limit to the brightness temperature which appeared to be significantly in excess of the inverse compton limit of @xmath27 k for most variable sources , the apparent violation of the inverse compton limit is now understood in terms of relativistic beaming . due to relativistic effects
, we observe apparent jet speeds , luminosities , and brightness temperatures which are related to the corresponding intrinsic quantities in the agn rest frame through the doppler factor , @xmath28 , the lorentz factor , @xmath14 , and the jet orientation , @xmath29 , with respect to the line of sight @xcite .
the apparent velocity transverse to the line of sight , @xmath30 , the apparent luminosity , @xmath31 , the apparent brightness temperature , @xmath32 and the doppler factor , @xmath28 , can be calculated from the lorentz factor , @xmath14 , @xmath29 , and the intrinsic luminosity , @xmath33 .
the apparent transverse velocity @xmath30 is given by @xmath34 where @xmath35 for small values of @xmath29 , because the radiating source is almost catching up with its own radiation , equation 5.1 shows that the apparent transverse can exceed the speed of light , which is commonly referred to as `` superluminal motion . ''
the apparent luminosity , @xmath31 , is given by @xmath36 where the doppler factor , @xmath28 , is @xmath37 and where @xmath33 is the luminosity that would be measured by an observer in the agn frame .
the quantity @xmath38 depends on the geometry and spectral index and is typically in the range between 2 and 3 .
the lorentz factor , @xmath14 , is given by @xmath39 quasars or agn with highly doppler boosted relativistic jets pointed nearly along the line - of - sight are often referred to as `` blazars . '' blazars are characterized by rapid flux density variability , apparent superluminal motion , and strong x - ray and @xmath14-ray emission .
high resolution observations of blazars provide unique insight to the process by which relativistic jets are accelerated and collimated in the region close to the smbh .
we want to understand : @xmath9 how and where is the relativistic beam accelerated and collimated into narrow jets ? are there accelerations or decelerations ?
do all parts of the jet move at the same speed ?
@xmath9 what causes the curvature of jets ?
does the flow follow a curved trajectory or is the motion ballistic and characteristic of a rotating nozzle ?
@xmath9 does the observed apparent velocity reflect the true bulk velocity of motion ?
what determines the jet velocity ? is the velocity related to other properties such as radio , optical , x , or @xmath14-ray luminosity ?
@xmath9 what is the maximum observed brightness temperature ? does it exceed the inverse compton limit ?
@xmath9 what is the energy production mechanism ? @xmath9 what can we learn from radio observations about the nature of the smbh ?
very long baseline observations made since 1971 have confirmed the apparent superluminal motion expected from highly relativistic bulk motion . since 1995 , the nrao very long baseline array ( vlba ) has been used to study the motions of a large sample of quasars and agn at 7 mm by a group from boston university @xcite and by the international mojave group @xcite .
more detailed information may be found on the respective web sites : boston university 7 mm program : * http://www.bu.edu/blazars/vlbaproject.html * mojave 2 cm program : * : //www.physics.purdue.edu/ mlister / mojave/ * the results of these programs may be summarized as follows .
@xmath9 radio loud quasars and agn show highly relativistic bulk motion with a broad distribution of apparent velocities . in general , the jets appear one sided , probably due to differential doppler boosting so that the approaching jet appears much brighter than the receding one .
each jet appears to have its own characteristic velocity but there is an appreciable spread in the apparent velocity of the different features within a given jet . the typical apparent velocity , @xmath40 corresponding to an intrinsic value of @xmath41 .
the maximum observed apparent velocity , @xmath42 corresponds to an intrinsic value of @xmath43 .
the parent jet population is mostly only mildly relativistic , but is under represented in flux density limited samples due to the effect of doppler boosting .
@xmath9 the jets with the fasted apparent velocities have the highest apparent luminosity , likely reflecting a correlation between intrinsic speed and intrinsic luminosity rather than simply being the result of doppler boosting ( cohen et al .
2007 ) .
@xmath9 apparent inward motions are uncommon and are likely the result of a feature moving outward along a curved trajectory approaches the line - of - sight so that the apparent separation from the jet base transverse to the line - of - sight appears to decrease with time .
@xmath9 individual jet features may show both apparent accelerations and decelerations . both the apaprent speed and direction of motion may change with time , but changes in speed are more common than changes in direction , indicating real changes in the lorentz factor
as features propagate down the jet .
in general the apparent speed is greater further down the jet , so that acceleration must take place at distances at least up to @xmath3 100 pc from the base of the jet ( homan et al .
2009 , 2014 ) .
@xmath9 many jets show a curved structure and in a few cases there is evidence of an oscillatory behavior .
sometimes the flow appears to follow pre - existing channels ; other times the flow appears ballistic as from a rotating nozzle , perhaps due to precession possibly resulting from a binary black hole pair ( lister et al .
2013 ) .
@xmath9 in some cases the direction of ejection appears to vary within a well defined cone forming what appears to be an edge brightened jet ( lister et al .
2013 ) such as shown by the jet in the nearby radio galaxy m87 where there is sufficient linear resolution to resolve the jet transverse to its structure @xcite .
@xmath9 there appears to be a relation between radio and gamma ray emission .
there is statistical evidence that radio outbursts follow a @xmath14-ray event by @xmath3 1 month , but it has been difficult to convincingly establish a one to one correlation between individual radio and @xmath14-ray events @xcite .
as described above , inverse compton scattering limits the maximum observed brightness temperature , t @xmath44 . at the inverse compton limit
, the energy contained in relativistic particles greatly exceeds that in the magnetic field which is perhaps not unreasonable in a very young source .
if the particle and magnetic energies are in equilibrium , then the corresponding brightness temperature is only @xmath45 .
the observed brightness temperature may be calculated from @xcite @xmath46 where k is the boltzman constant , @xmath29 the angular size in milliarcsec , s is the flux density in janskys , and @xmath47 the wavelength in cm .
the resolution of a radio interferometer , is given by the ratio of the observing wavelength , to the interferometer baseline , d ; or @xmath29 = @xmath47/d .
putting this back into eqn .
6.1 gives @xmath48 so the maximum brightness temperature that may be measured depends only on the flux density and baseline length , and is independent of wavelength . for ground based observations with a maximum baseline of @xmath49 km ,
the highest brightness temperatures which can be reached are @xmath50 k. recent observations with the russian radioastron space vlbi satellite have suggested lower limits to brightness temperatures of 3c 273 and other sources @xmath0 k ( kellermann et al .
2014 , kovalev 2014 ) , or at least two to three orders of magnitude greater than the limit set by inverse compton cooling .
several explanations are possible @xcite .
\1 . for a relativistically
beamed source the apparent brightness temperature is boosted by a factor @xmath28 .
to explain the high observed brightness temperatures in this way would require doppler factors , @xmath51 @xmath52 to @xmath16 .
but typical observed values of @xmath53 with maximum observed values @xmath54 , and for 3c 273 , @xmath55 ( lister et al .
possibly the bulk flow which is related to the doppler boosting might be much greater than the pattern flow observed by the vlba , but cohen et al .
( 2007 ) have shown that this is unlikely .
. the observed emission may be coherent such as observed in pulsars or the sun including possible stimulated synchrotron emission .
\3 ) the radio emission might be the result of synchrotron emission from protons rather than electrons which would enhance the upper limit to the brightness temperature by about the ratio of the proton to electron mass or more than a factor of 1000 .
however , proton synchrotron radiation would require a magnetic field strength more than @xmath56 times stronger than needed for electron synchrotron radiation of the same strength at the same wavelength .
\4 ) there may be a continuous acceleration of relativistic particles which balances the energy losses due to inverse compton cooling .
about ten percent of quasars and bright elliptical galaxies are strong radio sources with radio luminosity , @xmath57 and are thought to be driven by accretion onto a smbh .
the weaker radio sources , @xmath58 are mostly due to star formation in the host galaxy . the observed properties of radio jets can be interpreted in terms of a highly relativistic outflow from a central engine driven by a smbh of up to @xmath18 solar masses .
but , many questions remain .
@xmath9 most quasars and agn are not strong radio sources . why are only @xmath3 10% of quasars strong radio sources , although all quasars presumably contain a smbh to account for their extraordinary optical luminosity .
@xmath9 how do smbhs generate relativistic jets ?
@xmath9 how are the jets confined and shaped as they propagate away from the smbh ?
what are the relative roles of velocity shear , hydrodynamic turbulence , shocks , and plasma instabilities in shaping the form and kinematics of relativistic jets ?
@xmath9 why do only some jets produce @xmath14-rays ?
how and where are the @xmath14-rays produced ? what is the relation between radio and @xmath14-ray emission ?
@xmath9 is there evidence for binary black hole pairs ?
see the paper by @xcite in this volume .
@xmath9 if confirmed , observations of 3.3 ghz sky brightness combined with the density of faint sources detected in deep vla observations suggest the possible existence of new population of faint radio sources not due to star formation or to agn and unrelated to any known galaxy population ?
@xmath9 are there other emission processes which play a role beside incoherent synchrotron radiation ?
the national radio astronomy observatory is operated by associated universities , inc . under cooperative agreement with the national science foundation .
i am indebted to many colleagues , especially ron ekers and members of the mojave team for numerous discussions that have contributed to this paper .
|
although the radio emission from most quasars appears to be associated with star forming activity in the host galaxy , about ten percent of optically selected quasars have very luminous relativistic jets apparently powered by a smbh which is located at the base of the jet .
when these jets are pointed close to the line of sight their apparent luminosity is enhanced by doppler boosting and appears highly variable .
high resolution radio interferometry shows directly the outflow of relativistic plasma jets from the smbh .
apparent transverse velocities in these so - called `` blazars '' are typically about 7c but reach as much as 50c indicating true velocities within one percent of the speed of light .
the jets appear to be collimated and accelerated in regions as much as a hundred parsecs downstream from the smbh .
measurements made with earth to space interferometers indicate apparent brightness temperatures of @xmath0 k or more .
this is well in excess of the limits imposed by inverse compton cooling .
the modest doppler factors deduced from the observed ejection speeds appear to be inadequate to explain the high observed brightness temperatures in terms of relativistic boosting .
|
estimation of power spectral density ( spectrum ) of a continuous time , mean square continuous , stationary stochastic process is a classical problem . generally the estimation is based on finitely many samples of the process .
it is well known that if the spectrum is compactly supported ( bandlimited ) , then it can be estimated from uniformly spaced samples , provided the sampling is done at the nyquist rate or faster ( kay , 1999 ) . for sampled non - bandlimited processes , or bandlimited processes sampled at sub - nyquist rate ,
the problem of aliasing leads to biased estimation .
for this reason , it is sometimes argued that a non - bandlimited spectrum can never be estimated consistently from uniformly spaced samples at any fixed sampling rate ( shapiro and silverman , 1960 ; masry , 1978 ) .
consequently , some researchers have turned to non - uniform sampling schemes such as stochastic sampling and periodic non - uniform sampling .
masry ( 1978 ) proved the consistency of some spectrum estimators based on stochastic sampling schemes , under appropriate conditions that allow the underlying spectrum to be non - bandlimited .
for bandlimited processes , it has been shown that periodic non - uniform sampling at sub - nyquist average rate can lead to consistent spectrum estimation ( marvasti , 2001 ) .
it is important to note that the argument of inconsistency of spectrum estimators computed from uniformly spaced samples is based on the assumption that the sampling rate remains fixed even as the sample size goes to infinity .
however , when one has the resources to increase the sample size indefinitely , one would like to use some of those resources to sample faster , rather than being constrained by a fixed sampling rate .
in fact , it has been shown ( srivastava and sengupta , 2010 ) that if this constraint is removed , and the sampling rate is allowed to increase suitably as the sample size goes to infinity , then the smoothed periodogram can be a consistent estimator of a non - bandlimited spectral density .
it should be noted that uniform sampling is generally far easier to implement than non - uniform sampling .
for this reason , the fact of consistency of a spectrum estimator computed from uniformly spaced samples of a non - bandlimited process is noteworthy .
this fact also gives rise to further questions about the estimator , such as its convergence in distribution , and construction of asymptotic confidence intervals for the spectrum , based on the estimator .
these questions have so far not been addressed through asymptotic calculations that allow the sampling rate to go to infinity .
this is what we propose to do in this paper .
the asymptotic approach chosen here ( referred to as ` shrinking asymptotics ' by fuentes , 2002 ) was also adopted by other authors ( e.g. , constantine and hall , 1994 ; hall et al . , 1994 ;
lahiri , 1999 ) , although the asymptotic distribution of the smoothed periodogram has not been studied previously .
this approach is different from the ` fixed - domain asymptotics ' or ` infill asymptotics ' approach ( chen et al . , 2000 ;
stein , 1995 ; zhang and zimmerman , 2005 ; lim and stein , 2008 ) which , in the present case , would have required that the time - span of the original continuous - time data ( before sampling ) remains fixed as the sampling rate goes to infinity .
let @xmath0 be a vector - valued mean square continuous stationary stochastic process , having zero mean .
we denote the components of the process @xmath1 by @xmath2 for @xmath3 , and the variance - covariance matrix of the process @xmath1 at lag @xmath4 by @xmath5 where @xmath6~\mbox{for}~a_1,a_2\in\{1,2,\ldots , r\}.\ ] ] the spectral and cross - spectral density matrix of the process @xmath1 is denoted by @xmath7 where @xmath8 in this paper , we consider the following estimator of @xmath9 for @xmath10 : @xmath11}(\lambda),\ ] ] where @xmath12 is a covariance averaging kernel , @xmath13 is the kernel bandwidth , @xmath14 is the sampling rate and @xmath15 is the indicator of the event @xmath16 . in section 2 ,
we establish the consistency of the spectrum estimator ( [ est ] ) for non - bandlimited processes , which is a generalization of a result of srivastava and sengupta ( 2010 ) to the case of multivariate time series .
it paves the way for our main result on the asymptotic distribution of the estimator , given in section 3 .
section 4 contains some discussion on optimal rates of convergence . in section 5
, we investigate the question as to how large the sample size should be in order for the applicability of the asymptotic distribution and the resulting pointwise confidence intervals .
we look for answers through a monte carlo simulation study and report the findings .
all the proofs are given in the appendix .
in order to establish the consistency of the estimator @xmath17 given in ( [ est ] ) , we make a few assumptions on the process @xmath18 , the kernel @xmath12 and the sequences @xmath13 and @xmath14 . the function @xmath19 , defined over the real line as @xmath20 .
assumption 2 .
the covariance averaging kernel function @xmath12 is continuous , even , square integrable and bounded by a non - negative , even and integrable function having a unique maximum at 0 .
further , @xmath21 .
the kernel window width is such that @xmath22 as @xmath23 .
the sampling rate is such that @xmath24 and @xmath25 as @xmath26 .
note that assumption 4 implies that @xmath27 as @xmath23 . in order to establish convergence of the variance - covariance matrix ,
we need a further assumption involving cumulants . recall that the @xmath28-th order joint cumulant of the random variable @xmath29 is given by @xmath30 where the summation is over all partitions @xmath31 of size @xmath32 , of the index set @xmath33 .
the fourth moment @xmath34 $ ] is finite for all @xmath35 , while the fourth order cumulant function defined by @xmath36\ ] ] does not depend on @xmath37 , and this function , denoted by @xmath38 , satisfies @xmath39 where @xmath40 are all continuous , even , nonnegative and integrable functions over the real line , which are non - increasing over @xmath41 for all @xmath42 .
note that the cross spectral density is , in general , complex valued .
thus , the proposed estimator @xmath17 can be represented as the vector @xmath43 im\left(\widehat{\phi}_{a_1a_2}(\lambda)\right ) \end{array } \right),\ ] ] where @xmath44}(\lambda),\\ & & \hskip-20pt im(\widehat{\phi}_{a_1a_2}(\lambda))\\ & = & \frac{1}{2\pi n\rho_n}\sum_{t_1=1}^{n}\sum_{t_2=1}^{n}k(b_n(t_2-t_1))x_{a_1}\left(\frac{t_1}{\rho_n}\right)x_{a_2}\left(\frac{t_2}{\rho_n}\right)\sin\left(\frac{(t_2-t_1)\lambda}{\rho_n}\right ) 1_{[-\pi\rho_n,\pi\rho_n]}(\lambda).\\\end{aligned}\ ] ] the covariance between two complex - valued random variables is often defined as the trace of the @xmath45 cross - covariance matrix of the random vectors formed by their real and imaginary parts ( brockwell and davis , 1991 ) . in the case of the pair @xmath46 , the limiting covariance according to this notion can be easily be computed from theorem 2 .
theorem 1 and theorem 2 together establish the consistency of any vector of estimators having elements of the form @xmath17 .
we will make an additional assumption about the underlying process in order to prove the asymptotic normality of the estimator .
the process @xmath1 is strictly stationary ; all moments of the process exist , i.e. , @xmath47<\infty$ ] for each @xmath48 and for all @xmath49 ; and for each @xmath50 and each @xmath48 , the @xmath51th order joint cumulant denoted by @xmath52 satisfies @xmath53 where @xmath54 are continuous , even , nonnegative and integrable functions over the real line , which are non - increasing over @xmath55 . note that assumption 5a is stronger than assumption 5 .
the following theorem describes the asymptotic behaviour of the joint cumulants of the estimators @xmath17 for @xmath56 . in the present case
, a cumulant defined as in ( [ cumulant ] ) may be complex - valued . .
_ under assumptions 14 and 5a , a vector of real and imaginary parts of estimated spectra or cross - spectra converges weakly as follows .
@xmath57 \stackrel{d}{\rightarrow}n_{2j}(0,\sigma),\ ] ] where @xmath58 , and the elements of @xmath59 are defined in accordance with theorem 2 . _
the foregoing theorem only shows that the vector estimator , after appropriate mean adjustment and scaling , converges weakly to a multivariate normal distribution .
however , weak convergence around the true vector of spectra and cross - spectra remains to be established .
note that @xmath60\right)\\ & + \sqrt{nb_n}\left(e[\widehat{\phi}_{a_1a_2}(\lambda)]-\phi_{a_1a_2}(\lambda)\right ) .
\end{split}\ ] ] we make some further assumptions on the smoothness and the rate of decay of the spectrum and the shape of the kernel function in order to obtain the rate of convergence of the bias @xmath61-{\phi_{a_1a_2}}(\lambda)$ ]
. the function @xmath62 , defined over the real line as latexmath:[\[g_{qa_1a_2}(t)=\sup_{|s|\ge @xmath56 , for some positive number @xmath64 greater than 1 .
the power spectral density is such that , for all @xmath56 and for some @xmath65 , @xmath66 for some non - negative number @xmath67 . for any kernel @xmath12 ,
let us define @xmath68 for each positive number @xmath69 such that the limit exists .
the characteristic exponent of the kernel is defined as the largest number @xmath69 , such that the limit exists and is non - zero ( parzen , 1957 ) . in other words ,
the characteristic exponent is the number @xmath69 such that @xmath70 is @xmath71 .
the characteristic exponent of the kernel @xmath12 is a number , for which assumption 1a holds .
note that assumption 1a implies assumption 1 , and also that @xmath72 is @xmath73 $ ] times differentiable , where @xmath73 $ ] is the integer part of @xmath64 .
thus , the number @xmath64 indicates the degree of smoothness of the spectral density .
if assumption 1a holds for a particular value of @xmath64 , then it would also hold for smaller values .
the number @xmath74 indicates the slowest rate of decay of the various elements of the power spectral density matrix .
the following are two interesting situations , where assumption 1b holds . 1
. the real and imaginary parts of the components of the power spectral density matrix are rational functions of the form @xmath75 , where @xmath76 and @xmath77 are polynomials such that the degree of @xmath77 is more than degree of @xmath76 by at least @xmath74 .
note that continuous time arma processes possess rational power spectral density .
the function @xmath78 has the following smoothness property : @xmath78 is @xmath74 times differentiable and the @xmath79 derivative of @xmath78 is in @xmath80 .
theorem 5 shows that the second term in ( [ asym1 ] ) would go to zero if the sampling rate @xmath14 satisfies additional conditions .
the sampling rate is such that @xmath81 and @xmath82 as @xmath23 .
note that , whenever assumption 3 holds , assumption 4a is stronger than assumption 4 . with this assumption ,
the expected values of the estimators in theorem 4 can be replaced by the respective true values .
we are now in a position to optimize the rates of @xmath13 and @xmath14 so that @xmath83 tends to 0 as fast as possible under the conditions of theorem 6 .
it has been shown in srivastava and sengupta ( 2010 ) that under the assumptions of theorems 2 and 5 , the optimal rate of convergence for mean square consistency of the estimator ( [ est ] ) is given as @xmath84=o\left(n^{-\frac { 2pq}{p+q+2pq}}\right),\ ] ] which corresponds to the choices @xmath85 theorem 7 shows that the optimal rate of weak convergence of the estimator @xmath17 is slower than the square root of the optimal rate corresponding to mean square consistency .
it is important to note that for every fixed value of @xmath64 , the number @xmath74 , which indicates rate of decay of the spectrum , can be increased indefinitely by continuous time low pass filtering with a cut off frequency larger than the maximum frequency of interest .
there are well - known filters such as the butterworth filter , which have polynomial rate of decay of the transfer function with specified degree of the polynomial , that can be used for this purpose . for fixed @xmath64 ,
the best rate of weak convergence given in theorem 7 , obtained by allowing @xmath74 to go to infinity , happens to be @xmath86 .
the rate of weak convergence crucially depends on the number @xmath64 , the assumed degree of smoothness of the spectrum .
the stronger the assumption , the faster is the rate of convergence .
the rate corresponding to @xmath87 ( weakest possible assumption ) is @xmath88 , assuming that @xmath74 can be allowed to be very large .
for very large @xmath64 ( very strong assumption ) and large @xmath74 , the rate approaches @xmath89 .
with a view to investigating the applicability of the asymptotic results reported in section 3 to finite sample size , we consider the bivariate continuous time linear process @xmath90
x_2(t)\\ \end{array}\right)=\left(\begin{array}{c } \int_{-\infty}^{t}h_{1}(t - u)z_1(u)du+\int_{-\infty}^{t}h_{2}(t - u)z_2(u)du\\[2ex ] \int_{-\infty}^{t}h_{3}(t - u)z_1(u)du+\int_{-\infty}^{t}h_{4}(t - u)z_3(u)du\\ \end{array}\right)\ ] ] where @xmath91 , @xmath92 are independent continuous time white noise and @xmath93 for @xmath94 .
the elements of the spectral density matrix @xmath95 are defined as follows ( hoel et al . , 1972 ) .
@xmath96 we simulate this bivariate process with the choices @xmath97 , @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102 , @xmath103 and @xmath104 .
note that for this process , assumption 1a holds with @xmath105 and assumption 1b holds with @xmath106 .
for the purpose of estimation , we make these assumptions with @xmath107 and @xmath108 . in accordance with this choice of @xmath64 ,
we use the second order kernel function @xmath109}(x).\ ] ] we also use the rates @xmath110 and @xmath111 .
we estimate the bivariate spectrum matrix for frequencies in the range @xmath112 $ ] at intervals of @xmath113 ( i.e. , @xmath114 uniformly spaced grid points ) .
we subsequently compute the normalized statistics @xmath115}}\right),\\ t_4(\lambda)&=&\sqrt{nb_n}\left(\frac{im(\widehat\phi_{12}(\lambda))-im(\phi_{12}(\lambda ) ) } { \sqrt{b[\widehat\phi_{11}(\lambda)\widehat\phi_{22}(\lambda)-\{re(\widehat\phi_{12}(\lambda))\}^2 + \{im(\widehat\phi_{12}(\lambda))\}^2]}}\right)\left[1 - 1_{\{0\}}(\lambda)\right],\\\end{aligned}\ ] ] in accordance with theorem 2 . according to theorem 6
, the asymptotic distribution of each of these four statistics is standard normal .
this procedure is repeated for 500 simulation runs . by regrading the values of the above statistics for the different simulation runs as four data sets of size 500 each
, we calculate the kolmogorov - smirnov test statistic ( shorak and wellner , 1986 ) for these data sets , and the corresponding p - value .
this procedure is repeated for the 301 frequency values mentioned above .
the percentage of p - values ( across 301 frequency values ) exceeding the number 0.05 are reported in table 1 , for sample sizes @xmath116 , 1000 , 10000 and 100000 .
the table shows that for each statistic , the percentage approaches the ideal value of 95 very slowly as @xmath117 increases .
we now turn to computation of confidence limits of the power spectral density . for each frequency value , we compute the 95% asymptotic confidence intervals of @xmath118 , @xmath119 , @xmath120 and @xmath121 from the statistics @xmath122 , @xmath123 , @xmath124 and @xmath125 , assuming that the latter have the standard normal distribution .
subsequently , we compute the fraction of times ( out of 500 simulation runs ) the confidence limits capture the true value of the function .
these percentages are plotted against the frequency , for sample sizes @xmath116 , 1000 , 10000 and 100000 , in figure 1 .
it is seen that the observed fraction approaches the ideal coverage probability ( 0.95 ) for larger sample sizes . since there is a discontinuity of the asymptotic variance function at the point @xmath126 , while the estimated spectrum is constrained to be continuous , some anomalous behaviour in the neighbourhood [ cols="^ , > , > , > , > " , ] table 1 .
observed percentage of frequencies ( in the range 0 to @xmath127 ) for which p - values of the kolmogorov - smirnov statistics for testing normality of @xmath118 , @xmath119 , @xmath120 and @xmath121 are greater than 0.05 ( ideal percentage is 95% ) .
empirical coverage probability ( based on 500 simulation runs ) of pointwise confidence intervals of @xmath118 , @xmath119 , @xmath120 and @xmath121 for sample sizes 100 , 1000 , 10000 and 100000 . of
the the point @xmath126 is expected .
this results in substantially lower values of the empirical coverage probability in this region .
however , this region of anomaly is observed to shrink as the sample size increases .
it would be interesting to note that the empirical coverage probability is reasonably close to the ideal coverage probability for most frequency values when the sample size as small as 1000 , even though table 1 indicates that the asymptotic distribution is not applicable at this sample size .
we denote by @xmath128 a function that bounds the covariance averaging kernel @xmath12 as in assumption 2 .
further , we denote @xmath129 by @xmath130 .
we shall show that the bias of the estimator @xmath131 given by ( [ est ] ) converges to @xmath132 uniformly over @xmath133 $ ] for any @xmath134 , @xmath135 such that @xmath136 .
note that @xmath137 = \frac{1}{2\pi \rho_n}\sum_{u=-(n-1)}^{n-1}\left(1-\frac{|u|}{\rho_n}\right)k(b_nu)c_{a_1a_2 } \left(\frac{u}{\rho_n}\right)e^{-\frac{iu\lambda}{\rho_n}}1_{[-\pi\rho_n,\pi\rho_n]}(\lambda).\ ] ] consider the simple function @xmath138 , defined over @xmath133\times\mathbb{r}$ ] , by @xmath139}(\lambda)1_{\left(\frac{u-1}{\rho_n},\frac{u}{\rho_n}\right]}(x).\end{aligned}\ ] ] observe that @xmath140 $ ] .
define the function @xmath141 , over @xmath133\times\mathbb{r}$ ] , by @xmath142 for any @xmath143 , let @xmath144 be the smallest integer greater than or equal to @xmath145 .
note that the interval @xmath146 $ ] contains the point @xmath147 and @xmath148 . for sufficiently large @xmath117
, we have from assumptions 3 and 4 , @xmath149}(\lambda).\end{aligned}\ ] ] proving the uniform convergence of @xmath150 $ ] over the finite interval @xmath151 $ ] amounts to proving @xmath152 uniformly over @xmath151 $ ] .
observe that @xmath153 , which is continuous . by virtue of the continuity of the limiting function , ( [ alt ] )
is equivalent to proving that @xmath154 converges continuously over this interval ( resnick , 1987 ) , i.e. , for any sequence @xmath155 , @xmath156 where @xmath157 $ ] . by continuity of the function @xmath158 with respect to @xmath147 and @xmath159
, we have from assumptions 3 and 4 , for any fixed @xmath147 , @xmath160 note that from assumptions 1 and 2 , we have the dominance @xmath161}(x)\le m g_{a_1a_2}(x),\end{aligned}\ ] ] where @xmath19 is the function described in assumption 1 .
thus , by applying the dominated convergence theorem ( dct ) , we have ( [ star ] ) .
hence , @xmath162\rightarrow \phi(\lambda)$ ] uniformly on @xmath151.\hfill\box$ ] now consider the function @xmath166 . by using the transformations @xmath167 , @xmath168 and @xmath169
, we have @xmath170 the range of the four summations on the right hand side is described by the set of inequalities @xmath171 and @xmath172 , which is equivalent to the inequalities @xmath173 and @xmath174 .
therefore , the expression for @xmath166 simplifies to @xmath175 where @xmath176 by writing the cosine functions in terms of complex exponentials , we have @xmath177 where the four terms correspond to the four summands appearing within braces in the last factor on the right hand side of ( [ eq54 ] ) . by using the results of lemmas 1 and 2 given below
, we have the convergence @xmath178and similar arguments show that @xmath179 for the function @xmath180 , one can similarly use the transformations @xmath167 , @xmath181 and @xmath182 , interchange the order of summation and expand the cosine functions in terms of complex exponentials to obtain @xmath183 by using similar arguments as in the case of @xmath184 , it can be shown that @xmath185 finally , for the term @xmath186 , we use the transformations @xmath187 , @xmath188 and @xmath189 and interchange the order of summations to have @xmath190 from assumptions 2 and 5 , we have @xmath191 now consider the function @xmath138 defined over @xmath192 as @xmath193}(x ) . \end{split}\ ] ] observe that @xmath194 and @xmath195 is dominated by @xmath196 . by applying dct ,
we have @xmath197 thus , the upper bound of @xmath198 given by ( [ i3bound ] ) is @xmath199 .
assumption [ as4 ] ensures that @xmath198 converges to zero uniformly . by combining all these terms ,
we have the convergence of @xmath200 $ ] as given in the theorem . convergence of the other three covariances follow from a similar argument .
+ @xmath201 proof of lemma 1 . consider a compact subset @xmath202 of the set @xmath203 .
consider the simple function @xmath138 , defined over @xmath204 by @xmath205}(x_1 ) 1_{\left(\frac{(u_2 - 1)}{\rho_n},\frac{u_2}{\rho_n}\right]}(x_2)1_{\left(\frac{(u_3 - 1)}{\rho_n},\frac{u_3}{\rho_n}\right]}(x_3 ) .
\end{split}\ ] ] so that @xmath206 define @xmath207 , @xmath208 and @xmath209 as the smallest integers greater than or equal to @xmath210 , @xmath211 and @xmath212 , respectively .
thus , @xmath213\times\left(\frac{u_{2n-1}(x_2)}{\rho_n},\frac{u_{2n}(x_2)}{\rho_n}\right]\times \left(\frac{u_{3n-1}(x_3)}{\rho_n},\frac{u_{3n}(x_3)}{\rho_n}\right]$ ] and @xmath214 as @xmath215 . since @xmath22 and @xmath216 as @xmath23 , we have , for any point @xmath217 and large enough @xmath117 , the inequalities @xmath218 , i.e. , @xmath219 .
thus , for sufficiently large @xmath117 , we have @xmath220 observe that , under assumptions 1,3 and 4 , the function @xmath221 converges to the function @xmath141 , defined over @xmath204 by @xmath222 observe also that @xmath223 is a continuous function in @xmath224 .
as in the proof of theorem 1 , we prove the convergence of the left hand side of uniformly on @xmath202 , by showing that for any sequence @xmath225 , @xmath226 for @xmath227 .
the latter convergence follows , through assumption 1 and 2 and the dct , from the dominance @xmath228 and the convergence of the integrand , which holds because of the continuity of @xmath229 , @xmath230 and the kernel and the exponential functions .
hence , @xmath231 converges as stated uniformly on the compact set @xmath232 proof of lemma 2 .
let @xmath202 be any compact subset of the set @xmath203 .
consider the simple function @xmath138 , defined over @xmath204 by @xmath233}(x_1 ) 1_{\left(\frac{(u_2 - 1)}{\rho_n},\frac{u_2}{\rho_n}\right]}(x_2)1_{\left(\frac{(u_3 - 1)}{\rho_n},\frac{u_3}{\rho_n}\right]}(x_3 ) .
\end{split}\ ] ] so that @xmath206 an argument similar to that used in the proof of lemma 1 shows that for @xmath234 and sufficiently large @xmath117 , @xmath235 where @xmath207 , @xmath208 and @xmath209 are the smallest integers greater than or equal to @xmath210 , @xmath236 and @xmath212 , respectively .
for obtaining the uniform convergence of @xmath184 , consider @xmath237 where the function @xmath238 is defined over @xmath204 by @xmath239 we will show the uniform convergence of the right hand side of by considering the two terms separately . for the first term , we follow the route taken in the proof of theorem 1 , i.e. , show that for any sequence @xmath240 , @xmath241 . for this purpose ,
we write the above integral as @xmath242 is defined over @xmath204 by @xmath243 now observe that @xmath244 where @xmath245 since @xmath246 as @xmath215 , we have @xmath247 since from assumption 1 and 2 , we have the dominance @xmath248 by applying the dct , we have @xmath249 turning to the second term on the right hand side of ( [ 2part ] ) , observe that for any fixed @xmath250 , @xmath251 thus , @xmath252 and so @xmath253 from assumption 1 and 2 , we have the dominance @xmath248 which leads us , through another use of the dct , to the convergence of the second integral of ( [ 2part ] ) .
this establishes that the first term on the right hand side of ( [ twostep ] ) converges to 0 .
we only have to deal with the second term .
let @xmath254 in order to establish the uniform convergence of @xmath255 over @xmath202 , it is enough to show that @xmath256 for any sequence @xmath257 , where @xmath258 . by using the reimann - lebesgue lemma
, we have @xmath259 .
thus , the second term on the right hand side of ( [ twostep ] ) also converges to 0 .
hence , @xmath184 converges to 0 uniformly on @xmath202
as @xmath260 @xmath261 can be written as @xmath262 it follows that @xmath263 now @xmath264 where the summation is over all _ indecomposable _ ( brillinger , 2001 ; leonov and shiryayev , 1959 ) partitions @xmath265 , such that @xmath266 , @xmath267 , of the table @xmath268 and @xmath269 , @xmath267 .
since the partition @xmath270 is indecomposable , we have @xmath271 define @xmath272 note that @xmath273 for @xmath267 .
then the joint cumulant of @xmath274 , @xmath275 , @xmath276 , @xmath277 given by is absolutely bounded by @xmath278\right|\cdots \left|k[b_n(u_{2l-1}+t_{p_{2l-1}}'-u_{2l}-t_{p_{2l}}')]\right|\\ & \qquad\times\left| c_{a_{j_{11}}a_{j_{12}}\ldots a_{j_{1k_1}}}\left(\frac{u_{j_{11}}}{\rho_n},\ldots,\frac{u_{j_{1,k_1 - 1}}}{\rho_n } \right)\ldots c_{a_{j_{p1}}a_{j_{p2}}\ldots a_{j_{pk_p}}}\left(\frac{u_{j_{p1}}}{\rho_n},\ldots,\frac{u_{j_{p , k_p-1}}}{\rho_n } \right)\right|,\end{split}\ ] ] where @xmath279 is that member of the set @xmath280 which satisfies @xmath281 for @xmath282 .
we will now show that the set @xmath283 has @xmath284 linearly independent elements .
note that the set @xmath285 consists of differences of pairs of elements of the set @xmath286 .
so the set @xmath285 can have at most @xmath284 linearly independent differences .
suppose that the set @xmath285 has exactly @xmath287 linearly independent differences for some @xmath288 .
denote the @xmath287 independent differences of the set @xmath285 by @xmath289 where @xmath290 .
let , if possible , @xmath291 , and consider a difference @xmath292 for @xmath293 which is linearly independent of the elements of the set @xmath294 . since the partition @xmath270 is indecomposable , the sets @xmath295 and @xmath296 _ communicate _
( leonov and shiryayev , 1959 ) .
therefore , there exists an index set @xmath297 with @xmath298 , which is a proper subset of @xmath280 , such that @xmath299 , @xmath300 and the pairs @xmath301 are _ hook _ ( leonov and shiryayev , 1959 ) .
consequently , there exist indices @xmath302 such that for @xmath303 , one of the points @xmath304 and @xmath305 belongs to @xmath306 and the other belongs to @xmath307 .
it follows that for @xmath303 , @xmath308 is in @xmath285 , and hence , they can be written as linear combinations of the members of @xmath294 .
note that for @xmath303 , @xmath309 is equal to either @xmath308 or @xmath310 . thus , @xmath311 can be written as a linear combination of the members of @xmath294 .
this fact contradicts the assumption that @xmath292 is linearly independent of the elements of the set @xmath294 .
therefore , @xmath312 can not be larger than 1 .
this proves that the set @xmath285 can not contain fewer than @xmath284 linearly independent differences .
consider the @xmath284 linearly independent elements of the set @xmath294 , where @xmath313 , and define @xmath314 using the above transformation , and by replacing the @xmath315 sums over indices @xmath316 by @xmath284 sums overs the indices @xmath317 , we find that the joint cumulant given in ( [ 2ndbound ] ) is bounded from above by @xmath318 the above simplification has been made by taking into account the upper bound for @xmath319 copies of @xmath12 and conservative estimates of the ranges of summation of @xmath317 .
now one can rewrite the expression in ( [ 3rdbound ] ) as follows .
@xmath320\ldots\left[\sum_{v_{p-1}=-3n}^{3n}k(b_n v_{p-1})b_n\right]\\ & \times\left\{\sum_{u_{j_{11}}=-(n-1)}^{n-1}\ldots\sum_{u_{j_{1,k_1 - 1}}=-(n-1)}^{n-1}\left|c_{a_{j_{11}}a_{j_{12}}\ldots a_{j_{1k_1}}}\left(\frac{u_{j_{11}}}{\rho_n},\ldots , \frac{u_{j_{1,k_1 - 1}}}{\rho_n } \right)\right|\frac{1}{\rho_n^{k_1 - 1}}\right\}\ldots\\ & \times\left\{\sum_{u_{j_{p1}}=-(n-1)}^{n-1}\ldots\sum_{u_{j_{p , k_{p}-1}}=-(n-1)}^{n-1}\left|c_{a_{j_{p1}}a_{j_{p2}}\ldots a_{j_{pk_p}}}\left(\frac{u_{j_{p1}}}{\rho_n},\ldots , \frac{u_{j_{p , k_p-1}}}{\rho_n } \right)\right|\frac{1}{\rho_n^{k_p-1}}\right\}. \end{split}\ ] ] consider the simple function @xmath138 defined over @xmath192 by @xmath321}(x).\ ] ] note that @xmath322 , and from assumption 2 we have the dominance @xmath323 . by applying the dct
, we have @xmath324 this fact establishes the convergence of the sums over @xmath317 .
consider the simple function @xmath325 defined over @xmath326 by @xmath327}(x_1)\ldots 1_{\left(\frac{u_{j_{1,k_{1}-1}}-1}{\rho_n},\frac{u_{j_{1,k_{1}-1}}}{\rho_n}\right]}(x_{k_1 - 1 } ) .
\end{split}\ ] ] note that @xmath328 from assumption 5a , we have that the function @xmath325 is bounded by an integrable function .
thus , by applying the dct , we have @xmath329 likewise , we have the convergence for the remaining @xmath284 sets of summations . using these above convergence results , the upper bound of ( [ 2ndbound ] ) given in ( [ 4thbound ] )
can be written as @xmath330 where @xmath331 are appropriate constants .
the summation is over the finite number of indecomposable partitions , and the worst - case value of the partition size @xmath315 is @xmath332 .
therefore , the upper bound is @xmath333 .
this completes the proof of theorem 3.@xmath201 note that the first moment of the random vector on the left hand side of ( [ meancorrected ] ) is zero and the second moment converges in accordance with theorem 2 .
further , @xmath334 for any set of constants @xmath335 . from the above fact and theorem 3 , for all @xmath48 , the absolute value of the @xmath51th order joint cumulant of the random vector on the left hand side of ( [ meancorrected ] ) is bounded from above by an @xmath336 term . according to assumption 3 , this upper bound tends to 0 as @xmath117 tends to infinity .
this completes the proof.@xmath201 the result can be proved along the lines of the proof of theorem 3 of srivastava and sengupta ( 2010).@xmath201 the weak convergence of the first term on the right hand side of ( [ asym1 ] ) follows from theorem 4 . on the other hand ,
the second term can be written , in view of theorem 5 , as @xmath337-\phi_{a_1a_2}(\lambda_1)\right ) = \sqrt{nb_n}\left(o\left((\rho_nb_n)^{q}\right)+o\left(\frac{\rho_n}{n}\right)+o\left(\frac{1}{\rho_n^{p}}\right)\right).\ ] ] under assumption 3 , @xmath338 therefore , under assumptions 3 and 4a , the right hand side of ( [ asymrate ] ) goes to zero as @xmath23 .
this completes the proof.@xmath201 note that under assumption 4a , we have @xmath339 and @xmath340 from ( [ 2nd ] ) and ( [ 1st ] ) , we have @xmath341 the right hand sides of ( [ rate1 ] ) and ( [ rate2 ] ) are increasing and decreasing functions , respectively , of @xmath14 .
assumption 3 , together with ( [ 2nd ] ) , indicate that @xmath14 goes to infinity as @xmath117 goes to infinity .
the rate given by ( [ rate1 ] ) will be unduly slow if @xmath14 goes to infinity too slowly , while the rate given by ( [ rate2 ] ) will be unduly slow if @xmath14 goes to infinity too fast . at either event , @xmath342 will have a sub - optimal rate of convergence to zero .
it follows that @xmath342 has the fastest convergence to zero if @xmath343 this condition requires that @xmath344 . for this rate of @xmath14 , ( [ rate1 ] ) implies that @xmath345 this completes the proof.@xmath201
brillinger , d. r. ( 2001 ) . _ time series data analysis and theory_. philadelphia : siam .
lahiri , s. n. ( 1999 ) .
asymptotic distribution of the empirical spatial cumulative distribution function predictor and prediction bands based on a subsampling method .
_ probab . theory related fields _ 114 , 55-84 .
srivastava , r. and sengupta , d. ( 2010 ) .
consistent estimation of non - bandlimited spectral density from uniformly spaced samples . to appear in _ ieee trans .
_ , ( preprint available in ` arxiv:0906.5045v2 ` ) .
|
it is well known that if the power spectral density of a continuous time stationary stochastic process does not have a compact support , data sampled from that process at any uniform sampling rate leads to biased and inconsistent spectrum estimators . in a recent paper
, the authors showed that the smoothed periodogram estimator can be consistent , if the sampling interval is allowed to shrink to zero at a suitable rate as the sample size goes to infinity . in this paper ,
this ` shrinking asymptotics ' approach is used to obtain the limiting distribution of the smoothed periodogram estimator of spectra and cross - spectra .
it is shown that , under suitable conditions , the scaling that ensures weak convergence of the estimator to a limiting normal random vector can range from cube - root of the sample size to square - root of the sample size , depending on the strength of the assumption made .
the results are used to construct asymptotic confidence intervals for spectra and cross spectra .
it is shown through a monte - carlo simulation study that these intervals have appropriate empirical coverage probabilities at moderate sample sizes .
_ keywords : _ power spectral density , spectrum estimation , smoothed periodogram , shrinking asymptotics , asymptotic confidence interval .
|
we consider index - coding problems , first introduced by birk and kol @xcite , where a sender communicates with multiple receivers simultaneously through a shared broadcast medium .
the aim is to find the shortest codeword that the sender needs to broadcast in order for each receiver , knowing some of the messages broadcast by the sender a priori , to obtain its requested message .
index - coding problems have been receiving much attention lately due to its equivalence to network - coding problems @xcite .
each index - coding problem instance can be fully described by a directed or an undirected graph .
bar - yossef et al .
@xcite characterized the optimal index codelength for graphs of the following types : ( a ) directed and acyclic , ( b ) undirected and perfect , ( c ) undirected odd holes of five or more vertices , and ( d ) undirected odd anti - holes of five or more vertices .
in general , the index - coding problem remains open to date , though lower and upper bounds have been obtained @xcite . a lower bound
is given by the number of vertices in a maximum acyclic induced subgraph ( mais ) @xcite .
the _ minrank _ function @xcite of the graph gives an upper bound ( i.e. , achievability ) , and it also gives the optimal _ linear _ index codelength .
both the mais lower bound and the minrank upper bound are np - hard to compute @xcite , and both have been shown to be loose in some instances @xcite .
this implies that linear index codes , though having practical advantages of simplifying encoding and decoding , are not necessarily optimal . in this paper ,
we extend existing results to a new class of problem instances : we show that if an mais is formed by removing two or fewer vertices , then the mais lower bound is achievable using linear index codes , meaning that linear index codes are optimal for this class of index - coding problems . to this end , we show that this class of graphs must contain some special configurations ; by proposing a new coding scheme on these special configurations , we are able to construct the required optimal index code .
this incidentally characterizes a class of ( infinitely many ) graphs where the minrank upper bound and the mais lower bound coincide .
let there be @xmath0 receivers , @xmath1 , and each receiver @xmath2 requests a message @xmath3 from the sender .
the sender knows all the messages , @xmath4 .
it encodes @xmath5 into a length-@xmath6 codeword @xmath7 .
it then broadcasts the codeword to all receivers noiselessly to allow each receiver @xmath2who knows some prior side information @xmath8to decode its requested message , i.e. , @xmath9 , for all @xmath10 . here , @xmath11 is the index code , and @xmath6 is the index codelength . the aim is the find the minimum index codelength , denoted by @xmath12 .
each index - coding problem instance is completely specified by @xmath13 for all @xmath10 .
it can also be fully described by a directed graph @xmath14 , consisting of a set of vertices @xmath15 and a set of arcs @xmath16 .
an arc from vertex @xmath2 to vertex @xmath17 exists , denoted by @xmath18 , if and only if receiver @xmath2 knows @xmath19 , or equivalently , @xmath20 . by definition
, there is no self loop or parallel arc . for an arc @xmath21 ,
vertex @xmath2 is the _ tail _ and @xmath17 the _
head_. we term this graphical representation _ side - information graph_. each vertex @xmath2 in @xmath14 represents both message @xmath22 and receiver @xmath2 . for the special case where for each @xmath18 , there exists an arc @xmath23 , the index - coding problem instance can also be represented by an undirected graph @xmath24 , consisting of the same set of vertices @xmath25 , and a set of edges @xmath26 , where @xmath27 if and only if @xmath18 , which also means @xmath23 .
we denote by @xmath28 the optimal ( or minimum ) index codelength of an index - coding problem instance represented by graph @xmath14 , which can be directed or undirected .
bar - yossef et al .
have shown the following lower bound @xcite : @xmath29 where @xmath30 is the number of vertices in a maximum acyclic induced subgraph ( mais ) of @xmath14 .
an mais is obtained by finding an acyclic vertex - induced subgraph that has the largest number of vertices . for any @xmath14 , denote an mais by @xmath31 , and the set of vertices removed by @xmath32 , where @xmath33 is the number of removed vertices .
note that @xmath33 for each @xmath14 is fixed , but the mais may not be unique . for a _ vertex - induced _ subgraph ,
when we remove a vertex @xmath34 from a graph @xmath14 , we remove all outgoing arcs from and all incoming arcs to @xmath34 , but we must keep all other remaining arcs in @xmath14 .
in this paper , we derive the optimal index codelength , together with the corresponding optimal index code , for a class of graphs , for any message alphabet size @xmath35 .
recall that for any directed graph @xmath14 , the number of vertices we need to remove to obtain an mais is denoted by @xmath33 . with this
, we now state the main result of this paper .
[ theorem-2 ] for any @xmath14 where @xmath36 , the optimal index codelength is given by @xmath37 consider the index - coding problem instance depicted in figure [ fig : bipartite](a ) as an example .
an mais can be formed by removing @xmath38 ( the choice is not unique ) . from theorem
[ theorem-2 ] , we have @xmath39 . an optimal index code is @xmath40 ( which is also not unique ) , where @xmath41 denotes addition modulo @xmath42 .
this problem instance is not covered by the results by bar - yossef et al .
@xcite , as the side - information graph is neither acyclic nor undirected .
neely et al .
@xcite and tehrani et al .
@xcite have shown the following achievability : [ lemma ] if a directed graph @xmath14 contains @xmath43 vertex - disjoint cycles , then the index codelength of @xmath44 is achievable . for figure [ fig : bipartite](a )
, their scheme only achieves @xmath45 , which is strictly suboptimal .
recently , shanmugam et al .
@xcite have shown that an upper bound of @xmath12 is given by the _ local chromatic number _ of the complement graph of @xmath14 , denoted by @xmath46 .
the local chromatic number of @xmath46 is the maximum number of colors in any , minimized over all proper coloring of the undirected counterpart ( by ignoring the arc direction ) of @xmath46 . from figure
[ fig : bipartite](c ) , we see that this scheme achieves @xmath47 .
bipartite graphs are also used to represent a more general in fact , the most general class of index - coding problem instances where a message can be requested by more than one receiver ( cf .
side - information graphs ) .
neely et al .
* theorem 1 ) found @xmath12 for all acyclic bipartite graphs ; yu and neely @xcite found @xmath12 for all planar bipartite graphs . the bipartite graph that represents our example is depicted in figure [ fig : bipartite](b ) , which contains cycles . also , ignoring the arc direction , we _ contract _
the edges @xmath48 to obtain a complete bipartite graph on three and three vertices , commonly denoted by @xmath49 .
since any planar graph can not contain a @xmath49 minor , figure [ fig : bipartite](b ) is not planar .
tehrani et al .
@xcite have proposed a _
packet decomposition _
scheme to obtain an upper bound on bipartite graphs .
achievability of the scheme was derived on the assumption that @xmath50 is a large finite field . even if the results can be shown to hold for smaller alphabets
, the scheme can only achieve @xmath45 for the instance in figure [ fig : bipartite](a ) .
consequently , our results strictly extend the existing results .
recently , we built on the results of this paper to characterize @xmath12 for all @xmath14 up to five vertices @xcite .
the optimal _ linear _ index codelength for any graph @xmath14 is given by its minrank value @xcite .
characterizing graphs having a certain minrank value is hard ; dau et al .
@xcite managed to characterize all undirected graphs whose minrank value is @xmath51 or @xmath52 , and all directed graphs whose minrank value is 2 or @xmath53 .
they are , however , unable to characterize directed graphs whose minrank value is @xmath52 or @xmath51 . for any directed graph @xmath14 whose @xmath30 equals @xmath52 or @xmath51 , we show in this paper that linear index codes are optimal , meaning that @xmath30 equals its minrank .
so , we have incidentally characterized a subset of directed graphs whose minrank equals @xmath52 or @xmath51 .
as @xmath30 is a lower bound to @xmath12 , we only need to prove achievability . recall that @xmath33 is the minimum number of vertices we need to remove from @xmath14 to make it acyclic , i.e. , to obtain an mais .
firstly , suppose that @xmath33 = 0 .
sending all messages uncoded achieves the index codelength @xmath53 , and we have .
next , suppose that @xmath33 = 1 .
the directed graph @xmath14 must contain at least one cycle ; otherwise , @xmath54 .
it follows from lemma [ lemma ] that @xmath55 is achievable .
lastly , @xmath56 .
there are two possibilities for @xmath14 : 1 .
there exist two vertex - disjoint cycles , or 2 .
there are no two vertex - disjoint cycles . for case ( i )
, it again follows from lemma [ lemma ] that @xmath57 is achievable .
the savings of two symbols ( compared to sending all @xmath53 symbols uncoded ) can be achieved using a cyclic code on each disjoint cycle .
for example , for a cycle of length @xmath58 , say @xmath59 , we send the following @xmath60 coded symbols , thereby saving one symbol : @xmath61 .
since each receiver @xmath2 knows at least one other symbol in the cyclic code , it can decode its required @xmath22 . is to show that if @xmath62 and condition ( ii ) is true , then @xmath14 must contain a subgraph @xmath63 shown above . here , every arrow represents a path , which is denoted by a capital letter .
the paths do not share common vertices except the end points .
vertices in each path is denoted by the corresponding small letter , indexed in the direction of the arcs , e.g. , path @xmath64 is @xmath65 .
all paths except @xmath66 , @xmath67 , and @xmath68 must contain one or more arcs.,width=283 ] the main challenge of theorem [ theorem-2 ] is to show that for case ( ii ) , even though we can not find two vertex - disjoint cycles , we can still save two symbols . to this end , we will show that if @xmath62 and if condition ( ii ) above is true , then there exists a subgraph in @xmath14 of a certain configuration , stated as follows : [ lemma : special - subgraph ] if @xmath62 , and there are no two vertex - disjoint cycles ( i.e. , condition ( ii ) ) , then @xmath14 must contain a subgraph ( not necessarily an induced subgraph ) shown in figure [ fig : three - cycles ] .
we will design a special code on this subgraph to show that @xmath51 is indeed achievable .
we will first present the code in the next section , and then prove lemma [ lemma : special - subgraph ] in section [ sec : lemma ] .
we now propose a new coding scheme that achieves the codelength @xmath51 for case ( ii ) .
we need to show that each receiver @xmath69 can decode its intended message , i.e. , @xmath22 .
we will propose a code for the subgraph in figure [ fig : three - cycles ] , denoted by @xmath63 , and send the rest of the messages ( which correspond to the vertices in @xmath70 ) uncoded .
this means all receivers @xmath71 can decode their intended messages , and we only need to show that all receivers @xmath72 can also decode their intended messages .
we propose the following coding strategy : for each vertex @xmath73 , with all its denoted by @xmath74 , we send the code symbol @xmath75 . for each path in @xmath63 ( denoted by a capital letter ) , we denote the vertices therein by its corresponding small letter , indexed in the direction of the arcs . for example , path @xmath64 is @xmath65 . in figure
[ fig : three - cycles ] , we have @xmath76 , i.e. , the first vertex in path @xmath64 is the last vertex in path @xmath68 , which is also the first vertex in path @xmath77 .
we use the above coding strategy to send the code symbol for all vertices in @xmath63 except @xmath78 and @xmath79 . by design , all receivers except @xmath78 and @xmath79can decode their requested messages , as each receiver @xmath80 knows a priori the messages corresponding to the head of all outgoing arcs from @xmath80 .
so , we only need to show that receivers @xmath78 and @xmath79 can decode their respective requested messages , @xmath81 and @xmath82 .
we start with receiver @xmath83 .
knowing @xmath84 a priori , it decodes along path @xmath77 ( i.e. , @xmath85 ) , to get @xmath86 , and continues along path @xmath66 to get @xmath87 . in the event that path @xmath66 has zero arc , it can also obtain @xmath88 , which is @xmath89 . also knowing @xmath90 a priori , receiver @xmath78 decodes along path @xmath64 to get @xmath91 ,
from which it can decode backward along path @xmath92 to get @xmath93 . having decoded @xmath88 earlier and now @xmath93 , it obtains @xmath94 from @xmath95 . with @xmath94 , it decodes along path @xmath96 and then path @xmath68 to get @xmath97 . in the event that path @xmath68 has zero arc ,
the receiver would have obtained @xmath98 earlier . for receiver @xmath79 , knowing @xmath99 , it decodes along path @xmath100 to get @xmath101 .
it then decodes backward along path @xmath96 to get @xmath94 .
also knowing @xmath102 , receiver @xmath79 decodes along path @xmath103 , and then path @xmath66 to get @xmath104 .
in the even that path @xmath66 has zero arc , receiver @xmath79 would have obtained @xmath105 earlier .
knowing both @xmath106 and @xmath94 , the receiver obtains @xmath93 from @xmath107 .
it then decodes along path @xmath92 , and then path @xmath67 to get the requested @xmath108 .
if path @xmath67 has zero arc , the receiver can also obtain @xmath109 . in this coding scheme , we send one symbol for each vertex ( coded symbols for @xmath110 , and uncoded symbols for the rest ) except for @xmath78 and @xmath79 .
we have shown that this index code satisfies the decoding requirements of all receivers , meaning that @xmath51 is achievable .
we first give an intuition for lemma [ lemma : special - subgraph ] , by showing that there must exists three interlinked cycles in @xmath14 , in subsection [ sec : intuition ] . in subsections [
sec : figure ] to [ sec : figure - end ] , we prove that these three interlinked cycles must assume the configuration in figure [ fig : three - cycles ] .
let @xmath112 , i.e. , vertices @xmath113 and @xmath114 are removed from @xmath14 to get an mais .
we first show the following : [ proposition:3-cycles ] there exist three cycles in @xmath14 , each containing either @xmath113 , @xmath114 , or both @xmath113 and @xmath114 .
every cycle must contain @xmath113 , @xmath114 , or both .
otherwise , removing @xmath113 and @xmath114 will not give an acyclic induced subgraph .
suppose that there is only one cycle in @xmath14 .
removing any vertex from the cycle gives an acyclic induced subgraph .
hence , @xmath115 .
( contradiction ) suppose that there are only two cycles in @xmath14 .
note that these two cycles can not be vertex - disjoint , as per condition ( ii ) above .
so , these two cycles must shared at least one vertex , and removing only this shared vertex gives an acyclic induced subgraph , i.e. , @xmath115 . (
contradiction ) so , there must exist at least three cycles .
we further show some properties of these three cycles : [ proposition : no - common ] there exist three cycles in @xmath14 , where 1 .
any two cycles must have at least one common vertex , and 2 .
the three cycles do not have any common vertex .
it follows from proposition [ proposition:3-cycles ] that there are at least three cycles . as no two cycles are vertex - disjoint , we have property 1
. arbitrarily select one cycle , say @xmath116 .
consider every other cycle @xmath117 , and denote the set of common vertices between @xmath118 and @xmath116 as @xmath119 .
since every @xmath118 shares some vertex with @xmath116 , we have @xmath120 .
now suppose that @xmath121 , meaning that some vertex is shared among all cycles . then removing only this vertex from @xmath14
would have resulted in an acyclic subgraph ( contradiction ) .
so , there must exist two cycles , say @xmath122 and @xmath123 , where @xmath124 .
selecting @xmath116 , @xmath122 , and @xmath123 gives property 2 .
denote the subgraph formed by the three cycles in proposition [ proposition : no - common ] by @xmath63 .
we have the following : [ proposition : subgraph ] the subgraph @xmath63 , formed by the three cycles in proposition [ proposition : no - common ] , satisfies both the following : ( 1 ) we can not find two vertex - disjoint cycle in @xmath63 , and ( 2 ) we need to remove two not fewer vertices to make @xmath63 acyclic . since @xmath14 can not contain two vertex - disjoint cycles , so does any of its subgraphs .
we have property 1 .
denote by @xmath43 the minimum number of vertices we need to remove to make @xmath63 acyclic . from proposition [ proposition : no - common ] , there is no common vertex among the three cycles .
so , removing any one vertex will not disconnect all three cycles simultaneously , i.e. , @xmath125 . on the other hand
, we only need to remove two vertices , @xmath126 , to make @xmath14 acyclic .
so , removing @xmath127 from @xmath63 will definitely make it acyclic , i.e. , @xmath128 .
so , we have property 2 .
note that these three cycles , @xmath63 , capture all the constraints we impose on @xmath14 in lemma [ lemma : special - subgraph ] .
we will proceed to show that @xmath63 must assume the configuration in figure [ fig : three - cycles ] .
we will build the configuration from a cycle , say @xmath122 , in @xmath63 .
we call it the _ center cycle_. we re - label the vertices in @xmath63 such that the vertices in @xmath122 are in ascending order in the direction of the arcs , i.e. , @xmath129 , where the choice of vertex 1 is arbitrary . for any path
@xmath130 that originates from vertex @xmath131 and terminates at vertex @xmath132 , i.e. , @xmath133 , we refer to all @xmath134 as _ inner vertices_. here , we allow @xmath135 ; in such a case , @xmath130 is a cycle .
we first show the following : [ proposition : outside - cycle ] consider the subgraph @xmath63 and the cycle @xmath122 in the subgraph .
every arc not in @xmath122 belongs to some _ outer path _ , defined as a path that originates from a vertex in @xmath122 and terminates at a vertex ( which can be the same vertex ) in @xmath122 , but with all arcs and all inner vertices ( if exists ) not in @xmath122 .
since @xmath63 is constructed by three cycles , any arc , say @xmath21 , not in @xmath122 must belong to either @xmath123 or @xmath136 ( or both ) .
furthermore , from proposition [ proposition : no - common ] , @xmath123 and @xmath136 must each share some vertex with @xmath122 .
hence , @xmath21 must belong to an outer path that originates from @xmath122 and terminates at @xmath122 .
note that the outer paths can not form any cycle outside @xmath122 . otherwise , we have two vertex - disjoint cycles .
it follows from proposition [ proposition : outside - cycle ] that @xmath63 consists of only a cycle @xmath122 and outer paths ( from @xmath122 and back to @xmath122 ) .
figure [ fig : redraw](a ) shows an example of @xmath63 where @xmath122 is marked with thick arrows and all outer paths thin arrows .
we now prove a key proposition for proving lemma [ lemma : special - subgraph ] .
[ proposition : outer - path ] remove vertex 1 in @xmath122 .
there exists another cycle in @xmath63 if and only if there is an outer path from some @xmath137 to some @xmath138 , where @xmath139 .
[ the _ only if _ part : ] we remove vertex 1 .
if there is another cycle , then there is a vertex ( not vertex 1 ) in @xmath122 that has a path back to itself ( this is because any cycle must share some vertex with @xmath122 ) .
this can not happen if every outer path terminates at a higher - indexed vertex ( we can ignore all outer paths that originate or terminate at vertex 1 as the vertex has been removed ) .
so , there must exist an outer path with @xmath139 .
[ the _ if _ part : ] clearly , if @xmath135 , we have another cycle formed by the outer path
. otherwise , i.e. , @xmath140 , the outer path and the path along @xmath122 from @xmath132 to @xmath131 form a cycle . see figure [ fig : looping](a ) for an example .
next , we define a _ looping _ outer path as an outer path that originates and terminates at the same vertex in @xmath122 .
the graph @xmath63 can be categorized as follows : * there exists one or more looping outer path ( case 1 ) , or * there is no looping outer path ( which we will further divide into cases 2 and 3 ) .
we will show that in any case , we have figure [ fig : three - cycles ] . is drawn with thick lines , and the second outer path ( denoted as @xmath130 ) from @xmath131 to @xmath132 , blue lines . to get another cycle after removing vertex 1
, we must have that @xmath141 , as shown in subfigure ( a ) . however , there are two vertex - disjoint cycles in subfigure ( a ) .
so , @xmath130 must touch the looping outer path , as shown in subfigure ( b ) . taking the segment of @xmath130 from @xmath122 to the looping outer path , and that from the looping outer path back to @xmath122 , we have subfigure ( c ) .
we can re - draw the path from @xmath142 to @xmath132 and that from @xmath143 to @xmath142 in subfigure ( c ) to get subfigure ( d ) , where we have drawn the new center cycle with thick lines.,width=313 ] suppose that there exists a looping outer path from and to vertex @xmath144 .
this incurs no loss of generality as the choice of vertex 1 is arbitrary . removing vertex 1 disconnects cycle @xmath122 and the cycle formed by the looping outer path .
recall that we need to remove two vertices to disconnect all cycles in @xmath63 .
so , there must exist another cycle in @xmath63 . from proposition [ proposition : outer - path ] , there exists another outer path @xmath130 from @xmath137 to @xmath138 , where @xmath139 .
the outer path @xmath130 must share some vertex with the looping outer path ; otherwise there exist two cycles as shown in figure [ fig : looping](a ) .
re - label the inner vertices of the looping outer path in ascending order , as follows : @xmath145 , where @xmath58 is the number of inner vertices .
it follows that the sequence of vertices shared by @xmath130 and the looping outer path ( in the order of the direction of @xmath130 ) must be in ascending order ( see figure [ fig : looping](b ) ) ; otherwise , a cycle forms outside @xmath122
. see figure [ fig : looping](c ) .
consider only the following segments of @xmath130 : ( i ) from @xmath131 to the vertex where @xmath130 first touches the looping outer path , denoted by @xmath146 ; and ( ii ) the vertex where @xmath130 leaves the looping outer path , denoted by @xmath143 , to @xmath132 .
it follows that @xmath147 . by construction ,
all paths in figure [ fig : looping](c ) do not share inner vertices , i.e. , they touch only at end points .
finally , re - draw figure [ fig : looping](c ) to get figure [ fig : looping](d ) , which is isomorphic to figure [ fig : three - cycles ] .
note that vertices 1 , @xmath131 , and @xmath146 must be unique .
we have shown that if there is a looping outer path , then we have the configuration in figure [ fig : three - cycles ] , where path @xmath66 has zero arc , paths @xmath67 and @xmath68 possibly have zero arc ( if @xmath148 and/or @xmath149 ) , and all other paths must contain at least one arc .
as in subfigure ( a ) , i.e. , a center cycle @xmath122 and outer paths from @xmath122 and back to @xmath122 .
subfigure ( b ) shows the coverage of an outer path , i.e. , vertices in @xmath122 in the gray area _ excluding _ the two end points .
subfigure ( c ) shows that when multiple outer paths originate from one vertex , we consider only the outer path with the largest coverage , i.e. , the dotted path from @xmath131 to @xmath132 .
the outer paths in subfigure ( d ) provide full coverage . ] for a non - looping outer path from vertex @xmath150 to @xmath151 , we say that the vertices from @xmath131 to @xmath132 ( in the direction of the arcs in @xmath122 ) but excluding @xmath131 and @xmath132 is _ covered _ by this outer path .
see figure [ fig : redraw](b ) for an example . for the purpose of this paper ,
we exclude outer paths with strictly smaller coverage , or multiple outer paths with equal coverage .
referring to figure [ fig : redraw](c ) , consider an outer path that originates from @xmath131 .
suppose that it has multiple paths back to @xmath122 .
we consider only the path ( back to @xmath122 ) that has the _
largest coverage_. similarly , for any path that terminates at @xmath132 , we consider only the path ( leaving @xmath122 ) that has the largest coverage . by doing this
, each path that we consider has a unique originating vertex and a unique terminating vertex .
we now show the following property : [ proposition : full - coverage ] if there is no looping outer paths in @xmath63 , then all largest - covering outer paths must , together , provide full coverage for the cycle @xmath122 . in other words , every vertex in @xmath122 must be covered by some outer path(s ) .
consider any vertex @xmath152 .
re - label @xmath80 as vertex 1 , and other vertices @xmath153 in ascending order in the arc direction .
remove vertex 1 from @xmath63 .
there must exist another cycle .
it follows from proposition [ proposition : outer - path ] that an outer path @xmath130 from @xmath131 to @xmath132 must exist , where @xmath154 ( @xmath155 since there is no looping path ) , meaning that this outer path must cover vertex 1 .
we can safely ignore other outer paths that provide smaller or equal coverage , because if @xmath130 does not cover vertex 1 , then none of the ignored outer paths does . since the choice of @xmath80 is arbitrary ,
we have proposition [ proposition : full - coverage ] .
for example , the outer paths in figure [ fig : redraw](d ) provides full coverage for @xmath122 , but the outer paths in figures [ fig : redraw](b)(c ) do not . removing one
uncovered vertex from @xmath122 makes @xmath63 acyclic .
now , we consider @xmath63 that consists of the cycle @xmath122 and all outer paths that provide the largest coverage ( i.e. , we remove all other arcs have gives smaller or equal coverage ) .
we are ready to proceed with cases 2 and 3 : * there is no looping outer path , and no two outer paths have any common inner vertex ( case 2 ) .
* there is no looping outer path , and there exist two outer paths sharing the same inner vertex ( case 3 ) .
we will show that we can always choose three outer paths to provide full coverage .
first , note that one outer path can not provide full coverage . now , suppose that we can find two outer paths providing full coverage .
we show in figure [ fig:2-coverage](a ) that we can always form two vertex - disjoint cycles .
so , this also can not happen .
with @xmath156 outer paths providing full coverage can be converted to @xmath157 outer paths providing full coverage .
( c ) if the blue and the right black outer paths ( non - adjacent ) give overlapping coverage , then the right red outer path is actually redundant , i.e. , @xmath158 outer path is sufficient to give full coverage , instead of @xmath159.,width=321 ] next , suppose that we can find three outer paths providing full coverage , we have figure [ fig : three - cycles ] . as there is no looping outer path , the nine paths in figure [ fig : three - cycles ] each have one or more arcs .
finally , we show that if we can find @xmath156 outer paths providing full coverage , we can always modify the cycles such that @xmath160 outer paths provide full coverage .
we illustrate this in figure [ fig:2-coverage](b ) .
we do the following : 1 . combine the dashed blue arrows to be the new @xmath122 .
2 . combine the two adjacent ( red ) outer paths , and the red arc in @xmath122 that connects the two red outer paths ( i.e. , @xmath130 , which can be of zero length ) into a new outer path .
3 . remove the two green paths in @xmath122 .
each green path must contain at least one arc ; otherwise , the outer paths can not provide full coverage . note that by doing this , the new graph still retains the structure of a cycle with outer paths covering it .
the new graph has @xmath157 outer paths providing full coverage .
this reduction is always possible as the coverage of two non - adjacent outer paths does not overlap , illustrated in figure [ fig:2-coverage](c ) . by repeating this step , starting from any @xmath156 outer paths
, we can find a graph with @xmath161 or @xmath162 outer paths .
as @xmath161 is not possible , we will always get a graph with @xmath162 outer paths providing full coverage , i.e. , figure [ fig : three - cycles ] . let the two outer paths that share some common inner vertex be @xmath130 and @xmath163 , and one of the shared inner vertices be @xmath164 .
further , let the originating and terminating vertices of @xmath130 be @xmath165 and @xmath166 respectively , and those of @xmath163 be @xmath167 and @xmath168 . here , @xmath169 and @xmath170 as there is no looping outer path , and @xmath171 and @xmath172 as no two outer paths have the same originating or terminating vertices .
( @xmath173 ) , and the green line the outer path @xmath163 ( @xmath174),width=302 ] now , the coverage of @xmath130 and @xmath163 can be either ( a ) non - overlapping , ( b ) overlapping once , or ( c ) overlapping twice , as shown in figure [ fig : overlapping - coverage ] . the red line shows the coverage of @xmath130 , and the green line @xmath163 . by definition
, there is a subpath from @xmath165 to @xmath164 along @xmath130 and another subpath from @xmath164 to @xmath166 along @xmath130 .
the two subpaths must be vertex - disjoint , except @xmath164 , as there is no cycle in @xmath130 .
similarly , we have two vertex - disjoint paths from @xmath167 to @xmath164 , and from @xmath164 to @xmath168 , both along @xmath163 .
this means , there is an subpath from @xmath165 to @xmath168 through @xmath164 , and another from @xmath167 to @xmath166 through @xmath164 .
so , @xmath175 , @xmath176 , as there is no looping outer path , and hence @xmath165 , @xmath166 , @xmath167 , and @xmath168 are distinct .
suppose that we have figure [ fig : overlapping - coverage](a ) .
the largest - covering outer path from @xmath165 should terminate at @xmath168 , and that from @xmath167 at @xmath166 .
the outer path from @xmath165 to @xmath168 and that from @xmath167 to @xmath166 should have been chosen .
this means the largest - covering paths actually overlap twice , i.e. , we should have figure [ fig : overlapping - coverage](c ) .
suppose that we have figure [ fig : overlapping - coverage](b ) .
the outer path from @xmath165 to @xmath168 , through @xmath164 , gives the largest coverage , and it would have been chosen .
so , we can only have the configuration in figure [ fig : overlapping - coverage](c ) , where the coverage overlaps twice . the coverage from @xmath165 to @xmath168
is smaller than that from @xmath165 to @xmath166 .
so , the largest - covering outer path from @xmath165 was correctly identified .
similarly , the largest - covering outer path from @xmath167 terminates at @xmath168 .
we will now show that we can always get figure [ fig : three - cycles ] from figure [ fig : overlapping - coverage](c ) . recall that there is a subpath from @xmath165 to @xmath164 and another subpath from @xmath164 to @xmath168 , and these two subpaths are vertex - disjoint , except @xmath164 .
we denote the outer path from @xmath165 to @xmath168 ( through @xmath164 ) by @xmath177 ( drawn with a thick dashed line ) . next ,
recall that there is a subpath from @xmath167 to @xmath164 , and another from @xmath164 to @xmath166 .
so , the subpath from @xmath167 to @xmath164 must meet @xmath177 .
denote the vertex it first meets @xmath177 as @xmath178 .
similarly , the subpath from @xmath164 to @xmath166 must share some common vertices with @xmath177 ( at least vertex @xmath164 ) .
let the last shared vertex be @xmath179 . with this construction , @xmath177 ,
the subpath from @xmath167 to @xmath178 , and the subpath from @xmath179 to @xmath166 are vertex - disjoint , except at @xmath179 and @xmath178 .
we now re - draw figure [ fig : overlapping - coverage](c ) as follows : let the path from @xmath168 to @xmath165 along @xmath122 ( drawn with a thick solid line ) plus path @xmath177 ( drawn with a thick dashed line ) be the center cycle , and let the subpaths ( drawn with dotted arrows ) ( i ) from @xmath165 to @xmath168 along @xmath122 , ( ii ) from @xmath179 to @xmath166 , and ( iii ) from @xmath167 to @xmath178 be the three outer paths .
note that only @xmath179 and @xmath178 can co - locate .
this is isomorphic to figure [ fig : three - cycles ] , with path @xmath66 possibly having zero arc ( if @xmath180 ) . combining the cases 13 ,
we have lemma [ lemma : special - subgraph ] . @xmath181
we have solved a new class of index - coding problems , characterized by their side - information graphs .
we have shown that for any side - information graph whose maximum acyclic induced subgraph ( mais ) can be formed by removing two or fewer vertices , the optimal index codelength ( i ) equals the order of the mais , and ( ii ) is achievable by linear index codes .
we proved this by constructing a special subgraph that the side - information graph must contain , and design a linear index code on it .
we then show that the liner index code achieves the mais lower bound .
y. birk and t. kol , `` coding on demand by an informed source ( iscod ) for efficient broadcast of different supplemental data to caching clients , '' _ ieee trans .
inf . theory _ ,
52 , no . 6 , pp . 28252830 , june 2006 . + ( the conference version of this paper appeared in _ infocom _ 1998 ) s. el rouayheb , a. sprintson , and c. georghiades , `` on the index coding problem and its relation to network coding and matroid theory , '' _ ieee trans .
inf . theory _ ,
56 , no . 7 , pp .
31873195 , july 2010 .
m. f. wong , m. langberg , and m. effros , `` on a capacity equivalence between network and index coding and the edge removal problem , '' in _ proc .
inf . theory ( isit ) _ ,
istanbul , turkey , july 712 2013 , pp . 972976 .
r. m. karp , `` reducibility among combinatorial problems , '' in _ complexity of computer computations _
, r. e. miller and j. w. thatcher , eds.1em plus 0.5em minus 0.4emplenum press , new york - london , 1972 , pp . 85104 .
|
we study index - coding problems ( one sender broadcasting messages to multiple receivers ) where each message is requested by one receiver , and each receiver may know some messages a priori .
this type of index - coding problems can be fully described by directed graphs .
the aim is to find the minimum codelength that the sender needs to transmit in order to simultaneously satisfy all receivers requests .
for any directed graph , we show that if a maximum acyclic induced subgraph ( mais ) is obtained by removing two or fewer vertices from the graph , then the minimum codelength ( i.e. , the solution to the index - coding problem ) equals the number of vertices in the mais , and linear codes are optimal for this index - coding problem .
our result increases the set of index - coding problems for which linear index codes are proven to be optimal .
|
recently , the understanding of many - body quantum systems has dramatically progressed .
nowadays we are achieving an amazing degree of control over larger and larger systems @xcite .
therefore , verification during each stage of experimental procedures is of utmost importance ; quantum tomography is the appropriate tool for that purpose .
the goal of quantum tomography is to reconstruct the state of a system by performing multiple measurements on identically prepared copies of the system .
once the experimental data are extracted , a numerical procedure determines which density matrix fits best the measurements .
this estimation can be performed using different approaches , such as maximum likelihood estimation @xcite , or bayesian methods @xcite .
however , tomography becomes harder as we explore more intricate systems .
if we look at the simple , yet illustrative case of @xmath0 qubits , which will serve as the consistent thread in this paper , one has to make at least @xmath1 measurements in different bases before one can claim to know everything about an _ a priori _ unknown system .
with such an exponential scaling in the number of qubits , it is clear that current methods rapidly become intractable for present state - of - the - art experiments . as a result , more sophisticated tomographical techniques are called for .
new protocols try to simplify the process by making an educated guess about the nature of the state .
among other assumptions , this includes rank deficiency @xcite , extra symmetries @xcite , or gaussianity @xcite .
while all these approaches are extremely efficient , their pitfall is that when the starting guess is inaccurate , they produce significant systematic errors . here , we pursue a different approach , inspired by a notion from statistical mechanics : coarse graining @xcite .
this operation transforms a probability density in phase space into a `` coarse - grained '' density that is a piecewise constant function , a result of density averaging in cells .
this is the chief idea behind the renormalization group @xcite , which allows a systematic investigation of the changes of a physical system as viewed at different scales . in our case
, we consider a system of qubits and look at the associated phase space , which turns out to be a discrete grid of @xmath2 points .
we assign to each suitably defined line in phase space a specific rank - one projection operator representing a pure quantum state . for each point of the grid ,
a suitable quasi - probability as the wigner function can be directly computed from the measurement of the states associated with the lines passing through that point .
we coarse grain by combining groups of these lines into thick lines , which we will show to be lines in the phase space of an effectively smaller system .
our coarse - grained phase spaces are endowed with many nice properties .
most notably , our procedure systematically and naturally reveals a subset of measurements which one could use to perform incomplete tomography .
in addition , using the coarse - grained points and lines , we show that one can define a discrete wigner function in largely the same way as it is defined in the original space . when plotted , the coarse functions resemble smoothed out versions of the originals , preserving many of their prominent visual features .
a qubit is a two - dimensional quantum system , with hilbert space isomorphic to @xmath3 .
it is customary to choose two normalized orthogonal states , say @xmath4 , as a computational basis .
the unitary matrices @xmath5 generate the pauli group @xmath6 , which consists of all the pauli matrices plus the identity , with multiplicative factors @xmath7 @xcite . for @xmath0 qubits ,
the corresponding hilbert space is the tensor product @xmath8 .
a compact way of labeling both states and elements of the corresponding pauli group @xmath9 is by using the finite field @xmath10 . in appendix
[ sec : galois ] we briefly summarize the basic notions of finite fields needed to proceed .
let @xmath11 , @xmath12 , be an orthonormal basis in the hilbert space @xmath13 ( henceforth , field elements will be denoted by greek letters ) .
the elements of the basis can be labeled by powers of a primitive element @xmath14 ( i.e. , a root of an irreducible primitive polynomial ) : @xmath15 .
now the equivalent version of ( [ eq : sigmas ] ) is @xcite @xmath16 so that @xmath17 which is the discrete counterpart of the weyl - heisenberg algebra for continuous variables @xcite . here
, the additive character @xmath18 is defined as @xmath19 $ ] and the trace of a field element ( we distinguish it from the trace of an operator by the lower case `` tr '' ) is defined in appendix [ sec : galois ] .
moreover , @xmath20 and @xmath21 are related through the finite fourier transform @xcite @xmath22 so that @xmath23 .
the operators ( [ xzgf ] ) generate the pauli group @xmath9 of @xmath0 qubits and , with a suitable choice of basis , they can be factorized into a tensor product of single - qubit pauli operators . to this end , it is convenient to consider @xmath10 as an @xmath0-dimensional linear space over @xmath24 .
it is spanned by an abstract basis @xmath25 , so that given a field element @xmath26 the expansion @xmath27 allows us the identification @xmath28 .
the basis @xmath29 can be chosen to be orthonormal with respect to the trace operation ; i.e. , @xmath30 .
this is a self - dual basis , which always exist for the case of qubits . in this way
, we associate each qubit with a particular element of the self - dual basis : qubit@xmath31 . using this basis , we have the factorization @xmath32 where @xmath33 and @xmath34 are the corresponding expansion coefficients for @xmath35 and @xmath36 in the self - dual basis .
we next recall @xcite that the grid defining the phase space for @xmath0 qubits can be appropriately labeled by the discrete points @xmath37 , which are precisely the indices of the operators @xmath20 and @xmath21 : @xmath35 is the `` horizontal '' axis and @xmath36 the `` vertical '' one . in this grid
we can introduce the set of displacements @xmath38 where @xmath39 is a phase required to avoid plugging extra factors when acting with @xmath40 .
a sensible choice for the case of qubits is @xmath41 , which ensures the hermiticity of the displacement operators .
in addition , we impose @xmath42 and @xmath43 , which means that the displacements along the `` position '' axis @xmath26 and the `` momentum '' axis @xmath44 are not associated with any phase . these displacement operators shift phase space points , so the action of @xmath45 maps @xmath46 , justifying their designation .
note that we still have to fix the sign of the phase @xmath47 .
we choose the phase as @xmath48 where @xmath49 , which ensures that the operators defined in eq . below are rank - one projections . on the phase space grid
one can introduce a variety of geometrical structures with much the same properties as in the continuous case @xcite .
the simplest are the straight lines passing through the origin ( also called rays ) , with equations @xmath50 the rays have a very remarkable property : the monomials @xmath51 belonging to the same ray commute , and thus , have a common system of eigenvectors @xmath52 , @xmath53 where @xmath54 is fixed and @xmath55 is the corresponding eigenvalue , so @xmath56 are eigenstates of @xmath20 ( displacement operators labeled by the ray @xmath57 , which we take as the horizontal axis ) . the projection operators associated with the lines of equal slope are the projections onto these eigenvactors . indeed ,
we have that @xmath58 and , in consequence , they are mutually unbiased bases ( mubs ) @xcite .
now suppose for each ray we disregard the origin @xmath59 , whose monomial is the identity operator .
this leaves us with @xmath60 commuting operators .
if we then consider the whole bundle of @xmath1 rays ( which are obtained by varying the `` slope '' @xmath61 over all of @xmath62 ) , we can construct a complete set of mub operators arranged in a @xmath63 table @xcite . to round up the scenario , we need to represent states in phase space .
the discrete wigner function @xcite is the appropriate tool .
it can be considered as an invertible mapping @xmath64 \ , , \label{eq : wigdn}\ ] ] so that @xmath65 the operational kernel is defined as @xmath66 which , in view of equation ( [ ftcomp ] ) , can be interpreted as a double fourier transform of @xmath67 .
one can check that this kernel has all the good properties @xcite : it is hermitian , normalized and covariant under the pauli group . as a result , for each point on the grid , the corresponding value of the wigner function can be computed from the probabilities of measuring the pure states associated with the lines passing through that point .
as heralded in the introduction , our goal is to tailor a procedure that allows us to coarse grain the phase space of a multiqubit system ; i.e. , to break it down into simpler sub - components . to this end
, we consider the number @xmath0 of qubits to be composite , i.e. @xmath68 . let @xmath69 be a basis of @xmath70 with respect to @xmath71 .
we define @xmath72 i.e. , the subspace made of linear combinations of basis elements @xmath73 with coefficients in the base field @xmath74 .
we can use this set @xmath75 , which we henceforth refer to as the initial coset , to decompose the field @xmath70 into cosets : @xmath76 the coarse - grained space will be labeled according to these cosets .
we can imagine the process of coarse graining as partitioning the grid @xmath77 in such a way that we superimpose a grid of size @xmath78 on top , with each superimposed point indexed by cosets rather than field elements in the original grid .
each point in the coarse grid then contains a sub - grid the same size as @xmath79 . to provide some intuition for this ,
we show a visual example of this process in action in fig .
[ colourfulpanels ] .
our procedure for coarse - graining the grid arises naturally from consideration of the line structure of phase space .
we will use the _ thin _ lines in @xmath70 to create _
thick _ lines in the coarse phase space , by grouping together lines having the same slope , and with intercepts in the same coset .
we write thin lines in the big field @xmath70 as @xmath80 , where @xmath61 is the slope , and @xmath81 is the intercept . a large , coarse - grained line is denoted as @xmath82 , where now the intercept is a whole coset . to each line in the fine - grained phase space we can assign a projector @xmath83 , constructed as a linear combination of the displacement operators .
we choose as our convention for the rays @xmath84 the all - positive sum @xmath85 these lines are eigenstates with eigenvalue @xmath86 for all displacement operators in the sum .
projectors with nonzero intercepts are obtained by conjugating that of the ray with an appropriate displacement operator .
the coarse lines are produced by grouping together lines with intercepts in the same coset : @xmath87 the possible choices of slope for these lines will be limited to elements of the subfield @xmath74 , as these have natural analogues between the two fields . as discussed in more detail in appendix [ sec : lineop ] , the coarse rays of eq .
( [ coarse_untransformed ] ) can be simplified and rewritten as the sum of displacement operators @xmath88 d(\alpha , \lambda \alpha ) \ , .
\label{master_eq}\ ] ] one can check here that the inner sum over the elements of @xmath75 will cause some of the displacement operators to vanish .
the sum in brackets in eq .
is either zero or a positive constant .
hence , the projection associated to the thick lines are a sum over a subset of the displacement operators associated with the thin lines .
this leads us to the key idea of our work : rather than measuring all the displacement operators , we measure only those which are present in the rays of the coarse - grained space .
we note here that the choice of @xmath75 is not unique , and will ultimately determine the resultant set of displacement operators .
for example , a special case occurs when the dimension of the system is square . in this case
, we can consider the relationship between the fields as a quadratic field extension , i.e. when @xmath89 . in this case
we can partition @xmath90 into @xmath91 .
we can then choose the initial coset as the copy of the subfield @xmath92 : @xmath93 where @xmath14 is a primitive element of @xmath70 and we use the notation @xmath94 for 0 .
the subsequent cosets are obtained additively from this subfield using the representatives @xmath95 .
finally , the coarse - grained phase space inherits a coarse - grained wigner function .
a coarse kernel can be constructed by grouping together kernel operators from the same coset , i.e. @xmath96 desired properties of a wigner function all follow from the original kernel .
as was the case with the displacement operators , differing choices of the subset @xmath75 will lead to differing wigner functions .
colours are indicative of particular coarse rays . the left image coarse grains by taking @xmath97 , whereas the lower image uses the subfield @xmath98 . ]
we illustrate the previous ideas with some relevant examples .
we have written a python software package capable of generating all the following results , which we make available online @xcite the first nontrivial instance we can have is the case of two qubits , so dimension @xmath99 . using the irreducible primitive polynomial @xmath100
, we have that @xmath101 .
the self - dual basis is @xmath102 , and we use it to produce the displacement operators .
another basis for @xmath103 is @xmath104 .
taking all scalar multiples of @xmath105 from the prime field gives us @xmath106 .
we then obtain @xmath107 .
for each ray , we can list the operators which survive in the inner sum over @xmath75 in eq .
( [ master_eq ] ) .
moreover , we can label the points of the coarse - grained grids by those displacement operators . disregarding the identity operator ,
the resulting set @xmath108 constitutes the appropriate measurements to be performed to determine which coarse - grained line they are in .
they are essentially pauli measurements on one of the two qubits in the system .
alternatively , the dimension is a square , so we can choose as our initial coset the subfield @xmath109 : @xmath110 .
this yields the second coset @xmath111 .
we once again compute the surviving operators using eq . .
the final result now is @xmath112 . here
, we see that we are making a measurement with the same pauli operator on both qubits , thereby capturing the full correlations between the two qubits .
figure [ dim4_plane_rays ] shows both partitioning methods side by side .
system down to dimension @xmath113 .
( left panel ) coarse graining using the basis @xmath114 .
the resultant measurements are unitarily equivalent to a case where two of the qubits remain untouched .
( right panel ) resultant operators when the coarse - graining uses the initial basis @xmath115 .
here we obtain the interesting result that all resultant operators commute . ]
system down to dimension @xmath99 .
the left panel contains the surviving operators from the general basis method , the right panel from choosing the subfield as @xmath75 .
the cosets are listed in eqs .
( [ dim16cosets_general ] ) and ( [ dim16cosets_subfield ] ) respectively . in the case of the left panel ,
these operators are unitarily equivalent to a set where two qubits are untouched and the 2-qubit mub operators are applied to the rest .
the right panel has no such transformation . ]
our next example is the case of dimension 8 .
we choose @xmath14 a root of the irreducible primitive polynomial @xmath116 , and obtain a self - dual basis @xmath117 .
an obvious choice for a basis of @xmath118 is a polynomial basis @xmath114 . to construct @xmath75
, we must take all possible linear combinations of @xmath14 and @xmath119 with coefficients in @xmath109 .
this produces @xmath120 we obtain the second coset by adding the remaining subfield element @xmath121 to @xmath75 : @xmath122 the traces of all elements in @xmath75 are @xmath123 , and the traces for all elements in @xmath124 are @xmath121 .
the surviving four operators are shown in fig .
[ dim8_dim2coarse ] . using a clifford transformation
, we can `` trace out '' two of the qubits . the sequence of cnot gates : cnot@xmath125 cnot@xmath126 cnot@xmath127 cnot@xmath128 transforms the set into @xmath129
, so we see that this partitioning is , after a global change of basis , equivalent to measuring each pauli on only a single qubit .
if we choose instead the basis @xmath130 to build our cosets , we get a more interesting result : @xmath131 the operators that survive have the form @xmath132 , @xmath133 , yielding the operators in fig .
[ dim8_dim2coarse ] , which all commute . in this case
, we are already ignoring one of the three qubits .
however , it is not possible to find a clifford which will trace out a remaining one as was the case with the polynomial basis case .
so , in a sense , using this partitioning we are ignoring fewer qubits than before .
dimension @xmath134 is perhaps the first really interesting case .
first of all , we can consider it in two ways : @xmath135 , or @xmath136 . essentially , to do the partitioning
, we can look at @xmath137 as a quartic extension over @xmath109 , or a quadratic extension over @xmath138 .
we consider the quadratic case , so we can coarse grain in two ways .
we work with @xmath137 as constructed by the irreducible primitive polynomial @xmath139 over @xmath109 , and @xmath140 over @xmath138 where we denote a primitive element of @xmath138 as @xmath141 .
we know from eq .
( [ subfield_in_big_field ] ) that @xmath142 , where @xmath14 is the primitive element in @xmath137 .
then @xmath138 in @xmath137 can be written as @xmath143 .
for the general case , we choose the basis @xmath104 .
taking all @xmath138-multiples of @xmath14 , we obtain @xmath144 .
the full set of cosets is : @xmath145 proceeding in the standard way , and taking into account that a self - dual basis is @xmath146 , we obtain the operators in fig .
[ dim16_dim4_general ] .
what is ( un)interesting about these operators is that we can transform them all into operators which completely ignore two of the qubits . in particular , consider the following sequence of operations : cnot@xmath147 cnot@xmath148 cnot@xmath128 cnot@xmath149 cnot@xmath150 .
application of this to the operators of the first panel of fig .
[ dim16_dim4_general ] yields a new set of operators where the last two qubits contain only @xmath151 , and the first two qubits contain the full set of mub operators on two qubits .
alternatively , we can choose our initial coset as the subfield , and the coset representatives as @xmath152 .
we obtain the cosets @xmath153 using eq .
( [ master_eq ] ) we get the table shown in the right panel of fig .
[ dim16_dim4_general ] . unlike in the previous case
, there is no transformation which will lead to us ` tracing out ' two of the qubits .
however , we can bring these operators into a more basic form by applying the sequence cnot@xmath126 cnot@xmath150 .
the resultant operators have the property that on the first two qubits , we only have @xmath154 , and on the last two qubits only @xmath155 , so that they all commute
. to conclude , we present some of the coarse - grained wigner functions we obtain using our method .
those in dimensions 4 and 8 are somewhat trivial , so we focus on dimension 16 . wigner functions for the states @xmath156 and @xmath157 are presented in fig .
[ double_bell_wf ] .
recall in section [ sec : qpwig ] that we could associate the elements of @xmath62 with a basis in our hilbert space .
then , in the coarse wigner functions , when we group the field elements into cosets , we can consider this also as grouping together the associated basis states .
hence , the probabilities in these wigner functions become distributed over the cosets which contain the constituent basis states of our target state . as a result ,
the coarse wigner functions resemble ` smoother ' versions of the original one to varying degrees .
compared to the continuous wigner function , the discrete wigner function is an adolescent formulation , slowly developing into adult maturity .
discrete phase space imposes several new challenges , which leads to an intricate mapping of the wigner function .
our coarse graining procedure shows a way to facilitate our understanding when the number of qubits is high .
while it is always possible to ignore part of the system and to determine the full wigner function of the resulting reduced density matrix , our approach allows more choices regarding which information of the whole system is measured . in another extremal case , the coarse - grained wigner function is completely determined by a set of commuting operators that can be measured simultaneously
. however , several open questions remain
. an obvious next step would be to extend the coarse graining procedure to multi - qudit systems .
furthermore , knowing the coarse - grained function , does there exist another subset of measurements which will allow us to zoom in on specific areas of it and gain more information ?
a logical first choice would be to extend the set of measurements such that they include all operators that correspond to slopes in the subfield .
for example , in the dimension @xmath134 case , we would measure all operators for the rays @xmath158 and @xmath159 , rather than just three from each .
this strategy would allow us to optimize measurements in a very subtle way .
work along these lines is in progress .
o.d.m . is funded by canada s nserc .
she is also grateful for hospitality at the mpl .
iqc is supported in part by the government of canada and the province of ontario . l.l.s.s .
acknowledges financial support from the spanish mineco ( grant no .
fis2015 - 67963-p ) .
in this appendix we briefly recall some background needed for this paper .
the reader interested in more mathematical details is referred , e.g. , to the excellent monograph by lidl and niederreiter @xcite .
a commutative ring is a nonempty set @xmath160 with two binary operations , called addition and multiplication , such that it is an abelian group with respect to addition , and the multiplication is associative .
the most typical example is the ring of integers @xmath161 , with the standard sum and multiplication . on the other hand ,
the simplest example of a finite ring is the set @xmath162 of integers modulo @xmath163 , which has exactly @xmath163 elements .
a field @xmath164 is a commutative ring with division , i.e. , such that 0 does not equal 1 and all elements of @xmath164 except 0 have a multiplicative inverse ( note that 0 and 1 here stand for the identity elements for the addition and multiplication , respectively , which may differ from the familiar real numbers 0 and 1 ) . elements of a field form abelian groups with respect to addition and multiplication ( in this latter case , the zero element is excluded ) .
note that the finite ring @xmath165 is a field if and only if @xmath166 is a prime number .
the characteristic of a finite field is the smallest positive integer @xmath166 such that @xmath167 and it is always a prime number .
any finite field contains a prime subfield @xmath165 and has @xmath168 elements , where @xmath163 is a natural number .
moreover , the finite field containing @xmath169 elements is unique up to isomorphism and is called the galois field @xmath170 .
we denote as @xmath171 $ ] the ring of polynomials with coefficients in @xmath172 . if @xmath173 is an irreducible polynomial of degree @xmath163 ( that is , one that can not be factorized over @xmath172 ) , the quotient space @xmath174/p(x)$ ] provides an adequate representation of @xmath170 .
its elements can be written as polynomials that are defined modulo the irreducible polynomial @xmath173 .
the multiplicative group of @xmath170 is cyclic and its generator is called a primitive element of the field . as a trivial example of a nonprime field , we consider the polynomial @xmath175 , which is irreducible over @xmath24 . if @xmath14 is a root of this polynomial , the elements @xmath176 form the finite field @xmath177 and @xmath14 is a primitive element .
a basic map is the trace @xmath178 the image of the trace is always in the prime field @xmath165 and satisfies @xmath179 in terms of it we define an additive character as @xmath180 \
, , \ ] ] which possesses two important properties : @xmath181 any finite field @xmath170 can be also considered as an @xmath163-dimensional linear vector space over its prime field @xmath182 . given a basis @xmath183 , ( @xmath184 ) in this vector space , any field element can be represented as @xmath185 with @xmath186 . in this way , we map each element of @xmath170 onto an ordered set of natural numbers @xmath187 . two bases @xmath188 and @xmath189 are dual when @xmath190 a basis that is dual to itself is called self - dual
. a self - dual basis exists if and only if either @xmath166 is even or both @xmath163 and @xmath166 are odd .
there are several natural bases in @xmath170 .
one is the polynomial basis , defined as @xmath191 where @xmath14 is a primitive element .
an alternative is a normal basis , constituted of @xmath192 the appropriate choice of basis depends on the specific problem at hand .
for example , in @xmath177 the elements @xmath193 are both roots of the irreducible polynomial . the polynomial basis is @xmath194 and its dual is @xmath195 , while the normal basis @xmath196 is self - dual .
here we present the derivation of our equation for the surviving displacement operators .
we begin by considering the projectors for the rays , @xmath197 as mentioned in sec .
[ sec : qpwig ] , the projectors for the shifted lines can be obtained by applying an appropriate displacement operator to induce a transformation .
let us ignore for now the ray with infinite slope , @xmath158 .
then for the rest of the rays , we can shift them vertically by applying the displacement operators of the form @xmath198 : @xmath199 where we recall the convention that all the phases @xmath200 . here
, we can make further use of the commutation relation in eq . .
we obtain @xmath201 it is then straightforward to see that the thick rays , which are obtained by summing over all intercepts @xmath81 in coset @xmath202 , can be written as @xmath203 d(\alpha , \lambda \alpha).\ ] ] finally , we mention that for the infinite slope the analysis proceeds in exactly the same way , but that the lines are translated by displacement operators of the form @xmath204 and eq .
gives us @xmath205 instead .
|
we develop a systematic coarse graining procedure for systems of @xmath0 qubits .
we exploit the underlying geometrical structures of the associated discrete phase space to produce a coarse - grained version with reduced effective size .
our coarse - grained spaces inherit key properties of the original ones .
in particular , our procedure naturally yields a subset of the original measurement operators , which can be used to construct a coarse discrete wigner function .
these operators also constitute a systematic choice of incomplete measurements for the tomographer wishing to probe an intractably large system .
= 1
|
the teichmller space of a surface admits several natural metrics , starting with the teichmller metric introduced by teichmller in 1939 , followed by the weil - petersson metric introduced by weil in 1958 , and by many others .
it is probable that the third most active subject in the metric theory of teichmller space is now that of the metric introduced by thurston in 1985 and which is called now thurston s metric .
the three metrics w mentioned , along with several others defined on teichmller space are extremely interesting and they were studied from various points of view : the infinitesimal structure ( finsler or riemannian ) , the geodesics , the convexity properties , the boundary structure , etc .
some difficult questions concerning these properties were solved and others remain open and make the subject a living one . in the present paper ,
when we consider the theory for surfaces with boundary , we mean the non - reduced theory , that is , the homotopies that we consider , in the definition of the equivalence relation defining teichmller spaces , do not fix the boundary of the surface pointwise .
in general , the metrics on teichmller space were defined for closed surfaces or surfaces with punctures ( or distinguished points ) .
some of these metrics , like the teichmller metric , admit a straightforward generalization for surfaces with boundary , but other metrics do not generalize as such for surfaces with boundary , and one has to modify their definition .
one example is thurston s metric , whose modification , the so - called arc metric , was studied in the papers @xcite , @xcite , @xcite , @xcite .
we shall recall the definition of the arc metric below .
several basic questions concerning this metric still resist .
for instance , it is unknown whether it is finsler , its isometry group is still not identified , and it is unknown whether this metric coincides with the lipschitz metric .
( all these results are known for surfaces without boundary . ) in the present paper , we construct new families of geodesics for this metric . in order to state more precisely the results , we start with some notation .
let @xmath0 be a surface of finite type with nonempty boundary .
the hyperbolic structures on @xmath0 that we consider are those for which the boundary components are closed geodesics .
the lengths of the boundary components of the surface are not fixed .
a simple closed curve on @xmath0 is _ essential _ if it is neither homotopic to a point nor to a puncture ( but it can be homotopic to a boundary component ) .
an _ arc _ on @xmath0 is the image of a compact interval of @xmath1 by a proper embedding , that is , the interior of the arc is embedded in the interior of @xmath0 and the images of its endpoints are on the boundary of @xmath0 . in this paper ,
when we deal with homotopies of arcs , we only consider homotopies that are relative to @xmath2 , that is , they keep the endpoints of the arc on the boundary of the surface ( but they do not necessarily fix pointwise the points of that boundary ) .
an arc is _ essential _ if it is not homotopic to an arc contained in @xmath2 .
we shall also use the following notation : @xmath3 is the set of homotopy classes of essential arcs on @xmath0 .
@xmath4 is the union of the set of homotopy classes of essential arcs on @xmath0 and of the set of homotopy classes of closed curves homotopic to a boundary component of @xmath0 .
@xmath5 is the set of homotopy classes of essential simple closed curves on @xmath0 .
suppose that the surface @xmath0 is equipped with a hyperbolic structure @xmath6 . in any equivalence class @xmath7 , there is a unique geodesic arc whose length is minimal among the arcs in that class relative the boundary .
this geodesic arc makes at its endpoints right angles with the boundary of @xmath0 .
we denote by @xmath8 the length of this geodesic arc , and we call it the _ geodesic length _ of @xmath9 for the hyperbolic metric @xmath6 .
likewise , for any element @xmath9 of @xmath10 , we denote by @xmath8 the length of its unique geodesic representative for the hyperbolic metric @xmath6 .
let @xmath11 be the teichmller space of @xmath0 .
we view @xmath11 as the space of homotopy classes of hyperbolic structures on @xmath0 with geodesic boundary , and where the lengths of the boundary components are not fixed .
the _ arc metric _
@xmath12 on @xmath11 is defined by the formula @xmath13 where @xmath6 and @xmath14 are hyperbolic structures on @xmath0 .
this is an asymmetric metric ( that is , it satisfies all the axioms of a metric except the symmetry axiom ) .
it was introduced in @xcite and it is an analogue for surfaces with boundary of thurston s asymmetric metric on the teichmller space of a surface without boundary ( possibly with cusps ) @xcite .
it was shown in @xcite ( proposition 2.13 ) that one obtains the same metric by using the same formula but taking the supremum there over @xmath4 instead of @xmath15 . in other words
, we also have @xmath16 the arc metric is also studied in the papers @xcite , @xcite , @xcite and @xcite .
we now recall the definition of the lipschitz metric on the teichmller space of a surface with boundary .
the definition is the same as the one of the lipschitz metric defined by thurston on teichmller spaces of surfaces without boundary .
one first defines the _ lipschitz constant _ of a homeomorphism @xmath17 between two metric spaces by the formula @xmath18 the homeomorphism @xmath19 is said to be _ lipschitz _ if its lipschitz constant is finite .
given an ordered pair of hyperbolic structures @xmath6 and @xmath14 on @xmath0 , the _ lipschitz distance _ between them ( and between the corresponding points in the teichmller space @xmath11 ) is defined as @xmath20 where the infimum
is taken over all homeomorphisms @xmath21 in the homotopy class of the indentity of @xmath0 . in the case of surfaces without boundary ,
thurston s metric and the lipschitz metric coincide .
this is a result of thurston in @xcite .
it is unknown whether in the case of surfaces with boundary the lipschitz and the arc metrics coincide . in thurston s theory for surfaces without boundary developed in @xcite , maps between ideal triangles are the building blocks for the construction of geodesics for thurston s metric on teichmller space .
the geodesics obtained are the so - called stretch lines , " and thurston proves that any two points in the teichmller space of @xmath0 are joined by a concatenation of stretch lines .
it is possible to construct a class of geodesics for the @xmath12-metric ( we call them @xmath12-geodesics ) using thurston s method for stretch lines .
for this , a complete maximal geodesic lamination is needed ( in the language of @xcite , this will be the geodesic lamination which is maximally stretched ) ; for instance , we can take a lamination whose leaves spiral along the boundary components of the surface @xmath0 .
we then apply thurston s method decribed in @xcite for the construction of stretch lines using the stretch maps between ideal triangles and gluing them over all the ideal triangles that are the connected components of the complement of the lamination . to see that the one - parameter family of surfaces obtained in this way is a @xmath12-geodesics
, one can double the surface @xmath0 along its boundary components and consider the resulting one - parameter family of hyperbolic structures on the doubled surface @xmath22 .
this is a geodesic for thurston s metric on @xmath22 , and the restriction to @xmath0 of this one - parameter family of deformations of hyperbolic structures is a geodesic for the @xmath12- and the @xmath23-metrics on the teichmller space of the surface with boundary .
the @xmath12-geodesic defined in this way is of a special type ( that is , not all @xmath12-geodesics are obtained ) .
along this geodesic , the lengths of all the geodesic boundary components of @xmath0 are multiplied by the same constant factor , cf .
@xcite .
another construction of @xmath12-geodesics in the setting of surfaces with boundary is obtained by taking as building blocks right angled hexagons instead of ideal triangles .
we shall introduce canonical maps between such hexagons which depend on a parameter @xmath24 , and by varying @xmath24 we obtain geodesics for the arc metric .
each right - angled hexagon belongs to such a family of deformations , and between each pair of hexagons of the same family there is a lipschitz map that realizes the best lipschitz constant in its homotopy class relative to the boundary .
gluing the hexagons along pieces of their boundaries , we obtain deformations of the hyperbolic surfaces that are geodesics for the arc metric .
these lines are geodesics for both the lipschitz metric and the arc metric and the lipschitz metric on the teichmller space of the surface with boundary coincide on these lines . by gluing surfaces with boundary along their boundary components , we obtain geodesics for the thurston metric that are different from the stretch lines .
this generalizes the set of results obtained in the two papers @xcite and @xcite .
for the construction of lipschitz maps between right angled hexagons , we shall divide each such hexagon @xmath25 into three regions ( one such region may possibly be empty ) which carry natural coordinates ; that is , the points in such a region are parametrized by pairs of real numbers . the coordinates in each region
are induced by a pair of orthogonal foliations @xmath26 and @xmath27 on @xmath25 , which we now define .
we consider three pairwise non - consecutive edges @xmath28 of a right angled hexagon @xmath25 , and we call them the _ short edges_. the three lengths @xmath29 of @xmath28 may or may not satisfy the triangle inequality . up to a permutation of the indices , there are three cases : 1 .
type i : @xmath30 satisfy the three triangle inequalities .
type ii : @xmath31 .
type iii : @xmath32 .
we now give another characterization of the three distinct types . consider three non - intersecting geodesic lines @xmath33 in the hyperbolic plane ,
relatively positioned so that for each geodesic @xmath34 , the other two geodesics @xmath35 are lying on the same side of @xmath34
. then each pair @xmath36 @xmath37 has a common perpendicular geodesic segment @xmath38 @xmath39 .
this common perpendicular segment , which we call a short edge , is unique .
we denote by @xmath40 , the three geodesic lines that contain the edges @xmath28 respectively .
the six geodesic lines @xmath41 enclose a hexagonal region ( figure [ hex ] ) .
in hyperbolic geometry , two geodesics @xmath34 and @xmath42 having a common perpendicular are said to be _ hyper - parallel_. thus , for the time being , we exclude the possibility that some @xmath34 and @xmath42 are asymptotic to each other .
( classically , in the latter case , @xmath34 and @xmath42 are said to be _
parallel_. ) the side opposite to a short edge @xmath43 will be denoted by @xmath44 , and called a _ long edge _
( figure [ hex ] ) .
there exists a unique point @xmath45 which is equidistant from @xmath40 .
the three types listed above correspond respectively to the cases where the point @xmath45 lies inside the right angled hexagon @xmath25 ( type i ) , or on one of the long edges @xmath46 ( type ii ) , or outside @xmath25 ( type iii ) ; cf .
figure [ center ] .
one should be aware of the fact that interchanging the short and long edges of a given right angled hexagon makes the resulting center @xmath45 different from the original center unless the hexagon has a @xmath47-rotational symmetry . from now on ,
when we refer to a right angled hexagon , it will be understood that a choice of the short and long edges has been made .
the right angled hexagon @xmath25 is naturally equipped with a partial measured foliation ( that is , a measured foliation whose support is a _
subsurface _ of @xmath25 ) which is a union of three foliated regions with disjoint interiors , @xmath48 , where each leaf of each of these foliations is a segment in the locus of equidistant points to one of the three long edges @xmath49 . in the hyperbolic plane ,
equidistant points to geodesics are classically called hypercycles , and we shall use this terminology . note that one of these foliations might be empty .
this occurs in type iii , where the triangle inequality among the lengths @xmath50s of the short edges @xmath43s is not satisfied .
the three types of foliations corresponding to types i , ii and iii are represented in figure [ foliation ] from left to right respectively .
the supports of the foliations @xmath51 by hypercycles have a natural structure of rectangles and these foliations are equipped with transverse measures induced from the lebesgue measure on the boundaries of the supports , whose total masses @xmath52 satisfy the equations @xmath53 with an appropriate and clear meaning when one of the @xmath54s is negative ( see figure [ negative ] ) .
we denote the foot of @xmath45 on the geodesic @xmath55 by @xmath56 .
we then have a tripod @xmath57 whose edges have equal lengths @xmath58 , and where each edge @xmath59 meets @xmath60 perpendicularly .
in type i , corresponding to the case to the left picture of figures [ center ] and [ foliation ] , @xmath56 divides each short edge @xmath43 at an inner point , for each @xmath61 . in this case
@xmath52 are all positive . in type ii ,
two of the @xmath56s , say @xmath62 and @xmath63 , coincide with endpoints of @xmath64 and @xmath65 respectively . in this case
, we have @xmath66 and the foliation @xmath25 is of the type depicted on the middle hexagon of figure [ foliation ] .
finally , type iii corresponds to the case where two of the @xmath56 , say @xmath62 and @xmath63 , divide the short edges @xmath64 and @xmath65 externally . in this case
we have @xmath67 , and the foliation @xmath25 is as depicted in the right hexagon of figure [ foliation ] . in each case , the non - foliated region in @xmath25 is called the _ central region _ of that hexagon , and it will be denoted by @xmath68 . for type
i , this region is bounded by three pieces of hypercycles distant from @xmath46 by @xmath69 . for types ii and iii , the central region is bounded by two hypercycles and one boundary geodesic segment . for type ii , we have @xmath66 and the geodesic boundary of the central region is the long edge @xmath70 . in the type
i case , as the distance from a long edge @xmath46 of length @xmath71 increases from zero to @xmath54 , the leaves of @xmath51 move from the long edge @xmath46 to one of the sides of the central region , which is itself a leaf of @xmath51 .
we further extend the foliations of the rectangular pieces to the central regions by adding leaves which are also pieces of hypercycles so that we get a singular foliation of the whole hexagon @xmath25 , as indicated in the left picture of figure [ extend ] . in this picture
the point @xmath45 is as before equidistant to the three lines @xmath40 , and it is joined to the short sides by geodesic segments . in the cases of type ii and iii , we have singular foliations of the type depicted in the middle and right pictures of figure [ extend ] respectively , where the point @xmath45 is on a long edge in the middle picture , while it lies outside the hexagon in the right picture .
we now look into the type i case more closely using the poincar disc model and taking the point @xmath45 at the center of this disc .
the central region is divided into three pentagons @xmath72 by the geodesic tripod @xmath73 centered at @xmath45 , with the three edges having the same hyperbolic length @xmath12 , each edge meeting one of the short edges perpendicularly .
we have @xmath74 .
each pentagon is foliated by pieces of hypercycles which are equidistant from the long edges @xmath46 of @xmath25 . in this way
, the entire hexagon @xmath25 is foliated as a union of three foliated regions which we call @xmath51 , @xmath75 . for each @xmath76 ,
the leaves of @xmath51 fill out the pentagon @xmath77 .
see figure [ penta ] .
we denote the right angled quadrilateral @xmath78 by @xmath79 .
the pentagon @xmath77 has a @xmath80 symmetry across the angle bisector of @xmath81 and @xmath82 @xmath83 .
we denote the vertex angle of @xmath77 at @xmath45 by @xmath84 .
we equip each pentagon @xmath77 with a second foliation @xmath85 transverse to @xmath51 and whose leaves are the fibers of the nearest point projection map @xmath86 .
the leaves of this foliation are geodesics which make right angles with the leaves of @xmath26 .
the type ii hexagon is a limiting case of the type i hexagon , where the two edges @xmath87 and @xmath88 form an angle @xmath89 , and thus two edges of the tripod are aligned . in this case ,
one of the foliated rectangles as well as one of the subregions of the central region are collapsed to the union of the two edges of the tripod @xmath57 , so that the hexagon is covered by the remaining two foliated regions . in other words
, @xmath25 is a union of two pentagons @xmath90 and @xmath91 whose central vertex angles at @xmath45 satisfy @xmath92 .
the type iii hexagon is represented in figure [ negative ] . here
, the region @xmath25 is a proper subset of two pentagons @xmath90 and @xmath91 ( these are the left and right halves of the hexagon pictured ) with its central angles satisfying @xmath93 , or equivalently @xmath94 . the right angled hexagon is a union of pentagons @xmath77 each of which is foliated by two mutually perpendicular foliations @xmath51 and @xmath85 . in
what follows , we shall concentrate the discussion on type i hexagons , while type ii and iii hexagons will be considered as needed . in a type
i hexagon , we parameterize the leaves of the foliation @xmath51 by @xmath95 as follows . for each @xmath76 ,
the region @xmath77 is foliated by leaves @xmath96 where @xmath97 is the long edge @xmath46 , @xmath98 is the side of the central region in @xmath77 , and @xmath99 is the origin @xmath45 ( a degenerate leaf ) . in between , for @xmath100 , we interpolate the leaves proportionally to the hyperbolic distance from the long edge @xmath46 . for @xmath101 ,
the interpolation is done proportionally to the hyperbolic length along the edges of the geodesic tripod @xmath57 .
as for the leaves of @xmath27 , we parameterize them by @xmath102 so that @xmath103 and @xmath104 are parts of the short edges @xmath105 with @xmath106 of length @xmath54 sandwiching the long edge @xmath46 , and @xmath107 is the geodesic segment from the origin @xmath45 meeting the long edge @xmath46 perpendicularly at its midpoint and bisecting the central angle @xmath84 at @xmath45 .
this geodesic segment is shared by the two congruent quadrilaterals @xmath79 and @xmath108 with @xmath109 . in between , the interpolation is done proportionally to hyperbolic length along the long edge @xmath46 . with this pair of foliations @xmath110 where the leaves are parameterized by @xmath111 and @xmath112 ,
the hexagon has a coordinate system .
namely each point @xmath113 is identified with an ordered pair @xmath114 .
note that along the gluing edges of the three pentagons @xmath115 and @xmath91 , the parameters @xmath111 and @xmath116 with @xmath117 are compatible , that is , @xmath118 if @xmath119 lies on the edge of the tripod between @xmath77 and @xmath120 .
a type ii and iii hexagon @xmath25 has a similar coordinate system , which is just the restriction of the three pentagons @xmath77s to the hexagon @xmath25 .
first note that a congruent pair of hyperbolic geodesic quadrilaterals with three right angles with opposite orientations form a geodesic pentagon with four right angles ( see figures [ nh1 ] and [ qq ] below ) . with this in mind , we recall a set of classical formulae ( @xcite ) for a hyperbolic quadrilateral with three right angles , relating the side lengths and the non - right angle @xmath121 of the quadrilateral .
such a quadrilateral is usually called a trirectangular quadrilateral .
we represent the quadrilateral in the poincar disc model of the hyperbolic plane .
we assume that @xmath122 is positioned at the origin @xmath45 , with two edges being radial segments , one of which , the side @xmath123 , having length @xmath12 and the other one , and the side @xmath124 having length @xmath125 where @xmath23 is the length of the edge @xmath126 opposite to the side @xmath124 ( see figure [ nh2 ] ) .
finally , we denote the length of the side @xmath127 by @xmath128 .
we note that this quadrilateral is uniquely determined by the two values @xmath122 and @xmath12 .
the trigonometric formulae are the following : @xmath129 in these formulae , the angle @xmath130 is the right angle at the vertex @xmath68 . in the book @xcite , the above formulae are given for a quadrilateral with two consecutive right angles , with the angle @xmath68 not necessarily @xmath130 .
we have rewrote them in the special case @xmath131 .
now suppose that the acute non - right angle is expressed as @xmath132 with @xmath133 .
the set of formulae then becomes @xmath134 by noting that @xmath135 , @xmath136 , @xmath137 and @xmath138 , we can rewrite them as @xmath139 - \cosh d \sinh [ -(l + h ) ] \cos \alpha \\ \label{e3.1}\sinh d & = & \sinh \ell \cosh ( l+h ) - \cosh \ell \sinh ( l + h ) \cos \frac{\pi}{2 } \\
\label{e4.1 } \cos \alpha & = & \sin \frac{\pi}{2 } \sinh ( -l ) \sinh \ell - \cos \frac{\pi}{2 } \cosh ( -l ) \\ \label{e5.1 } \cos \frac{\pi}{2 } & = & \sin \alpha \sinh ( -l ) \sinh d - \cos \alpha \cosh ( -l ) \\ \label{e6.1 } \cosh [ -(l+h ) ] & = & - \sinh \ell \sinh d + \cosh \ell \cosh d \cosh ( -l ) \\ \label{e7.1 } \cosh ( -l ) & = & \sin \alpha \cosh [ -(l+h ) ] .
\end{aligned}\ ] ] formally , in comparing to the equations ( [ e1 ] , [ e2 ] , [ e3 ] , [ e4 ] , [ e5 ] , [ e6 ] , [ e7 ] ) , this set seems to describe the shape of quadrilateral with three right angles and one obtuse angle @xmath133 and negative lengths @xmath140 and @xmath141 , even though there is no such hyperbolic quadrilateral as the non - right - angle of the trirectangular quadrilateral in the hyperbolic plane is always acute , and there are no _ negative _ side lengths .
these algebraic expressions , however , can be interpreted as follows . as the isometry type of a trirectangular quadrilateral
is uniquely determined ( up to orientation ) by @xmath142 and the side length @xmath143 , for a fixed @xmath12 , consider the family of quadrilaterals obtained by increasing @xmath122 .
when @xmath122 approaches @xmath130 , the side @xmath127 of length @xmath128 converges to the side @xmath123 , and the sides @xmath126 and @xmath124 collapse to the points @xmath144 and @xmath45 respectively .
thus , the quadrilateral becomes degenerate and we have @xmath145 and @xmath146 in equations ( [ e1 ] ) , ( [ e2 ] ) and ( [ e3 ] ) . for @xmath133 with the same @xmath143
, we identify the situation by equations ( [ e1.1 ] , [ e2.1 ] , [ e3.1 ] , [ e4.1 ] , [ e5.1 ] , [ e6.1 ] , [ e7.1 ] ) with a quadrilateral with vertex angle @xmath132 appearing on the other side of @xmath123 , the mirror image of the corresponding quadrilateral with the angle @xmath147 across the side @xmath123 .
the negative lengths @xmath140 and @xmath141 refer to the reflective symmetry across the side @xmath123 .
the situation is depicted in figure [ sign ] as well as in figure [ negative ] . in this sense ,
the set of equations ( [ e1 ] , [ e2 ] , [ e3 ] , [ e4 ] , [ e5 ] , [ e6 ] , [ e7 ] ) describes the moduli of quadrilaterals with a fixed @xmath143 and variable vertex angle @xmath122 , where @xmath122 varies in @xmath148 $ ] , where the corresponding quadrilaterals are possibly degenerate , and with the opposite orientations with respect to the symmetry across the side @xmath123 .
for a given trirectangular quadrilateral @xmath149 with three right angles at the vertices @xmath150 and one acute angle @xmath122 at @xmath45 , we consider the trirectangular quadrilateral @xmath151 which is the mirror image of @xmath152 by the reflection thorough the straight edge @xmath124 .
we denote the corresponding vertices of @xmath153 by @xmath154 , @xmath155 and @xmath156 .
the union of @xmath152 and @xmath151 is a pentagon @xmath157 ( figure [ qq ] ) , where all the interior angles are right , except at the vertex @xmath45 whose interior angle is @xmath158 .
now consider the situation where there are three pentagons @xmath115 and @xmath91 satisfying the compatibility condition @xmath159 with @xmath160 for each @xmath76 , and @xmath161 , a common value which we denote by @xmath12 .
the three pentagons can be glued together via the identifications of the edges emanating from the origin @xmath45 , @xmath162 to produce a right angled hexagon @xmath25 .
this is the type i case described before , and it is depicted in figure [ nh1 ] . the condition @xmath160 is violated when @xmath163 and thus @xmath164 .
this is the case when @xmath165 degenerates to the line segment @xmath166 , and the hexagon is just @xmath167 . finally when @xmath168 , then @xmath169 , and as described in the previous section , the quadrilateral @xmath170 with its vertex angle @xmath132 appears on the other side of @xmath171 and @xmath153 on the other side of @xmath172 .
consequently the pentagon @xmath173 overlaps with @xmath90 and @xmath91 .
this is the type iii picture .
note that the compatibility condition among @xmath174 is still intact as the vertex angle of @xmath165 at @xmath45 is the sum of those of @xmath90 and @xmath91 : @xmath175 also note that by using the original @xmath168 in the set of equations ( [ e1 ] , [ e2 ] , [ e3 ] , [ e4 ] , [ e5 ] , [ e6 ] , [ e7 ] ) , the resulting negativity of @xmath176 can now be understood with the observation we made in the case of the type iii setting , in which the triangle inequality among @xmath177 and @xmath178 is violated .
we incorporate the poincar unit disc model for the hyperbolic plane and we obtain a new set of formulae that include the euclidean geometry of the unit disc . with the notation of figure [ nh2 ] , with @xmath179 being the euclidean distance from @xmath180 to @xmath181 in the poincar disc , and @xmath182 being the euclidean distance from @xmath180 to @xmath183 ,
we can write @xmath12 and @xmath14 as functions of @xmath179 and @xmath182 : @xmath184 we have the following formulae for the quadrilaterals @xmath185 and @xmath186 : @xmath187 we have the following relations : @xmath188 and @xmath189 equation ( [ e7 ] ) can be written as @xmath190 by applying the angle - addition formula , the equality becomes @xmath191 on the other hand , equation ( [ e5 ] ) shows that @xmath192 which in turn says @xmath193 by combining these equations , we have the following relation satisfied by the two euclidean parameters @xmath182 and @xmath179 : @xmath194 factorizing in equation ( [ e8 ] ) , we get @xmath195 = 0\ ] ] which in turn implies @xmath196 formula ( [ f : combining ] ) , which states that for @xmath122 constant , the two parameters @xmath179 and @xmath182 are linearly related , has the following geometric interpretation .
consider an arc of hypercycle which is a set of points at hyperbolic distance @xmath23 from the side @xmath127 .
the arc starts at the vertex @xmath144 and meets the side @xmath124 at a point , which we call @xmath197 , which is hyperbolic distance @xmath14 away from @xmath45 .
the geometry of the poincar disc implies that the arc is a euclidean circular arc , and we call the region surrounded by the line segment @xmath123 , the arc @xmath198 and the side @xmath199 , the central region .
note when @xmath200 , then @xmath201 , indicating that the central region appears on the other side of @xmath123 compared to the case @xmath202 . as the moduli of convex quadrilaterals with three right angles has two parameters @xmath122 and
@xmath12 , for a fixed angle @xmath122 , we have a one - parameter family of right angled quadrilaterals , with isometry type determined by @xmath12 , or alternatively , by @xmath182 .
the linear relation ( [ f : combining ] ) between the values @xmath179 and @xmath182 says the following : [ central ] for a fixed angle @xmath122 , as the value of @xmath12 varies , the central region changes its shape via euclidean homotheties , centered at the origin @xmath45 , where the scaling is given by the value @xmath203 with @xmath204 .
now we come back to the hexagonal setting where three pentagons , or equivalently , six quadrilaterals , are combined .
type i hexagon is presented in figure [ nh1 ]
. , which is the common length of the three branches of the tripod @xmath57.,width=226 ] equation ( [ e1 ] ) says that @xmath205 by taking a ratio of the equality for @xmath76 and @xmath206 , we have @xmath207 a quantity independent of @xmath12 .
hence for different values of @xmath12 and @xmath208 , @xmath209 for @xmath210 .
this equality implies the following geometric statement : suppose the angles @xmath174 are fixed .
as the value of @xmath12 varies , the ratios among @xmath211 , @xmath212 and @xmath213 of the right angled hexagons remain invariant .
in other words , the ratios of weighted lengths ( where taking the weighted length means replacing it by the hyperbolic cosine of that length ) are invariant .
we concentrate on the deformations such that the long edges @xmath46 get longer , or , equivalently , the scale of the central region gets larger .
we can call these deformations _ forward _ deformations .
namely the forward deformation of @xmath25 makes _ long edges longer and short edges shorter .
_ in the limit of such a deformation , as @xmath214 , one obtains ideal triangles .
this convergence @xmath214 is studied using the formula @xmath215 regardless of the cases @xmath121 or @xmath200 .
note that @xmath216 and @xmath217 are equivalent , and that @xmath218 means that the long edge is being pushed to the geometric boundary of the poincar disc .
hence the one - parameter family of forward deformations provides a canonical path for the right angled hexagon asymptotically converging to an ideal triangle . also , in terms of elongating the short side by a factor @xmath24 ( here @xmath219 ) we can see from an equality obtained from the trigonometric formulae ( [ e1 ] ) and ( [ e5 ] ) , namely , @xmath220 that @xmath221 approaches zero as @xmath222 along the forward deformation .
let us consider the special
symmetric " case where @xmath223 , as this is the case treated in detail in @xcite .
then six copies of the trirectangular quadrilateral ( up to orientation ) placed together form a right angled hexagon with @xmath47-rotational symmetry , which has side lengths @xmath224 and @xmath225 appearing alternatively . inserting the values @xmath226 and @xmath227 in the formula ( 1 )
, we obtain @xmath228 where @xmath229 . solving this for @xmath182 , we get @xmath230 .
this value of @xmath182 is obtained by letting @xmath128 approach zero , or equivalently by letting the right angled hexagon converge to an ideal triangle via backward " deformations .
the value of @xmath179 then is @xmath231 .
when deformed in the forward direction , the hexagon also converges to the ideal triangle , this time with @xmath232 and @xmath233 .
note that the forward deformation is defined for all @xmath234 , while in general the backward deformation is not .
let us recall a few elements from @xcite . a map @xmath19 between two metric spaces is said to be _ contracting _ if @xmath235 , and _ weakly _ contracting
if @xmath236 if @xmath19 is a self - map of class @xmath237 of a convex domain @xmath238 of the hyperbolic plane , then one can compute the norm of its differential at each point @xmath239 of @xmath238 , @xmath240 setting @xmath241 we have the following the quantity @xmath242 provides an upper bound for the global lipschitz factor . in other words , we have @xmath243 in what follows , we consider maps between hyperbolic surfaces .
the following statement will be used later in relating two hyperbolic surfaces .
[ lip.const ] suppose there exist two pairs of orthogonal foliations @xmath244 and @xmath245
which are preserved by @xmath19 ( that is , @xmath19 sends any leaf of @xmath246 onto a leaf of @xmath247 and any leaf of @xmath248 onto a leaf of @xmath249 ) .
if @xmath19 is @xmath24-lipschitz along the leaves of @xmath246 and contracting along the leaves of @xmath248 , then @xmath19 is @xmath24-lipschitz ( that is , @xmath250 . ) furthermore , if the map @xmath19 stretches the leaves of @xmath246 by a constant factor @xmath24 and if there exists an open geodesic segment contained in a leaf of @xmath246 which is sent to a geodesic segment contained in a leaf of @xmath247 , @xmath24 is the best lipschitz constant for @xmath19 .
the proof follows from an argument in @xcite ( p. 65 - 66 ) which we reproduce here .
consider the coordinate system @xmath251 defined by the orthogonal grid formed by the leaves of @xmath246 and @xmath248 , where @xmath122 and @xmath252 are suitably chosen so that the coordinate vector fields @xmath253 are an orthonormal basis for each tangent plane , and are linearly sent to an orthogonal basis by the differential @xmath254 .
if @xmath255 , then @xmath256 let @xmath257 a vector field with @xmath258 .
the norm of the differential @xmath259 is computed as follows .
first , we have @xmath260 from this , it follows that @xmath261 which in turn gives @xmath262 now as the map @xmath19 is @xmath24-lipschitz along the leaves of @xmath246 , we have @xmath263 at any point of @xmath238 . on the other hand as the map
@xmath19 is contracting along the leaves of @xmath248 , we have @xmath264 at any point of @xmath238 .
thus we have @xmath265 .
given three non - negative numbers @xmath266 , let @xmath25 be a right - angled hexagon in the hyperbolic plane with pairwise non - consecutive side lengths @xmath267 .
such a hexagon is uniquely determined up to isometry by these three sides ( see e.g. @xcite ) .
we call these sides the _ long _ edges of @xmath25 .
we denote the lengths of the diametrically facing three pairwise non - consecutive sides ( called the _ short _ edges ) by @xmath268 .
if some @xmath71 is equal to @xmath180 , we consider that the corresponding side is at infinity . for any @xmath269 ,
we denote by @xmath270 the right - angled hexagon with long edges of lengths @xmath271 , with @xmath272 , for @xmath273 , determined by @xmath274 for @xmath273 , we define @xmath275 and we set @xmath276 note that @xmath277 is determined by @xmath25 and @xmath24
. we also note that from its definition , @xmath277 is an increasing function of @xmath24 and that when @xmath24 varies from @xmath278 to @xmath279 , so does @xmath277 . in the rest of this section ,
we construct a @xmath277-lipschitz map @xmath280 which is lipschitz extremal in its homotopy class in the sense that the lipschitz constant of any other map between the two given right angles hexagons sending each edge of @xmath25 to the corresponding edge of @xmath270 is at least equal to @xmath277 . making @xmath24 ( or , equivalently ,
@xmath277 ) vary from 1 to @xmath279 , we obtain a family of marked right angled hexagons which is a geodesic for the lipschitz metric on the space of right angled hyperbolic hexagons a natural analogue of the metric defined above in ( [ lip ] ) . by gluing hexagons along their sides ,
we obtain families of geodesics for the arc metric for surfaces with nonempty boundary ( and with variable lengths of the boundary components ) .
gluing surfaces with boundary along their boundary components and assembling the maps between them , we obtain geodesics for the thurston metric of surfaces without boundary ( possibly with punctures ) .
the maps @xmath281 between the right angled hexagons hexagons that we construct preserve the two pairs of orthogonal pairs of partial foliations @xmath282 and @xmath283 of @xmath25 and @xmath270 respectively that we constructed in
[ s : geo ] .
we construct a @xmath277-lipschitz map @xmath285 where @xmath270 is obtained by stretching each long edges @xmath71 of @xmath25 by the factor @xmath286 , such that @xmath281 satisfies the following conditions : 1 .
@xmath281 sends each leaf of @xmath26 to a leaf of @xmath287 affinely ( with respect to the natural arc - length parametrization ) and it sends each leaf of @xmath27 to a leaf of @xmath288 .
2 . the central region
is sent to the central region . using the coordinates we established on each foliated pentagon @xmath77 ( the foliated rectangle with part of the foliated central region ) ,
the map @xmath281 is defined by sending the point on @xmath25 represented by @xmath289 for some @xmath76 to the point in @xmath270 represented by the same coordinates @xmath289 .
this map is clearly a homeomorphism between @xmath25 and @xmath270 .
it sends the leaves of @xmath26 and @xmath27 on @xmath25 to those on @xmath270 , and the central region to the central region .
we will show a few properties of the map @xmath281 , concluding that this map is lipschitz - optimal between the pair of right - angled hexagons . here
the optimality means that the lipschitz constant of any other map from @xmath25 to @xmath270 is at least @xmath277 .
[ th : k ] the map @xmath281 is an optimal @xmath277-lipschitz map from @xmath25 to @xmath270 for all @xmath269 .
we first recall a classical result in hyperbolic geometry stated as a lemma : [ lem : part ] a part of the hypercycle @xmath290 with @xmath291 , which projects onto a geodesic segment of length @xmath292 on the short edge @xmath43 via the nearest point projection map @xmath86 , has hyperbolic length @xmath293 , which is given explicitly by @xmath294 we use the following elementary fact from the hyperbolic geometry of the upper - half plane model : the ray through the origin of slope @xmath295 is a set of points which are equidistant of @xmath296 from the geodesic represented by the positive @xmath297-axis . thus the hypercycle
is represented by the slanted ray .
the fibers of the nearest point projection map from the slanted ray to the positive @xmath297-axis are ( subsets of ) concentric circles centered at the origin . hence in order to prove the lemma it suffice to compute the hyperbolic stretching factor between the geodesic line segment between @xmath298 and @xmath299 with @xmath300 , and the hypercycle segment represented by the line segment between @xmath301 and @xmath302 .
we carry out this computation by comparing the hyperbolic norm of two paths @xmath303 and @xmath304 defined over @xmath305 $ ] .
recall that the length element in the upper - half plane is given by @xmath306 then the hyperbolic norms of the tangent vectors are given by @xmath307 by taking the ratio and substituting @xmath308 by @xmath309 , it follows that @xmath310 a number independent of @xmath179 .
this proves that the nearest point projection induces a stretching between the geodesic and the hypercycle by the factor of @xmath311 .
lemma [ lem : part ] implies that each leaf @xmath312 of the foliations @xmath51 , for @xmath291 , is obtained by stretching the geodesic segment @xmath313 by the constant factor @xmath314 and that each leaf @xmath315 of the foliations @xmath316 , for @xmath291 , is obtained by stretching the geodesic segment @xmath317 by the constant factor @xmath318 where @xmath319 is the length corresponding to @xmath54 in the right angle hexagon @xmath270 .
we note that @xmath320 .
as the map @xmath281 sends the leaf @xmath312 to @xmath315 , and the long edge @xmath46 of length @xmath71 to the long edge @xmath321 of length @xmath322 , the lipschitz constant of @xmath281 restricted to the leaf @xmath312 is given by @xmath323 where the inequality follows from @xmath320 .
concerning the foliation @xmath85 , whose leaves are perpendicular to the leaves of @xmath51 , for @xmath102 , each leaf @xmath324 of length @xmath54 , namely the part of the leaf in the quadrilateral @xmath325 , is sent to @xmath326 of length @xmath319 , and thus the lipschitz constant is @xmath327 our next task is to obtain a control on the lipschitz constant of @xmath281 in the central region of the hexagon .
we divide the central region @xmath68 into the three sectors @xmath328 , whose interface is the geodesic tripod @xmath57 centered at the origin @xmath45 of the poincar disc , as we discussed in [ s : geo ] .
let @xmath329 be the edge length of the tripod in the right angled hexagon @xmath270 , spanning the central region @xmath330 . from the preceding discussion
, we have @xmath331 . using the formula ( 1 ) in [ s : hyp ] , @xmath332
we obtain the comparison between @xmath12 and @xmath329 @xmath333 this ratio is bounded above by @xmath334 ; namely @xmath335 .
this follows from the convexity , positivity , and monotonicity of the function @xmath336 for @xmath337 .
recall that when we introduced the parameter @xmath111 on the central region , it was defined so that the parameterization on the edges of the geodesic tripod @xmath57 are proportional to the hyperbolic arc - length as @xmath111 varies over @xmath338 $ ] .
hence the lipschitz constant of the map @xmath281 restricted to the tripod , which is thus equal to @xmath339 , is strictly less than @xmath277 .
the region @xmath340 is foliated by the hypercycles equidistant from the long edge @xmath46 , which are the leaves of @xmath312 for @xmath341 , and the discussion used to obtain the lipschitz constant bound of @xmath281 on @xmath79 is also valid .
namely we have @xmath342 as for the lipschitz constant of @xmath281 restricted to the leaves of @xmath85 in the central region @xmath340 , we recall that the leaves @xmath324 are parameterized proportionally to the arc length of the pair of the tripod edges @xmath81 and @xmath82 @xmath83 of equal length @xmath12 , as @xmath111 varies in @xmath343 $ ] .
let @xmath344 be the hyperbolic distance between the point @xmath345 and the long edge @xmath46 .
first note that the function @xmath344 is constant in @xmath112 where it is defined .
as @xmath112 varies over @xmath346 $ ] , the range of @xmath347 varies , and it takes the maximal value for @xmath348 when the endpoint of the leaf @xmath107 is at the center @xmath45 .
we define a function @xmath349 in @xmath350 $ ] by restricting the function @xmath347 to the tripod edge @xmath81 ( or equivalently to @xmath351 ) and subtracting the constant @xmath54 from it .
namely the value of @xmath352 is the distance between the point @xmath353 on @xmath81 ( or equivalently to @xmath351 ) and the boundary hypercycle of the central region @xmath340 . by observing how the leaves @xmath312 intersect with the tripod edges @xmath81 modelled on the poincar disc , with the center @xmath45 of @xmath25 identified with the origin
, we see that the term @xmath354 is equal to @xmath355 where @xmath356 is the angle between the edge @xmath357 and the leaf @xmath358 .
note that @xmath359 and @xmath360 monotonically decreases as @xmath111 increases , and goes down to the value @xmath361 .
it then follows that the derivative of @xmath362 in @xmath111 is monotonically increasing and that @xmath363 the latter limit occurs at the vertex @xmath45 . in other words ,
the function @xmath362 is convex in @xmath111 . defining the function @xmath364 for @xmath270 analogously
, we have similarly : @xmath365 we now claim that the function @xmath366 is increasing in @xmath367 $ ] and bounded above : @xmath368 namely , the claim says that the biggest stretch by @xmath281 along the leaves of @xmath85 occurs at the center of the hexagon on the leaf @xmath107 .
the upper bound @xmath369 is obtained from the ratio between the separation distances of leaves for @xmath287 and @xmath26 near the center of the hexagon , identified with the origin of the poincar disc .
this follows from the observation that for a given value of @xmath370 , we have @xmath371 , which comes from the comparison of the behavior of the poincar metric of the disc , between the two euclidean - homothetic regions @xmath68 and @xmath330 . geometrically , when looking at the distributions of the leaves of @xmath312 and @xmath315 on @xmath68 and on @xmath372 , which are mutually euclidean - homothetic on the poincar disc , the leaves for @xmath316 are more tightly packed than the leaves of @xmath51 near @xmath373 .
hence for a particular pair of leaves @xmath312 and @xmath374 $ ] , we have the inequality @xmath371 .
this comparison in turn shows that as a convex function of @xmath111 , @xmath349 is more convex than another convex function @xmath375 .
this implies the claim about the ratio @xmath376 being increasing and bounded by @xmath377 .
this in turn says that the rate of change @xmath378 is bounded above by @xmath334 , that is , as the map @xmath281 send @xmath340 onto @xmath379 , the stretching factor of the leaves of @xmath85 is strictly less than @xmath334 . combining these results ,
we conclude that for @xmath380 , @xmath381 now the map @xmath281 sends the central region @xmath68 to @xmath330 with each open set @xmath77 to @xmath382 with the norm @xmath383 of the differential bounded by @xmath277 and it sends the tripod @xmath57 to the tripod @xmath384 with constant stretch ratio of @xmath385 as noted above .
those three regions @xmath77 can be glued along the tripod @xmath57 , and the stretching lipschitz constant remains less than @xmath24 , proving that @xmath386 is @xmath277-lipschitz .
so far we have considered a single right - angled hexagon , and its forward deformations .
we consider now a hyperbolic surface @xmath0 with boundary which is partitioned by a maximal sub - system of disjoint geodesic arcs into right - angled hexagons .
such a partition is usually called an ideal triangulation @xmath57 on @xmath0 ( see e.g. @xcite . ) we denote the combinatorial data by @xmath387 .
the gluing edges of the hexagons whose union constitutes the surface @xmath0 are designated to be the long edges , and the edges appearing as part of the boundary components of @xmath0 are designated to be the short edges .
we use the notation of [ s : geo ] . by stretching each long edge @xmath46 of a hexagon by a factor @xmath388 satisfying equation ( [ eq : k_i ] )
, the canonical deformations between hexagons define a map , which we still denote @xmath389 , from the hyperbolic surface with boundary @xmath0 to the hyperbolic surface with boundary @xmath390 obtained by gluing the image hexagons . here , @xmath277 is the maximum among the stretching factors @xmath334 , with @xmath76 indexing the long edges of all the right angled hexagons .
this map @xmath281 makes each boundary component shorter .
the lipschitz map @xmath389 is an optimal @xmath277-lipschitz map in the sense that it is @xmath277-lipschitz , and that for any @xmath391 , there is no @xmath392-lipschitz map from @xmath0 to @xmath390 .
we use lemma [ lip.const ] and we note that when two hexagons @xmath393 and @xmath394 are glued along a long edge @xmath395 , the foliations @xmath396 and @xmath397 are glued together to form a new orthogonal pair of foliations on the union of the image hexagons . the constructed maps between the hexagons satisfy the properties required by the proposition , and the hexagons glue together as well . the following theorems are proved in the same way as theorem 7.3 of @xcite ; we do not repeat the arguments here . in the following statements ,
we take the point of view where each map @xmath389 is an element , which we also denote by @xmath398 , of the teichmller space of @xmath0 , that is , a marked hyperbolic structure .
the lipschitz and the arc metric on @xmath11 coincide on the path @xmath398 , @xmath400 .
more precisely , we have , for any @xmath401 , @xmath402 , [ ( s , f_{k_{2}})])= l([(s , f_{k_{1 } } ) ] , [ ( s , f_{k_{2}})])\ ] ] where @xmath12 and @xmath23 are the arc and lipchitz metrics respectively .
we note that the resulting geodesics for a surface @xmath0 with boundary are distinct from the geodesics constructed by thurston using the stretch line " construction where the ideal triangles spiral along the boundary components of @xmath0 , as mentioned in the introduction . for surfaces with more than one boundary component
, the distinction can be easily seen from the fact that thurston s path causes the lengths of all boundary components change by the same factor , whereas our deformation causes the length of each boundary components to change at different factors in general .
there is a relation between our coordinates on right angled hexagons and coordinates constructed by luo in @xcite on the teichmller space of a surface with boundary .
luo calls his coordinate functions _
radius coordinates_. in our context , they correspond to the sum of the @xmath54s . to be more precise , consider two right angled hexagons @xmath393 and @xmath394 , sharing a long edge of the same length , which we call the edge @xmath403 .
the edge @xmath403 borders two strips , each foliated by leaves equidistant from the edge @xmath403 bordered on the other end by the sides of the central regions in @xmath393 and @xmath394 . in our notation
, the widths of the strips , namely the distances from the shared edge to the respective central regions , are called @xmath404 and @xmath405 .
the radius coordinate of luo is equal to @xmath406 by setting @xmath407 to be the set of long edges of an ideal triangulation @xmath57 with respect to a hyperbolic metric on a surface @xmath0 with totally geodesic boundary , we can consider the set of functionals @xmath408 on the space of hyperbolic metrics of the surface with totally geodesic boundary .
luo calls @xmath409 the radius invariant of @xmath395 , and @xmath410 the radius coordinate system of @xmath387 .
given an ideal triangulation @xmath387 on a compact surface @xmath0 with boundary , each hyperbolic metric with totally geodesic boundary on @xmath0 is determined up to isotopy by its radius coordinates .
furthermore , the image of the map @xmath410 is a convex polytope satisfying the following two properties : for each fundamental edge cycle @xmath411 @xmath412 and for each boundary edge cycle @xmath413 corresponding to the boundary component of length @xmath128 , @xmath414 finally , we mention that alessandrini and disarlo have work in progress in which they define generalized stretch lines " for surfaces with boundary and they prove that every two hyperbolic structures can be joined by a geodesic segment which is a finite concatenation of such generalized stretch lines , cf . @xcite .
n. acampo and a. papadopoulos , notes on non - euclidean geometry , in _ strasbourg master class on geometry _ ( a. papadopoulos , ed .
) irma lectures in mathematics and theoretical physics vol .
18 , european mathematical society , zurich , 2012 , p. 1 - 182 .
|
we construct one - parameter families of right - angled hexagons in the hyperbolic plane such that each right - angled hexagon belongs to such a family , and between each pair of hexagons in the same family we describe a lipschitz map that realizes the best lipschitz constant in its homotopy class relative to the boundary .
this produces new geodesics for the arc metric on teichmller spaces of hyperbolic surfaces with nonempty boundary and as a by - product , we get new geodesics for thurston s metric on teichmller spaces of hyperbolic surfaces without boundary .
the results generalize results obtained in the two papers @xcite and @xcite .
ams mathematics subject classification : 32g15 ; 30f30 ; 30f60 ; 53a35 .
keywords : teichmller space ; arc metric ; thurston s metric ; deforming hyperbolic hexagons .
|
central to the study of galaxy evolution is the formation and evolution of the supermassive black holes in their nuclei .
accretion onto these black holes can provide the energy source for active galactic nuclei , which in turn impact the evolution of galaxies ( silk 2005 ) .
the processes believed to play a role in the formation of seed black holes , from which large black holes may form through accretion , involve 1 ) dynamical friction and collision processes in dense young stellar clusters ( portegies zwart et al .
2004 ) ; 2 ) seeds as the remnants of popiii stars ( e.g. , bromm et al . 1999 , abel et al .
2000 , yoshida et al .
2003 ) ; 3 ) accretion of low angular momentum material and gravitational instability in primordial disks ( e.g. , koushiappas , bullock & dekel 2004 ; lodato & natarajan 2007 ) ; and 4 ) the ( singular ) collapse of massive pre - galactic halos ( bromm & loeb 2003 ; spaans & silk 2006 ) .
the growth of these seed black holes to larger sizes involves accretion that roughly follows an eddington rate and requires the incorporation of feedback effects ( silk & rees 1998 , wyithe & loeb 2003 , di matteo , springel & hernquist 2005 ) .
see pelupessy , di matteo & ciardi ( 2007 ) for an asessment on the difficulties that seed black holes have to grow at the eddingtion rate .
understanding the growth of these black holes is also important because there appears to be a scaling relation between bulge and black hole mass , with about @xmath10 of the bulge mass tied up in the central black holes ( magorrian et al .
1998 , ferrarese & merritt 2000 , hring & rix 2004 ) . in this work
, it is investigated how one can observe a population of these putative black holes in the early universe , at redshifts @xmath11 , through high temperature molecular lines that are driven by x - rays and that are accessible to the atacama large millimeter array ( alma ) , which covers the 300 @xmath12 m to 3 mm wavelength range .
we consider a high redshift halo which already contains , by assumption , a seed black hole .
suppose further that an initially metal - free hydrogen gas is cooled by lyman @xmath13 to @xmath14 k and settles in the center of a halo close to the black hole .
part of the gas is likely to experience a period of popiii star formation .
if popiii star formation occurs , then the short life times of primordial stars ensure that the accreted gas is quickly polluted by a modest amount of dust and metals .
if no star formation takes place , then the gas will remain metal - free . in either case
, the accretion process will lead to the emission of x - rays that impact the thermal , ionization an chemical balance of the gas in the halo , leading to an x - ray dominated region ( xdr , maloney et al .
1996 ; meijerink & spaans 2005 ) . for simplicity the black hole is taken to radiate at the eddington luminosity . for a baryonic number density @xmath15 @xmath16 today , hydrogen mass @xmath17 , halo masses of @xmath18 @xmath1 and a characteristic size scale of @xmath19 , one has a typical mean density and column of @xmath20 ^ 3 $ ] @xmath16 and @xmath21 ^ 2(m_h/10 ^ 9m_\odot)^{1/3}$ ] @xmath22 , respectively .
the work of mo , mao & white ( 1988 ) shows that the subsequent formation of a disk occurs , with a collapse factor of @xmath23 .
this yields densities that exceed @xmath24 @xmath16 within 1 kpc .
the above cosmology provides the boundary conditions for the ambient density and column density of individual models , and the metallicity is put in by hand .
the models of meijerink & spaans ( 2005 ) and meijerink et al .
( 2007 ) are used to compute the thermal , chemical and ionization balance of the irradiated gas self - consistently , for one - dimensional constant density slabs of gas .
the multi - zone escape probability method of poelman & spaans ( 2005 , 2006 ) has been used to compute the line intensities presented here .
the same radiative transfer is performed in relevant atomic ( fine - structure ) and molecular ( rotational and vibrational ) cooling lines .
cloud type `` a '' from meijerink et al .
( 2007 , their table 1 ) is adopted , which is 1 pc in size .
the interested reader is referred to the cited papers for a detailed description of all physical processes involved .
there are four free parameters in the models : hydrogen density , hydrogen column density , metallicity and x - ray flux .
the latter parameterizes unknowns like the accretion rate , turbulent viscosity and the spectral shape of the x - ray radiation . for definiteness , a power law radiation field with @xmath25 is adopted for energies @xmath26 between 1 and 100 kev , appropriate for a self - absorbed seyfert nucleus .
this slope , if it is between @xmath27 and @xmath28 does not significantly impact the chemistries of h@xmath2 , co and h@xmath7 ; see also meijerink & spaans ( 2005 ) for the case of a 1 kev thermal spectrum . solar elemental abundance ratios are adopted . as long as [ o
/ c]@xmath29 this does not influence the co results .
chemical equilibrium is assumed . at densities of @xmath30 @xmath16 and for high x - ray fluxes ,
collisional and radiative time scales are short compared to the free - fall time .
the chemical network comprises a few thousand reactions between 154 species with sizes up to 4 atoms ( woodall et al .
polycyclic aromatic hydrocarbons and small grains are included in the charge balance and are assumed to scale with the elemental carbon abundance . following the above cosmology
, we consider column densities of @xmath31 @xmath22 and densities of @xmath32 @xmath16 .
the x - ray flux takes values of @xmath33 erg s@xmath34 @xmath22 .
results are shown in the figures 1 , 2 and 3 for a density of @xmath35 @xmath16 .
values below this density but above @xmath36 @xmath16 were found to lead to similar signal strengths for h@xmath2 and h@xmath7 . for co ,
line intensities smaller by a factor of @xmath37 were found .
recall that the eddington luminosity is @xmath38 erg s@xmath34 .
so @xmath39 erg s@xmath34 @xmath22 corresponds to a @xmath40 @xmath1 black hole that emits @xmath41 erg s@xmath34 through a surface with a radius of about 100 pc .
the level to which popiii star formation pollutes the center of the primodial galaxy with metals through supernova explosions is a free parameter .
metallicities between @xmath10 and 1.0 solar are considered since @xmath10 solar is quite like a zero - metallicity gas as far as h@xmath2 and co are concerned , and supersolar values appear unlikely for the bulk of the very high redshift gas .
dust grains , with standard milky way properties ( mathis et al . 1977 ) , are included and their abundance is assumed to scale with the overall metallicity . the velocity dispersion of the gas has a thermal contribution , equal to @xmath42 km / s , and a turbulent contribution , equal to @xmath43 km / s , for individual gas clouds on the scale of 1 pc . in the dense gas considered here , turbulence is expected to be maintained at a level similar to that in active galaxies .
the main goal is to compute the expected columns of h@xmath2 , co and h@xmath7 , and their associated emission strengths . for simplicity ,
a fiducial column of @xmath44 @xmath22 is adopted for the predicted line emissivities .
this is driven by the theoretical considerations above , but also by recent observations of massive galaxies at @xmath45 ( daddi et al . 2007a , b ) .
a significant fraction ( 20 - 30% ) of the systems appear to contain heavily obscured agn with columns in excess of @xmath46 @xmath22 .
these massive systems appear to be a somewhat later stage of the concurrent bulge - black hole mass forming systems studied here .
the x - ray luminosities of these lower redshift counterparts is @xmath47 erg s@xmath34 in the 2 - 8 kev band .
the precise value of the total obscuring hydrogen column does not impact our emissivity results as long as it exceeds @xmath48 @xmath22 , i.e. , includes all @xmath49 k gas . in the absence of dust grains , h@xmath2 is formed in the gas phase through the h@xmath50 route , h@xmath50+h@xmath8h@xmath2+e@xmath50 .
this leads to high abundances of h@xmath2 , @xmath51 , even for low metallicities ( figure 1 ) .
so , contrary to far - ultraviolet ( fuv , 6 - 13.6 ev ) illumination , x - ray irradiation constitutes a form of positive feedback for h@xmath2 ( haiman et al .
the models with a modest metallicity of @xmath52 , and thus dust grains , follow the h@xmath2 formation prescription as in cazaux & spaans ( 2004 ) , which includes both physisorbed and chemisorbed hydrogen atoms .
the columns of h@xmath2 that are reached at low metallicities are as large as @xmath48 @xmath22 .
this while temperatures exceed 100 k over the bulk of this column , and reach @xmath36 k at its edge , sufficient to excite the @xmath53 k s(0 ) line at 28 @xmath12 m and @xmath54 k s(1 ) line at 17 @xmath12 m .
the s(1 ) line is typically weaker than the s(0 ) line , while both lines are optically thin and in lte .
high molecular gas temperatures are reached in xdrs because of efficient ionization and coulomb heating , rather than photo - electric heating from dust grains ( meijerink & spaans 2005 ) .
interestingly , an increase in metallicity decreases the h@xmath2 line strength .
this is a direct consequence of enhanced cooling in fine - structure lines and a lower resulting temperature .
hence , the pure rotational h@xmath2 lines are particularly well suited to detect the earliest stages of black hole accretion , prior to significant metal pollution by star formation .
metallicities below @xmath10 were found to lead to quite similar h@xmath2 emissivities since all h@xmath2 is formed in the gas phase and fine - structure cooling is modest .
figure 2 ( see section 4 for its details ) shows rest frame line intensities of @xmath55 erg s@xmath34 @xmath22 sr@xmath34 at a metallicity of @xmath10 solar . for the numbers below
, the alma sensitivity tool on the eso website has been used and the concordance model , with the latest wmap3 results , is adopted . with a source size of @xmath56 , or 600 pc at @xmath57 , this yields a s(0 ) flux density of @xmath58 jy , for a fiducial galaxy center velocity dispersion of 20 km / s ( @xmath59 pc from a @xmath60 central mass ) .
for the s(0 ) line from @xmath57 , so at 0.45 mm , and for 10 km / s spectral resolution , this is detectable with alma ( 50 antennas , beam size = source size ) at the @xmath61 level in 4 hours of integration .
hence , despite the fact that the line falls in the less sensitive band 9 , spectrally resolved detection is possible .
the pure rotational h@xmath2 lines are intrinsically very bright , because of the strong contribution from the x - ray driven h@xmath50 route , and are accessible to alma for redshifts above 10 , for s(0 ) 28 @xmath12 m , and above 16 for s(1 ) 17 @xmath12 m .
figure 1 , for a density of @xmath35 @xmath16 , shows that the h@xmath7 abundance increases with ionization parameter @xmath62 for metallicities below @xmath63 of solar .
a higher x - ray ionization rate leads to a larger h@xmath7 abundance , through h@xmath2 secondary ionizations , and boosts the formation rate h@xmath2+h@xmath64@xmath8h@xmath7 .
the h@xmath2 abundance is thus key since no h@xmath7 can be formed without it .
too large values of @xmath62 lead to a decrease in the h@xmath7 abundance , driven by dissociative recombination .
the best candidate for an h@xmath7 detection is the optically thin ( 2,2)@xmath8(1,1 ) line at 95 @xmath12 m , as suggested by pan & oka ( 1986 ) .
the critical density of the ( 2,2)@xmath8(1,1 ) transition is about @xmath65 @xmath16 and the excitation energy @xmath66 k. collisions between h@xmath7 and electrons have been included in the non - thermal rotational excitation of h@xmath7 ( faure et al .
2006 ) . in all , figure 2 shows that one reaches a maximum rest frame intensity of @xmath67 erg s@xmath34 @xmath22 sr@xmath34 , at low metallicity and strong x - ray irradiation , with typical abundances of @xmath68 . with a source size of @xmath69 , or about 1.8 kpc at @xmath70 ,
this yields a flux density of @xmath71 mjy , for a fiducial galaxy center velocity dispersion of 20 km / s .
for @xmath72 at 1 mm ( from 95 @xmath12 m ) , this is barely detectable with alma at the @xmath73 level , in 48 hours of integration ( 50 antennas , beam size = source size ) , and only if the line is unresolved .
[ [ co ] ] co ~~ any metallicity larger than @xmath10 solar leads to significant , @xmath74 , abundances of co. x - rays , because of their large energy , do not dissociate co directly .
fuv photons are produced through collisional excitation of h and h@xmath2 by electrons , followed by radiative decay .
this uv flux is generally modest and thus co can survive even in strong x - ray radiation fields .
given the high molecular gas temperatures in xdrs , @xmath75 k , rotational levels with @xmath76 are excited ( meijerink et al .
2006 , 2007 ) .
this very high @xmath77 co emission requires densities @xmath78 @xmath16 because the critical densities of these lines are about @xmath40 @xmath16 .
metallicities in excess of @xmath63 of solar further raise the co emissivities , even though the higher abundances of c , o , si and fe also enhance the cooling of the gas ( santoro & shull 2006 ) .
the co emissivities below @xmath10 solar are negligible .
one finds from figure 2 that , at a metallicity @xmath79 solar , rest frame co line intensities reach @xmath4 erg @xmath22 s@xmath34 sr@xmath34 , which yields a flux density of @xmath80 jy for a @xmath56 source size , or 850 pc at @xmath72 , and a fiducial galaxy center velocity dispersion of 20 km / s . for the rest frame peak in the co line spectral energy distribution ( sed ) at 3000 ghz ( see figure 2 ) , 10 km / s velocity resolution and for @xmath72 , alma detects such a source at @xmath81 in 8 hours of integration ( 50 antennas , beam size = source size ) .
spectrally resolved detection is possible .
alma covers 0.3 to 3.0 mm , fortuitously matching most of the x - ray driven co line sed for @xmath11 .
figure 2 shows how the rest frame spectral line distribution of co , pure rotational h@xmath2 and h@xmath7 evolves with irradiation and metallicity for an alma beam that is filled with 1 pc clouds at @xmath30 @xmath16 .
these clouds are in keplerian orbit around the central black hole and are randomly distributed over a spherical region with a linear size of 1 kpc in such a way that there is about one cloud along each line of sight .
this central region is slowly enriched in metals as indicated .
the total column density is on average @xmath82 @xmath22 along each line of sight .
the total intensity is then found by a ray - trace on the level populations of the individual clouds .
more complicated geometries are not an issue as long as the cloud covering factor is of the order of unity and the bulk of the emitted x - rays are absorbed .
overall , x - ray fluxes of @xmath29 erg s@xmath34 @xmath22 , corresponding to black hole masses of @xmath83 @xmath1 and a region with a size of @xmath84 pc yield detectable lines , provided the black hole is radiating at its eddington luminosity .
this mass value is comparable to that of black holes present in in cosmological simulations of @xmath85 halos collapsing at @xmath86 ( peluppessy et al .
black holes that are emitting below the eddington luminosity by a factor of @xmath87 are detectable with alma only if their masses exceed @xmath88 @xmath1 .
also , the eddington time of @xmath89 yr is 25% of the hubble time at @xmath72 , yielding a fair probability for detection of these systems in the alma frequency window .
finding sources is best done in the continuum ( e.g. the james webb space telescope ) , with alma follow - up . since the molecular lines are optically thin , they trace the kinematics of all irradiated gas .
multiple co and h@xmath2 lines further allow one to determine the ambient density , temperature and x - ray flux .
[ oi ] 63 @xmath12 m and [ cii ] 158 @xmath12 m fine - structure lines can be important coolants of x - ray irradiated gas as well . at a metallicity of @xmath5 , fine - structure line cooling contributes about half of the total cooling .
these lines could also be observed with alma , but fuv or cosmic ray irradiation boosts them as well ( meijerink et al .
2007 ) , diminishing their diagnostic value for tracing black hole accretion .
in addition , the 149 @xmath12 m ( @xmath90 , @xmath91 ) rotational line of heh@xmath92 is shown in figure 2 , because it was suggested by maloney et al .
( 1996 ) as a useful xdr tracer .
it is found that this line is typically much weaker than the h@xmath2 and co lines , and heh@xmath92 abundances do not exceed @xmath93 .
fast ( @xmath94 km / s ) shocks can lead to similar co and h@xmath2 emissivities , albeit over much smaller regions so that beam dilution would be an issue .
also , shocks should produce co line profiles with strong non - gaussian wings . at redshifts of a few ,
co emission with upper levels @xmath95 from massive tori in active galaxies is also accessible to alma .
this has been studied by kawakatu et al .
( 2007 ) for the case where fuv photons drive the chemistry .
narayanan et al . ( 2008a ) look at simulations of @xmath96 quasars in @xmath97 @xmath1 halos and find that the co is highly excited by starbursts , peaking at @xmath98 .
narayanan et al . ( 2008b ) further find that agn - driven winds may leave signatures in the co line emission profile in the form of high velocity peaks at a few times the circular velocity .
also , lintott & viti ( 2006 ) and meijerink et al .
( 2007 ) find an increase in hcn emission with x - ray flux , but this effect is suppressed at low metallicity and for densities @xmath99 @xmath16 .
the x - ray driven lines presented here complement these efforts .
finally , figure 3 shows a co comparison between an agn ( xdr ) and a starburst ( pdr = photon dominated region ) model , for the same impinging flux by energy of 100 erg s@xmath34 @xmath22 , typical of a @xmath100 erg s@xmath34 seyfert nucleus or @xmath40 b0 stars within a 200 pc region .
solar metallicity is assumed , merely because it favors the pdr .
it is obvious that star formation can never compete with an xdr , for the same illuminating flux by energy , in terms of the very high @xmath77 co line intensities that are produced .
|
many present - day galaxies are known to harbor supermassive , @xmath0 @xmath1 , black holes .
these central black holes must have grown through accretion from less massive seeds in the early universe .
the molecules co and h@xmath2 can be used to trace this young population of accreting massive black holes through the x - ray irradiation of ambient gas .
the x - rays drive a low - metallicity ion - molecule chemistry that leads to the formation and excitation of co and h@xmath2 in @xmath3 k gas .
h@xmath2 traces very low metallicity gas , @xmath4 solar or less , while some pollution by metals , @xmath5 solar or more , must have taken place to form co. strong co @xmath6 and h@xmath2 s(0 ) & s(1 ) emission is found that allows one to constrain ambient conditions .
comparable line strengths can not be produced by fuv or cosmic ray irradiation .
weak , but perhaps detectable , h@xmath7 ( 2,2)@xmath8(1,1 ) emission is found and discussed .
the models predict that black hole masses larger than @xmath9 @xmath1 can be detected with alma , over a redshift range of 5 - 20 , provided that the black holes radiate close to eddington .
|
the x - ray appearance of most nearby ( @xmath6 mpc ) galaxies is determined by bright x - ray binaries @xcite , such as high - mass x - ray binaries ( hmxbs ) and low - mass x - ray binaries ( lmxbs ) .
because of their high x - ray luminosity ( @xmath7 ) notable fraction of them can be resolved with moderately deep _
chandra _ exposures , and hence hence they can be studied in full particulars ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
underneath bright x - ray binaries diffuse emission is present in all types of galaxies .
similarly to the galactic ridge x - ray emission @xcite , a part of this emission originates from the collective emission of faint ( @xmath8 ) compact x - ray sources , mostly from active binaries ( abs ) , cataclysmic variables ( cvs ) , young stars , and young stellar objects ( ysos ) . based on the analysis of a few nearby early - type galaxies ,
it was suggested that the overall luminosity of abs and cvs is tightly correlated with the stellar mass of the galaxy ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
additionally , in all types of galaxies truly diffuse emission from warm ( @xmath9 kev ) ionized interstellar medium ( ism ) is also present , whose measure , among others , depends on the mass of the host galaxy , albeit with large dispersion @xcite .
the total observed unresolved emission is the combination of these , and possibly other , components . [ cols="^,^,^,^,^,^,^,^,^,^,^ " , ] _ note .
_ columns are as follows .
( 1 ) contamination subtracted @xmath0 ratios in the @xmath1 kev energy range . ( 2 )
total k - band luminosity from the 2mass archive .
( 3 ) references are : @xmath10 @xcite
@xmath13 @xcite .
( 4 ) central velocity dispersion from hyperleda catalog ( http://leda.univ-lyon1.fr/ ) .
( 5 ) radial velocity relative to m87 . ( 6 ) distance from m87 .
( 7 ) globular cluster specific frequency .
refererences are : @xmath14 @xcite
[ tab : list3 ] main parameters of the virgo galaxies are summarized in table [ tab : list3 ] , listing their @xmath0 ratio , k - band luminosity , age , stellar velocity dispersion , radial velocity with respect to m87 , offset from the center of m87 , and globular cluster ( gc ) specific frequency .
we use m87 to characterize the position and velocity of the galaxy with respect to the intracluster medium . from the table ,
one may conclude that galaxies with anomalous @xmath0 ratios have several common properties : ( i ) are members of the virgo cluster of galaxies ; ( ii ) are massive , their total k - band luminosities are in the range of @xmath16 ; ( iii ) have relatively small velocity with respect to the intracluster gas @xmath17 km / s ; ( iv ) are old , with the stellar age in the @xmath18 gyrs range @xcite .
based on our , admittedly limited , sample , it appears that each of these properties is essential .
indeed , the following galaxies , missing one of the above properties , have perfectly normal @xmath0 ratios : ( i ) a similarly massive and old virgo galaxy m89 , but having a large radial velocity with respect to m87 , @xmath19 km / s ; ( ii ) a massive field galaxy ngc3585 with a total mass of @xmath20 and relatively young age ; ( iii ) two massive , @xmath21 , but apparently young virgo galaxies ngc4365 and ngc4526 with the age of 3.6 and 1.7 gyrs , respectively @xcite , one of them having small ( @xmath22 km / s ) and the other large ( @xmath23 km / s ) velocity .
interestingly , there seems to be no dependence on the radial distance from the center of virgo cluster , which may be explained by the somewhat irregular spatial distribution of its intracluster gas . as one can see from the table [ tab : list3 ] , the four galaxies with anomalously high @xmath0 ratios have the highest values of the globular clusters specific frequency .
therefore their lmxb populations have a larger contribution of x - ray binaries formed dynamically in globular clusters , which could , hypothetically , result in enhanced x - ray luminosity in the hard band , due to unaccounted contribution of faint sources formed in globular clusters .
this is further discussed in the section [ sec : faint_src ] ) and it is shown that gc sources can not explain the anomalous hard emission . because of the limited size of our sample , it is unclear , whether the observations presented above reflect true properties or are a result of a chance coincidence . with this in mind , we proceed with the discussion of various mechanisms which may be responsible for the excess hard emission . in the left panel of fig .
[ fig : allspec ] we show the energy spectra of all studied galaxies combined in three groups : 1 ) the galaxies showing anomalous @xmath0 ratios m49 , m60 , m84 , and ngc4636 , 2 ) all remaining early - type galaxies and 3 ) all late - type galaxies .
to facilitate comparison , the spectra are normalized to @xmath24 and are projected to the distance of @xmath25 mpc .
firstly , fig .
[ fig : allspec ] shows that at every energy the spectra of the `` anomalous '' ellipticals is higher by @xmath26 orders of magnitude .
secondly , it is obvious that the enhanced hard @xmath0 ratios are not a result of the soft component `` leaking '' into the @xmath1 kev band .
instead , it requires an additional hard component in the spectrum .
such a component , if of thermal origin , should have a temperature of about @xmath27 kev , as illustrated by the right - hand panel in fig.[fig : allspec ] .
we fitted the spectrum with a model consisting of two mekal models and a power law component .
the latter accounts for the population of unresolved lmxbs and has the slope fixed at @xmath28 , and the normalization fixed at the value corresponding to the estimated luminosity from unresolved lmxbs .
the best - fit temperatures of the thermal components are @xmath29 kev and @xmath30 kev .
the hotter component can be replaced by a power law with the photon index of @xmath31 , giving similar fit quality . in both models ,
the fit is statistically unacceptable , mainly because of the poor approximation at low energies , indicating that the spectrum of the soft component is more complex .
the surface brightness profiles of unresolved emission are presented in fig [ fig : profiles ] . in the left four panels we show for each galaxy the soft and hard band profiles along with the k - band light distribution .
the contribution of the warm ism emission and unresolved lmxbs are subtracted from the hard band profiles .
the emission from the former was computed from the soft band profiles using the best fit temperature of the soft component from table [ tab : list2 ] .
the emission from unresolved lmxbs was calculated from the k - band light distribution , the normalizations were determined as described in sect . [ sec : xraybin ] .
the amplitude of these two main contaminating factors is investigated in the right hand panels , where we show the total hard band emission , before subtracting ism and lmxb contributions ( but instrumental and cosmic background removed ) and the distributions of hard band emission from ism and unresolved lmxbs . as one can see from the right hand panels , the hard band emission exceeds by a factor of about @xmath32 the level of contamination ( except for m84 where unresolved lmxbs contribute @xmath33@xmath34 of the total emission in the hard band ) .
this is much larger than the possible uncertainty in their determination .
this confirms that the observed anomalous hard x - ray emission is not the result of these contaminating factors .
we remind that the contribution of ism and unresolved lmxbs is subtracted from the hard band luminosities plotted in fig .
[ fig : lxlk ] and fig .
[ fig : early_virgo ] .
although there are some differences between the two distributions , the soft and hard band emissions have roughly the same spatial extent , approximately similar to the spatial extent of the k - band emission .
the best fit temperature of the hard component , @xmath35 kev is close to the temperature of the intracluster gas in virgo ( e.g. * ? ? ?
* ; * ? ? ?
* ) , suggesting that the excess emission may be due to intracluster gas accreted into the gravitational potential of a massive galaxy ( e.g. * ?
this interpretation is qualitatively consistent with the four abovementioned properties of the galaxies with anomalous hard emission . indeed ,
( i ) such mechanism can work only for galaxies located in clusters ; ( ii ) galaxies must be sufficiently massive to confine the @xmath33@xmath27 kev gas ; ( iii ) galaxies should move sufficiently slowly with respect to the gas ; ( iv ) some time is needed in order to accrete sufficiently massive x - ray halo .
one of the major facts , of such interpretation would have to explain , is the short cooling time of the gas .
given the gas density and temperature determined from the spectral fits , its cooling time inside 1 kpc does not exceed @xmath36@xmath37 gyr in any of the four galaxies .
another difficulty of this interpretation is the nearly identical radial profiles in the soft and hard bands .
indeed , in the hydrostatic equilibrium , the hotter gas component is expected to have broader spatial distribution than the cooler one , as the scale heights for gas with @xmath38@xmath39 kev and @xmath40@xmath41 kev are by factors of @xmath33@xmath42 different .
this , however , is not observed the soft and hard band emission appear to fill the same volume .
on the other hand , the ism may have a complex multiphase thermal structure , with gas of different temperatures present at all radii .
existence of such multiphase ism in massive elliptical galaxies has been proposed , for example , by @xcite and @xcite .
based on x - ray spectroscopy of ngc1399 and ngc5044 they suggested that these galaxies host multiphase ism .
the temperature of the cooler component was found to be @xmath43 kev , whereas the temperature of the hotter component was consistent with the intracluster medium ( icm ) temperature , i.e. similar to our results .
it remains to be seen , whether accretion of hot icm can explain anomalous hard emission from the four massive galaxies .
a major advance in testing this hypothesis can be achieved by means of high resolution x - ray spectroscopy . in principle , it is possible that the hard emission is due to an enhanced population of faint stellar sources , such as abs and cvs .
this would explain the hard spectrum of the excess emission , consistent with the power law with the photon index of @xmath44@xmath45 and the fact that hard band profiles approximately follow the distribution of the k - band light .
although this assumption can not be ruled out entirely , we stress that the numbers of resolved lmxbs in these galaxies are in very good agreement with the prediction based on the average scaling relations for lmxbs @xcite .
the stellar ages of the galaxies under consideration are also entirely normal for elliptical galaxies . for these reasons
, it seems implausible that in these galaxies the ratio of @xmath46 is @xmath47 times higher than in other early - type galaxies from our sample .
the high specific frequency of globular clusters ( gcs ) observed in m49 , m60 , m84 , and ngc4636 may suggest , that the excess hard band emission could be due to unaccounted contribution of undetected faint sources in gcs .
these sources may be undetected , but relatively luminous lmxbs , whose luminosities fall just below the _ chandra _ source detection threshold , as well as much fainter sources with luminosities in the @xmath48 range ( cvs , quiescent lmxbs , millisecond pulsars etc )
. this , however , does not seem to be a viable possibility for the following reasons .
the method of the xlf renormalization is based on the numbers of all observed bright lmxbs , irrespective of their association with gcs .
therefore the possibly enhanced specific frequency of lmxbs in these galaxies , due to the contribution of systems dynamically formed in gcs , is automatically taken into account by this procedure and is included in the calculation of the luminosity of unresolved lmxbs .
in fact , the shapes of the luminosity functions of field and gc lmxbs are markedly different : in the latter the fraction of faint sources ( @xmath49 ) is @xmath36@xmath50 times smaller @xcite .
therefore this procedure may slightly oversubtract the contribution of undetected gc lmxbs .
it is also worth to note that the specific frequency of resolved x - ray binaries in these galaxies is not much different , within a factor of @xmath33@xmath51 , from other galaxies in our sample , which fact speaks against significantly enhanced population of ( bright ) x - ray binaries .
the possibly enhanced population of faint sources ( @xmath52 ) can not be measured directly in external galaxies but their contribution can be constrained . from _
_ observations of gcs in the milky way , the total luminosity of faint sources even in those of them with the largest encounter rates , e.g. ngc6440 and 47 tuc @xcite does not exceed @xmath53 erg / s .
the latter number was obtained by integrating the luminosity functions of faint sources in ngc6440 @xcite and in 47 tuc @xcite and refers to the @xmath54 kev band .
the numbers of globular clusters in massive ellipticals is @xmath55 , therefore their total luminosity in the soft band does not exceed @xmath33@xmath56 erg / s .
the soft - to - hard band ratios for faint globular cluster sources are in the @xmath33@xmath57 range , but brighter sources tend to have softer spectra @xcite .
thus an upper limit on the hard band luminosity is in the @xmath36@xmath58 erg / s range .
this is about @xmath42 orders of magnitude smaller than the hard band excess luminosity detected in galaxies under consideration , @xmath33@xmath59 erg / s .
an even much less likely possibility is that the enhanced x - ray emission is due to population of young starts and yso .
using the scaling relation @xmath4sfr@xmath60 obtained in sect .
[ sec : latetype ] we conclude that a star - formation rate of @xmath61/yr is required in order to maintain @xmath62 kev luminosity of @xmath63 ( table [ tab : list2 ] ) .
this values are appropriate for intensely star - forming galaxies and are unrealistic for ellipticals of @xmath64 gyrs age @xcite .
we investigated the properties of unresolved x - ray emission in a broad sample of nearby early - type and late - type galaxies based on archival _ chandra _ data . after removing the contribution of resolved and unresolved x - ray binaries we measured @xmath0 ratios in the @xmath65 kev and @xmath1 kev bands and compared them with the solar neighborhood values . complementing this data with the spectral and spatial information we concluded that the unresolved x - ray emission originates from , at least , four distinct components . 1 . the population of faint unresolved sources associated with old stellar population . based on the data for elliptical galaxies we obtained a scaling relation for its @xmath62 kev luminosity : @xmath66 with the rms@xmath67 , which number is in excellent agreement with the value obtained by @xcite for the solar neighborhood .
the @xmath0 for the soft band and for the late - type galaxies can not be determined unambiguously because of the contribution of ism and young stellar sources . for three low - mass gas poor ellipticals we obtained in the soft band @xmath44@xmath68 , which is by a factor of @xmath33@xmath27 smaller than the solar neighborhood value .
2 . in all galaxies , warm ism with the temperature in the @xmath69 kev range
is present . in our sample ,
the gas temperatures in late - type galaxies tend to be lower than in early - types galaxies of the same mass .
similar to the results found in previous studies ( e.g. * ? ? ?
* ) , the amount and luminosity of the gas generally scales with the stellar mass of the host galaxy , albeit with the large scatter . the scale factor and
the scatter appear to increase with the mass of the galaxy .
3 . the population of unresolved young stars and ysos in late - type galaxies .
the x - ray emission of this component approximately scales with the sfr of the host galaxy with the average @xmath4sfr@xmath70 .
the most unexpected result of our study is the detection of anomalous emission in the @xmath1 kev band from four old and massive virgo ellipticals ( m49 , m60 , m84 , ngc4636 ) .
we could not offer an unambiguous explanation of this emission
. a plausible interpretation may be that it is a result of accretion of the intracluster gas in the gravitational potential of the massive galaxy . in this scenario , the two component spectra on one hand and the similarity of surface brightness profiles on the other ,
would point at the multiphase nature of ism in these galaxies .
high resolution x - ray spectroscopy may shed further light on the origin of the anomalous emission .
_ acknowledgements . _
the authors thank bill forman , christine jones , and ralph kraft for critical discussions of these results and the anonymous referee for the careful reading of the manuscript and constructive comments .
this research has made use of _ chandra _ archival data provided by the _
chandra _ x - ray center .
the publication makes use of software provided by the _
chandra _ x - ray center ( cxc ) in the application package .
the _ spitzer space telescope _ is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
anders , e. & grevesse , n. , 1989 , gecoa , 53 , 197 bavouzet , n. , dole , h. , le floch , e. , caputi , k. i. , lagache , g. , kochanek , c. s. , 2008 , a&a , 479 , 83 bell , e. f. , 2003 , apj , 586 , 794 bogdn , . ,
gilfanov , m. , 2008 , mnras , 388 , 56 bogdn , . & gilfanov , m. , 2010 , a&a , 512 , 16 bogdn , . & gilfanov , m. , 2011 , mnras , 412 , 401 bhringer , h. & briel , u. g. , schwarz , r. a. , voges , w. , hartner , g. & trmper , j. , 1994 , nature , 368 , 828 boroson , b. ; kim , d .- w . & fabbiano , g. , 2011 , apj , 729 , 12 brighenti , f. , mathews , w. g. , humphrey , p. j. & buote , d. a. , 2009 , apj , 705 , 1672 buote , d. a. , 2002 , apj , 574 , 135 buote , d. a. , lewis , a. d. , brighenti , f. & mathews , w. g. , 2003 , apj , 594 , 741 david , l. p. , jones , c. , forman , w. , vargas , i. m. & nulsen , p. , 2006 , 653 , 207 denicolo , g. , terlevich , r. , terlevich , e. , forbes , d. a. & terlevich , a. , 2005 , mnras , 358 , 813 de vaucouleurs , g. 1975 , apjs , 29 , 193 dickey , j. m. , lockman , f. j. , 1990 , ara&a , 28 , 215 dirsch , b. , schuberth , y. & richtler , t. , 2005 , a&a , 433 , 43 fabbiano , g. , 2006 , ara&a , 44 , 323 fazio , g. g. , et al .
, 2004 , apjs , 154 , 10 freedman , w. l. , et al . , 2001 , apj , 553 , 47 gallagher , j. s. , garnavich , p. m. , caldwell , n. , kirshner , r. p. , jha , s. w. ; li , w. , ganeshalingam , m. & filippenko , a. v. , 2008 , apj , 685 , 752 gilfanov , m. , 2004 , mnras , 349 , 146 grimm , h .- j . , gilfanov , m. & sunyaev , r. , mnras , 339 , 793 heinke , c. o. , grindlay , j. e. , edmonds , p. d. , cohn , h. n. , lugger , p. m. , camilo , f. , bogdanov , s. & freire , p. c. , 2005 , apj , 625 , 796 irwin , j. a. & sarazin , c. l. , 1996 , apj , 471 , 683 irwin , j. a. , athney , a. e. & bregman , j. n. 2003 , apj , 587 , 356 jarrett , t. h. , chester , t. , cutri , r. , et al . , 2003 , aj , 125 , 525 jensen , j. b. , tonry , j. l. , barris , b. j. , et al . 2003 , apj , 583 , 712 karachentsev , i. d. , karachentseva , v. e. , huchtmeier , w. k. & makarov , d. i. , 2004 , aj , 127 , 2031 koyama , k. , hamaguchi , k. ; ueno , s. , kobayashi , n. & feigelson , e. d. , 1996 , pasj , 48 , 87 leonard , d. c. , et al . , 2002 , aj , 124 , 2490 li , z. & wang , q. d. , 2007 , apj , 668 , 39 macri , l. m. 2001 , apj , 549 , 721 mateo , m. l. 1998 , ara&a , 36 , 435 mathews , w. g. & brighenti , f. , 2003 , ara&a , 41 , 191 mineo , s. , gilfanov , m. & sunyaev , r. , arxiv1105.4610 moretti , a. , campana , s. , lazzati , d. & tagliaferri , g. , 2003 , apj , 588 , 696 osullivan , e , forbes , d. a. & ponman , t. j. , 2001 , mnras , 328 , 461 osullivan , e. & ponman , t. j. , 2004 , mnras , 349 , 535 pellegrini , s. & fabbiano , g. , 1994 , apj , 429 , 105 peng , e. w. , et al . , 2008 , apj , 681 , 197 pooley , d. , et al . , 2002 ,
apj , 573 , 184 pooley , d. , et al .
, 2003 , apj , 591 , l131 rieke , g. h. , et al . , 2004 , apjs , 154 , 25 revnivtsev , m. , sazonov , s. , gilfanov , m. , churazov , e. & sunyaev , r. , 2006 , a&a , 452 , 169 , 178 revnivtsev , m. , churazov , e. , sazonov , s. , forman , w. & jones , c. , 2008 , a&a , 490 , 37 snchez - blzquez , p. , gorgas , j. , cardiel , n. & gonzlez , j. j. , 2006 , a&a , 457 , 809 sazonov , s. , revnivtsev , m. , gilfanov , m. , churazov , e. & sunyaev , r. , 2006 , a&a , 450 , 117 shtykovskiy , p. & gilfanov , m. , 2005 , mnras , 362 , 879 stanek , k. z. , garnavich , p. m. , 1998 , apj , 503 , 131 swartz , d. a. , ghosh , k. k. , tennant , a. f. & wu , k. , 2004 , apjs , 154 , 519 terlevich , a .
i. & forbes , d. a. , 2002 , mnras , 330 , 547 trinchieri , g. , et al . , 2008 , apj , 688 , 1000 tonry , j. l. , dressler , a. , blakeslee , j. p. , ajhar , e. a. , fletcher , a. b. , luppino , g. a. , metzger , m. r. & moore , c. b. , 2001 , apj , 546
, 681 zhang , z. , et al . , 2011 , a&a , in press ,
|
we explore the nature of unresolved x - ray emission in a broad sample of galaxies of all morphological types based on archival _ chandra _ data . after removing bright compact sources , we study @xmath0 luminosity ratios of unresolved emission , and compare them with the solar neighborhood values .
we conclude that unresolved emission is determined by four main components , three of which were known before : ( i ) the population of faint unresolved sources associated with old stellar population . in early - type galaxies , their @xmath1 kev band luminosity scales with the stellar mass with @xmath2 ; ( ii ) the ism with @xmath3 kev present in galaxies of all types . because of the large dispersion in the gas content of galaxies , the size of our sample is insufficient to obtain reliable scaling law for this component ; ( iii ) the population of unresolved young stars and young stellar objects in late - type galaxies .
their @xmath1 kev band luminosity scales with the star - formation rate with @xmath4sfr @xmath5 ; ( iv ) in four old and massive virgo ellipticals ( m49 , m60 , m84 , ngc4636 ) we find anomalously high x - ray emission in the @xmath1 kev band .
its presence has not been recognized before and its nature is unclear .
although it appears to be stronger in galaxies having stronger ism component , its existence can not be explained in terms of an extrapolation of the warm ism spectrum .
association with virgo cluster of galaxies suggests that the excess emission may be due to intracluster gas accreted in the gravitational well of a massive galaxy .
we investigate this and other possibilities .
galaxies : elliptical and lenticular , cd galaxies : irregular galaxies : spiral x - rays : galaxies x - rays : stars x - rays : ism
|
in anechoic environments , audio source separation with a microphone array can be achieved by beamforming and post - filtering @xcite . on the other hand , in reverberant environments , reverberation causes significant degradation in the quality and intelligibility of the beamformer output , especially if the room impulse responses ( rir ) are not taken into account or are not known precisely .
this is a serious limitation because in practice , rirs are usually not known with precision and they can vary wildly with respect to the positions of the microphone / source pair .
a related problem addressing reverberation without explicit rir knowledge is called blind dereverberation @xcite . a successful class of methods for blind dereverberation fall under the framework of multi - channel linear prediction ( mclp ) @xcite .
these methods require a microphone array and employ linear prediction to extract information about the properties of reverberation . in this paper
, we propose to employ mclp , which is primarily used for single source dereverberation , as a pre - processing step in source separation .
we show theoretically and demonstrate numerically that such a scheme not only dereverberates the sources but also helps in the subsequent separation . in order to make this statement more precise ,
let us briefly discuss an mclp formulation . consider an array of microphones , recording a source in a reverberant room .
recent mclp formulations @xcite select one of the microphones as the reference and use delayed versions of the whole set of observations to linearly predict the reference signal .
the residual of prediction , which may also be interpreted as the ` innovations ' component , forms the dereverberated estimate of the source . in our scenario
, we have multiple sources recorded with a microphone array and we would like to separate and dereverberate the sources . for this , suppose we apply mclp for each possible choice of the reference microphone .
if we have @xmath0 microphones in the array , this results in @xmath0 hopefully dereverberant but mixed observations .
we say ` hopefully dereverberant ' because the use of mclp has previously been considered and justified solely for the case of a single source @xcite .
further , even if the mclp outputs thus obtained are dereverberated , in order to achieve separation , we have to know the relation between the recordings . in this paper , we show that , under a mild assumption on the rirs , mclp actually suppresses reflections other than the direct source and preserves the phase of the direct source .
therefore , the mclp preprocessing step essentially reduces the reverberant separation problem to an anechoic one .
this greatly simplifies the separation step because we can now use the known array geometry to estimate the direction of arrival and perform separation using this estimate .
we argue that this effect is not dependent on the following doa estimation and separation algorithms . in that sense , the proposed mclp stage complements currently available source separation algorithms , improving their performance .
consequently , we do not have any reservations about which method to use following the mclp stage .
nevertheless , for the sake of completeness , we include a brief description of the geometric source separation ( gss ) algorithm @xcite followed by a post - filter .
literature on dereverberation / source separation with a microphone array is vast . here
, we focus on blind methods for dereverberation and source separation for arrays with a known geometry . for source separation ,
provided the directions of arrival ( doa ) of the sources are known , and reverberation is limited , beamformers like the minimum variance distortionless response ( mvdr ) beamformer , along with a post - filter @xcite can be effective . in mvdr ,
one aims to minimize the energy of the reconstructed signals , while preserving the signals from the doas .
however , reverberation causes a source to contribute to different directions , reducing the effectiveness of such approaches .
geometric source separation ( gss ) @xcite addresses this problem by also enforcing statistical independence of the reconstructed sources .
this is achieved by selecting the beamformer weights so as to minimize a cost function that consists of the sum of a term imposing geometric constraints based solely on the array geometry and a term that promotes the statistical independence of the sources ( motivated by blind source separation techniques @xcite ) . in practice ,
gss outputs are still reverberant for moderate to highly reverberant rooms .
gss has been complemented with a post - filter in @xcite to devise a high performance source - separation method that can operate in real time .
an alternative approach , based on the assumption that the activity of the different sources do not overlap in the time - frequency domain @xcite , has been proposed in @xcite .
however , the model in @xcite assumes that the room impulse responses are relatively short , compared to the analysis window size used in the stft . despite its simplicity , blind dereverberation is a challenging problem due mainly to the characteristics of room impulse responses . for a general overview and discussion of relatively recent methods ,
we refer to the edited book @xcite . here , we focus on a successful class of methods that can be collected under the heading ` multichannel linear prediction ' ( mclp ) , as they are more directly related to the content of our work .
these methods usually fit in a schema as follows .
given multiple observations @xmath1 , for @xmath2 , one of the observations is chosen as the reference .
then , this reference observation is linearly predicted by the delayed version of all of the observations .
once the linear prediction coefficients are determined , this information has been used differently in prior work .
for instance , in @xcite , the authors use the coefficients to estimate the power of late reverberations , which are subsequently reduced by spectral subtraction . in more recent work
@xcite , the residual of linear prediction is taken as the dereverbed audio signal .
prior work also differentiates based on the criteria used for determining the weights used in linear estimation . in @xcite , the authors propose to select the linear estimation weights so as to minimize the energy of the residual . in @xcite ,
the source signal is modeled as a time - varying gaussian process and the weights are selected so as to maximize the likelihood of such a model .
more recently , @xcite employ a sparsity- based model in the time - frequency domain and select the weights so that the residual minimizes a sparsity promoting cost function . an interesting approach that employs mclp to achieve suppression of late reverberation and source separation
is presented in @xcite .
the authors use mclp to suppress late reverberation , using a statistical framework similar to that in @xcite .
in contrast to @xcite , a time - varying scenario with multiple sources is considered and the mclp weights are determined along with the statistics of the late reverberant part of the sources in an iterative manner .
our proposal in this paper is of a complementary nature and is likely to improve the performance of existing separation algorithms .
we present an inventive way to employ mclp as a pre - processing step for blind source separation using a microphone array .
we show theoretically that , under certain assumptions , such a pre - processing step converts the reverberant source separation problem into a non - reverberant one , which in turn can be solved by existing methods such as @xcite outlined above .
such an approach is inherently different than previous uses of mclp which aim to dereverb a single source as in @xcite .
our proposal also differs from that of @xcite because @xcite requires the source statistics to cancel the late reverberation and thus resorts to a joint estimation of the linear prediction weights and source statistics .
this approach in @xcite requires an iterative procedure and is handled with an em algorithm .
in contrast , we propose to employ the mclp step once , in order to reduce the reverberant problem to an approximately non - reverberant one .
in addition to theoretical justification , we demonstrate the validity of the claims using real recordings . for the sake of completeness
, we include a description of gss along with a simple post - filter , to be applied following the mclp step , in order to achieve separation .
however , after mclp , other methods that assume short rirs , or operate under a free propagation model may also be used , as noted above .
sequences and continuous functions of time are denoted with a small letter as in @xmath3 or @xmath4 , respectively .
the dft and stft of sequences are denoted with a capital letter as @xmath5 and @xmath6 respectively which one is referred to will be clear from the context .
the @xmath7-transform of sequences are denoted with calligraphic letters as in @xmath8 .
we refer to @xmath8 as ` causal ' if it is a polynomial of @xmath9 . for a causal @xmath8 ,
we denote the constant @xmath10 also as @xmath11 . throughout the paper
, @xmath12 denotes the number of sources and @xmath0 denotes the number of microphones in the array .
throughout the paper , we are interested in far - field processing .
thus , the distance of the source to the microphones is assumed to be significantly greater than the array size .
our proposed formulation / algorithm is in the stft domain .
however , for simplicity of exposition , we first introduce a model in the continuous - time domain and then present an approximation of this model in the stft domain .
we introduce the time - domain model in three stages , where complexity increases with each stage .
we start with the case of a single source , single microphone . in the second stage , we consider the single source , multiple microphone case . finally , we discuss the multiple source , multiple microphone case , which is the main scenario of interest in this paper . for a given microphone - source pair ,
let @xmath13 denote the signal produced by the source .
the observation @xmath14 can be approximated as , @xmath15 where @xmath16 is the room impulse response ( rir ) for the given source - microphone positions .
now suppose that the distance from the source to the microphone is @xmath17 meters and the speed of sound in the environment is @xmath18 m / sec .
then , the acoustic signal emitted by the source reaches the microphone after a delay of @xmath19 seconds .
we declare the _ distance compensated room impulse response _ ( dcir ) to be @xmath20 notice that for practical purposes , we can work with the dcir instead of the rir for a single microphone - source setup . for multiple microphones or sources ,
we need to take into account relative distances , as discussed next .
consider a setup consisting of a microphone array and a source as shown in fig .
[ fig : setup ] .
suppose we set the center of the array as the origin for a coordinate system .
for this origin , let the vector @xmath21 denote the position of the the @xmath22 microphone .
also , let @xmath23 denote the unit vector in the direction of the source .
finally , let @xmath24 .
note that @xmath25 denotes the coordinate of the projection of the @xmath22 microphone s position to the subspace spanned by @xmath23 .
the induced time - delay is @xmath26 .
notice that @xmath27 may be positive or negative , i.e. , may represent an advance or delay .
finally , let the dcir from the source to the @xmath22 microphone be denoted as @xmath28 .
then , we model the observation at the @xmath22 microphone , namely @xmath1 , as , @xmath29 the far field assumption amounts to the following : if the distance of the source to the center of the array is @xmath17 , then the distance of the source to the @xmath22 microphone is approximately @xmath30 . under the far - field assumption
, the described model preserves the ability of the array to differentiate different directions .
we next add another layer to this model by considering multiple sources .
suppose now that there are @xmath12 sources and @xmath0 microphones .
let the @xmath31 source signal be denoted by @xmath32 . also , let the unit vector in the direction of the @xmath31 source be denoted by @xmath33 . for the @xmath22 microphone , positioned at @xmath21 , we set @xmath34
. finally , let @xmath35 denote the dcir between the @xmath31 source and the @xmath22 microphone .
then , we model the observation at the @xmath22 microphone as , @xmath36 notice that , the absolute delay from the @xmath31 source is not included in the model
. however , this is not a careless omission , because this delay is in fact not known and is not estimated in practice .
therefore , the described model is suitable for practical far - field processing scenarios . , and @xmath37 respectively .
the center of the array , marked with a circle , is taken to be the origin . the position vector for the @xmath38 microphone
is projected onto the space spanned by the unit vector pointing in the direction of the source to obtain the relative coordinate @xmath25 .
note that @xmath39 , @xmath40 ( not shown ) are positive whereas @xmath41 , @xmath42 ( not shown ) are negative .
[ fig : setup ] ] we work in the stft domain for reasons to be briefly clarified in the sequel .
we are specifically interested in the representation of an lti system in the stft domain .
consider a convolution equation as , @xmath43 let @xmath44 , @xmath45 denote the stfts of @xmath46 and @xmath37 respectively , obtained using a window @xmath47 .
we differentiate between two cases , following the discussion in @xcite ( see lemma 1 in @xcite ) .
a. [ item : contained ] if the support of @xmath48 is contained in the support of @xmath49 , then @xmath50 where @xmath5 denotes the dft of @xmath48 . b. if the support of @xmath48 is significantly larger than the support of @xmath49 , then @xmath51 for some @xmath52 , where @xmath53 denotes convolution with respect to the time variable @xmath54 , defined as @xmath55 we remark that @xmath52 is not necessarily the stft of @xmath48 , but is chosen so as to make the approximation as close as possible
see @xcite for a discussion .
we remark that , as a special case of , if @xmath56 is small enough , then the stft of @xmath57 can be approximated as @xmath58 .
therefore , the model introduced in can be approximated in the stft domain as @xmath59 such a model is preferred over the time domain model because prior or penalty functions that enforce ( structured ) sparsity of the audio signals are typically defined in the stft ( or some other time - frequency ) domain . if we use the time domain convolutional model in an iterative algorithm , we have to work in the time domain to undo the effects of reverberation , and the stft domain to take advantage of the signal prior . switching from the time domain representation to the stft representation at each iteration increases the computational cost significantly . working with a convolutional model directly in the stft domain , such as , allows to carry out the iterations in the stft domain , eliminating the need to apply the forward and inverse stft operations at each iteration .
in this section , we discuss the implications of using multi - channel linear prediction ( mclp ) as a preliminary step for the source separation problem . we derive conditions under which mclp transforms the original blind separation / dereverberation problem to a non - blind one .
we start the discussion with a single source and then consider extensions to multiple sources .
consider a signal @xmath37 that is observed via different channels as @xmath60 where @xmath61 denotes the impulse response of the @xmath22 channel . in the @xmath7-domain , this can be written as , @xmath62 we would like to find causal filters @xmath63 ,
, @xmath64 such that @xmath65 mint @xcite ensures that such filters exist provided that @xmath66 s are co - prime , i.e. , they do not have a common non - constant factor . even though it is relieving to know that such filters exist
, finding the filters is still a challenging problem when @xmath61 s are not known .
mclp addresses this challenge . in mclp ,
the idea is to estimate @xmath37 by subtracting from @xmath67 ( for a chosen reference index @xmath68 ) delayed weighted combinations of @xmath1 s .
this is equivalent to using in a set of filters of the form @xmath69 for a positive integer @xmath70 ( determining the amount of delay ) and causal @xmath71 s . in @xcite ,
the selection of the filters is guided by the assumption that the original @xmath37 is sparse .
more specifically , suppose @xmath72 denotes the 1d time - series derived from a frequency band of the observations , that is @xmath73 for some @xmath74 .
the filters @xmath75 are found by solving the following minimization problem .
@xmath76 where @xmath77 is a sparsity promoting function
. we will have more to say about @xmath78 , but first we would like to stress a point about the choice .
mint ensures the existence of dereverberating filters but does not guarantee that they can be of the form .
below , we propose modified conditions which ensure the existence of such @xmath79 s .
another issue is the role of the reference index @xmath68 .
it turns out that repeating the linear prediction for different @xmath68 values leads to an interesting outcome , summarized in prop .
[ prop : observations ] below . in order to derive prop .
[ prop : observations ] , we need an auxiliary result ( prop . [ prop : modmint ] ) and an assumption . [ prop : modmint ] suppose that the @xmath80 polynomials @xmath81 , @xmath82 , are co - prime . then , for @xmath83 , there exist a set of filters of the form such that @xmath84 for a constant @xmath18 .
further , for a set of causal filters of the form , if holds , then @xmath85 .
see appendix [ app : modmint ] .
consider now the application of this result to the model .
assume that there is a single source , i.e. , @xmath86 .
for a fixed frequency @xmath74 , let @xmath87 under the hypotheses of prop .
[ prop : modmint ] , if , for @xmath88 , mclp can recover a set of filters satisfying , then prop .
[ prop : modmint ] implies that , @xmath89 now , if further , @xmath90 then , @xmath91 for some constant @xmath92 , independent of @xmath68 ( actually , @xmath93 , for any @xmath94 ) .
we remark that the difference of and is that the rhs of depends only on the direction of arrival , whereas this is not the case in . in other words
, mclp eliminates reflections from different directions .
the assumption in states that , around the origin , distance compensated impulse responses are similar .
this assumption is key to the development in this paper and we state it explicitly for later reference .
we will verify this assumption experimentally in section [ sec : exp ] .
[ assm : main ] for a given source - microphone array pair , let @xmath95 denote the stft representation of the distance compensated impulse response for the @xmath22 microphone . then , @xmath96 for @xmath97 and all @xmath74 .
we state the foregoing observations as a proposition . [
prop : observations ] for a fixed frequency @xmath74 , let @xmath98 . assume also that the collection of polynomials @xmath81 , @xmath82 are co - prime .
then , for @xmath83 , there exist a set of filters of the form such that @xmath99 for a constant @xmath100 .
further , for a set of filters of the form , if holds , and assumption [ assm : main ] is in effect , then @xmath101 for a constant @xmath92 , independent of @xmath68 . next , we extend the discussion to the case of multiple sources .
consider an observation model with @xmath12 sources @xmath102, @xmath103 and @xmath104 observations defined as , @xmath105 in the @xmath7-domain , this can be written as , @xmath106 for this model , mint also ensures that we can find filters @xmath107 , for @xmath108 , @xmath109 such that @xmath110 provided that the ( smith ) canonical form of @xmath111 ( see appendix [ app : smith ] or @xcite , sec.vi.2 , defn.3 ) has a non - zero constant ( i.e. , no polynomial higher than zero order ) diagonal .
however , unlike the single source case , we can not use filters of the form to achieve .
we have the following extension in this case .
[ prop : mclp2 ] for a fixed reference microphone index @xmath112 , let @xmath113 be the matrix obtained by removing the @xmath114 row of @xmath111 . if the greatest common divisor of all the @xmath12-minors of @xmath113 is a non - zero constant , then , for @xmath83 , there exist a set of filters of the form such that @xmath115 for a constant vector @xmath116 ( i.e. , containing no polynomial of order greater than zero ) .
further , for a set of filters of the form , if holds , then the components of @xmath18 satisfy , @xmath117 see appendix [ app : mclp2 ] . [ rem : equivalence ] for an @xmath118 matrix @xmath111 with @xmath119 , it can be shown that , the greatest common divisor of all the @xmath12-minors of @xmath111 is a non - zero constant if and only if the smith canonical form of @xmath111 has a non - zero constant diagonal .
this equivalence is shown in prop . [
prop : equivalence ] in appendix [ app : smith ] .
therefore , we can replace the hypothesis in prop . [
prop : mclp2 ] with an alternative , stated in terms of the smith canonical form of @xmath113 .
when we apply this result to our observation model with @xmath12 sources , we obtain the following proposition . [ prop : main ] for a fixed frequency @xmath74 , let @xmath120 for @xmath121 , @xmath122 .
also , let @xmath111 be defined as in , and @xmath113 be defined as in prop .
[ prop : mclp2 ] .
finally , let @xmath73 .
if the greatest common divisor of all the @xmath12-minors of @xmath113 is a non - zero constant , then , for @xmath83 , there exist a set of filters of the form such that @xmath123 for some constants @xmath124 .
further , for a set of filters of the form , if holds , and assumption [ assm : main ] is in effect , then we have @xmath125 where @xmath92 is a constant independent of @xmath68 .
see appendix [ app : main ] .
this proposition suggests that , in a scenario with multiple sources , mclp , if successful , suppresses reflections and preserves the direct signal even when there exist multiple sources from different directions .
therefore , after the mclp step , the recordings appear as if they are recorded with the same microphone array , but in a non - reverberant environment .
this in turn allows to read off the directions of the sources with a simple procedure like music @xcite .
once the source directions are determined , we can then achieve separation by using _ known _ manifold vectors of the microphone array .
the discussion after prop .
[ prop : main ] sets the stage for the proposed framework .
the main building blocks of the framework are shown in fig .
[ fig : framework ] .
we first input the @xmath0 observations @xmath126 , recorded in a reverberant environment with unknown reverberation , to an mclp block .
this block applies mclp for every possible choice of the reference microphone .
the output of this block is @xmath0 dereverberated signals @xmath127 .
these signals are input to a doa estimation algorithm ( such as music @xcite ) and the directions of the @xmath12 sources are determined . then these estimated directions are used to separate the sources in @xmath128 using a non - blind source separation method . the estimated sources are denoted as @xmath129 in fig . [ fig : framework ] . in the following ,
we briefly present the algorithm we used for a practical implementation of the components .
however , we remark that following mclp , the problem turns into a non - reverberant source separation problem ( with known array geometry ) and one can use any method of choice for this task . ,
@xmath130 , which are @xmath0 reverberant observations of @xmath12 sources .
source directions are estimated under the assumption of known array geometry .
the estimated source directions are used in a non - blind source - separation algorithm .
[ fig : framework ] ] we use the formulation given in sec .
[ sec : multisourcemclp ] .
we work in the stft domain and process the frequency channels separately . for a fixed value of @xmath74 , we set @xmath73
. then for each value of @xmath112 , the dereverbed estimate @xmath131 is defined as , @xmath132 where @xmath133 denote filters determined by solving @xmath134 we use a weighted @xmath135 norm as the sparsity promoting function @xmath77 .
more precisely , for @xmath136 , we set @xmath137 where @xmath138 is a given weight vector .
we update the weight vector @xmath49 as the iterations progress , to improve the sparsity promoting property of @xmath78 .
we remark that for mclp , a similar formulation has been discussed in @xcite , along with an algorithm based on the alternating direction method of multipliers ( admm ) algorithm @xcite .
we employ the douglas - rachford ( dr ) algorithm @xcite for solving this problem . to that end
, we first define the matrices @xmath139 for @xmath140 such that @xmath141 and as @xmath54 ranges over the domain of the observation @xmath142 .
notice that @xmath139 is toeplitz and has @xmath143 columns and as many rows as the number of samples in @xmath142 . by concatenating @xmath139 s , we construct the data matrix @xmath144 as , @xmath145 similarly , we construct the variable vector @xmath146 as @xmath147 where @xmath148 denotes the column vector with @xmath149 . also , let @xmath150 denote the column vector with @xmath151 .
using these , we can rewrite the problem as , @xmath152 we split variables and introduce @xmath7 in order to make the formulation feasible for employing dr . for this formulation ,
the douglas - rachford algorithm is of the form outlined in algorithm [ algo : dr ] .
we remark that in this algorithm , at each @xmath153 iterations , the weights @xmath49 used to compute @xmath154 are updated , so as to improve the sparsity inducing property of the penalty function , following @xcite .
we refer to @xcite for more intricate update strategies .
input data @xmath150 , @xmath144 , a constant for weight updates @xmath155 initialize @xmath156 for all @xmath94 update @xmath49 every @xmath153 iterations @xmath157
@xmath158 @xmath159 @xmath160 @xmath161 @xmath162 for all @xmath94 @xmath163 because audio is a wideband signal , narrowband doa estimation techniques like music @xcite , or esprit @xcite are not directly applicable
. however , simple extensions of music to wideband scenarios exist and we employ one of them . in a nutshell , the idea is to estimate the doa for each frequency separately and average the estimates .
this approach is referred to as the ` incoherent signal subspace method ' ( issm ) in the recent literature
@xcite see @xcite for early work . despite its simplicity
, issm is noted to perform well in high snr @xcite .
subspace methods for doa estimation make use of two essential components , namely the covariance matrix of the data and the manifold vector , determined by the array geometry . for a fixed frequency @xmath74 , and time instance @xmath54 ( in the stft domain )
, we estimate the correlation matrix @xmath164 as @xmath165 where @xmath166 in this definition , @xmath167 $ ] is a smoothing ( or forgetting ) factor justified by the assumption that the sources move slowly .
let us now define the manifold vector .
recall from sec .
[ sec : tdmodel ] that @xmath21 denotes the position of the @xmath22 microphone with respect to the center of the array .
also , let @xmath168 denote the unit vector in the direction @xmath169 .
finally , let @xmath18 denote the speed of sound , @xmath170 denote the sampling frequency and @xmath171 denote the relative delay ( in samples ) of the @xmath22 microphone for a source in the direction @xmath169 .
the manifold vector is defined as @xmath172 assuming there are @xmath12 sources , let @xmath173 denote the @xmath174 unitary matrix whose columns are the eigenvectors of @xmath164 corresponding to the smallest @xmath175 eigenvalues .
the doa function is defined as @xmath176 for each time instance @xmath54 , the @xmath12 dominant peaks of @xmath177 with respect to @xmath169 are taken to be the directions of the sources .
for source separation , we use the geometric source separation ( gss ) method @xcite complemented with a simple post - filter .
we assume that the sources are stationary , so that the estimated directions are constant with respect to time .
extension to moving sources is straightforward , at least in principle , and will not be further discussed .
let @xmath178 , for @xmath179 denote the directions of the sources determined by the method described in sec .
[ sec : doa ] . for each @xmath74
, we define a matrix @xmath180 as , @xmath181 we model the mclp outputs as @xmath182 this model is not exact because despite the significant dereverberation by mclp , @xmath183 still contains reflections other than the direct signal . for this model , gss proposes to estimate @xmath45 as @xmath184 , where the @xmath185 weight matrix @xmath186 is determined by solving a minimization problem . to describe the problem ,
let us denote @xmath187 for a given weight matrix @xmath188 .
also , let @xmath189 denote the empirical autocorrelation matrix of @xmath190 , obtained by averaging @xmath191 over the time variable @xmath54 . given this notation , @xmath186
is defined as the solution of @xmath192 where @xmath193 denotes the frobenius norm .
thanks to the differentiability of both terms in the cost function , gradient descent can be applied for obtaining a stationary point . for a fixed @xmath74
, we remark that the gradient @xcite of the complex valued @xmath194 is given as @xcite , @xmath195 where @xmath196 denotes the empirical autocorrelation matrix of @xmath183 , obtained by averaging @xmath197 over the time variable @xmath54 .
the gss algorithm aims to achieve source separation by considering the best linear combination of the inputs . in practice
, separation performance can be improved by post - filtering .
specifically , we view the separated signals as contaminated with noise and eliminate this noise using a soft - threshold , where the time - frequency varying threshold value depends on an estimate of the remaining noise variance for each time - frequency sample .
more specifically , if @xmath198 , , @xmath199 denotes the @xmath12 source estimates obtained by gss , for each value of the triplet @xmath200 , we regard @xmath201 as an unbiased estimate of @xmath202 with variance @xmath203 .
the post - filtered estimate is formed as , @xmath204 where @xmath205 is a parameter . in order to realize such a post - filter
, we need to estimate @xmath203 .
we estimate @xmath203 as , @xmath206 where @xmath207 denotes the @xmath208 entry of @xmath209 .
while this expression may make intuitive sense , further justification is provided in appendix [ app : pf ] .
we also remark that this is a generalized version of the post - filter employed in @xcite .
our primary claim in this paper is that an mclp step suppresses reflections other than the direct source , under a certain assumption ( assumption [ assm : main ] ) .
the experiments in this section are geared towards demonstrating the claims and verifying assumption [ assm : main ] .
there are three experiments . in the first ,
we verify assumption [ assm : main ] using real impulse responses obtained in a classroom . in the second experiment ,
we demonstrate that an mclp preprocessing step eliminates reflections other than the direct source , by showing how the doa function changes after the mclp step .
finally , the third experiment validates our claim about the effect of employing an mclp preprocessing step in a source separation problem .
all of the experiments are performed on real data . in order to obtain the data
, we used a circular microphone array with eight omni - directional microphones .
the configuration and the definition of doa , namely @xmath169 , are as shown in fig .
[ fig : micsetup ] .
the microphones are distributed uniformly on a circle of radius 5 cm . in the experiments
, we use an stft with a smooth window of length 64 ms , and hop - size is 16 ms . the window is selected such that the stft is self - inverting ( i.e. , a parseval frame ) .
the sampling frequency for the audio signals is 16 khz . for the mclp block ,
the parameters ( see algorithm [ algo : dr ] ) are chosen as @xmath210 , @xmath211 , @xmath212 , @xmath213 , @xmath214 .
we run the algorithm for 2000 iterations .
some of the audio signals used in the experiments , as well as those from experiments not contained in the paper can be listened to at `` http://web.itu.edu.tr/ibayram/mclp/ '' .
denotes the angle of arrival .
in particular , @xmath215 radians if the source is in the direction of @xmath216 , @xmath217 radians if the source is in the direction of @xmath218 , etc .
[ fig : micsetup ] ] [ exp : assm ] in order to verify assumption [ assm : main ] , we used estimated impulse responses from different doas .
the impulse responses are estimated using the method described in @xcite , using loudspeakers placed at known doas in the far field . knowing the doas
, we computed the distance compensated impulse responses and obtained filters @xmath95 that approximate their action in the stft domain . for each @xmath54 and @xmath74 value , we computed the variance of @xmath95 s with respect to @xmath94 ( recall @xmath94 takes eight different values ) , which we denote as @xmath219 .
we normalize @xmath219 by the total power at the location @xmath54 , @xmath74 , defined by @xmath220 .
thus for each @xmath54 and @xmath74 value , we obtained a number @xmath221 .
we remark that assumption [ assm : main ] is valid if and only if @xmath222 for all @xmath74 . in fig .
[ fig : corr ] , we show ( in decibels ) the fields @xmath223 for different doas , namely @xmath224 , @xmath225 and @xmath226 radians .
observe that for all doas and almost all @xmath74 shown , @xmath227 ( the first column in each image ) is significantly lower than @xmath223 for @xmath228 ( the remaining columns ) .
therefore , assumption [ assm : main ] is seen to be valid for practical purposes . : ` correlation figures ' ( see the text for a description ) , in db , for the circular array and different directions of arrival .
( a ) doa = 0 rad , ( b ) doa = @xmath225 rad , ( c ) doa = @xmath226 rad .
[ fig : corr ] ] [ exp : doa ] in this experiment , we demonstrate the effect of an mclp front - end on estimating the doa .
the source signal is produced with a loudspeaker located at @xmath217 radians in a reverberant environment .
the spectrogram of the signal observed at the first microphone is shown in fig .
[ fig : doaspect]a .
notice that the harmonics are seriously smeared . for this signal
, mclp significantly removes the effects of reverberation see fig .
[ fig : doaspect]b . : spectrograms of the signals used in the experiment .
( a ) the observation at the first microphone .
( b ) mclp output when the first microphone is used as the reference .
[ fig : doaspect ] ] in order to visualize the effects of mclp on doa estimation , we show the doa function @xmath229 ( defined in ) averaged over the time variable @xmath230 in fig .
[ fig : doa ] , for both the reverberant observations and mclp outputs .
we remark that averaging is meaningful here , since the source is stationary . for comparison
, we also normalize the averaged doa functions so that their peak is unity .
the correct position of the source is also indicated with a dotted line .
we see that the doa function is clearly improved by the application of mclp .
the variance is clearly reduced and the function is closer to an impulse located at the correct position . :
the doa functions @xmath229 ( see ) averaged over time ( i.e. , @xmath54 ) for the reverberant observations and mclp outputs .
we also apply a normalization to each doa function so that the peak is equal to unity.[fig : doa ] ] [ exp : ss ] in this experiment , we demonstrate the improvement gained by employing mclp as a pre - processing step in a blind separation setup .
we use the same microphone array as in exp .
[ exp : doa ] .
we use the source from exp .
[ exp : doa ] as the first source .
the second source is produced by a loudspeaker placed at @xmath231 radians ( so that the sources are separated by @xmath232 radians ) .
the spectrogram of the observation at the first microphone is shown in fig .
[ fig : spectmixed]a .
the output of mclp when the first microphone is taken as the reference is shown in fig .
[ fig : spectmixed]b . : ( a ) spectrogram of the two mixture signal observed at the first microphone .
( b ) mclp applied to the observations , when the first microphone is taken as the reference .
[ fig : spectmixed ] ] the average ( over time ) doa functions obtained from the original recordings and mclp outputs are shown in fig .
[ fig : mixmusic ] .
we observe that mclp has considerably sharpened the functions .
also , without mclp , the peaks of the doa functions are offset by some amount from the actual directions , which would lead to a biased estimate of the true doa .
: the doa functions @xmath229 ( see ) averaged over time ( i.e. , @xmath230 ) for the reverberant observations and mclp outputs .
we also apply a normalization to each doa function so that the peak is equal to unity.[fig : mixmusic ] ] the source separation results are shown in fig .
[ fig : spectseparate ] .
perceptually , without mclp , especially for the second source , we can clearly hear the first source , despite the post - filter .
further , reverberation is not eliminated .
these artifacts are eliminated using mclp . :
spectrograms of the sources separated with and without mclp .
( a , b ) source-1,2 obtained with gss and a post - filter , without mclp .
( c , d ) source-1,2 obtained when mclp is applied first .
[ fig : spectseparate ] ] in order to quantify the results , we also report pesq and sir scores @xcite .
pesq gives an objective perceptual distance between an input speech signal and a reference . in order to obtain meaningful distances
, we use the multiple channel recordings of source-1 and source-2 , prior to mixing , to extract the reference signals .
we take the @xmath22 microphone as reference and apply mclp to the multiple microphone recordings to obtain the references for each source .
we then compute the pesq scores for the methods described in this paper .
we finally average over @xmath94 to obtain a reliable pesq score .
the resulting scores are tabulated in table [ table : pesq1 ] .
we remark that the inclusion of an mclp step ( denoted as lp in the table ) improves the pesq scores noticeably .
the sir scores for separation with different combinations of the discussed methods are also provided in table [ table : pesq1 ] .
we observe that including an mclp preprocessing step also improves significantly the separation quality measured by sir .
.pesq and sir scores for exp .
[ exp : ss ] .
( lp : multichannel linear prediction , g : geometric source separation , p : post filter)[table : pesq1 ] [ cols="^ " , ]
in this paper , we advocated the use of mclp as a pre - processing step for blind reverberant source separation .
we provided theoretical justification for such a scheme and demonstrated numerically that it can significantly improve the performance of readily available separation algorithms .
the current algorithm for mclp is computationally intensive .
however , for a practical scenario with slowly moving sources , if we have approximate mclp weights at a certain time , we think it would be possible to update the weights in a computationally efficient manner by warm - starting the optimization procedure .
therefore , the proposed scheme has potential for real - time applications as well .
another issue involves the correlation between the weights obtained for different reference microphones . in the current formulation
, mclp is applied independently for each choice of the reference microphone .
however , we would expect some correlation between these weights , especially if the microphones are close with respect to the room dimensions , which is usually the case for a microphone array . we plan to explore the implications of such correlation in future work .
without loss of generality , suppose @xmath233 in the following . since @xmath234 , , @xmath235 are co - prime , we can find causal polynomials @xmath236 ,
, @xmath237 such that @xmath238 now set , @xmath239 and @xmath240 then for @xmath241 as in , and @xmath83 , we have @xmath242 for the converse , observe that for causal @xmath243 and @xmath244 , the polynomial @xmath245 does not have any constant terms , i.e. , @xmath246 .
observe also that @xmath247 .
thus if holds , then @xmath248
in this appendix , we briefly present , for the sake of completeness , the ( smith ) canonical form of a polynomial matrix , along with some of its properties . for a more detailed exposition
, we refer to the fine treatment in sec .
vi.13 of gantmacher s book @xcite .
we refer to a quantity or matrix as non - zero constant if it is a polynomial of order zero .
let us start by introducing three _ elementary row operations _ on a polynomial matrix @xmath249 : 1 .
multiplication of a row with a non - zero constant .
2 . addition , to the @xmath250 row , the product of the @xmath22 row and a polynomial @xmath251 .
interchange of two rows .
these operations can be performed by left multiplication with a square matrix whose determinant is a non - zero constant .
we refer to matrices that can be used to perform elementary operations as _ elementary matrices _ in this appendix .
we also remark that similar operations can be defined on the columns instead of rows .
elementary column operations can be performed by right - multiplication with elementary matrices .
we say that two polynomial matrices @xmath252 and @xmath253 are equivalent if @xmath252 can be transformed to @xmath253 by elementary row / column operations .
let us now introduce the ( smith ) canonical form .
[ def : smith ] an @xmath118 ( with @xmath254 ) diagonal polynomial matrix @xmath249 of rank @xmath68 with diagonal @xmath255 for @xmath256 is said to be in ( smith ) canonical form if 1 .
@xmath257 if @xmath258 , @xmath259 if @xmath260 , 2 .
@xmath261 for @xmath258 , 3 .
@xmath255 divides @xmath262 for @xmath263 .
a fundamental result is that we can transform any polynomial matrix to one in canonical form via elementary row / column operations .
each @xmath264 polynomial matrix is equivalent to a unique matrix in canonical form .
in general , the canonical form is reached by performing both row and column operations . however , we are interested in row operations only , for the development in this paper .
we have the following result .
[ prop : smith ] suppose an @xmath118 matrix @xmath265 has rank @xmath12 and the diagonal of its canonical form is constant .
then , @xmath265 can be reduced to its canonical form via row operations .
this result is a corollary of the discussion in @xcite but is not stated explicitly .
we provide a proof for the sake of completeness .
we need a few intermediate results for the proof .
all of the non - obvious claims stated before the proof of prop .
[ prop : smith ] can be found in @xcite .
given an @xmath118 polynomial matrix @xmath266 , let @xmath267 denote the greatest common divisor of all minors of @xmath265 of order @xmath68 , normalized such that the leading coefficient is unity .
notice that @xmath268 is divisible by @xmath269 .
therefore , for @xmath270 , the ratios @xmath271 are polynomials .
we also set @xmath272 .
it can be shown that @xmath268 s are invariant under elementary row / column operations .
this property transfers to @xmath273 as a consequence of its definition .
suppose now that the canonical form of @xmath266 consists of a diagonal matrix with elements @xmath274 as in defn .
[ def : smith ] .
it turns out that @xmath275 , giving us an alternative definition of @xmath276 .
we need a final result before proving prop .
[ prop : smith ] , which can be shown by a simple elimination argument .
[ prop : u ] using elementary row operations , any polynomial matrix @xmath277 can be reduced to an upper triangular matrix @xmath278 such that , at each column of @xmath278 , if the polynomial on the diagonal is non - zero , then the degree of any off - diagonal polynomial is strictly smaller than the degree of the polynomial on the diagonal .
we are now ready for the proof of prop .
[ prop : smith ] .
suppose the canonical form of @xmath265 is @xmath279 .
observe that the only non - zero minor of @xmath279 of order @xmath12 is given by the product of its diagonals , which is unity .
therefore , the invariant polynomial @xmath280 of @xmath265 is a constant .
suppose we reduce @xmath265 to an upper triangular form @xmath281 as in prop .
[ prop : u ] using elementary row operations only .
observe that the only non - zero minor of @xmath281 of order @xmath12 is given by the product of its diagonals .
let us denote this non - zero minor by @xmath282 .
but since this is the only non - zero minor of order @xmath12 , and elementary matrices have constant determinants , it follows that we must have @xmath283 for some constant @xmath18 .
thus , @xmath282 is a constant .
therefore , @xmath281 is a constant diagonal matrix with a non - zero diagonal .
each constant on the diagonal can be made unity by an elementary row operation , yielding the canonical form @xmath279 . in a final proposition
, we show the equivalence stated in remark [ rem : equivalence ] .
[ prop : equivalence ] suppose @xmath265 is an @xmath118 polynomial matrix with @xmath119 .
the greatest common divisor of all the @xmath12-minors of @xmath265 is a non - zero constant if and only if the smith canonical form of @xmath265 has a non - zero constant diagonal . as in the discussion above , let @xmath267 denote the greatest common divisor of all minors of @xmath265 of order @xmath68 , where the leading coefficient is unity , @xmath271 for @xmath270 , with @xmath272 . recall that @xmath273 s form the diagonal in the smith canonical form of @xmath266 .
suppose now that @xmath284 .
this implies that @xmath285 for all @xmath286 .
therefore the diagonal of the canonical form of @xmath265 is the identity matrix . for the converse , suppose the canonical form of @xmath265 has a constant diagonal
this means that @xmath287 s are constant for @xmath286 . but
this implies that @xmath288 for @xmath289 .
without loss of generality , suppose @xmath233 in the following . by remark [ rem : equivalence ] , which is proved in prop .
[ prop : equivalence ] , we first note that the gcd of the @xmath12-minors of @xmath290 is a non - zero constant if and only if the canonical form of @xmath290 has a non - zero constant diagonal . thus , by prop . [
prop : smith ] , we can find a @xmath291 polynomial matrix @xmath292 such that @xmath293 for @xmath294 , let us define the polynomials @xmath295 and set @xmath296 also , set @xmath297 . then for @xmath241 defined as in , we have @xmath298 for the converse , observe that , for causal @xmath243 and @xmath299 , the polynomial @xmath300 does not have any constant terms , regardless of @xmath230 , i.e. , @xmath301 for all @xmath230 .
observe also that @xmath247 .
thus if holds , then , @xmath302
we remark that follows by expressing in the time - domain .
let us consider .
if holds , and @xmath303 s satisfy , then by , we have @xmath304 by assumption [ assm : main ] and , it follows that @xmath305 where @xmath92 is independent of @xmath68 and @xmath230 . using , and the properties of convolution
, we finally obtain @xmath306
in order to justify the variance expression in , we will make use of prop .
[ prop : pf ] below . for simplicity , this proposition uses a seemingly independent notation .
we will identify the elements in the proposition with our separation scenario right after the proposition . recall that a complex valued gaussian random variable @xmath307 is said to be circular @xcite and have variance @xmath308 if its real and imaginary parts are independent and gaussian random variables with variance @xmath309 .
[ prop : pf ] suppose @xmath310 is an unknown constant and we observe @xmath311 , for @xmath312 , where @xmath313 , @xmath314 and @xmath315 s are circularly symmetric , independent , unit variance , zero - mean gaussian random variables .
the uniformly minimum variance unbiased ( umvu ) estimator for @xmath310 is of the form @xmath316 where @xmath317 with @xmath318 we also have that , @xmath319 is an unbiased estimator for the variance of @xmath320 .
it can be shown , by taking into account the gaussian pdf of @xmath321 , that a complete sufficient statistic for @xmath310 is @xmath322 observing that @xmath323 , it follows by the rao - blackwell theorem that the umvue is as given in for @xmath324 as in and @xmath325 as in .
it can also be checked that the variance of @xmath320 is , @xmath326 for the @xmath22 source , if we regard all of the remaining sources as noise terms with a gaussian distribution , then the optimal estimate of the @xmath22 source is obtained by a linear combination of the observations .
we assume that gss approximately finds these optimal weights . in that case
, the expression in gives an estimate of the variance of these unwanted components ( or , ` noise ' ) that remain after gss .
p. l. combettes and j .- c .
proximal splitting methods in signal processing . in h.
h. bauschke , r. s. burachik , p. l. combettes , v. elser , d. r. luke , and h. wolkowicz , editors , _ fixed - point algorithms for inverse problems in science and engineering_. springer , new york , 2011 .
j. m. valin , j. rouat , and f. michaud .
enhanced robot audition based on microphone array source separation with post - filter . in _
2004 ieee / rsj international conference on intelligent robots and systems ( iros ) ( ieee cat .
no.04ch37566 ) _ , volume 3 , pages 21232128 vol.3 , sept 2004 .
valin , j. rouat , and f. michaud .
microphone array post - filter for separation of simultaneous non - stationary sources . in _ proc .
conf . on acoustics , speech and signal proc .
( icassp ) _ , 2004 .
|
a class of methods based on multichannel linear prediction ( mclp ) can achieve effective blind _
dereverberation _ of a source , when the source is observed with a microphone array .
we propose an inventive use of mclp as a pre - processing step for blind _ source separation _ with a microphone array .
we show theoretically that , under certain assumptions , such pre - processing reduces the original blind reverberant source separation problem to a non - reverberant one , which in turn can be effectively tackled using existing methods .
we demonstrate our claims using real recordings obtained with an eight - microphone circular array in reverberant environments .
microphone array , source separation , blind dereverberation , multichannel linear prediction , beamforming , post - filtering .
|
it has recently been suggested @xcite that fermionic alkaline - earth atoms @xcite can be used to efficiently simulate condensed - matter systems with spin and orbital degrees of freedom .
one of those many - body systems is the one - dimensional kondo - lattice model @xcite with its intricate interplay between kondo screening and magnetic order @xcite .
particularly , the regime of strong antiferromagnetic local exchange coupling @xmath0 is accessible to the experiment @xcite . the purpose of the present paper is to demonstrate that quantum confinement effects resulting from strong @xmath0 in a _ spin - diluted _ system can effectively result in a _
weak _ indirect magnetic rkky @xcite interaction .
this is achieved by exploiting the characteristics of a novel `` inverse '' indirect magnetic exchange ( iime ) mechanism that has been proposed recently @xcite .
we study the distance dependence of the effective iime coupling in a one - dimensional prototypical two - impurity model by means of strong - coupling perturbation theory and density - matrix renormalization group ( dmrg ) .
we argue that the spin - diluted kondo lattice opens a new field where the complex , cooperative as well as competitive interplay between the kondo effect @xcite and different kinds of indirect magnetic exchange mechanisms can be studied in quantum - confined geometries .
the conventional rkky exchange between two impurity spins is mediated by itinerant conduction electrons and leads to an effective , indirect magnetic coupling @xmath1 .
this coupling is obtained from second - order perturbation theory in the local exchange interaction @xmath0 between the local spins and the conduction electrons .
it is oscillatory in the distance @xmath2 of the spins , @xmath3 , for a non - interacting one - dimensional metallic host system given by a tight - binding model with nearest - neighbor hopping @xmath4 at half - filling . on the other hand ,
if @xmath0 is much larger than @xmath4 , a strong - coupling variant of rkky exchange @xcite can be derived perturbatively in powers of @xmath5 . if @xmath0 is antiferromagnetic , rkky exchange typically competes with the emergence of the kondo effect @xcite which is responsible for the individual screening of impurity spins by the conduction electrons .
the kondo temperature @xmath6 is the corresponding energy scale of the crossover to the screened regime and can be converted by the fermi velocity @xmath7 into a length scale @xmath8 which may be interpreted as the extension of a kondo screening cloud .
the weak - coupling regime is dominated by rkky exchange , while the kondo regime is realized at strong couplings @xmath0 . here , we consider the competition between kondo effect and rkky exchange in _ confined _ systems with open boundaries .
i.e. , rather than discussing effects of the particular form of a trapping potential , infinitely large potential barriers are assumed for simplicity .
the spectral gap @xmath9 of the confined , non - interacting host system near the fermi edge acts as a cutoff for the characteristic kondo correlations@xcite in case of @xmath10 .
this also considerably affects the competition between the rkky exchange and the kondo effect @xcite and can lead to exotic ground states and even to a reappearance of a kondo regime for @xmath11 .
unconventional finite - size effects are also found in the strong - coupling limit .
although the kondo effect quenches the impurity spins for @xmath12 , it has been realized that it may also _ help _ to generate magnetism in the case of systems with diluted impurity spins @xcite , or , more generally , diluted correlated impurity sites @xcite .
namely , almost local kondo singlets can result in an efficient additional confinement of the itinerant electrons .
this implies the formation of local moments in the otherwise non - interacting conduction electron system .
moreover , these local moments are coupled magnetically by virtual excitations of the magnetically inert kondo singlets .
this constitutes an `` inverse '' indirect magnetic exchange ( iime ) @xcite where the roles of the conduction electrons and of the impurity spins are essentially interchanged .
the strong - coupling kondo physics is necessarily local or almost local and induces , via the iime mechanism , a short - range coupling between local conduction - electron magnetic moments .
as is shown here , the iime mechanism also applies to situations where the additional confinement extends over larger regions .
an interesting question is if this can be employed to generate a situation where effective rkky - like interactions couple magnetic moments over larger distances : can one make use of the strong - coupling limit and of a strong and almost local kondo effect to implement an effective weak - coupling model free of any competing kondo screening ?
how does the effective magnetic coupling depend on the distance between the moments and on the bare coupling @xmath0 ?
these questions are actually part of a more general approach to understand the complex interplay between magnetic exchange interactions and the kondo effect in quantum - confined systems .
and @xmath13 at a distance @xmath14 couple via a strong antiferromagnetic exchange @xmath0 to the local spins @xmath15 and @xmath16 at sites @xmath17 and @xmath18 of a one - dimensional chain of @xmath19 sites .
@xmath4 is the nearest - neighbor hopping in the half - filled system of conduction electrons .
b ) effective model for @xmath12 : the impurity spins form local kondo singlets with @xmath15 and @xmath16 .
this leads to local - moment formation at the edge sites @xmath20 and @xmath21 .
an `` inverse '' indirect magnetic exchange with coupling constant @xmath22 emerges and mediates a ferromagnetic coupling of the edge spins @xmath23 and @xmath24 to the rest of the conduction - electron sea .
c ) effective rkky and effective central - spin model @xcite : for even @xmath19 , i.e. for odd distance @xmath25 between the edge spins in the effective model , the low - energy model is a two - spin antiferromagnetic rkky model with rkky exchange @xmath26 . for odd @xmath19 ( @xmath27 even ) ,
the low - energy model is a central - spin model where @xmath28 and @xmath29 couple ferromagnetically with strength @xmath22 to the ( delocalized ) spin @xmath30 of the singly occupied fermi orbital @xmath31 .
d ) low - energy spectrum : for even @xmath19 ( odd @xmath27 ) , the ground state is a spin singlet , @xmath32 is the singlet - triplet excitation energy . for odd
@xmath19 ( even @xmath27 ) , the ground state is a spin quartet . for
non - interacting conduction electrons this quartet is degenerate with a doublet .
, scaledwidth=30.0% ] to study those questions , we consider a one - dimensional prototype model as depicted in fig .
[ fig : model]a
. two impurity spins @xmath33 and @xmath13 ( with quantum number @xmath34 ) are strongly coupled locally via an antiferromagnetic exchange @xmath12 to the local conduction - electron spins @xmath35 and @xmath36 of a system of @xmath37 itinerant and non - interacting conduction electrons .
the conduction - electron system is half - filled , i.e. @xmath38 where @xmath19 is the number of lattice sites .
the two resulting local kondo singlets at strong @xmath0 are located at a distance @xmath14 , thereby defining an intermediate `` central region '' of sites @xmath39 .
the hopping of the conduction electrons is @xmath40 between non - degenerate orbitals on nearest - neighboring sites @xmath41 of the lattice . throughout the paper ,
all energies are given in units of the nearest - neighbor hopping , i.e. , @xmath42 fixes the energy scale .
the hamiltonian is @xmath43 here , @xmath44 annihilates an electron at site @xmath45 with spin projection @xmath46 , and @xmath47 is the local conduction - electron spin at @xmath48 , where @xmath49 is the vector of pauli matrices and @xmath50 .
let us first concentrate on even lattice sizes @xmath51 .
systems with odd @xmath19 are discussed in sec .
[ sec : ferro_dist ] .
we also focus on the weak - coupling limit first . for @xmath11 ,
the ground state is a total spin singlet ( @xmath52 ) . here
, the kondo correlations are cut by the finite system size@xcite , i.e. , @xmath10 , since the kondo temperature is exponentially small , @xmath53 .
the chemical potential @xmath54 falls into the finite - size gap @xmath9 between fully occupied and unoccupied energy levels of the non - interacting conduction - electron system .
consequently , the impurity spins @xmath55 and @xmath56 are effectively decoupled from the host system at low energies .
using perturbation theory in @xmath0 , one then obtains an rkky hamiltonian @xmath57 featuring an antiferromagnetic coupling @xmath58 , where @xmath59 is the nonlocal static susceptibility of the non - interacting conduction - electron system .
one finds that @xmath60 for a one - dimensional system , where @xmath14 is the distance between the impurities
. however , in the strong - coupling regime @xmath61 , eq . can not be the effective low - energy hamiltonian anymore , since the impurity spins are compensated by individual kondo effects and @xmath62 .
perturbation theory , however , allows to study the remaining inter - impurity ground - state correlations for @xmath12 , yielding the envelope function @xcite @xmath63 this result can be regarded as a strong - coupling variant of rkky theory .
furthermore , it can be interpreted as a signature of the fact that each impurity is located in the exterior of the kondo cloud of the other impurity , as the low - energy physics of the single - impurity kondo model is essentially the one of a fermi liquid with @xcite @xmath64 for @xmath65 .
one may explicitly derive the effective low - energy hamiltonian by using strong - coupling degenerate perturbation theory , where hopping to and off a local kondo singlet is treated as a weak perturbation .
the resulting effective iime hamiltonian excludes the ( high - energy ) local kondo singlets at @xmath66 and @xmath67 and operates on the remaining conduction - electron degrees of freedom only @xcite .
up to fourth order in @xmath5 , for impurity distances @xmath68 , and disregarding unimportant constant energy shifts , one obtains @xmath69 , \nonumber \\\end{aligned}\ ] ] where @xmath70 is the iime coupling constant and @xmath71 the tight - binding hamiltonian of the remaining conduction - electron sea of the central region .
furthermore , @xmath72 is the particle number of electrons with spin projection opposite to @xmath73 , and @xmath74 denotes the isospin @xcite at site @xmath48 , defined as @xmath75 .
together with the su(2 ) spin symmetry , the isospin su(2 ) symmetry constitutes the so(4 ) symmetry of the effective hamiltonian at half - filling .
analyzing @xmath76 ( eq . )
, one may imagine two different scenarios for the low - energy physics .
on the one hand , we may expect the formation of non - magnetic edge isospins or charge fluctuations at the edge sites . on the other hand
, it also appears plausible that stable magnetic edge spins evolve for increasing @xmath12 .
the interplay of the antiferromagnetic isospin coupling and the ferromagnetic spin coupling ( third and fourth term in eq . ) alone would result in the formation of isospins at the edge sites @xmath20 and @xmath21 .
this is prevented , however , by the repulsive local hubbard - like interaction term ( second term ) which favors the formation of magnetic moments at the edges and suppresses doubly occupied or empty sites , i.e. the formation of isospins .
nonetheless , the formation of magnetic edge spins could still be hampered by the charge fluctuations triggered by the spin - isospin interaction term ( fifth term ) .
we have performed exact diagonalization calculations for small lattices @xmath77 .
it is found that the first scenario is not realized , since @xmath78 for @xmath61 , e.g. @xmath79 for @xmath80 and @xmath81 .
this shows that the isospin exchange term and the spin - isospin interaction term in eq .
are not relevant for the low - energy spectrum of our model .
hence , we may think of it as an effective model with two @xmath34 edge spins ( see fig .
[ fig : model]b ) which are ferromagnetically coupled with a weak effective interaction strength @xmath22 to the remaining electrons of the central region .
an rkky - like mechanism may then be responsible for an indirect coupling of both spins @xmath23 and @xmath24 , resulting in an effective antiferromagnetic two - spin model due to their odd distance @xmath25 .
in fact , the exact - diagonalization calculations for @xmath82 predict a total spin - singlet ground state .
to gain further analytical understanding of the edge spin coupling , one would ideally have to perform an eighth - order strong - coupling perturbation theory .
this quite demanding task can be circumvented by taking the effective hamiltonian @xmath76 as starting point for a second renormalization step and performing standard perturbation theory @xcite in powers of @xmath83 since @xmath84 .
one should note that the onsite interaction in @xmath76 ( second term ) makes the central region correlated at the interfaces between the kondo singlets and the central region , i.e. @xmath85 and @xmath86 .
since the spins , developing at the chain edges for strong @xmath0 , are only weakly coupled to the central region ( @xmath84 ) , finite - size effects play an important role at sufficiently large @xmath0 : the finite - size gap @xmath87 of the remaining conduction - electron system near the fermi energy is the largest energy scale appearing in @xmath76 ( eq . ) .
consequently , we may safely neglect the additional interactions at the interface sites for a moment , as they become relevant only on much smaller energy scales @xmath88 and will give small corrections only . with this idea ,
the unperturbed hamiltonian is given by @xmath89 and the perturbation by @xmath90 .
single - particle excitations of the non - magnetic central region require at least an energy of @xmath91 , i.e. , this is a so - called off - resonance situation ( see ref .
hence , there is no contribution linear in @xmath22 and second - order perturbation theory predicts an effective rkky model @xcite @xmath92 illustrated in the left panel of fig .
[ fig : model]c .
the rkky exchange coupling is given by @xmath93 where @xmath94 is the local weight of the state with momentum @xmath95 of the remaining conduction - electron sea at site @xmath48 and @xmath96 its energy .
@xmath97 has the typical fermi - liquid dependence @xcite on @xmath22 and @xmath27 at large distances @xmath98 and is antiferromagnetic as the edge - spin distance @xmath25 is assumed as odd in our model .
the corresponding energy spectrum of @xmath99 is shown in the left panel of fig .
[ fig : model]d , containing a singlet ground state and excited triplet states .
as a function of the edge spin distance @xmath27 on a double - logarithmic scale for different couplings @xmath0 as indicated .
dots : dmrg results for even @xmath19 and odd edge spin distance @xmath25 , i.e. antiferromagnetic coupling . colored dashed lines : spin gap as given by second - order perturbation theory in @xmath83 ,
black dashed line : expected @xmath27-dependence of @xmath100 for large @xmath27 and large @xmath0 .
inset : @xmath0-dependence of the spin gap . for different
system sizes @xmath19 , the dmrg data follow the expected power law @xmath101 at strong @xmath0 . the nearest - neighbor hopping @xmath42 fixes the energy scale . ] for a numerical check of the distance and @xmath0 dependence predicted by the proposed effective hamiltonian @xmath99 ( eqs . and ) , we apply a standard implementation of density - matrix renormalization group@xcite ( dmrg ) based on matrix product states and exploiting the two u(1 ) symmetries of the hamiltonian , i.e. conservation of the total particle number and the @xmath102-component of the total spin .
we compute the effective edge - spin coupling @xmath97 , given at strong @xmath0 by the spin gap between singlet and triplet states @xmath103 @xmath104 is the ground state in the hamiltonian block with total particle number @xmath37 and total spin @xmath105 , which can be addressed by selecting either the @xmath106 or the @xmath107 block .
@xmath100 has been calculated for different @xmath0 between 2.5 and 15 to study the crossover to the strong - coupling regime , where @xmath99 is suggested to be valid and where @xmath108 .
the results are shown in fig .
[ fig : gap ] as functions of the edge - spin distance @xmath27 ( dots connected by solid lines ) .
indeed , one finds that the asymptotic behavior of the dependence of @xmath100 on the edge - spin distance @xmath27 is given by a power - law @xmath109 for @xmath110 and @xmath80 and @xmath111 , @xmath112 has almost converged to @xmath113 ( see black dashed line ) from below .
this is a strong evidence for the validity of the suggested two - spin iime model ( eq . ) and indicates that the effective on - site interactions at @xmath85 and @xmath86 have negligible influence .
decreasing @xmath0 for fixed @xmath27 drives the system out of the iime limit , and consequently also @xmath112 departs from the predicted value of @xmath114 , e.g. @xmath115 for @xmath116 and @xmath117 .
additional insights can be obtained by comparing the numerical results to the spin gaps from perturbation theory ( eq . ) , which are shown as colored dashed lines in fig .
[ fig : gap ] .
the distance dependence is well recovered by the dmrg results for stronger couplings @xmath118 . while this is also the case at weaker couplings such as @xmath119 for small @xmath27 , one observes increasing deviations from the perturbative results with increasing system size .
this can be attributed to friedel oscillations induced by the open boundaries of the system which lead to a strongly site - dependent local density of states near the system boundaries . in particular
, the local density of states near the fermi energy at the impurity positions @xmath120 and @xmath121 strongly decreases with increasing system size @xmath19 .
since the kondo temperature @xmath6 is a very sensitive function of the local density of states , @xmath6 likewise decreases .
this tends to increase the kondo screening cloud and makes the starting point of our analysis based on local kondo clouds progressively worse .
our picture is also supported by the @xmath0-dependence of @xmath100 , displayed in the inset of fig .
[ fig : gap ] . in line with the expectations from the effective two - spin model ( eq . )
, we find that the spin gap is described by @xmath122 for large @xmath0 with increasing deviations from this relation for smaller @xmath123 . at
weak coupling strengths @xmath11 the model ( eq . )
is in the rkky regime , which also gives rise to an effective two - spin model , yet composed of the two impurity spins @xmath124 .
their antiferromagnetic indirect coupling @xmath125 leads to a singlet ground state and determines @xmath100 .
on the other hand , the iime limit ( fig .
[ fig : model]b and the left panel of fig .
[ fig : model]c ) , characterized by the effective two - edge - spin model with @xmath126 , is realized for strong couplings @xmath118 as discussed above . only then kondo clouds are sufficiently local to induce stable local moments at the chain ends .
the crossover region between these regimes , however , lacks a formulation as an effective two - spin model .
it can rather be described as a fermi liquid with vanishing spin gap , with two separate kondo screening clouds , and a paramagnetic region in - between .
as a function of the edge spin distance @xmath27 on a double - logarithmic scale for different couplings @xmath0 as indicated .
dots : dmrg results for odd @xmath19 and even edge - spin distance @xmath25 , i.e. ferromagnetic coupling . colored dashed lines : spin gap as given by first - order perturbation theory in @xmath83 ,
black dashed line : expected @xmath27-dependence of @xmath100 for large @xmath27 and large @xmath0 .
inset : @xmath0-dependence of the spin gap .
the dmrg data are approximately described by the expected power law @xmath127 at strong @xmath0 . the nearest - neighbor hopping @xmath42 fixes the energy scale . ]
we complete our study by investigating the emerging low - energy model for odd lattice sizes @xmath19 , i.e. even @xmath27 ( see right panel of fig . [
fig : model]c ) .
clearly , the low - energy spectrum of the non - interacting central region , which builds up at strong @xmath0 , now contains a singly occupied fermi level due to the odd number of sites . performing the same type of perturbation theory as above for weak coupling @xmath84
, shows that the effective ferromagnetic rkky interaction between the two edge spins is exceeded by a linear - in-@xmath22 coupling between single edge spins and the spin of the delocalized fermi electron @xcite of the form @xmath128 the coupling is ferromagnetic and given by @xmath129 thus , rather than an rkky model , the system renormalizes to a central - spin model @xcite , as sketched in the right panel of fig .
[ fig : model]c , with the delocalized spin of the fermi electron @xmath130 as the central spin .
the ground state of the ferromagnetic central - spin model is a total spin quartet ( @xmath131 ) .
the @xmath0 as well as the distance dependence of the effective coupling constant @xmath132 can be checked using the dmrg by computing the gap @xmath133 see fig .
[ fig : model]d ( right panel ) .
note that the spin gap of the central - spin model is related to @xmath132 as @xmath134 . to get @xmath100 as a difference of two ground - state energies in sectors with different @xmath135 but possibly equal @xmath136
, the dmrg algorithm must be extended slightly . making use of the matrix - operator representation of the hamiltonian
, we have implemented an additional interaction term @xmath137 for @xmath138 , this allows us to target blocks with a given total - spin quantum number @xmath139 .
our calculations show that there is an incidental degeneracy of the spin - quartet ground state of the central - spin model with a spin - doublet state which is different from the excited @xmath140 state of the central - spin model .
this doublet , however , can be shifted to higher energies by adding a weak repulsive hubbard term , @xmath141 to the conduction - electron system .
namely , the lieb - mattis theorem @xcite dictates that the total ground - state spin of the _ correlated _ ( @xmath142 ) model must have @xmath131 .
we have used weak hubbard interaction strengths of @xmath143 up to @xmath144 and have checked that this does not significantly affect the results obtained for @xmath145 . the results in fig .
[ fig : gap_fm ] indicate that the linear @xmath22 dependence of @xmath100 , suggested by perturbation theory ( eq . ) , is indeed realized at strong couplings @xmath118 ( see also inset of fig .
[ fig : gap_fm ] ) .
this behavior along with the considerably larger gaps distinguishes the ferromagnetic case clearly from the antiferromagnetic case discussed in the preceeding sections .
there is also a remarkable agreement between the numerical and the perturbative data with regard to the distance dependence of @xmath100 ( fig .
[ fig : gap_fm ] ) . unlike the antiferromagnetic case , even at relatively small couplings such as @xmath116 and for all considered system sizes @xmath19 , @xmath100 scales as @xmath146 ,
( [ eq : delta_s_onr_pt ] ) .
larger energy gaps make the effective central - spin model in the ferromagnetic case more robust as compared to the effective low - energy model in the antiferromagnetic case . for extremely large system sizes @xmath19 ( not accessible here ) and for still strong enough @xmath0
, we expect that the finite - size physics , emerging here as effective central - spin model , is replaced by a conventional rkky interaction obtained within the thermodynamic limit @xcite .
the reason is that the linear - in-@xmath22 contribution to the effective low - energy hamiltonian will be less and less important with increasing @xmath19 as compared to higher - order terms , which indicates the breakdown of finite - size perturbation theory .
quantum - confined multi - impurity kondo systems exhibit a complex interplay of different magnetic exchange mechanisms competing or cooperating with the kondo effect .
our present study has addressed a prototypical model where the emergence of local bound states in the strong kondo - coupling regime ( @xmath12 ) , i.e. the formation of local kondo singlets , strongly confines the conduction - electron mobility .
we have demonstrated that this leads to the formation of local spin moments in the conduction - electron system .
for the one - dimensional system studied here , moments are created at the edges of the chain .
virtual excitations of the local kondo singlets mediate an indirect coupling @xmath147 of these edge spins to the local magnetic moments at the edges of the remaining one - dimensional conduction - electron system .
this inverse indirect magnetic exchange ( iime ) is the strong - coupling analog of the well - known indirect rkky exchange which operates at weak @xmath148 .
an interesting observation made here is that the central part of the conduction - electron system can mediate an effective mutual coupling of the edge spins which can be understood by _ weak - coupling _ rkky - like perturbation theory as for @xmath149 the effective iime coupling @xmath150 is weak .
as evidenced by numerical dmrg results and by analytical insights from strong - coupling ( in @xmath0 ) and weak - coupling ( in @xmath22 ) perturbation theory , we have seen that odd - even effects are crucial to understand the total ground - state spin , the elementary excitation gap and the distance as well as the @xmath0 dependence of the effective magnetic coupling between the edge spins .
namely , depending on the number of lattice sites @xmath19 , two very different effective low - energy models are obtained : for odd distances @xmath151 between the edge spins , an effective antiferromagnetic two - spin model emerges which is composed of the two edge spins and coupled by an effective rkky interaction @xmath152 . in this case , where rkky emerges from iime , the spin gap is given by @xmath153 . for even distances ,
on the other hand , we find an effective central - spin model where the two edge spins couple ferromagnetically to the delocalized spin of the electron in the highest occupied spin - degenerate one - particle state of the conduction - electron system .
this induces ferromagnetic correlations between the two edge spins , and the ground state is essentially given by a spin quartet , but there is no indirect ferromagnetic coupling , i.e. this case is fundamentally different from a ferromagnetic rkky model .
this is also reflected in a spin gap @xmath154 which is linear rather than quadratic in @xmath22 .
quantum confined kondo systems with a wide range of accessible model parameters can be simulated by ultracold fermi atoms trapped in optical lattices .
we believe that the systematic study of magnetic exchange interactions and kondo correlations in those systems is only at the very beginning .
the largely different physics of effective low - energy models that are obtained by slightly different geometries or slightly different impurity - spin configurations offers an exciting perspective for tailoring physical properties in the experiment .
further theoretical work may e.g. address the low - energy physics of stacked kondo singlets where the number of intermediate kondo singlets between two edge spins is varied .
quite generally , not much is known on the crossover from a single or few kondo impurities to the dense kondo lattice in one- and higher - dimensional quantum - confined systems .
|
the diluted kondo lattice model is investigated at strong antiferromagnetic local exchange couplings @xmath0 , where almost local kondo clouds drastically restrict the motion of conduction electrons , giving rise to the possibility of quantum localization of conduction electrons for certain geometries of impurity spins .
this localization may lead to the formation of local magnetic moments in the conduction - electron system , and the inverse indirect magnetic exchange ( iime ) , provided by virtual excitations of the kondo singlets , couples those local moments to the remaining electrons .
exemplarily , we study the one - dimensional two - impurity kondo model with impurity spins near the chain ends , which supports the formation of conduction - electron magnetic moments at the edges of the chain for sufficiently strong @xmath0 . employing degenerate perturbation theory as well as analyzing spin gaps numerically by means of the density - matrix renormalization group , it is shown that the low - energy physics of the model can be well captured within an effective antiferromagnetic rkky - like two - spin model ( `` rkky from iime '' ) or within an effective central - spin model , depending on edge - spin distance and system size .
|
there is accumulating evidence that natural landslides exhibit scaling properties @xcite , including the area and volume of the slope failures @xcite , and the amount of rainfall required for the initiation of landslides in a region @xcite .
the scaling properties of landslides are revealed by power law dependencies , and are considered evidence of the critical state of landscape systems dominated by slope wasting phenomena @xcite it is known that , regardless of the physiographic or the climatic settings , the probability ( or frequency ) density of event landslides increases with the area of the landslide up to a maximum value , known as the rollover " , after which the density decays along a power law @xcite . the length scale for the rollover , and the rapid decay along a power law , are conditioned by the mechanical and structural properties of the soil and bedrock where the landslides occur @xcite , and are independent of the landslide trigger @xcite .
the probability ( or frequency ) density of the landslide volume obeys a negative power - law , with a scaling controlled by the type of the landslides @xcite .
the dependence of landslide volume on landslide area was also shown to obey a distinct scaling behavior over more than eight orders of magnitude @xcite .
rainfall is a recognized trigger of landslides , and early investigators have recognized that empirical rainfall thresholds can be established to determine the amount of rainfall required to initiate landslides in a region @xcite .
different types of empirical thresholds that use combinations of rainfall measurements obtained from the analysis of rainfall events that resulted ( or did not result ) in landslides were proposed in the literature , including mean intensity - duration ( @xmath3 ) and rainfall cumulated event - duration ( @xmath4 ) thresholds , and their variations @xcite . with a few exceptions
@xcite all the empirical rainfall thresholds are represented by power law models , indicative of the self - similar behavior of the rainfall characteristics responsible for landslide occurrence . despite the abundant empirical evidence ,
the reasons for the scaling behaviors of landslide phenomena are poorly known , and only a few attempts were made to interpret the empirical evidences with deterministic or physically based models @xcite . in this paper , we show that a relatively simple , physically based model that describes the stability / instability conditions of slopes forced by rainfall , when applied to a sufficiently large geographical area produces results that are in agreement with two known scaling properties of landslides , namely : ( i ) the rainfall conditions that result in unstable slopes , which match regional empirical @xmath3 thresholds for possible landslide occurrence @xcite , and ( ii ) the frequency distribution of the area of the patches of terrain predicted as unstable by the model , which matches the statistics of landslide area for event landslides @xcite . the paper is organized as follows . in section [ sec : data ] , we describe the geographical area in central italy where we have conducted our experiments ( fig . [ fig01 ] ) , and in section [ sec : model ] we provide general information on the transient rainfall infiltration and grid - based regional slope - stability analysis code ( trigrs , version 2.0 ; @xcite ) that we adopted for the experiments .
next , in section [ sec : iddep ] , we compare the @xmath3 conditions capable of producing slope instability in the study area predicted by trigrs , with empirical rainfall @xmath3 thresholds for possible landslide occurrence in central italy .
this is followed by a comparison of the probability density of the area of the patches of terrain predicted as unstable by trigrs in the study area with the probability density of natural landslides in the same general area , and by a discussion about the possible relations of this work to other existing approaches for the description of landslide scaling phenomena .
we conclude , in section [ sec : concl ] , summarizing the lessons learnt .
the upper tiber river basin ( utrb ) extends for 4,098 km@xmath5 in central italy , with elevation in the range from 163 m at the basin outlet to 1571 m along the divide between the adriatic sea and the tyrrhenian sea ( fig .
[ fig01 ] ) . in the area
the landscape is hilly or mountainous , with open valleys and intra - mountain basins . in the mountains and the hills ,
the morphology is conditioned by lithology and the attitude of the bedding planes .
climate is mediterranean , with most of the precipitation falling from october to december and from february to april @xcite . for the utrb two digital representations of the terrain elevation ( dem )
were available to us .
a coarser dem , with a ground resolution of 25 @xmath6 25 m , was obtained through the linear interpolation of elevation data along contour lines shown on 1:25,000 topographic base maps @xcite . a finer dem , with a ground resolution of 10 @xmath6 10 m , was prepared by the italian national institute for geophysics and volcanology through the interpolation of multiple sources of elevation data @xcite . five lithological complexes , or groups of rock units , crop out in the utrb @xcite , including , from younger to older : ( a ) recent fluvial and lake deposits , which crop out mostly along the valley bottoms , ( b ) unconsolidated and poorly consolidated sediments pertaining to a continental , post - orogenic sequence , pliocene to pleistocene in age , ( c ) allochthonous rocks , lower to middle miocene in age , ( d ) sediments pertaining to the tuscany turbidites sequence , eocene to miocene in age , and to the umbria turbidites sequence , miocene in age , and ( e ) sediments pertaining to the the umbria - marche stratigraphic sequence , lias to lower miocene in age .
-0.5 cm = 0.35 soils reflect the lithological types , and range in thickness from less than 20 cm to more than 1.5 m. landslides are abundant in the area , -0.5 cm = 0.35 and cover 523 km@xmath5 ( 12.8% of the catchment ) , for a total estimated landslide volume of 5.9@xmath7 m@xmath8 @xcite . for the numerical experiments we selected the area in the utrb where unconsolidated and poorly consolidated continental sediments crop out ( green area in fig . [ fig01 ] ) .
this is the lithological complex where shallow landslides are more frequent in the study area @xcite , and where the trigrs conceptual landslide scheme is best suited to model the stability / instability conditions of slopes forced by rainfall .
the geotechnical properties for the soils in this lithological complex used for the numerical modeling are listed in line ( b ) of table [ tab01 ] .
as noted , we adopted trigrs version 2.0 @xcite . the software implements a grid - based , spatially distributed slope stability model coupled with an infiltration model capable of simulating = 0.45 the infiltration of rainfall in the terrain , modulating the stability / instability conditions of the individual grid cells @xcite .
trigrs adopts a grid - based representation of a real landscape based on a dem , and uses local terrain characteristics as an input for the solution of a system of equations whose output is the factor of safety @xmath9 i.e. , a positive number representing the balance of the driving and the resisting forces acting in each grid cell . for stable conditions , where the resisting forces exceed the driving forces , @xmath10 .
@xmath11 represents the metastable condition where the driving and the resisting forces are equal , and @xmath12 represents the physically unrealistic condition where the driving forces exceed the resisting forces , and the slope fails . for the modeling , trigrs exploits an infinite slope approximation i.e. , a rigorous , lowest order approximation of a multi - dimensional landslide geometry .
this approximation is adequate where @xmath13 , where @xmath14 is the depth of the slip surface and @xmath15 is the landslide linear dimension @xcite .
an upper bound for @xmath14 can be identified on the basis of the lithological setting , where stronger rocks underly a weaker layer of soil or regolith .
this is an assumption for shallow landslides for which depth is significantly smaller than width , or length .
0.1 cm = 0.41 -0.1 cm [ fig04 ] for simplicity , trigrs neglects all the lateral stresses and the inter - cell forces , and the stability of each grid cell is governed solely by the balance of the vertical component of gravity , against the resisting stress due to the basal coulomb friction , plus the pore pressure @xcite .
failure occurs at depth @xmath16 , measured vertically from the topographic surface , if at that depth @xcite : [ effesse ] f_=f_+f_+f_=1 .
where , @xmath17 is the ratio of resisting to driving forces , and the individual components in eq .
( [ effesse ] ) can be written as a function of the local cell characteristics @xcite : [ effeffe ] f_&=&/ , + [ effedabliu ] f_&= & , + [ effeci ] f_&= & , where : @xmath18 is the soil cohesion , @xmath19 is the soil internal friction angle , @xmath20 is the wet soil unit weight , @xmath21 n m@xmath22 is the unit weight of water , and @xmath23 is the local angle of the terrain with respect to the horizontal .
the model assumes that rainfall modulates the = 1.05 water table height and groundwater flow parallel to the slope , and that sub - surface flow is one - dimensional @xcite . when rainfall hits the ground , due to water infiltration the driving force described by eq .
( [ effedabliu ] ) is modulated , @xmath9 varies with @xmath16 and @xmath24 , and @xmath25 is the pressure head as a function of depth @xmath16 and time @xmath24 . in eq .
( [ effedabliu ] ) , the effects of variation is determined for each cell by solving the richards equation @xcite : [ richeq ] = ( k()-1 ) , where @xmath26 is the soil water content , and @xmath27 is the soil hydraulic conductivity . in trigrs ,
( [ richeq ] ) is linearized and solved at discrete time steps and in the vertical coordinate .
the linearization procedure relies on the identification of two different time scales @xcite . for each grid cell
, @xmath28 is the upslope contributing area i.e. , the cumulated area of all the upslope cells draining in the considered cell , and @xmath29 is the soil diffusivity in the cell , where @xmath30 is the saturated hydraulic conductivity ( see table [ tab01 ] ) and @xmath31 is the specific water storage . the first time scale can be identified with @xmath32 as the time for lateral pore pressure transmission from the upstream area to the grid cell .
the second time scale is @xmath33 , the time needed for pore pressure transmission from the surface to the depth @xmath14 @xcite .
one can then build the length scale ratio @xmath34 : [ epsdef ] = = . under the condition @xmath35 , eq .
( [ richeq ] ) can be simplified identifying long - term and short - term response terms @xcite used in the numerical implementation of @xcite .
( [ epsdef ] ) is also used to identify the approximate limits of applicability of the model implemented by trigrs .
when rainfall intensity exceeds the local infiltration capacity , the excess water in the single grid cells is routed downslope to the nearest cells @xcite .
trigrs accepts as input complex rainfall histories ( i.e. , spatially and temporally varying ) , permitting a realistic modeling of the slope stability / instability conditions driven by real rainfall events .
[ fig02 ] shows the result of the trigrs modeling for two representative sub - basins in the study area .
the map portrays , superimposed on a shaded relief image showing the terrain morphology , the geographical distribution of the @xmath9 calculated using typical values for the mean = 1.05 rainfall intensity and the rainfall duration .
shades of colours , from red ( @xmath36 ) to green , show the values of @xmath9 calculated for each grid cell .
single unstable cells , or clusters of unstable cells with @xmath37 are shown in blue .
to investigate the relationship between the magnitude of the rainfall trigger ( i.e. , @xmath38 and @xmath39 ) we first partitioned the study area into sub - basins i.e. , hydrological ensembles of slope units , where a slope unit is a hydrological region bounded by drainage and divide lines @xcite . to obtain the sub - basins we exploited the 25 @xmath6 25 m resolution dem , and we used the ` r.watershed ` command in the grass gis , release 7.0 ( www.grass.org ) .
sub - basins with a surface area smaller than 2,500 m@xmath5 ( i.e. , four grid cells ) were identified and excluded from the analysis .
[ fig03 ] shows the statistical size - distribution of the slope units in the study area i.e. , where post - orogenic sediments crop out in the utrb ( green areas in fig .
[ fig01 ] ) .
next , we forced the individual sub - basins in the study area with a uniform rainfall of a given intensity @xmath38 , for a given period of time @xmath39 .
we then studied the rainfall conditions that have resulted in unstable cells in the sub - basins . for simplicity
, we considered only ( nearly ) mono - lithological sub - basins i.e. , sub - basins for which at least 50% of the area is covered by unconsolidated and poorly consolidated continental sediments .
this reduced the complexity that the presence of multiple rock types in a sub - basin may introduce .
we further considered unstable the sub - basins with at least 10% of the grid - cells with @xmath12 .
this is a reasonable assumption for rainfall - triggered landslides in the study area @xcite .
the results depend only weakly on the proportion of cells in a sub - basin necessary to consider a sub - basin as unstable .
the first set of experiments was conducted adopting the following procedure . for each sub - basin
, we started with a given ( reasonable e.g. , @xcite ) set of @xmath39 and mean @xmath38 conditions , and we checked if these conditions resulted in landslides i.e. , if at least 10% of the grid - cells in the sub - basin had @xmath12 .
next , we increased @xmath38 maintaining @xmath39 constant , and we checked if the new rainfall conditions had resulted in landslides .
the procedure was repeated for different , increasing rainfall durations . for the modeling , we used = 0.9 \(i ) _
i _ = 1 mm h@xmath40 to _ i _ = 200 mm h@xmath40 with steps of 2 mm h@xmath40 , and ( ii ) _ d _ = 1 hour to _ d _
= 100 hours , with steps of 1 hour for the 1 - 10 h range , and steps of 10 hours for the 10 - 100 h range . at the end of the procedure , for each of the considered sub - basins we obtained a set of rainfall @xmath41 conditions that had resulted in ( predicted ) slope instabilities ( 10% or more of the grid - cells with @xmath12 ) .
we plotted these rainfall conditions in a @xmath2 plot , in log - log coordinates . the result is shown in fig .
[ fig04 ] where each black line represents the threshold above which a single sub - basin in the study area is unstable .
visual inspection of fig .
[ fig04 ] reveals that the rainfall @xmath41 conditions that have resulted in unstable conditions in each sub - basin follow power laws that represent rainfall thresholds for slope instability in the sub - basins .
the lines are equivalent to rainfall thresholds for possible landslide occurrence @xcite .
not all the sub - basins in the study area resulted in an instability threshold curve , because : ( i ) a number of sub - basins contained less than 50% of continental , post - orogenic sediments , and were excluded from the analysis , ( ii ) the percentage of unstable cells in a sub - basins was less than 10% , regardless of the rainfall conditions , and where not considered as failed slopes , and ( iii ) the instability threshold curve was outside of the ranges of rainfall intensity and duration considered for the modeling .
further inspection of fig .
[ fig04 ] reveals that the threshold lines are well defined , and obey distinct power law trends for a significant range of rainfall durations ( @xmath42 hours ) known to initiate landslides @xcite .
significantly , the slope of the threshold curves is in agreement with the slope of empirical rainfall thresholds for possible landslide occurrence established for central italy @xcite and for italy @xcite , considering the uncertaintly associated with the empirical thresholds @xcite .
the position of the threshold curves in the @xmath2 plane depends on the location and geomorphological characteristics of the individual sub - basins , but all of the threshold curves lay above the empirical thresholds proposed in the literature and applicable to the study area ( e.g. , @xcite ) .
this is because empirical thresholds are defined as lower boundaries of @xmath38 , @xmath39 conditions in a large area , while in our case each curve corresponds to a specific ( local ) morpho - lithological setting .
we stress that the result was obtained using reasonable values for the model parameters , but without any attempt to fine - tune the parameters , and specifically the geotechnical parameters ( table [ tab01 ] ) .
also , no attempt was made to optimize the modeling = 0.9 parameters using e.g. , a probabilistic modeling approach @xcite .
we maintain that this indicates that the scaling behavior of the rainfall conditions responsible for shallow landslides in the utrb emerged from the physical modeling , and is dependent on the local geomorphological setting and the rainfall trigger . to further investigate the position of the obtained threshold curves in the @xmath2 plane with the local terrain slope and the soil depth , two key variables that control the modeling @xcite , we separated the threshold curves based on classes of terrain slope ( fig .
[ fig05 ] ) , and soil depth ( fig .
[ fig06 ] ) in the sub - basins .
visual inspection of fig .
[ fig05 ] reveals that the @xmath3 threshold curves for sub - basins characterized by steep slopes are lower that the thresholds for sub - basins characterized by gentle slopes .
this was expected @xcite .
analysis of fig .
[ fig05 ] indicates that thinner soils fail with less severe rainfall conditions ( lower thresholds ) than the thicker soils .
this was also expected @xcite .
we attribute the dispersion of the threshold lines in the different classes of terrain slope and soil depth to natural variability i.e. , to the terrain variability that characterizes the sub - basins in the study area . in figs .
[ fig05 ] and [ fig06 ] the maps show the geographical location of the sub - basins for which the threshold lines shown in the same panel were calculated .
inspection of the maps reveals that the sub - basins are distributed throughout the study area , with no geographical trend or bias .
this is an indication that the results are independent of a specific location or geomorphological setting .
we determined the probability density of the area of the patches of terrain predicted as unstable by trigrs , and we compared the probability density of the patches to the probability density of natural landslides in the utrb @xcite .
we defined a patch of unstable terrain a cluster of contiguous grid cells that individually have @xmath12 . to identify the unstable grid cells , we ran trigrs with a fixed rainfall duration @xmath39 = 1 hour , increasing the rainfall intensity @xmath38 from 25 to 100 mm h@xmath40 .
next , we repeated the calculations with a fixed rainfall intensity @xmath38 = 100 mm h@xmath40 , and we varied @xmath39 from 1 to 10 hours .
the remaining model parameters were kept constant for all the model runs ( see table [ tab01 ] ) .
results are summarized in fig .
[ fig07 ] . for the experiment shown in fig .
[ fig07 ] we used the finer resolution ( 10 = 0.34 @xmath6 10 m ) dem @xcite . visual inspection of fig . [ fig07 ] reveals a good agreement between the probability density of the area of the patches of unstable terrain , and the corresponding probability density of event landslides in the utrb @xcite , and with the general probability density curve for event landslides prosed by @xcite .
the agreement is best for medium to low rainfall intensities ( @xmath38 = 25 to 50 mm h@xmath40 ) and for medium rainfall durations ( @xmath39 = 5 h ) .
these rainfall conditions are known to initiate landslides in the utrb and in similar physiographic regions in umbria @xcite and in central italy @xcite .
higher rainfall intensities and shorter rainfall periods result in a poorer match between the modeled and the empirical probability densities . for small and very small unstable patches which we identify as single landslides , with area @xmath43 km@xmath5 , the probability density is lower than prescribed by a simple negative power law .
interestingly , the reduced density occurs at the same size @xmath44 for which the probability density of the natural landslides exhibits a distinct rollover , which was recognized to be a physical characteristic of populations of event landslides @xcite , and was attributed to the mechanical properties of the soil and regolith where the small landslides occur @xcite .
however , the smooth rollover for small and very small landslides typical of event landslides is not clearly reproduced by the model .
we can not exclude a priori that the result is related to the strategy used to decide the size of the patches of instability , or to the resolution of the dem used for the modeling . to investigate this possibility , we tested different strategies to identify the patches of unstable terrain , and we studied the effects of the different strategies on the probability density of the area of the patches . the simplest strategy consisted in considering all the grid cells in a sub - basin with @xmath9 @xmath45 1.0 that shared a boundary or a corner as belonging to a single landslide . in this strategy
, a single unstable cell not connected to any other unstable grid cell represents an unstable patch with the smallest possible area , @xmath44 = 100 m@xmath5 for a 10 @xmath6 10 m dem .
an alternative strategy consisted in smoothing the grid map of the computed @xmath9 values prior to searching for the contiguous cells . for the purpose
, we passed a 3 @xmath6 3 kernel over the grid map , and decided that the central pixel was unstable ( @xmath9 @xmath45 1.0 ) if at least two ( of the nine ) cells in the kernel had @xmath9 @xmath45 1.0 .
another strategy consisted in passing the same 3 @xmath6 3 kernel over the @xmath9 grid map , and in deciding that the central pixel is unstable if it has @xmath9 @xmath45 1 and if the average @xmath9 value of the nine pixels in the kernel was smaller than 5 , a figure determined heuristically through an iterative procedure .
the first strategy provided the best result ( fig .
[ fig07]a ) , with the other strategies deviating significantly from the landslide field data . for the first strategy , the agreement between the area of the modeled patches and the area of the event landslides in the utrb is best for @xmath46 km@xmath5 .
the density deviate slightly for larger areas ( @xmath47 km@xmath5 ) .
we attribute the deviation to the fact that very large unstable patches correspond to large , deep - seated landslides that are not modeled by trigrs .
we investigated the dependence of the model results to the ground resolution of the dem .
for the purpose , we repeated the model runs using the coarser 25 @xmath6 25 m dem .
the results are shown in fig .
[ fig08]b , and were obtained using the same rainfall intensity , rainfall duration , and geotechnical parameters , changing the resolution of the dem .
inspection of the plot shows that the probability density of the area of the patches of unstable grid cells obtained using the finer resolution ( 10 @xmath6 10 m ) dem follows nicely the probability density of the area of event landslides in the utrb , for @xmath48 km@xmath5 .
again , we attribute the observed deviation for larger areas to the fact that the very large patches correspond to deep - seated landslides that are not modeled by trigrs . instead , the results obtained using the coarser resolution ( 25 @xmath6 25 m ) dem follows a steeper power law trend that deviates from the probability density of the natural landslides in the utrb @xcite .
we repeated the simulations with the coarser dem using larger intensity and duration values , obtaining curves with a lower slope , with a limit to the correct one .
however the resulting probability densities as a function of @xmath44 ( not shown in fig .
[ fig08 ] ) were higher than the curves obtained using the finer resolution dem .
we conclude that the resolution of the coarser dem is insufficient to model the stability / instability conditions of shallow landslides in the study area .
lastly , we investigated the relationship between the size of the patches predicted unstable by the trigrs model ( fig .
[ fig03 ] ) and the size of the slope units in the utrb ( fig .
[ fig09 ] ) .
we find that the size distribution of the predicted unstable patches covers the same size range of the landslides mapped in the utrb , and that the relation between the size of the unstable patches and the size of the slope units depends on the ( synthetic ) @xmath3 rainfall conditions .
the differences between the modelled and the empirical distributions have multiple causes , including the method used to cluster the unstable cells , and the fact that our modeling results were obtained for the portion of the utrb where the post - orogenic sediments crop out ( area b in fig .
[ fig01 ] ) .
we attempted a comparison between the results obtained in our study area using the physically - based model trigrs @xcite , and other attempts conducted to exploit numerical , simplified ( i.e. , toy " ) , or conceptual models to simulate the rainfall conditions responsible for slope failures , and the frequency - size distribution of the landslides .
we concentrate on simplified , conceptual models , and we exclude the use of physically - based geo - mechanical models such as those proposed by @xcite or by @xcite to describe the frequency - size statistics of slope failures , or other physically - based spatially distributed slope stability models e.g. , shalstab @xcite , sinmap @xcite , shetran @xcite , starwars @xcite , geotop - fs @xcite , and chlt @xcite . we are not aware of any attempt to describe the scaling behaviour of rainfall thresholds for possible landslide occurrence executed using simplified or conceptual models .
the results obtained in this work represent a preliminary attempt to provide a physical background to the empirical rainfall thresholds for the prediction of possible landslide occurrence .
this is important for the application of rainfall thresholds in landslide warning systems @xcite .
investigators have attempted to simulate the self - similar scaling behaviour of landslide size using cellular automata ( ca ) models ( see e.g. , @xcite ) .
the interest in the use of ca models to investigate landslide sizes is motivated largely by the analogy between landslides in a landscape and avalanches in the sandpile " model , the first theoretical and practical realization of a self - organized critical ( soc ) system @xcite . in the sandpile model ,
the meta - stable condition is controlled by a critical slope , which gives rise to relaxation of the system i.e. , movements of single or multiple grains between cells that result in an avalanche - like behaviour .
soc systems are implemented numerically in a number of variants , choosing different dynamical variables or combination of variables to simulate the interactions between the slope and its time - dependent weakening @xcite . typically , the critical conditions are controlled by conservative or dissipative relaxation rules and by the boundary conditions , leading to an soc behaviour . when applied to the investigation of the statistics of landslide sizes , a limitation of ca models
consists in the fact that the exponents controlling the scaling of the distributions obtained through the ca models are significantly smaller ( e.g. , @xcite ) than the corresponding exponents obtained modeling empirical data taken from landslide inventory maps ( e.g. , @xcite ) .
the frequency distributions obtained from ca models are also very sensitive to the rules used to set up and run the models , and to the initial and boundary conditions @xcite .
@xcite hypothesized that an inverse cascade model could explain the power - law behavior of natural landslides and of soc models of slope instability , but did not demonstrate a link between the scaling of the empirical and modeled distributions . in a recent paper , @xcite proposed a criticality - inspired numerical model to explain the frequency - size distribution of large rockfalls , and to estimate the size and frequency of the largest possible failures in a region .
the model uses a dem to calculate local terrain slope , and exploits random perturbations to mimic natural fractures that control the location and size of the rockfalls .
results obtained for the alps revealed that the frequency distribution of the modelled failure volumes obeys a power - law with a scaling exponent in the range of the known empirical values @xcite , and exhibits a distinct `` rollover '' for small size failures .
@xcite proposed an alternative way for describing the scaling properties of landslide sizes exploiting non - extensive statistical mechanics , or tsallis statistics @xcite , a formal extension of the boltzmann - gibbs statistics .
this approach relies on the postulate that a system composed by two independent statistical systems has an entropy given by the entropies of the two subsystems plus a correction term dependent on a parameter that controls the degree of non - additivity . despite criticisms @xcite , the approach proved capable of describing quantitatively the frequency - size distributions typical of empirical landslide datasets , including the scaling exponents and the `` rollover '' observed for small landslide sizes @xcite . when applied to simulate the frequency - size statistics of landslides , the postulate of non - additivity of entropy implied by the tsallis statistics is interpreted with the redistribution of the soil materials in different ( larger ) volumes when a landslide occurs , which increases the entropy in the final system composed of a different distribution of the initial sub - systems .
the advantage of the approach is that the statistical distribution of the landslide size can be derived in an analytic way , starting from general principles @xcite . as discussed before ,
our results indicate that a relatively simple , physically - based model for simulating the stability / instability conditions of natural slopes forced by rainfall , when applied to a sufficiently large area , is for reproducing well the frequency area distribution of the natural landslides in the same area .
this proves that the observed self - similar scaling behaviour of the landslide areas , and its deviation from the main trend observed for small landslides , are the results of physical characteristics of the landscape and of the slope instabilities in it .
further , application of the physically - based model revealed the presence of both scaling properties of landslides i.e. , of the intensity of the driving forces ( the mean rainfall intensity and the rainfall duration responsible for the slope instabilities ) and of the magnitude of the consequences ( the relative proportion of landslides of different sizes ) .
the result was obtained selecting reasonable values for the model parameters , without changing or optimizing them .
this result indicates a functional link between rainfall and landslides , and opens to the possibility to quantify the link .
trigrs reproduced well the statistical distribution of landslide areas and the rainfall conditions that have resulted in landslides in our study area .
this is not a trivial result , because the landslide scaling properties captured by the numerical model are not keyed in the equations governing the model .
we maintain that the fact that the distribution of landslide size and the rainfall thresholds for landslide occurrence are reproduced well by the model is evidence that the model uses the correct dynamics and variables .
our results open to the possibility of establishing rainfall thresholds for [ cols="^,^,^,^,^,^,^ " , ] [ tab01 ] landslide occurrence for individual slopes or groups of slopes , reducing the geographical uncertainty associated with empirical rainfall thresholds , which are most commonly defined for large to very large areas @xcite .
a reduced geographical uncertainty of the thresholds may improve the performance of landslide warning systems based on rainfall thresholds .
further , the results open to the possibility of predicting the magnitude of landslide triggering meteorological events , and the impact of rainfall induced landslides on the landscape .
accurate information on the expected impact of landslides may be important for landscape evolution studies , and for improved landslide hazard and risk assessments .
we used the consolidated transient rainfall infiltration and grid - based regional slope - stability analysis code ( trigrs , version 2.0 ; @xcite ) to investigate two scaling properties of landslides , namely : ( i ) the rainfall intensity
duration ( @xmath3 ) conditions responsible for slope instability and landslide occurrence @xcite , and ( ii ) the probability density of the area of rainfall induced landslides @xcite . our results , obtained for a significantly large area in central italy where landslides are abundant and frequent @xcite , indicate that the physically based model was capable of reproducing the two landslide scaling behaviors .
more specifically , the numerical modeling revealed the followings : * the scaling exponent of the power - law curves describing the mean rainfall intensity vs. rainfall duration conditions that produced slope instability in the studied ( quasi ) mono - lithological sub - basins , where unconsolidated and poorly consolidated sediments crop out was almost identical to the scaling exponent of empirical @xmath3 rainfall thresholds for possible landslide occurrence defined for central italy @xcite , and for italy @xcite .
the finding is important because it reconciles the physically - based and the statistically - based approaches to the prediction of rainfall induced landslides @xcite .
this is relevant for landslide warning systems based on rainfall thresholds @xcite . * the probability density distribution of the area of the patches of grid cells predicted as unstable by the trigrs model ( i.e. , grid cells with @xmath12 ) in the study area was similar to the frequency density of the area of natural event landslides in the utrb @xcite , and to the general probability density distribution for event landslides proposed by @xcite , for areas @xmath49 km@xmath5 . for smaller areas ,
the probability density of the patches of unstable cells deviates from the power law trend .
this is similar to the rollover " observed in statistically complete populations of event landslides @xcite .
the finding is relevant because it provides a physical basis to the statistics of landslide areas in a large and geomorphologically complex region .
we consider the results of this work as a starting point for investigating the underlying mechanisms giving rise to the scaling properties of landslide phenomena .
we thank r.l . baum and j.w .
godt ( usgs ) for valuable discussion .
we are grateful to the editor ( t. oguchi ) , s. hergarten and an anonymous reviewer for their constructive comments .
ma was supported by grants provided by the regione umbria , under contract _ por - fesr umbria 2007 - 2013 , asse ii , attivit a1 , azione 5 _ , and by the _ dipartimento della protezione civile _ ,
99 aleotti , p. , 2004 .
a warning system for rainfall - induced shallow failures . engineering geology 7 , 247 - 265 .
bach kirschbaum , d. , adler , r.f . ,
hong , y. , kumar , s. , peters - lidard , c. , lerner - lam , a. , 2012 .
advances in landslide nowcasting : evaluation of a global and regional modeling approach . environmental earth sciences 66 , 1683 - 1696 .
bak , p. , tang , c. , wiesenfeld , k. , 1987 .
self - organized criticality : an explanation of 1/f noise .
physical review letters 59 , 381 - 384 .
bak , p. , tang , c. , wiesenfeld , k. , 1988 .
self - organized criticality . physical review a 38 , 364 - 374 .
baum , r. , savage , w. , godt , j.w . , 2008 .
trigrs - a fortran program for transient rainfall infiltration and grid - based regional slope - stability analysis , version 2.0 .
u.s . geological survey open - file report 2008 - 1159 , 75p .
brunetti , m.t . ,
guzzetti , f. , rossi , m. , 2009a .
probability distributions of landslide volumes .
nonlinear processes in geophysics 16 , 179 - 188 .
brunetti , m.t . ,
peruccacci , s. , rossi , m. , guzzetti , f. , reichenbach , p. , ardizzone , f. , cardinali , m. , mondini , a.c . , salvati p. , tonelli , g. , valigi , d. , luciani s. , 2009b .
a prototype system to forecast rainfall induced landslides in italy . in : proceedings of the first italian workshop on landslides , naples , 8 - 10 june 2009 , vol . 1 , 157 - 161 .
brunetti , m.t . ,
peruccacci , s. , rossi , m. , luciani , s. , valigi , d. , guzzetti , f. , 2010 .
rainfall thresholds for the possible occurrence of landslides in italy .
natural hazards and earth system sciences 10 , 447 - 458 .
burton , a. , bathurst , j. , 1998 .
physically based modelling of shallow landslide sediment yield at a catchment scale .
environmental geology 35(2 - 3 ) , 49 - 99 .
caine , n. , 1980 .
the rainfall intensity - duration control of shallow landslides and debris flow .
geografiska annaler a62 , 23 - 27 .
cancelli , a. , nova , r. , 1985 .
landslides in soil debris cover triggered by rainstorms in valtellina ( central alps , italy ) .
4th international conference and field workshop on landslides , tokyo , pp .
267 - 272 .
cannon , s. h. , ellen , s. , 1985 .
rainfall conditions for abundant debris avalanches , san francisco bay region , california .
california geology , 267 - 272 .
cardinali , m. , ardizzone , f. , galli , m. , guzzetti , f. , reichenbach , p. , 2000 .
landslides triggered by rapid snow melting : the december 1996-january 1997 event in central italy .
proceedings of the egs plinius conference on mediterranean storms , maratea , italy , pp .
439 - 448 .
cardinali , m. , antonini , g. , reichenbach , p. , guzzetti , f. , 2001 .
photo - geological and landslide inventory map for the upper tiber river basin .
cnr , gruppo nazionale per la difesa dalle catastrofi idrogeologiche , publication n. 2154 , scale 1:100,000 .
cardinali , m. , galli , m. , guzzetti , f. , ardizzone , f. , reichenbach , p. , bartoccini , p. , 2006 .
rainfall induced landslides in december 2004 in south - western umbria , central italy : types , extent , damage and risk assessment . natural hazards and earth system sciences 6 , 237 - 260 .
carrara , a. , cardinali , m. , detti , r. , guzzetti , f. , pasqui , v. , reichenbach , p. , 1991 .
gis techniques and statistical models in evaluating landslide hazard .
earth surface processes and landforms 16 , 427 - 445 .
chen , c .- c . , telesca , l. , lee , c .- t . ,
sun , y .- s . ,
statistical physics of landslides : new paradigm .
epl 95 , 49001 .
chen , c .- c . ,
telesca , l. , lee , c .- t . ,
sun , y .- s . ,
reply to comment by l.p .
epl 100 , 29002 .
chleborad , a.f . , 2003
. preliminary evaluation of a precipitation threshold for anticipating the occurrence of landslides in the seattle , washington , area .
us geological survey open - file report 03 - 463 .
crosta , g.b . ,
frattini , p. , 2003 .
distributed modelling of shallow landslides triggered by intense rainfall .
natural hazards and earth system sciences 3 , 407 - 422 .
dietrich , w.e . , montgomery , d.r .
shalstab : a digital terrain model for mapping shallow landslide potential .
ncasi ( national council of the paper industry for air and stream improvement ) technical report .
endo , t. , 1970 .
probable distribution of the amount of rain- fall causing landslides .
annual report , hokkaido branch , govern .
forest experiment station , sapporo , pp .
123 - 136 .
godt , j.w . ,
baum , r.l .
, chleborad a.f . , 2006 .
rainfall characteristics for shallow landsliding in seattle , washington , usa . earth surface processes and landforms 31 , 97 - 110 .
godt , j.w . , mckenna , p. , 2008
numerical modeling of rainfall conditions for shallow landsliding in seattle , washington .
reviews in engineering geology 20 , 121 - 136 .
godt , j.w . ,
schulz , w.h . ,
baum , r.l . , savage , w.z .
modeling rainfall conditions for shallow landsliding in seattle , washington .
reviews in engineering geology 20 , 137 - 152 .
govi , m. , sorzana , p.f . , 1980 .
landslide susceptibility as function of critical rainfall amount in piedmont basin ( northwestern italy ) .
studia geomorphologica carpatho - balcanica 14 , 43 - 60 .
guzzetti , f. , ardizzone , f. , cardinali , m. , galli , m. , reichenbach , p. , rossi , m. , 2008b .
distribution of landslides in the upper tiber river basin , central italy .
geomorphology 96 , 105 - 122 .
guzzetti , f. , ardizzone , f. , cardinali , m. rossi , m. , valigi , d. , 2009 .
landslide volumes and landslide mobilization rates in umbria , central italy . earth and planetary science letters 279 , 222 - 229 .
guzzetti , f. , carrara , a. , cardinali , m. , reichenbach , p. , 1999 .
landslide hazard evaluation : a review of current techniques and their application in a multi - scale study , central italy .
geomorphology 31 , 181 - 216 .
guzzetti , f. , reichenbach , p. , malamud , b.d .
turcotte , d.l . , 2002 .
power - law correlations of landslide areas in central italy . earth and planetary science letters 195 , 169 - 183 .
guzzetti , f. , peruccacci , s. , rossi , m. , stark , c.p .
rainfall thresholds for the initiation of landslides in central and southern europe .
meteorology and atmospheric physics 98 , 239 - 267 .
guzzetti , f. , peruccacci , s. , rossi , m. , stark , c.p .
the rainfall intensity - duration control of shallow landslides and debris flows : an update .
landslides 5 , 3 - 17 .
hergarten , s. , 2000 .
landslides , sandpiles and self - organized criticality . natural hazards and earth system sciences 3 , 505 - 514 .
hergarten , s. , 2002 .
self - organized criticality in earth systems springer - verlag , berlin , pp . 275 .
hergarten , s. , 2012 .
topography - based modeling of large rockfalls and application to hazard assessment .
geophysical research letters 39 , l13402 , doi:10.1029/2012gl052090 .
hergarten , s. , neugebauer , h.j . , 1998 .
self - organized criticality in a landslide model .
geophysical research letters 25 , 801 - 804 .
hergarten , s. , neugebauer , h.j .
self - organized criticality in two - variable models .
physical review e 61 , 2382 - 2385 .
katz , o. , aharonov , e. , 2006 .
landslides in vibrating sand box : what controls types of slope failure and frequency magnitude relations ?
earth and planetary science letters 247 , 280 - 294 .
klar , a. , aharonov , e. , kalderon - asael , b. , katz , o. , 2011 .
analytical and observational relations between landslide volume and surface area .
journal of geophysical research 116(f2 ) , f02001 , doi : 10.1029/2009jf001604 .
innes , j.l . , 1983 .
debris flows .
progress in physical geography 7 , 469 - 501 .
iverson , r.m . , 2000 .
landslide triggering by rain infiltration .
water resources research 36 , 1897 - 1910 .
larsen , i.j .
, montgomery , d.r . ,
korup , o. 2010 .
landslide erosion controlled by hillslope material .
nature geoscience 3 , 247 - 251 .
lee , l.p . ,
lan , h.x .
, wu , y.m . , 2012 .
comment on statistical physics of landslides : new paradigm by chen c .- c .
epl 100 , 29001 .
malamud , b.d . ,
turcotte , d.l . ,
guzzetti , f. , reichenbach , p. , 2004 .
landslide inventories and their statistical properties . earth surface processes and landforms 29 , 687 - 711 , doi:10.1002/esp.1064 .
malet , j .-
p . , van asch , t.w . , van beek , r. , maquaire , o. , 2005 .
forecasting the behaviour of complex landslides with a spatially distributed hydrological model .
natural hazards and earth system science 5 , 71 - 85 .
montgomery , d.r . , dietrich , w.e . , 1994 . a physically - based model for the topographic control on shallow landsliding .
water resources research 30 , 1153 - 1171 .
moser , m. , hohensinn , f. , 1983 .
geotechnical aspects of soil slips in alpine regions . engineering geology 19 , 185 - 211 .
newman , m.e.j . ,
power laws , pareto distributions and zipf s law .
contemporary physics 46 , 323 - 351 .
pack , r. , tarboton , d. , goodwin , c. , 1998 .
the sinmap approach to terrain stability mapping . in : 8th congress of the international association of engineering geology , vancouver , pp .
21 pelletier , j.d .
, malamud , b.d . , blodgett , t. , turcotte , d.l .
scale - invariance of soil moisture variability and its implications for the frequency - size distribution of landslides .
engineering geology , 5 , 255 - 268 .
peruccacci , s. , brunetti , m.t . ,
luciani , s. , vennari , c. , guzzetti , f. , 2012 .
lithological and seasonal control on rainfall thresholds for the possible initiation of landslides in central italy .
geomorphology 139 - 140 , 79 - 90 .
piegari , e. , cataudella , v. , di maio , r. , milano , l. , nicodemi , m. , 2006a .
finite driving rate and anisotropy effects in landslide modeling .
physical review e 73 , 026123 - 026129 .
piegari , e. , cataudella , v. , di maio , r. , milano , l. , nicodemi , m. , 2006b . a cellular automaton for the factor of safety field in landslides modelling .
geophysical research letters 33 , l01403 , doi:10.1029/2005gl024759 .
piegari , e. , di maio , r. , milano , l. , 2009 .
characteristic scales in landslide modelling .
nonlinear processes in geophysics 16 , 515 - 523 .
raia , s. , alvioli , m. , rossi , m. , baum , r.l . ,
godt , j.w . ,
guzzetti , f. , 2013 .
improving predictive power of physically based rainfall induced shallow landslides models : a probabilistic approach .
geoscientific model development discussion 6 , 1367 - 1426 .
richards , l.a . , 1931 .
capillary conduction of liquids in porous mediums .
physics 1 , 318 - 333 .
rigon , r. , bertoldi , g. , over , t. , 2006 .
geotop : a distributed hydrological model with coupled water and energy budgets .
journal of hydrometeorology 7 , 371 - 388 .
simoni , s. , zanotti , f. , bertoldi , g. , rigon , r. , 2008 . modelling
the probability of occurrence of shallow landslides and channelized debris flows using geotop - fs .
hydrological processes 22 , 532 - 545 .
stark , c.p . ,
guzzetti , f. , 2009 .
landslide rupture and the probability distribution of mobilised debris volumes .
journal geophysical research 114 , f00a02 , doi : 10.1029/2008jf001008 .
stark , c.p .
, hovius , n. , 2001 .
the characterization of landslide size distributions .
geophysical research letters 28 , 1091 - 1094 .
tarquini , s. , isola , i. , favalli , m. , mazzarini , f. , bisson , m. , pareschi , m.t . ,
boschi , e. , 2007 .
tinitaly/01 : a new triangular irregular network of italy .
annales geophysicae 50 , 407 - 425 .
tarquini , s. , vinci , s. , favalli , m. , doumaz , f. , fornaciai , a. , nannipieri , l. , 2012 .
release of a 10-m - resolution dem for the italian territory : comparison with global - coverage dems and anaglyph - mode exploration via the web .
computers and geosciences 38 , 168 - 170 .
tsallis , c. , 1988 . possible generalization of boltzmann - gibbs statistics . j. statist .
52 , 479 - 487 .
tsallis , c. , 1999 .
nonextensive statistics : theoretical , experimental and computational evidences and connections .
j. phys .
29 , 1 - 35 .
tsallis , c. , 2011 .
nonextensive statistical mechanics : applications to high energy physics .
epj web conf . 13 , 05001 , doi:10.1051/epjconf/20111305001 .
turcotte , d.l . ,
malamud , b.d . ,
guzzetti , f. , reichenbach , p. , 2002 .
self - organization , the cascade model , and natural hazards .
proceedings of the national academy of sciences 99 , supp . 1 , 2530 - 2537 , doi:10.1073/pnas.012582199 .
van den eeckhaut , m. , poesen , j. , govers , g. , verstraeten , g. , demoulin , a. , 2007 .
characteristics of the size distribution of recent and historical landslides in a populated hilly region . earth and planetary science letters 256 , 588 - 603 .
von ruette , j. , lehmann , p. , or , d. , 2013 .
rainfall - triggered shallow landslides at catchment scale threshold mechanics - based modeling for abruptness and localization .
water resources research 49 , 1 - 20 .
wieczorek , g.f . , 1987 .
effect of rainfall intensity and duration on debris flows in central santa cruz mountains , california . in : costa , j.e . , weiczorek , g.f .
( eds ) reviews in engineering geology 7 , geological society of america , boulder , colorado , pp .
|
[ sec : abstract ] natural landslides exhibit scaling properties revealed by power law relationships .
these relationships include the frequency of the size ( e.g. , area , volume ) of the landslides , and the rainfall conditions responsible for slope failures in a region .
reasons for the scaling behavior of landslides are poorly known .
we investigate the possibility of using the transient rainfall infiltration and grid - based regional slope - stability analysis code ( trigrs ) , a consolidated , physically - based , numerical model that describes the stability / instability conditions of natural slopes forced by rainfall , to determine the frequency statistics of the area of the unstable slopes and the rainfall intensity @xmath0
duration @xmath1 conditions that result in landslides in a region .
we apply trigrs in a portion of the upper tiber river basin , central italy .
the spatially distributed model predicts the stability / instability conditions of individual grid cells , given the local terrain and rainfall conditions .
we run trigrs using multiple , synthetic rainfall histories , and we compare the modeling results with empirical evidences of the area of landslides and of the rainfall conditions that have caused landslides in the study area .
our findings revealed that trigrs is capable of reproducing the frequency of the size of the patches of terrain predicted as unstable by the model , which match the frequency size statistics of landslides in the study area , and the mean rainfall @xmath2 conditions that result in unstable slopes in the study area , which match rainfall @xmath3 thresholds for possible landslide occurrence .
our results are a step towards understanding the mechanisms that give rise to landslide scaling properties .
|
the authors wish to acknowledge the financial support of the natural sciences and engineering research council of canada , the canadian institute for advanced research , the canadian foundation for innovation , and the canadian institute for photonic innovations .
|
we demonstrate tunable frequency - converted light mediated by a @xmath0 nonlinear photonic crystal nanocavity .
the wavelength - scale inp - based cavity supports two closely - spaced localized modes near 1550 nm which are resonantly excited by a 130 fs laser pulse .
the cavity is simultaneously irradiated with a non - resonant probe beam , giving rise to rich second - order scattering spectra reflecting nonlinear mixing of the different resonant and non - resonant components . in particular , we highlight the radiation at the sum frequencies of the probe beam and the respective cavity modes .
this would be a useful , minimally - invasive monitor of the joint occupancy state of multiple cavities in an integrated optical circuit .
to realize all - optical logic on an integrated semiconductor chip , it is necessary to develop a toolkit of logical operations ( gates ) which can manipulate optical signals .
these operations serve to generate , condition , and detect optical signals , and are most naturally applied in cavities , where photons can be localized for many optical cycles .
recent work has demonstrated that photonic crystal nanocavities can be fabricated with @xmath1 factors greater than @xmath2 , and are thus capable of storing photons for more than 1 nanosecond @xcite and realizing one - bit delays in bit streams of data @xcite .
cavities can also serve as channel drop filters to transfer a signal from one waveguide to another @xcite . in both classical and quantum optical protocols
, the ability to perform conditional logic requires nonlinear functionality @xcite .
recent advances on this front include all - optical switching via bistability and free - carrier tuning @xcite , dynamic @xmath1 factor control for the release of photons on demand @xcite , and adiabatic frequency conversion through the perturbation of cavity modes @xcite .
the latter approach is effective for generating small frequency shifts in an optical mode , but for larger shifts , the nonlinearity of the material can be exploited to generate new harmonics .
theoretical work has shown that , in principle , loss - less harmonic generation is feasible in doubly - resonant @xmath0 and @xmath3 cavities @xcite .
experimentally , we have shown in our previous work that second - harmonic generation and sum - frequency generation are possible at microwatt - scale powers by leveraging the high @xmath1 factor and small mode volume of a photonic crystal nanocavity @xcite . to perform conditional logic ,
it is important to be able to monitor mode occupation in a cavity . in this letter
, we use nonlinear mode mixing in a photonic crystal nanocavity to demonstrate a method to monitor the populations of modes in _ different _ cavities in a weakly perturbative fashion .
the experiment involves conditionally generating a signal dependent both on the occupation of one microcavity mode ( or many ) and the overlapping presence of a transient ( non - resonant ) signal in the cavity .
in addition to serving as a joint - state monitor , this technique could be used to generate new signals at a frequency of choice .
schematic optical circuit for the joint state nonlinear monitor .
( 1 ) cavity 1 drops channel @xmath4 from a multi - mode waveguide . ( 2 )
@xmath4 is coupled into cavity 2 ( non - resonantly ) .
cavity 2 supports two resonant modes , @xmath5 and @xmath6 , and when @xmath4 arrives , ( 3 ) new frequencies are conditionally generated at @xmath7 and/or @xmath8 , depending on the population of cavity 2 .
, width=226 ] our experiment is motivated by the schematic optical circuit of fig .
[ fig : tune ] .
the figure shows two cavities , * 1 * and * 2 * , with * 1 * acting as a single mode channel drop filter @xcite , and * 2 * acting to store photons at frequencies @xmath5 and @xmath6 .
signal @xmath4 is dropped from a multi - mode channel through the microcavity filter and into the middle waveguide .
when signal @xmath4 reaches cavity * 2 * , it can interact with modes @xmath5 and @xmath6 via the material second - order nonlinearity of the cavity .
new signals are generated at @xmath9 and/or @xmath10 that are conditional on the combined occupation state of cavities * 1 * and * 2*. our experiment illustrates the nonlinear physics involved in generating these conditional signals . the slab which hosts the two - mode phc nanocavity is a 230 nm thick [ 001]-oriented inp free - standing membrane mounted on a glass substrate .
it is excited by a train of @xmath11 130 fs pulses from a 80 mhz optical parametric oscillator ( opo ) ( spectra physics ) with a bandwidth of 100 @xmath12 and an average power of 60 @xmath13w that is focussed through a 100@xmath14 microscope objective .
using our well - established technique @xcite , the resonantly scattered light is collected in reflection , and detected in the cross - polarization with respect to the incident beam using a bomem fourier transform infrared spectrometer .
the second - order radiation is collected in transmission using a 40@xmath14 ( na=0.65 ) microscope objective , and detected using a grating spectrometer and a liquid - nitrogen cooled ccd detector . , width=226 ] we first consider the _ resonant _ linear and nonlinear response of the cavity ( taken to be cavity * 2 * in fig .
[ fig : tune ] ) .
the whole spectrum is excited simultaneously due to the wide bandwidth of the opo , and the modes show up as resonant peaks .
the linear spectrum is shown in blue in fig .
[ fig:2mode](b ) , clearly revealing the two modes of interest .
these modes are nt part of the usual in - bandgap spectrum of the l3-cavity , but rather are quasi - localized modes in the dielectric band continuum @xcite .
they suit our purposes well because of their close spectral spacing ( which allows simultaneous excitation by a single short pulse ) and relatively high @xmath1 factors . the nonlinear spectrum , superimposed on the same plot for comparison purposes ,
is shown in red .
the lowest and highest energy peaks are at exactly twice the frequencies of the microcavity modes evident in the linear spectra , and the central sharp feature , at 2 @xmath14 6315 @xmath12 , is at precisely their sum frequency .
the extra peak at the sum frequency therefore corresponds to the second - order nonlinear interaction of fields resonant in the two modes , and suggests an application . in an integrated optical circuit , detection of this second - order radiation
using narrow band filters would provide a weak , non - destructive means of monitoring the joint occupation state of the microcavity , providing a logical and gate operation .
although this functionality is demonstrated here in the case where a single short pulse is used to simultaneously populate the two modes , the sum - frequency signature would also occur if the modes were populated by independent sources . to demonstrate the conditional mode monitoring operation , we simultaneously irradiate the cavity with short , resonant pulses ( as above ) and a second _ non - resonant _ excitation beam consisting of longer , picosecond pulses tuned far off resonance with the cavity modes .
this two - color source is readily available from the unfiltered `` signal '' beam output of the opo when it is tuned near the degeneracy point ( where both signal and idler frequencies are close to half the pump frequency ) .
an example of this unfiltered spectrum when the signal is tuned to 6335 @xmath12 is shown in figure [ fig : sigidler](a ) .
the short opo signal pulses are accompanied by relatively long ( a few ps ) opo idler pulses .
the idler pulses in this spectrum are dominated by a single spectral feature at @xmath15 @xmath12 , but there are several smaller intensity peaks at lower frequency which are also revealed in the second - order nonlinear processes discussed below . the resonant scattering spectrum from the two - mode microcavity studied in fig .
[ fig:2mode ] , when excited by the source of fig .
[ fig : sigidler](a ) , is shown in fig .
[ fig : sigidler](b ) .
the resonant enhancement of the mode features is evident in comparison to the non - resonantly scattered idler .
( a ) spectrum of the opo laser beam , showing the femtosecond ( fs ) pulse at 6335 @xmath12 and the prominent picosecond ( ps ) pulse at the frequency of the idler .
a number of smaller features are also visible near the main idler peak .
( b ) resonant scattering spectrum from the microcavity studied in fig .
[ fig : idler ] , using the source as in ( a ) , but slightly detuned .
, width=188 ] spectra given by the interaction of a non - resonant , narrow linewidth idler pulse ; a broad , resonant laser pulse ; and a two - mode ppc microcavity .
the four second - order spectra ( red ) are plotted at half the measured energy .
the signal pulse is tuned to higher energy for each spectrum s@xmath16-s@xmath17 , and so the idler tunes to lower energy .
the amplitude of the low energy region has been multiplied by 10 for clarity .
the solid blue curve shows the laser spectrum scattered from a non - textured region of the sample , as in fig .
[ fig : sigidler](a ) .
the dashed blue curves schematically show the signal pulse spectrum tuning to higher energies.,width=340 ] we now use this spectrum to explore the scheme shown in fig .
[ fig : tune ] .
the off - resonant ps pulse from the idler plays the role of @xmath4 , while the two modes are associated with @xmath5 and @xmath6 .
figure [ fig : idler ] shows a series of four second - order spectra from the microcavity supporting the quasi - localized modes discussed in fig . [ fig:2mode](b ) irradiated by the two - color source .
the opo is tuned to four different values of the center wavelength to acquire the spectra , which show three principal groups of features that are marked by lines a , b , and c to guide the eye .
feature a and the broad background in group c ( the fit of which is plotted separately as a dashed blue line ) , are due to non - resonant second - harmonic generation of the ps and fs features , respectively , in the excitation spectra .
the three sharp features in group c that do not shift , and the two sharp features in group b that shift at half the rate of the excitation beam(s ) , are specific to the microcavity modes .
they all arise from second - order processes involving the fields `` trapped '' in at least one of the modes .
the three ( fixed ) peaks in group c correspond to the mode shg ( @xmath18 , @xmath19 ) and sfg ( @xmath20 ) features , as in figure [ fig:2mode ] .
the features in group b are then easily understood to result from the second - order radiation from the two field distributions trapped in the microcavity modes respectively interacting with the ps pulses that irradiate the cavity during the ring - down to produce peaks at @xmath9 and @xmath10 . to understand the difference in the shift rate between features a and b ,
consider a ps pulse at @xmath21 interacting with a microcavity mode at @xmath22 .
when the ps pulse is tuned from @xmath21 to @xmath23 , the second - order feature a shifts from 2@xmath21 to @xmath24 , which is a shift of @xmath25 , whereas feature b shifts from @xmath26 to @xmath27 , a shift of just @xmath28 .
the lifetime of each cavity mode is about 1 ps , and so the pulsed excitation probes the impulse response of the cavity ( i.e. there is minimal local field enhancement ) . the coupling efficiency into the cavity of the 60 @xmath13w average power resonant excitation beam is @xmath29 @xcite ; therefore , accounting for the 80 mhz repetition rate of the opo , approximately @xmath30 photons are resonantly coupled into the cavity with each pulse .
it is useful to consider what equivalent , critically - coupled cw - excitation or quasi - cw - excitation would create mode populations sufficient to yield an equivalent nonlinear response . in the cw limit , the local field intensity in the cavity is enhanced by a factor of @xmath1 . assuming the cw - beam is coupled into the cavity from a 600 nm @xmath14 200 nm single - mode ridge waveguide , the same nonlinear response could be achieved by a cw power of about 20 @xmath13w .
the processes illustrated here demonstrate that the fields stored in microcavity modes can be nonlinearly mixed with non - resonant signals to produce sum - frequency radiation . if the non - resonant fields were generated from light previously stored in a different microcavity , as in the schematic of fig .
[ fig : tune ] , this principle could be used to conditionally generate information at new frequencies which depend on the joint occupation state of two different microcavities .
|
as is well - known , how to solve effectively linear systems is a very important problem in scientific and engineering fields .
many of linear solvers have been researched such as gaussian elimination @xcite , relaxation methods @xcite , row - action iteration schemes @xcite and ( block ) krylov subspace @xcite .
recently , a low communication condensation - based linear system solver utilizing cramer s rule is presented in @xcite . as the authors stated that unique combination between cramer s rule and matrix condensation techniques yields an elegant parallel computing architectures , by constructing a binary , tree - based data flow in which the algorithm mirrors the matrix at critical points during the condensation process .
moreover , the accuracy and computational complexity of the proposed algorithm are similar to lu - decomposition @xcite . in this paper , we will continue research this kind of parallel algorithms and give some theoretical analysis and a generalized chi s determinant condensation process , which perfect the corresponding conclusions .
this paper is organized as follows . in second 2
, we will review some more general determinant condensation algorithms
sylvester s determinant identity , and then give theoretical basis on the above parallel computing architectures @xcite , which shows the negation in mirroring process is not necessary to arrive at the correct answer .
moreover , a more general scheme utilizing cramer s rule and matrix condensation techniques is also given .
in addition , the scheme suitable for parallel computing on the sylvester s identity is proposed in section 3 .
finally , a simple example is used to illustrate this new algorithm in section 4 .
throughout this section , we mainly consider an @xmath0 matrix @xmath1 @xmath2 with elements @xmath3 and determinant @xmath4 , also written @xmath5 .
recently , a chi condensation method @xcite is applied to solve large linear systems in @xcite .
in fact , the prototype of this method may be traced back to the following sylvester s determinant identity for calculating a determinant of arbitrary order in 1851 .
[ th2.1](sylvester s identity,@xcite ) .
let @xmath1 be an @xmath0 matrix over a commutative ring .
for a submatrix @xmath6 , @xmath7 of @xmath8 , set @xmath9.\ ] ] let @xmath10 , @xmath11 .
then @xmath12 specially when @xmath13 is an invertible matrix , we have that @xmath14 [ co2.2](chi s method , @xcite ) . for an @xmath0 matrix @xmath1 with @xmath15 ,
let @xmath16 be the @xmath17 matrix defined by @xmath18 then @xmath19 obviously , the above theorem [ th2.1 ] reduces a matrix of order @xmath20 to order @xmath21 to evaluate its determinant . repeating the procedure numerous times can reduce a large matrix to a small one , which is convenient for the calculation .
this process is called by condensation method @xcite . as an example of chi s condensation
, the paper @xcite considers the following @xmath22 matrix : @xmath23 in fact , the above condensation processes are not only used to evaluate determinants but also can be used to solve linear systems . for example , one can derive the following equivalence relation on the solution formula of linear systems .
[ th2.2](equivalence relation ) .
the linear system in the form @xmath24 ( where @xmath1 is an @xmath0 invertible coefficient matrix ) has the same corresponding solution as the linear system @xmath25 , where @xmath26 is defined as in theorem [ th2.1 ] , @xmath27^t}$ ] and @xmath28^t}$ ] . here
@xmath29,j = k + 1 , \ldots , n.\ ] ] according to theorem [ th2.1 ] or eq . , we know that there exists a constant @xmath30 between the determinant of @xmath8 and the determinant of @xmath31 , which is only dependent on the given submatrix @xmath13
. therefore , for any given submatrix @xmath13 , there also exists the same constant @xmath30 between the determinant of @xmath32 ( @xmath33 ) , the matrix @xmath8 with its @xmath34 column replaced by @xmath35 , and the determinant of @xmath36 .
thus , by cramer s rule , we have that @xmath37 the conclusion holds . obviously ,
when the submatrix @xmath13 is singular , the solution of linear systems can not be evaluated by this method . since interchanging the @xmath38th and @xmath20th rows and the @xmath39th and @xmath20th columns of linear systems has only an effect on the order of the unknowns @xmath40 , which has no effect on the whole solution @xmath41 .
therefore , we may obtain the following more general conclusion . for convenience , we firstly define the ordered index list @xmath42 for any positive integer @xmath20 . for two ordered index ( i.e. , for any @xmath43 )
lists @xmath44 and @xmath45 , we denote the corresponding complementary ordered index lists by @xmath46 and @xmath47 , respectively .
that is , @xmath48 .
[ th3.4 ] .
let @xmath1 be an @xmath0 matrix and @xmath49 be a fixed integer @xmath50 . @xmath51 and
@xmath52 are two ordered index lists .
we denote the corresponding submatrix , extracted from @xmath8 , as @xmath53 = a\left [ { \begin{array}{*{20}{c } } { { i_1 } , \ldots , { i_k}}\\ { { j_1 } , \ldots , { j_k } } \end{array } } \right ] \triangleq \left [ { \begin{array}{*{20}{c } } { { a_{{i_1}{j_1 } } } } & \cdots & { { a_{{i_1}{j_k}}}}\\ \vdots & \ddots & \vdots \\ { { a_{{i_k}{j_1 } } } } & \cdots & { { a_{{i_k}{j_k } } } } \end{array } } \right].\ ] ] suppose that the invertible submatrix @xmath54 $ ] in the theorem [ th2.1 ] , then the linear system @xmath24 has the same corresponding solutions as the linear system @xmath55 , where @xmath56 is defined as the subset of @xmath57 with the index coming from @xmath47 .
according to the above theorem , one can easily see that though the condensation process removes information associated with discarded columns , we may obtain certain variables values by controlling the elements in the set @xmath47 , see example [ e4.1 ] .
in addition , the matrix mirroring and the negation of matrix mirroring process in @xcite are also not necessary to arrive at the correct answer , since we may obtain the similar parallel computing process by condensing the index set @xmath47 from both sides ( left and right ) , see figure 1 .
@xmath58 [ fig1 ] similar to @xcite , copying occurs with the initial matrix and then each time a matrix is reduced in half .
an @xmath59 matrix is copied when it reaches the size of @xmath60 .
once the matrix is copied , there is double the work . in other words , two @xmath60 matrices
each require a condensation .
obviously , the amount of work for two matrices of half the size is much lower than that of one @xmath61 matrix , which avoids the @xmath62 growth pattern in computations .
this is due to the @xmath63 nature of the condensation process ( see @xcite ) .
similarly , one may consider a scenario in which the algorithm creates more than two matrices during each copying step , according to corollary [ th3.4 ] . on its computational complexity and more details ,
see @xcite .
the sylvester s identity [ th2.1 ] reduces a matrix of order @xmath20 to order @xmath21 when evaluating its determinant . since when @xmath64
, it is just the chi s method .
therefore , for convenience , we call the sylvester s identity a * k - chi s method * from now on .
as have been shown above , repeating the procedure numerous times can reduce a large matrix to a size convenient for the computations .
however , in order to condense a matrix from @xmath61 to @xmath65 , the core calculation is repeated @xmath66 times .
obviously , this is very expensive .
in fact , we may parallel computing each row of the matrix @xmath26 in , since @xmath13 or @xmath67 $ ] is common to each element of the row and may be calculated but once for each row via expanding by the last column .
for example , for the @xmath68-th row @xmath69 of @xmath31 , we may write @xmath70\\ = \left [ { \underbrace { a_0^{p,1},a_0^{p,2 } , \cdots , a_0^{p , k},|{a_0}|}_{k + 1 } } \right ] \cdot \left [ { \begin{array}{*{20}{c } } { { a_{1,k + 1}}}&{{a_{1,k + 2 } } } & \cdots & { { a_{1,n}}}\\ { { a_{2,k + 1}}}&{{a_{2,k + 2 } } } & \cdots & { { a_{2,n}}}\\ \cdots & \cdots & \cdots & \cdots \\ { { a_{p , k + 1}}}&{{a_{p , k + 2 } } } & \cdots & { { a_{p , n } } } \end{array } } \right ] , \end{array}\ ] ] where @xmath71 therefore , only @xmath49 determents @xmath72 ( @xmath73 ) and a common @xmath74 determent @xmath75 are needed for each row of the matrix @xmath26 . therefore , the matrix @xmath31 is essentially suitable for parallel computations since the each row of matrix @xmath31 may be independently computed by .
see example [ e3.3 ] below .
[ e3.3 ] consider the following four order determent @xmath76 let @xmath77 $ ] , then @xmath78 , and @xmath79\left [ { \begin{array}{*{20}{c } } 3&1\\ { - 1}&0\\ 1&5 \end{array } } \right ] = \left [ { \begin{array}{*{20}{c } } { 36}&{58 } \end{array } } \right];\\ { \alpha _ 4 } = \left [ { -\left| { \begin{array}{*{20}{c } } \mathbf{-3}&\mathbf{3}\\ 4&2 \end{array } } \right|,-\left| { \begin{array}{*{20}{c } } 1 & { - 2}\\ \mathbf{-3}&\mathbf{3 } \end{array } } \right|,\mathbf{10 } } \right]\left [ { \begin{array}{*{20}{c } } 3&1\\ { - 1}&0\\ 1&2 \end{array } } \right ] = \left [ { \begin{array}{*{20}{c } } { 61}&{38 } \end{array } } \right ] .
\end{array}\ ] ] therefore , @xmath80 from here , we note that only six @xmath81 determinants is needed .
however , chi s method will require fourteen @xmath81 determinants to be computed .
in addition , comparing with the gaussian elimination , our method increases only two multiplications .
but gaussian elimination method is not too suitable for parallel computing .
thus , the whole computational amount on the matrix @xmath26 will be much less than that involved in the old process of computation @xcite . concretely speaking ,
if we denote the total of multiplications / diversions on the @xmath49-order determinant @xmath82 by @xmath83 , then the total of multiplications / diversions by using the k - chi s method is about @xmath84 + km\left [ { k + 2k + \ldots + ( n - k ) } \right ] + \frac{{n - k}}{k}(m + 1 ) + m\\ \approx o\left ( { \frac{1}{3}(1 + \frac{1}{k}){n^3 } } \right )
. \end{array}\ ] ] similarly , the computational complexity of other algorithms is also described as follows , see table 1 .
[ t1 ] .comparisons of the computational complexity for different algorithms on determinant calculations in the nonparallel setting . [ cols="<,^,^",options="header " , ] from table 1 , we note that additions / subtractions on these algorithms are almost the same .
however , multiplications / diversions mainly depend on the parameter @xmath49 for k - chi s condensation method . but this does not show that the total computational complexity on k - chi s method is tending to decrease with the @xmath49 increasing , since the core loop of the k - chi s condensation method involves the calculation of @xmath74 determinants for each element of the matrix during condensation . normally , this would necessitate the standard computational workload to calculate the @xmath49-order determinant , i.e. , @xmath85 multiplications / divisions and @xmath85 additions / subtractions , using a method such as gaussian elimination @xcite . therefore , the parameter @xmath49 is not the better for the bigger number , see the following experimental results figure 1 and 2 on the @xmath86-order and @xmath87-order determinants , respectively .
the small subgraphs in fig . 1 and 2 show the optimal parameter @xmath49 value ranges .
for example , the optimal parameter @xmath49 is approximately ten for a @xmath87-order determinant . in addition , for matrices of different dimensions , we specifically compute the optimal parameters @xmath49 , we find the optimal parameter values increasing as the matrix dimension increases . but
this increase is still relatively slow , see fig .
[ fig1 ] and the number of multiplication / diversion for a 5000-order determinant on k - chi s condensation method.,width=710,height=664 ] [ fig2 ] and the number of multiplication / diversion for a 20000-order determinant on k - chi s condensation method.,width=748,height=664 ] [ fig3 ] and the dimension of matrices , left : only consider multiplications ; right : consider all computational complexity.,width=768,height=768 ] since the optimal parameter @xmath49 is usually small , by , we may normalize the each row of matrix @xmath26 by dividing the determinant of @xmath13 , which will further reduce the computational complexity of k - chi s condensation method , see example [ e4.1 ] .
as is well - known , the classical cramer s rule states that the components of the solution to a linear system in the form @xmath24 ( where @xmath1 is an @xmath0 invertible coefficient matrix ) are given by @xmath88 where @xmath40 is the @xmath89th unknown . in @xcite , an algorithm based on chi s condensation and cramer s rule for solving large - scale linear systems is achieved by constructing a binary , tree - based data flow in which the algorithm mirrors the matrix at critical points during the condensation process . however , according to the above corollary [ th3.4 ] , one may obtain certain unknowns values by * freely controlling the elements in the set * @xmath47 without matrix mirroring , see example [ e4.1 ] .
this also makes it more easily for more cpus to be used in computing process and even without any communication . at the same time
, the scheme also reduce the memory space .
[ e4.1 ] solve the following six - order linear system @xmath90\left [ { \begin{array}{*{20}{c } } { { x_1}}\\ { { x_2}}\\ { { x_3}}\\ { { x_4}}\\ { { x_5}}\\ { { x_6 } } \end{array } } \right ] = \left [ { \begin{array}{*{20}{c } } \textcolor[rgb]{0,0,1}{1}\\ \textcolor[rgb]{0,0,1}{-1}\\ \textcolor[rgb]{0,0,1}{1}\\ \textcolor[rgb]{0,0,1 } { - 1}\\ \textcolor[rgb]{0,0,1}1\\ \textcolor[rgb]{0,0,1 } { - 1 } \end{array } } \right].\ ] ] let @xmath91 and @xmath92 , then @xmath93 . denote @xmath94 } \right| = \left| { \begin{array}{*{20}{c } } 1&5&9\\ 2&0&0\\ 3&5&0 \end{array } } \right|.\ ] ] then , @xmath95 and @xmath96\left [ { \begin{array}{*{20}{c } } 3&7&{11}&\textcolor[rgb]{0,0,1}{1}\\ 0&0&9&\textcolor[rgb]{0,0,1 } { - 1}\\ 0&7&7&\textcolor[rgb]{0,0,1}{1}\\ \mathbf{0}&\mathbf{8}&\mathbf{5 } & { \textcolor[rgb]{0,0,1 } { \textbf{- 1 } } } \end{array } } \right]\\ & = & \left [ { \begin{array}{*{20}{c } } { 0,}&{-36,}&{-468,}&\textcolor[rgb]{0,0,1}{-180 } \end{array } } \right ] ; \end{array}\ ] ] @xmath97\left [ { \begin{array}{*{20}{c } } 3&7&{11}&\textcolor[rgb]{0,0,1}{1}\\ 0&0&9&\textcolor[rgb]{0,0,1 } { - 1}\\ 0&7&7&\textcolor[rgb]{0,0,1}{1}\\ \mathbf{0}&\mathbf{0}&\mathbf{3}&\textcolor[rgb]{0,0,1}{\textbf{1 } } \end{array } } \right]\\ & = & \left [ { \begin{array}{*{20}{c } } { 0,}&{0,}&{-1755,}&\textcolor[rgb]{0,0,1}{315 } \end{array } } \right ] ; \end{array}\ ] ] @xmath98\left [ { \begin{array}{*{20}{c } } 3&7&{11}&\textcolor[rgb]{0,0,1}{1}\\ 0&0&9&\textcolor[rgb]{0,0,1 } { - 1}\\ 0&7&7&\textcolor[rgb]{0,0,1}{1}\\ \mathbf{5}&\mathbf{3}&\mathbf{1}&{\textcolor[rgb]{0,0,1 } { \textbf{- 1 } } } \end{array } } \right]\\ & = & \left [ { \begin{array}{*{20}{c } } { 390,}&{-234,}&{-2132,}&\textcolor[rgb]{0,0,1}{20 } \end{array } } \right ] .
\end{array}\ ] ] therefore , we need only solve the condensed linear system @xmath55 , i.e. , @xmath99\left [ { \begin{array}{*{20}{c } } { { x_2}}\\ { { x_4}}\\ { { x_6 } } \end{array } } \right ] = \left [ { \begin{array}{*{20}{c } } \textcolor[rgb]{0,0,1}{-180}\\ \textcolor[rgb]{0,0,1}{315}\\ \textcolor[rgb]{0,0,1}{20 } \end{array } } \right].\ ] ] by cramer s rule or gaussian elimination , the solution of above sub - linear system is @xmath100 which is also the corresponding solution of original linear system . similarly , let @xmath101 and @xmath102 , then we may also obtain the solution of the unknown @xmath103 , @xmath104 and @xmath105 : @xmath106 in addition , we may continue condense the above @xmath107 , @xmath108 and @xmath109 . for example , we condense them from right side for @xmath93 . without loss of generality , we may let @xmath110 , @xmath111 , then @xmath112 and we have @xmath113\left [ { \begin{array}{*{20}{c } } { 390 } & { - 234}&\textcolor[rgb]{0,0,1}{20}\\ 0 & { - 36}&\textcolor[rgb]{0,0,1}{-180 } \end{array } } \right ] = \left [ { 182520 , - 32760,\textcolor[rgb]{0,0,1}{393120 } } \right];\\ \alpha _ 5 ' = \left [ { 1755 , - 2132 } \right]\left [ { \begin{array}{*{20}{c } } { 390 } & { - 234}&\textcolor[rgb]{0,0,1}{20}\\ 0&0&\textcolor[rgb]{0,0,1}{315 } \end{array } } \right ] = \left [ { 684450 , - 410670,\textcolor[rgb]{0,0,1}{- 636480 } } \right ]
. \end{array}\ ] ] by gaussian elimination , we obtain the solution of the above linear system : @xmath114 moreover , to further reduce the computational complexity of k - chi s condensation method , we may normalize the each row of matrix @xmath26 by dividing the determinant of @xmath13 .
for example , the above @xmath115 and @xmath116 may be written as @xmath117\left [ { \begin{array}{*{20}{c } } { 390}&{-234}&\textcolor[rgb]{0,0,1}{20}\\ 0 & { - 36}&\textcolor[rgb]{0,0,1 } { - 180 } \end{array } } \right ] = \left [ { -3510/41,630/41 , \textcolor[rgb]{0,0,1}{-7560/41 } } \right];\\ \alpha _ 5 ' = \left [ { 1755/ - 2132,\textbf{1 } } \right]\left [ { \begin{array}{*{20}{c } } { 390}&{-234}&\textcolor[rgb]{0,0,1}{20}\\ 0&0&\textcolor[rgb]{0,0,1}{315 } \end{array } } \right ] = \left [ { -26325/82,15795/82,\textcolor[rgb]{0,0,1}{12240/41 } } \right ] . \end{array}\ ] ] from the above example , we know that applying gaussian elimination method instead of cramer s rule to solve the small sub - linear system @xmath55 is also very convenient .
from the above discussion , one can see that unique utilization of matrix condensation techniques yields an elegant process that has promise for parallel computing architectures .
moreover , as was also mentioned in @xcite , these condensation methods become extremely interesting , since they still retain an @xmath118 complexity with pragmatic forward and backward stability properties when they are applied to solve large - scale linear systems by the cramer s rule or gaussian elimination . in this paper ,
some condensation methods are introduced and some existing problems on these techniques are also discussed . though the condensation process removes information associated with discarded columns , this makes the computation of linear systems become feasible by more freely parallel process . * acknowledgements . *
_ partial results of this paper were completed while the first author was visiting the college of william and mary in 2013 .
the first author is very grateful to professor chi - kwong li for the invitation to the college of william and mary . _
n. galoppo , n.k .
govindaraju , m. henson , and d. manocha .
lu - gpu : efficient algorithms for solving dense linear systems on graphics hardware .
proceedings of the acm / ieee sc2005 conference on high performance networking and computing , university of north carolina at chapel hill , 2005 .
|
the object of this paper is to introduce a new and fascinating method of solving large linear equations , based on cramer s rule or gaussian elimination but employing sylvester s determinant identity in its computation process .
in addition , a scheme suitable for parallel computing is presented for this kind of generalized chi s determinant condensation processes , which makes this new method have a property of natural parallelism .
finally , some numerical experiments also confirm our theoretical analysis .
_ keywords _ : sylvester s determinant identity ; cramer s rule ; chi s method ; parallel process .
|
micromegas ( for micro mesh gaseous structure ) is a parallel - plate detector invented by i. giomataris in 1995 @xcite .
it consists of a thin metallic grid ( commonly called mesh ) and an anode plane , separated by insullated pillars .
both structures define a small gap ( between 20 and 300 @xmath2 m ) , where primary electrons generated in the conversion volume are amplified , applying moderate voltages at the cathode and the mesh . this technology has proven to have many advantages like its high granularity , good energy and time resolution , easy construction , little mass and gain stability .
the first detectors were built by screwing two different frames : the anode plane and the metallic grid , under which the insullated pillars were electroformed . by applying voltages to both structures ,
the intense electric field pulled down the mesh and the flatness was thus defined by the height of the pillars , with an accuracy better than 10 @xmath2 m . the good flatness and parallelism between the anode and the mesh
was obtained only if the delicate operation of screwing was successful . to avoid this operation , two different technologies ( called bulk @xcite and microbulk @xcite )
have been developped so that the readout plane and the mesh formed a single integrated structure . in the microbulk technology ,
the raw material is a thin flexible polyimide foil with a thin copper layer on each side .
the foil is glued on top of a rigid substrate that provides the support of the micro - structure and carries the anode strips or pixels . in the first step ,
a thin photoresistive film is laminated on top of the kapton foil and is insolated by uv light to produce the required mask .
the copper is then removed by a standard lithographic process , generating the pattern of a thin mesh .
the polyimide between the mesh and the readout plane is then etched and removed under the mesh holes , but it remains almost intact elsewhere @xcite .
the resulting detector has a more homogeneous amplification gap and a better mesh than the classical micromegas , reducing the avalanche fluctuations and improving the energy resolution .
there is also a reduction of the border regions , eliminating dead zones and allowing its application in imaging with x - rays or neutrons .
moreover , the detector is made of kapton and copper , two of the materials with the best levels of radiopurity .
these materials define a level of contamination below pmts@xcite .
the only possible drawbacks of this technique are its lack of robustness and the limited maximum size ( @xmath3 @xmath4 ) , due to the actual equipment .
nevertheless , the improved detectors have been applied in several experiments like ntof @xcite , cast @xcite or next@xcite .
the microbulk technology shows the best energy resolution of mpgd detectors , as good as 11% fwhm at 5.9 kev in ar+5%ic@xmath0h@xmath1 , improving the values around 14% fwhm reached by proportional counters two decades ago @xcite .
the reference gas in tests was fixed by studies of that time , as a high gain and a good energy resolution was obtained . however , the improved performances of microbulk detectors give the oportunity to study in detail the influence of the gas mixture on these parameters and to verify the conclusions obtained with proportional counters . moreover ,
if microbulk detectors could work at higher gains ( @xmath5 ) keeping a good energy resolution , this technology could be used in sub - kev applications , like in synchroton radiation and dark matter searches where the low energy threshold is crutial . in this article , we present the performance of microbulk detectors in argon and neon - based mixtures .
the detectors tested respectively have a gap size of 50 and 25 @xmath2 m .
the first gap is the more common used in the new microbulk detectors and the second one is expected to show better performances at high pressure gases @xcite . in the second section , the setup and the procedure are described in detail .
the features of microbulk detectors in argon- and neon - based mixtures with three different quenchers are described in sections three and four . finally , we present the rose - korff gain model for micromegas , the dependencies obtained for some gas parameters and we discuss the validity of this model and its possible relation with the energy resolution .
we finish this report with some conclusions and an outlook of the actual developments .
the vessel used in these tests ( figure [ fig : setup2 ] , left ) was specifically designed to characterize a maximum of three micromegas detectors ( not necessary microbulk ) in the same gas conditions , with a maximum surface of @xmath6 mm@xmath7 .
it consists of an aluminium box , with an internal volume of @xmath8 mm@xmath9 and a wall thickness of 8 mm .
the chamber is equipped with two gas entrances to circulate the desired gas .
detectors are placed on a plastic support situated inside the vessel , which electrically isolates them from the inner walls . for each detector
, there is a mesh frame screwed to the plastic plate , which is used as a drift cathode and defines a drift distance of 10 mm .
the drift , mesh and anode voltages are respectively extracted by a high voltage ( shv ) and two signal ( bnc ) feed - throughs for each detector . on the top cap of the chamber , several holes allow the calibration by x - ray sources situated outside .
these holes are covered by an aluminized mylar films , that are scotched to the inner wall of the cap with kapton and create the gas - tightness of the vessel . during at least one hour , a flow of the desired mixture
was circulated by the chamber . in the case of mixtures with cyclohexane
, the base gas was forced to pass by a glass vessel , filled with this liquid quencher ( figure [ fig : setup2 ] , right ) .
the gas concentration was defined by the liquid temperature , which was kept constant by a refrigerator in which the vessel was immersed .
this temperature could not be higher than the ambient one to avoid possible quencher condensations inside the vessel .
for a temperature of 25@xmath10c , this condition fixes a maximum concentration of 11% .
the detectors were illuminated by an iron @xmath11fe source ( x - rays of 5.9 kev ) keeping the same gas flow .
the mesh voltage was typically varied from 200 to 600 v and the drift one from 280 to 5000 v , depending on the type of gas .
both voltages were powered independently by a caen n471a module .
the avalanche induces a negative signal in the mesh , which is read out by an ortec 142c preamplifier .
the preamplifier output was fed into an ortec 472a spectroscopy amplifier and subsequently into a multichannel analyzer amptek mca-8000a for spectra building . for reducing systematic errors
, each spectrum has a minimum of @xmath12 points .
two microbulk detectors respectively with a gap size of 25 and 50 @xmath2 m have been characterized .
they have a circular shape ( 35 mm of diameter ) and a single non - segmented anode .
their specific references are m50.70.35.04 and m25.50.25.09 , which respectively correspond to a pitch distance of 70 and 50 @xmath2 m and a hole diameter of 35 and 25 @xmath2 m .
these particular detectors have showed an energy resolution of 11.7% and 11.8% fwhm at 5.9 kev in ar + 5% ic@xmath0h@xmath1 , values which are close to the best value observed for microbulk detectors ( 11% fwhm ) .
the drift voltage was firstly varied in our tests for a fixed mesh voltage to obtain the electron transmission curve , that shows the mesh transparency to primary ionization .
these curves are shown for the argon - isobutane case and both detectors in figure [ fig : transariso ] .
microbulk detectors show a plateau of maximum electron transmission for a wide range of ratios of drift - to - amplification fields . at higher drift fields ,
the mesh stops being transparent for the primary electrons generated in the conversion volume and both the gain and the energy resolution degrade .
the specific maximum value for the ratio depends on the diffusion coefficient , being wider the plateau for higher quantities of the quencher .
this fact is in agreement with the shorter plateaus observed in pure gases @xcite and the wider plateaus measured in gases with low diffusion coefficients @xcite .
m in argon - isobutane mixtures .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m in argon - isobutane mixtures .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] however , the detector of 25 @xmath2m - thickness gap does not show the same behaviour , as shown in figure [ fig : transariso ] ( right ) . for low isobutane concentrations , there is a short plateau of maximum mesh transparency but it starts at relative high drift field . at high isobutane concentrations , this plateau disappears and there is a steady increase of the detector gain with the drift field . this effect has been verified by other microbulk of the same gap , a hole pitch between 50 and 100 @xmath2 m and a hole diameter between 20 and 30 @xmath2 m .
a possible explanation is that the mesh may not be completely transparent for electrons for the pitch and hole diameter considered .
indeed , the best values of the energy resolution are reached at low drift fields ( figure [ fig : eresdriftariso ] , right ) .
the dependence of the energy resolution with the ratio of fields is shown in figure [ fig : eresdriftariso ] for argon - isobutane mixtures . for a gap size of 50 @xmath2 m
, the energy resolution shows its best value for the same drift fields for which the electron transmission is the optimum .
there is also a dependence on the isobutane concentration . for values below 5% ,
the energy resolution is worse as not all the uv photons produced in the avalanche are absorbed @xcite . for values between 5 and 7%
, the energy resolution reaches its optimum and then degrades at higher isobutane concentrations .
there is an increase of the gain fluctuations due to a higher proportion of scattering processes than ionization ones @xcite .
m in argon - isobutane mixtures .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m in argon - isobutane mixtures .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] in the case of the 25 @xmath2m - thickness - gap detector , the energy resolution reaches its best values for the lowest drift fields .
this fact supports the hypothesis that the mesh is not totally transparent for electrons .
there is also a dependence on the isobutane concentrations , but not so strong as for the former detector .
the energy resolution degrades from 12% up to 14.5% fwhm when the isobutane concentration passes from 7% to 30% , which is less than the 19% fwhm measured with the 50 @xmath2m - thickness - gap detector .
the dependence of the peak position with the mesh voltage generates the gain curves , shown in figure [ fig : gainariso ] for argon - isobutane mixtures .
the maximum gain before the spark limit is respectively @xmath13 and @xmath14 for the detectors with a gap size of 50 and 25@xmath15 m .
this fact happens for all mixtures in the first case and only for isobutane concentrations over 10% for the second one .
apart from that , the gain curves for low quantities of isobutane show a deviation from the rose and korff gain model @xcite at high amplification fields .
this over - exponential behaviour is due to the low quencher concentrations , which can not avoid the production of secondary avalanches by the uv photons generated in the primary avalanche @xcite .
m ( right ) in argon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m ( right ) in argon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] the energy resolution has also a dependence with the amplification field ( figure [ fig : eresgainariso ] ) , and more specifically with the gain , which was already observed with proportional counters @xcite . at low gains ,
the energy resolution degrades because the signal starts being comparable to the noise . at high gains , the resolution worsens due to the increase of the gain fluctuations .
this effect disappears at high isobutane concentrations .
however , the energy resolution is also worse , due to the larger gain fluctuations @xcite .
m - thickness - gap ( right ) in argon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m - thickness - gap ( right ) in argon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] the results obtained for other quenchers are similar to the ones presented for argon - isobutane .
we will just focus on the main differences .
the first particularity ( shown in figure [ fig : transm50argon ] ) is that the plateau of maximum electron transmission is wider for mixtures with cyclohexane and narrower for ethane , in comparison to isobutane .
the argon - ethane result is consistent with its higher diffusion coefficient for the same percentage of quencher @xcite . in the case of cyclohexane , no data is available on its gas properties .
m - thickness - gap in argon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] m - thickness - gap in argon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] as shown in figure [ fig : gainarcycloeth ] , the maximum gain before the spark limit is respectively @xmath13 and @xmath16 for mixtures with ethane and cyclohexane and the 50 @xmath2m - thickness - gap detector . only for cyclohexane , gains higher than in argon - isobutane mixtures are reached .
we have also observed that the required voltage to reach the same gain is lower for cyclohexane , than for isobutane or ethane .
for instance , an amplification field of 65 kv / cm must be applied to reach a gain of @xmath14 in ar+5%ic@xmath0h@xmath17 .
this field is only 61 kv / cm in the case of cyclohexane and increases up to 72 kv / cm for ethane .
m - thickness - gap in mixtures of argon with ethane ( left ) and cyclohexane ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] m - thickness - gap in mixtures of argon with ethane ( left ) and cyclohexane ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] in table [ tab : eresargon ] we show the best values for the energy resolution obtained in each kind of mixture .
we first note that the optimum performance of the 50 @xmath2m - thickness - gap microbulk is reached with quencher concentrations around 5% . for the gap of 25 @xmath2 m
, this percentage rises up to 10% . nevertheless , the results are quite similar for all mixtures . from some measurements made with proportional counters in @xcite ,
it was concluded that quenchers whose ionization threshold were similar to the first metastable level of argon ( 11.5 ev ) should show a higher gain and a better energy resolution .
this improvement is related to a higher penning effect @xcite , i.e. , an energy transfer from the base gas atom to the quencher molecule , resulting in an increase of the avalanche ionization . however , this fact does not agree in the case of cyclohexane , as even if its ionization threshold ( 9.9 ev ) is a bit low , the gain is higher with no degradation on the energy resolution .
recent studies @xcite support the main role of the non - metastable levels in penning effect .
quenchers whose ionization threshold is lower than this level should show the same energy resolution .
the higher gain observed in argon - cyclohexane mixtures could be related to a little contribution of argon excimers , created for energies between 8.9 and 10.3 ev .
@xmath18 we finally present in figure [ fig : eresgainm50 ] the dependence of the energy resolution with the gain for the microbulk of 50 @xmath2m - thickness - gap in mixtures of argon with ethane and cyclohexane .
the maximum gain before the spark limit is respectively @xmath13 and @xmath16 .
note that in the case of cyclohexane there is a clear degradation of the energy resolution at high gains , which is absent for ethane .
m - thickness - gap in argon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] m - thickness - gap in argon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ]
micromegas detectors have been typically operated in argon - isobutane mixtures between 2 and 5% for a wide range of applications . this gas is well adapted for measurements in the 1 - 10 kev range , providing an excellent energy resolution and gains up to @xmath13 .
other gas mixtures are being studied to increase micromegas sensitivity in the sub - kev region . for this purpose ,
the signal to noise ratio must be increased and higher gains are needed , keeping a good energy resolution . a possible solution is the replacement of the base gas by neon , as the total charge per single avalanche is increased and approaches the rather limit ( @xmath19 electrons ) . however , the efficiency of the photon absorption is lower in this kind of mixtures , passing from 70% down to 10% at 6 kev , which limits the range of applications .
we present the main features of the two detectors studied in neon mixtures with the same quenchers used before ( isobutane , cyclohexane and ethane ) .
the electron transmission curves of both detectors show the same features observed in argon - based mixtures . for the case of the 50 @xmath2m - thickness gap
, the plateau of maximum mesh transparency widens as the quencher concentration is increased , as shown in figure [ fig : transeresneiso ] ( left ) for isobutane , and the energy resolution is correlated to this variable ( same figure , right ) .
the plateau is absent for the other detector but there is a narrow range of fields for which the energy resolution reaches its best values . comparing their performances to argon - based mixtures , we have observed narrower plateaus of maximum transmission for the gap of 50 @xmath2 m , which can be explained by the higher diffusion coefficients in neon @xcite .
m - thickness gap in neon - isobutane mixtures .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m - thickness gap in neon - isobutane mixtures .
the gain has been normalized to the maximum of each series .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] the dependence of the absolute gain with the amplification field is shown in figure [ fig : gainneiso ] .
gains higher than @xmath12 and @xmath13 are respectively reached for isobutane concentrations greater than 7 and 10% and for a gap of 50 and 25 @xmath2 m . moreover ,
the required field for a fixed gain does not increase with the quencher concentrations below 10% , as in argon - based mixtures .
we will try to interpret this difference in terms of gas properties in the next section . a second fact observed
is that higher fields are needed in these mixtures to get the same gain . as an example
, a field of 72 kv / cm must be applied in ar+10% ic@xmath0h@xmath1 to reach a gain of 10@xmath20 for the detector of 50 @xmath2m - thickness gap . in a neon based mixture ,
this value is 75 kv / cm .
this difference is greater for the other detector : respectively 140 kv / cm and 167 kv / cm for the same mixtures .
m ( right ) in neon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m ( right ) in neon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] an improvement in the energy resolution has been observed for neon - based mixtures and the 50 @xmath2m - thickness - gap detector , going from 11.7% fwhm for ar+5%ic@xmath0h@xmath1 down to 10.5% fwhm for ne+7%ic@xmath0h@xmath1 .
this fact can not be explained by the fluctuations of the primary ionization but by the contribution of the detector avalanches . as described in @xcite ,
the energy resolution of the detector can be expressed as @xmath21 where @xmath22 is the event energy , @xmath23 is the gas fano factor , @xmath24 is the mean energy necessary for creating an electron - ion pair in the gas and @xmath25 is the detector contribution . in neon - based mixtures , less
primary electrons are generated due to its higher mean electron - ion pair energy ( respectively 36.4 and 26.3 ev for neon and argon @xcite ) , which should worsen the energy resolution as the fano factor is similar to both gases ( 0.17 @xcite ) .
the detector contribution must be thus little for neon - based mixtures .
in fact , as noted in @xcite , less avalanche fluctuations are presented in light gases like helium or neon due to their higher ionization yield .
the energy resolution is also better at high gains , as shown in figure [ fig : eresgainneiso ] ( left ) . at a gain of @xmath12 ,
the energy resolution is 11.4% fwhm in ne+10%ic@xmath0h@xmath1 , better than the 15% fwhm reached in ar+11% cyclohexane .
similar values are also obtained with ethane or cyclohexane , as shown in table [ tab : eresneon ] , which may point out that it is mainly determined by the base gas .
m - thickness - gap ( right ) in neon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] m - thickness - gap ( right ) in neon - isobutane mixtures .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the isobutane concentration.__,title="fig:",width=283 ] in contrast , the energy resolution is worse in neon - based mixtures for the 25 @xmath2m - thickness - gap detector , passing from 11.6% fwhm in ar+10%ic@xmath0h@xmath17 to 12.6% fwhm in ne+25%ic@xmath0h@xmath1 .
worse values have also been obtained with the other quenchers , as shown in table [ tab : eresneon ] .
larger amplification gaps are in principle more adequate for light gases as they minimize the avalanche fluctuations produced by gap uniformaties @xcite .
another possible explanation is a small energy transfer between neon and the quencher due to the gap size @xcite , which may limit the penning effect .
@xmath26 we finally present in figure [ fig : eresgainnem50 ] the dependence of the energy resolution with the gain for the microbulk of 50 @xmath2 m gap in mixtures of neon with ethane and cyclohexane .
the maximum gain before the spark limit is respectively @xmath12 and @xmath27 .
in contrast to argon - based mixtures , there is no degradation of the energy resolution at high gains .
m - thickness - gap in neon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ] m - thickness - gap in neon - ethane ( left ) and -cyclohexane mixtures ( right ) .
the maximum gain of each curve was obtained just before the spark limit .
the percentage of each series corresponds to the quencher concentration.__,title="fig:",width=283 ]
the rose - korff model is a parametrization of the first townsend coefficient which neglects the possible secondary ionizations in the avalanche @xcite .
if we modelize micromegas as a parallel - plate detector and we suppose that the amplification field is uniform , the gain can be expressed by the formula @xmath28 where @xmath29 is the gain , @xmath30 is the amplification field , @xmath31 is the gap distance , @xmath32 is the electron mean free path and @xmath33 is the threshold energy for ionization .
therefore , this simple model can give information on gas properties .
this study was firstly made in @xcite .
the gain curves generated in argon- and neon - based mixtures have been fitted to the equation [ eq : rose ] and the resulting dependencies of the electron mean free path and the threshold energy for ionization are respectively shown in figures [ fig : fitgainmfp ] and [ fig : fitgainie ] .
as shown in figure [ fig : fitgainmfp ] , the electron mean free path decreases as the quencher concentration is increased and , for the same base gas and detector , there is almost no difference for low quencher quantities . at high concentration ,
mixtures with ethane show higher values than with isobutane . apart from that , neon - based mixtures show higher values than in argon - based ones . for instance , the electron mean free path is respectively 1.6 and 2.5 @xmath2 m in ar+5%ic@xmath0h@xmath1 and ne+5%ic@xmath0h@xmath1 for the 50 @xmath2m - thickness - gap detector .
these three facts can be explained by theory @xcite , even if the exact values do not match .
the electron mean free path is in general higher for light gases , for noble gases in comparison to hydrocarbons and for ethane compared to isobutane .
finally , we have also observed lower values for the gap size of 25 @xmath2 m , which can not be explained by gas properties . to confirm experimentally that the electron mean free path is shorter for heavier gases , we have revisited the gain curves generated for xenon - based mixtures in @xcite .
the electron mean free path ( figure [ fig : fitgainxenon ] , left ) is clearly lower than in other gases , as xenon is a heavier gas .
small differences are due to higher systematic errors in the gain measurements .
values have a relative systematic error of 20%.__,title="fig:",width=264 ] m @xcite .
values have a relative systematic error of 20%.__,title="fig:",width=264 ] the energy threshold for ionization has a different behaviour in argon- than in neon - based mixtures , as shown in figure [ fig : fitgainie ] .
for the first type , this parameter increases with the quencher quantity and is independent of the gap size .
for example , for a quencher percentage of 10% , the energy threshold is respectively 15.2 , 14.3 and 13.5 ev for ethane , isobutane and cyclohexane .
in contrast , this parameter shows a minimum in neon - based mixtures , which depends on both the quencher and the detector .
for all mixtures , values are higher for a gap size of 25 @xmath2 m .
these differences can be explained by the penning effect .
the transfer of energy from non - metastable levels of argon to the quencher molecule is complete for both detectors in argon - based mixtures and the energy threshold is only defined by the quencher . for neon , as the electron mean free path is longer , the transfer is limited by the gap size .
this limitation disappears at high quencher quantities and the energy threshold for ionization increases with quencher concentration , as in argon - based mixtures .
finally , the values obtained in neon are in general higher than in argon - based mixtures . for example
, this parameter is respectively 15.6 and 14.8 ev in ne+5% ethane and ar+5% ethane for the 50 @xmath2m - thickness - gap detector .
this fact is due to the higher values of this parameter for pure neon ( 21.6 ev ) than for pure argon ( 15.8 ev ) @xcite .
note that the values given in literature for the base gases and for the quenchers ( 11.7 and 10.8 ev for ethane and isobutane ) do not match with the ones obtained . as discussed in @xcite ,
this fact can be explained by the penning effect @xcite , which is quite important in mixtures with big differences between the ionization threshold of the quencher and the excitation levels of the base gas .
a good question is if the values obtained for the electron mean free path and the energy threshold for ionization could explain the energy resolution of a detector in a gas .
we can start our discussion using @xcite , where it was concluded that the fluctuations due to defects of the mesh and the gap flatness could be compensated chosing the idoneous gap for each gas type .
this optimum was obtained by differentiating the equation [ eq : rose ] , which results in @xmath34 we can interpret this minimum in terms of an optimum amplification field for a fixed gap .
for each gas , this field should be as low as possible to fastly reach an optimum detector performance and to avoid the degradation of the energy resolution by other effects .
therefore , the electron mean free path should be as high as possible and the values of the energy threshold for ionization should be little . as shown in figures [ fig : fitgainmfp ] and [ fig : fitgainxenon ] ( right ) , the first parameter is mainly determined by the base gas , being higher for neon than for argon or xenon .
the quencher concentration decreases its value , which in general degrades the energy resolution .
there is also a dependence on the gap size , as lower values are obtained for a gap of 25 @xmath2 m .
as we have shown , the gap size is only important in neon - based mixtures . in the case of the energy threshold for ionization ,
we have observed lower values for cyclohexane than for isobutane and ethane but there is no clear improvement in the energy resolution .
this variable could have thus little effect if the optimum amplification field is in the working range .
microbulk is a micromegas technology that has been used in nuclear ( ntof ) and astroparticles experiments ( cast , next ) .
it offers uniform and flexible structures with an excellent energy and time resolution , low background levels and low mass .
two detectors , with a gap distance of respectively 50 and 25 @xmath2 m , have been characterized in argon - based mixtures with ethane , isobutane and cyclohexane , using a @xmath11fe source .
the maximum gain before the spark limit is respectively @xmath13 and @xmath14 and the best energy resolutions obtained in argon - isobutane mixtures are 11.7% and 12.1% fwhm at 5.9 kev .
the plateau of maximum electron transmission depends on the quantity of quencher for the 50 @xmath2m - thick - gap detector and is almost inexistent for the gap of 25 @xmath2 m .
the same conclusions have been obtained with the other quenchers but the gain is higher in argon - cyclohexane mixtures and lower with ethane .
microbulk detectors have been also tested in neon - based mixtures with the same quenchers to reduce the energy threshold and to make micromegas sensitive to sub - kev energies .
possible applications are synchroton radiation and dark matter searches .
the maximum gain reached before the spark limit in argon - isobutane mixtures is respectively @xmath12 and @xmath27 for a gap size of 25 and 50 @xmath2 m and the best energy resolutions are 10.5% and 12.6% fwhm at 5.9 kev .
similar values for the energy resolution have been obtained with the other quenchers except for the detector of 25 @xmath2m - thickness gap and neon - based mixtures , whose values are worse . the last developments of microbulk technology are focused in creating detectors with small gaps ( 12.5 and 25 @xmath2 m ) to be used with gases at high pressure .
different hole pitchs and diameters are being studied .
the use of other quenchers with even lower ionization threshold like cyclohexene are also planned , in order to confirm the divergences observed in argon - cyclohexane between micromegas detectors and proportional counters .
we are thankfull to our colleagues of the university of zaragoza and irfu / cea - saclay , as well as to the rd51 collaboration for helpful discussions and encouragements .
f. i. acknowledge supports from the eurotalents program .
00 y. giomataris _ et al . _ ,
micromegas : a high - granularity position - sensitive gaseous detector for high particle - flux environments , _ nucl .
instrum .
meth . a _ * 376 * ( 1996 ) 29 .
i. giomataris _
et al . _ , micromegas in a bulk , _ nucl .
instrum . meth .
a _ * 560 * ( 2006 ) 405 . s. adriamonje _ et al .
_ , development and performance of microbulk micromegas detectors , _ jinst _ * 5 * ( 2010 ) p02001 .
iguaz _ et al .
_ , new developments of micromegas microbulk technologies , proceeding of the tipp 2011 conference , _ preprint : _ arxiv:1110.2641 .
s. cebrian _ et al .
_ , radiopurity of micromegas readout planes , _ astropart
* 34 * ( 2011 ) 354 .
n. colonna _ et al .
_ , neutron reasurements for advanced nuclear systems : the n_tof project at cern , accepted in _ nucl .
b_. s. aune _ et al .
_ , an ultra - low background detector for axion searches , _ j. phys .
* 179 * ( 2009 ) 012015 .
s. cebrian _ et al .
_ , micromegas readouts for double beta decay searches , _ jcap _ 10 ( 2010 ) 1010
. t. dafni _ et al .
_ , micromegas planes for the neutrinoless double beta decay search with next , _ j. phys .
* 309 * ( 2011 ) 012009 .
agrawal and b.d .
ramsey , use of propane as a quench gas in argon - filled proportional counters and comparison with other quench gases , _ nucl .
instrum . meth .
a _ * 273 * ( 1988 ) 331 . y. giomataris , development and prospects of the new gaseous detector micromegas , _ nucl .
instrum . meth . a _ * 419 * ( 1998 ) 239 .
h. shindler , private communication .
et al . _ ,
energy resolution of alpha particles in a micromegas detector at high pressure , _ nucl .
instrum . meth .
a _ * 608 * ( 2009 ) 259 .
iguaz _ et al .
_ , new results of microbulk detectors , talk in the 5th rd51 collaboration metting .
m. chefdeville , development of micromegas - like gaseous detectors using a pixel readout chip as collecting anode , phd thesis , university of amsterdam , 2009 .
bronic and b. grosswendt , gas amplification and ionization coefficients in isobutane and argon - isobutane mixtures at low gas pressures , _ nucl .
b _ * 142 * ( 1998 ) 219 .
h. schindler _ et al .
_ , calculation of gas gain fluctuations in uniform fields , _ nucl .
instrum . meth .
a _ * 624 * ( 2010 ) 78 .
m.e . rose and s.a .
korff , an investigation of the properties of proportional counters .
i , _ phys .
* 59 * ( 1941 ) 850 .
b.d . ramsey and p.c .
agrawal , xenon - based peenning mixtures for proportional counters , _ nucl .
instrum . meth .
a _ * 278 * ( 1989 ) 576 .
magboltz - transport of electrons in gas mixtures , http://magboltz.web.cern.ch / magboltz/. . sahin _ et al .
_ , penning transfer in argon - based gas mixtures , _ jinst _ * 5 * ( 2010 ) p05002 .
i.k . bronic and b. grosswendt , ionization yield formation argon - isobutane mixtures as measured by proportional - counter method , _ nucl .
b _ * 117 * ( 1996 ) 5 .
w. blum and l. rolandi , particle detection with drift chambers , springer - verlag , 1993 .
s. biagi , mip program , http://consult.cern.ch / writeup / magboltz/. d. atti _ et al .
_ , gain measurements in various gas mixtures and optimization of the stability , talk at the tpc jamboree , aachen , 2007 .
guo _ et al .
_ , 3d simulation of micromegas detector performance , _ chinese physics c _ * 34 * ( 2010 ) 482 .
l. laval , etude de melanges gazeux pour loptimisation du dtecteur micromegas , stage report at the cea / dsm , 1997 .
|
a recent micromegas manufacturing technique , so called microbulk , has been developed , improving the uniformity and stability of this kind of detectors .
excellent energy resolutions have been obtained , reaching values as low as 11% fwhm at 5.9 kev in ar+5%ic@xmath0h@xmath1 .
this detector has other advantages like its flexible structure , low material budget and high radio - purity .
two microbulk detectors with gaps of 50 and 25 @xmath2 m have been characterized in argon- and neon - based mixtures with ethane , isobutane and cyclohexane .
the results will be presented and discussed .
the gain curves have been fitted to the rose - korff gain model and dependences of the electron mean free path and the threshold energy for ionization have been obtained . the possible relation between these two parameters and the energy resolution
will be also discussed .
|
our main goals are to identify traces of hydrogen , helium and oxygen in type ib - c sne , and to identify any systematic similarities and differences correlating with other physical properties in the ccsne family .
we have selected a sample of 45 photospheric spectra of 20 ccsne objects . in the present paper
we show results for a few objects . a detailed analysis for the complete sample has been presented recently in elmhamdi et al .
2006 @xcite . for the purpose of our analysis
we make use of the parameterized supernova synthetic spectrum code `` synow '' .
the main task is to identify lines and put constraints on the velocities .
synow has a number of free input fitting parameters .
the most important are : 1 .
( @xmath0 ) , the optical depth of `` the reference line '' .
2 . ( @xmath1 ) : the underlying blackbody continuum temperature .
3 . ( @xmath2 ) : the velocity at the photosphere , estimated from the match with fe ii lines . the radial dependence of the line optical depths can be chosen to be either exponential with an @xmath3-folding velocity `` @xmath4 '' as free parameter ( i.e. @xmath5 ) , or a power - law with an index `` @xmath6 '' ( i.e. @xmath7 ) . detailed description of the code can be found elsewhere @xcite an important aspect to understand within the `` synow '' line formation logic is the `` detachment '' concept . when assigning a minimum ion velocity , @xmath8 , greater than the photospheric one @xmath2 ,
the ion is said to be `` detached '' from the photosphere , and consequently has a `` non - zero '' optical depth only starting at @xmath8 . in `` synow ''
the profile of a detached line has a flat - topped emission , and the absorption minimum is blueshifted by the detachment velocity .
an undetached line has a rounded emission peak .
a slightly detached line has a flat top but only over a small wavelength interval .
these possibilities are illustrated in fig .
1(left ) , where the code run adopts @xmath2=10000 km s@xmath9 and t@xmath10=6000 k , including only hydrogen balmer lines with @xmath11(h@xmath12 and having 3 different detachment velocities .
we introduce the parameter `` contrast velocity '' , defined as : @xmath13 .
we discuss as well the behaviour of a similar parameter , defined instead as a ratio , i.e. @xmath14 . : the `` synow '' run for balmer h i lines , using @xmath2=10000 km s@xmath9 and t@xmath10=6000 k. the resulted profiles for different detachment velocities are illustrated .
@xmath15 : the ssp fit , thin line , to the spectrum of sn 1990i near maximum .
the region around the 6250 trough is zoomed in the window.,title="fig : " ] : the `` synow '' run for balmer h i lines , using @xmath2=10000 km s@xmath9 and t@xmath10=6000 k. the resulted profiles for different detachment velocities are illustrated .
@xmath15 : the ssp fit , thin line , to the spectrum of sn 1990i near maximum .
the region around the 6250 trough is zoomed in the window.,title="fig : " ]
in the following we present results for some individual objects of the ccsne sample . * sn 1990i :* the event is one of the well sampled and exploited , both photometrically and spectroscopically , objects among type ib - c sne@xcite .
1(right ) compares the observed spectrum at maximum light to a synthetic spectrum ( ssp ) that has a velocity at the photosphere @xmath16 km s@xmath9 and t@xmath1714000 k. the introduced lines are he i , fe ii , sc ii , mg ii , o i , ca ii and h@xmath18 .
absorption troughs of p - cygni he i lines at 5876 , 6678 and 7056 are evident , @xmath11(hei)@xmath19 and @xmath8(hei)@xmath20 km s@xmath9 , although they are not simultaneously fitted ( i.e. their relative strengths ) .
this limitation is faced each time we analyze and fit he i lines using the lte approach .
a more precise analysis requires nlte treatment as he i lines are probably non - thermally excited by the decay products of @xmath21ni and @xmath21co @xcite .
apart from he i and h@xmath18 lines , the remaining lines are undetached .
the absorption minimum near 6250 is well fitted by h@xmath18 , with @xmath8(h@xmath18)=16000 km s@xmath9 and assigned a moderate optical depth of 0.6 .
we tested however other plausible alternative identifications to account for a similar feature seen in type ib sne , namely si ii 6355 , ne i 6402 and c ii 6580 lines . for sn 1990i ,
@xmath22 ( h@xmath18 ) is only 4000 km s@xmath9 , hence undetached ne i line is excluded because it is too blue to fit the observed feature .
si ii 6355 is obviously also too blue .
c ii 6580 remains then a plausible alternative for h@xmath18 , since its rest wavelength is slightly redder than h@xmath18 and therefore has its contrast velocity as a free parameter . in the window of fig .
1(right ) , we show the h@xmath18 and c ii fit cases .
c ii , with @xmath8(cii)=17000 km s@xmath9 and @xmath11(cii)@xmath23 , provides a fit as good as h@xmath18 .
but an acceptable c ii 6580 line would require an additional velocity of about 820 km s@xmath9 compared to h@xmath18 ( @xmath24 difference ) . furthermore , the he i lines have @xmath22(he i)@xmath252000 km s@xmath9 while @xmath22(cii)@xmath255000 km s@xmath9 .
this means that the c ii lines would need to be formed above the helium layer , which would be surprising , if not unphysical , in type ib sne .
sn 1990i presents therefore good evidence for an h@xmath18 feature .
it is not absolutely clearcut however that h@xmath18 is always responsible for the absorption at about 6300 seen in type ib sne .
the presence of the h@xmath26 balmer line would be of great support .
unfortunately the optical depth sufficient to fit the h@xmath18 trough is so small that the other balmer features are too weak to be unambiguously detectable ( @xmath11(h@xmath27 is about 7 times greater than @xmath11(h@xmath28 ) .
the second factor that makes h@xmath26 barely discernable is when the contrast velocity of h@xmath18 is high as can be clearly seen from fig .
1(left ) .
: ssp fits of sn 1990i spectra at 12d and 21d .
@xmath15 : ssp fit of sn 1991d around 3 weeks .
the top panel illustrates the ne i possibility , while h@xmath18 case is shown in the lower panel.,title="fig : " ] : ssp fits of sn 1990i spectra at 12d and 21d .
@xmath15 : ssp fit of sn 1991d around 3 weeks .
the top panel illustrates the ne i possibility , while h@xmath18 case is shown in the lower panel.,title="fig : " ] based on our investigation , the ions that generally might be encountered in shaping the `` 6000@xmath296500 '' wavelength range are : h@xmath18 , ne i , c ii , si ii , sc ii , ca i , he i , fe ii , si i and ba ii . here , we propose some selection criteria to decide on the presence of hydrogen for the 6300 feature : + @xmath30 undetached ne i 6402 line is rejected once its too blue to fit the 6300 trough or / and when the other ne i lines clearly introduce unwanted features .
similar reasoning applies to the si ii 6355 line . @xmath31 with its contrast velocity as a free parameter , c ii 6580 line could be a candidate for the 6300 trough ; nevertheless it is ruled out once it exceeds the he i contrast velocity .
the observed spectra at 12days and 21days are compared in fig .
2(left ) to the synthetic ones with ( @xmath2=10000 km s@xmath9 ; t@xmath10=5500 k ) and ( @xmath2=9500 km s@xmath9 ; t@xmath10=5400 k ) , respectively .
he i is still detached from the photosphere with @xmath22(he i)@xmath25 3000 km s@xmath9 at 12d and @xmath22(he i)@xmath25 2500 km s@xmath9 at 21days .
the he i 7065 seems to increase in strength relative to the he i lines at 5876 and 6678 , indicating the non - thermal excitation effects are changing but still existent .
for the `` 6000@xmath296500 '' wavelength region , we find that fe ii , he i and h@xmath18 are sufficient to reproduce the overall shape in the ssp at both 12days and 21days .
we have , however , tested the ne i , si ii and c ii possibilities as candidates for the 6250 trough .
they are all ruled out by means of our above criteria .
we therefore consider h@xmath18 to be the most likely explanation in sn 1990i . +
* sn 1991d :* the event is a peculiar case where its difficult to decide between h@xmath18 and ne i. the object has narrow features and lower velocities compared to other type ib events at similar phases @xcite . fig .
2(right ) illustrates the fit to the 21d spectrum ( @xmath32 km s@xmath9 and t@xmath177000 k=. the he i lines are evident through our ssp ( @xmath22(he i)= 1400 km s@xmath9 and @xmath11(he i)=1.8 ) , with the na i d feature ( @xmath11(nai)=4 ) contributing to the he i 5876 broad p - cygni profile .
the presence of na i lines is consistent with the good fit around 8100 .
the undetached ne i lines possibility is illustrated in the top panel ( @xmath11(nei)=2 ) , while the h@xmath18 case is tested in the bottom panel . except for adding ba ii lines in order to help the fit at @xmath336100 ,
the other ion parameters are kept unchanged .
the fit to the observed absorption near 6300 with h@xmath18 , @xmath22(h@xmath18)= 7400 km s@xmath9 and @xmath11(h@xmath18)=0.46 , is slightly better compared to the ne i case .
however the ssp , in the bottom panel , does not account for the observed features near 6630 and 6840 as does nei lines in the ssp of the top panel .
ne i remains hence a strong candidate in this type ib event .
: ssp fit of sn 1999ex at day 4 .
conspicuous line features are indicated .
@xmath15 : the 4days ir spectrum compared with ssp that contain only lines of he i.,title="fig : " ] : ssp fit of sn 1999ex at day 4 .
conspicuous line features are indicated .
@xmath15 : the 4days ir spectrum compared with ssp that contain only lines of he i.,title="fig : " ] + * sne 1999ex : * because of weak optical he i lines , the object was classified as an intermediate case between ib and ic sne , while the evident trough at @xmath336250 was attributed to si ii 6355 @xcite . in fig .
3(left ) the observed 4d spectrum is compared with the best fit that has @xmath34 km s@xmath9 and t@xmath175800 k . to check the consistency of he i identification we combine the ir spectrum with the optical one .
in fact assigning the following parameters : @xmath22(he i)= 1000 km s@xmath9 and @xmath11(he i)=2.35 , we obtain a good match to the observed he i profiles including the ir he i 10830 .
furthermore , unambiguous evidence for the presence of helium in this event is clear from the extended fit to other he i - ir lines , adopting equal @xmath8(he i ) and @xmath11(he i ) as in the optical part of the spectrum ( fig .
3-right ) .
the oi 7773 line with @xmath11=0.5 , on the other hand , is found to be not so strong and deep as is the case in most type ic sne @xcite .
these two facts point to a ib nature , more than a ic class . :
ssp fit of sn 2000h at day 19 .
@xmath15 : ssp fits of sn iip 1999em observed spectra at day -6(top ) and day 25(bottom ) .
the reported phases are normalized to maximum in type ib - c sne.,title="fig : " ] : ssp fit of sn 2000h at day 19 .
@xmath15 : ssp fits of sn iip 1999em observed spectra at day -6(top ) and day 25(bottom ) .
the reported phases are normalized to maximum in type ib - c sne.,title="fig : " ] as far as the trough around 6250 is concerned , we checked the si ii identification attributed by hamuy et al.(2002)@xcite .
we test as well the ne i possibility .
the undetached ne i fit introduces unwanted features in the spectrum .
the si ii 6355 is rather blue to account for the feature .
we note here that if one adopts c ii 6580 , then one needs to assign it a very high velocity , about 8000 km s@xmath9 higher than the one of he i. the most likely identification therefore remains h@xmath18 .
in fact the best fit is achieved using @xmath22(h@xmath18)= 8000 km s@xmath9 and @xmath11(h@xmath18)=1.45 , + * sn 2000h :* the sn is an example of a type ib event with a strong and deep trough near 6300 @xcite .
4(left ) compares the 19d spectrum with the best fit ssp ( @xmath35 km s@xmath9 and t@xmath174600 k ) .
lines of ca ii , [ o ii ] , o i , mg ii are seen to develop .
the optical he i lines are noticeably deep ( @xmath22(he i)= 2000 km s@xmath9 and @xmath11(he i)=5 ) .
we note the good match to the @xmath336250 trough with h@xmath18 , achieved using @xmath22(h@xmath18)= 7000 km s@xmath9 and @xmath11(h@xmath18)=2.5 .
no alternative identification to h@xmath18 , that would be logically acceptable , has been found .
in addition , the h@xmath18 identification is supported by the presence of the absorption notch near 4660 , well accounted for by h@xmath26 in our resulting synthetic spectra . + * sn iip 1999em : * fig .
4(right ) displays two spectra during the photospheric phase of the typical type ii - plateau sn 1999em @xcite .
the phases are normalized to maximum light in type ib - c sne .
the best fit synthetic spectra have @xmath34 km s@xmath9 and t@xmath1710000 k ( -6d ; upper panel ) and @xmath32 km s@xmath9 and t@xmath176400 k ( 25d ; lower panel ) . at the earliest phase
only undetached lines of hydrogen balmer , he i and a weak contribution from ca ii are sufficient to reproduce the most conspicuous features superimposed on the `` hot '' continuum .
the fit with na i d , at this phase , is poor compared to he i 5876 .
the he i contribution , in type ii sne , is found to be important shortly after the explosion .
in fact in sn 1987a , the he i 5876 feature was clearly present during the first few days , then rapidly faded and disappeared completely around 1 week after the explosion , when na i d starts to emerge @xcite . at -6d ,
balmer hydrogen features are evident with @xmath36 for h@xmath18 .
the corresponding strong and broad p - cygni feature can not be produced completely by the `` synow '' code .
this is because in our ssp treatment we are adopting a resonance scattering source function . at day 25 ,
4(right - bottom ) , h@xmath18 is narrower and has @xmath22= 1000 km s@xmath9 and @xmath37
. the envelope temperature decreases and many lines emerge at this phase .
apart from the hydrogen lines ( slightly detached ) , all the lines are undetached .
based on results presented here and in the larger sample study of elmhamdi et al .
( 2006)@xcite , hydrogen is found to manifest its presence in a different way in the ccsne sample .
sne of type ii and iib show strong and broad h@xmath18 p - cygni emission component , which is absent in type ib objects .
indeed , the h@xmath18 p - cygni profile would lose its obvious emission component when it is highly detached .
h i balmer lines have large optical depths , allowing them to be distinctly visible , contrary to normal type ib sne , with the exception of some cases with deep and conspicuous h@xmath18 troughs that present the signature of h@xmath26 as well .
two factors make h@xmath26 barely discernible in type ib : the optical depth found to fit h@xmath18 is small and the contrast velocity of h@xmath18 is high .
the opposite is seen in type ii and iib events . in fig . 5 , upper panel , we report the resulting photospheric velocities from our best fits for the whole ccsne sample @xcite
. data for sn 1987a , corresponding to fe ii 5018 absorption , are also shown for comparison ( dashed line @xcite ) .
the plot indicates the low velocity behaviour of type ii sne , both 1987a and 1999em .
sn iib 1993j follows somewhat similar behaviour as sn ic 1994i in having higher velocities .
type ib sne appear to display a different velocity evolution .
the scatter seems to increase at intermediate phases ( around 20@xmath2930 days ) .
this fact can be simply due to the paucity of available observations outside that range . around day 20 , for example
, a scatter as high as 5000 km s@xmath9 is measured .
sne 1990i and 1998dt belong to a class with the higher `` @xmath2 '' , while objects such as sne 1991d and 1996aq have the lowest estimated velocities , approaching even type ii objects .
the remainder of the ib events follow a similar trend , namely the one described by branch et al .
( 2002)@xcite .
the middle and bottom panels in fig . 5 display the evolution the of h@xmath18 contrast velocities .
an increasing trend of `` @xmath22(h@xmath18 ) '' is found going from type ii to iib to ib sne .
while in sn iib 1993j and sn iip 1999em the line is found to be either undetached or slightly detached , it is highly detached in type ib events . moreover ,
the `` @xmath22(h@xmath18 ) '' is found to increase within the first 15 days , reaching values as high as 8000 km s@xmath9 , and then follows an almost constant evolution .
up to @xmath3360 days , type ib sne are found to have hydrogen down to 11000 - 12000 km s@xmath9 .
sne ii appear to have hydrogen down to lower velocities ( @xmath335000 km s@xmath9 in sn 1999em ) .
in addition , hydrogen in type ib sne is found to have very small optical depths independently of the contrast velocity .
an approximate method to estimate the hydrogen mass from the fit results invokes the amount needed to fill a uniform density sphere of radius `` @xmath38 '' at an epoch @xmath39 since explosion . the required ion mass can be given by ( @xcite ) : @xmath40 where @xmath39 in days and @xmath41 is in 10@xmath42 km s@xmath9 .
although non - thermal excitation and `` nlte '' effects may be also important for hydrogen , this simple approach seems to give reasonable estimates . for sn ib 1990i around maximum light a value of 0.02 @xmath43 is computed .
if we adopt a representative h@xmath18 optical depth of 0.5 at day 20 with an h@xmath18 velocity restriction similar to sn 1990i , the estimated hydrogen mass is of the order 0.015 @xmath43 .
events with higher velocity widths and/or deeper h@xmath18 troughs would eject larger amounts .
a thin layer of hydrogen , ejected at high velocities down to 11000 - 12000 km s@xmath9 , appears to be present in almost all the studied type ib events .
worth - noticing here is that the presence of h@xmath18 in type ic events was recently re - addressed by branch et al .
( 2006 ) @xcite .
a large type ic sample study is clearly needed to assess this issue .
however , based on these preliminary investigations the progenitor nature of the stripped - envelope sne should be revised .
an important by - product results of our fit analysis concerns the behaviour of the o i 7773 line . at intermediate phases , it seems that type ib objects tend to have low optical depths in this line , while sne ic are found to display the deepest profile .
sne iib an ii , at similar phases , are the objects with the lowest o i 7773 optical depth .
the stronger and deeper permitted oxygen lines at early phases of sne ic spectra might imply that they are less diluted by the presence of a helium envelope .
indeed one might expect oxygen lines to be more prominent for a `` naked '' c / o progenitor core . despite the paucity of well sampled ccsne observations , two observational aspects tend to reinforce this belief : @xmath44 , the forbidden lines , especially [ oi]6300 , 6364 , seem to appear earlier following a sne sequence `` ic@xmath29ib@xmath29iib@xmath29ii '' .
in fact the oxygen line emerges at an age of 1 - 2 months in type ic sn 1987 m @xcite .
sn ic 1994i displayed evidence for the line at an age of 35 - 50 days @xcite . while in sn ib 1990i it was hinted at the 70day spectrum @xcite .
in other type ib sne it appears earlier than in sn 1990i . in sn iib 1993j ,
a transition object , the line was visible in the 62day spectrum @xcite . in sne ii
, however , the line appears later : around day 150 in sn 1987a @xcite and after day 138 in sn 1992h @xcite .
@xmath45 , the nebular emission line looks decreasing in breadth following the sne sequence above .
more investigations concerning the oxygen issue in ccsne are under progress ( elmhamdi & danziger , in preparation ) .
9 barbon r. , benetti s. , cappellaro e. et al .
1995 , @xmath46 , * 110 * , 513 benetti s. , branch d. , turatto m. et al .
2002 , @xmath47 , * 336 * , 91 branch d. , benetti s. , kasen d. et al .
2002 , @xmath48 , * 566 * , 1005 branch d. , jeffery d.j .
, young t.r .
et al . 2006 , _
`` astro - ph/0604047 '' _ catchpole r. m. , whitelock p. a. , feast m. w. et al .
1988 , @xmath47 , * 231 * , 75 clocchiatti a. , benetti s. , wheeler j. c. al .
1996a , @xmath49 , * 111 * , 1286 clocchiatti a. , wheeler j. c. , brotherton m. s. et al .
1996b , @xmath48 , * 462 * , 462 elmhamdi a. , danziger i. j. , chugai n. n. et al .
2003 , @xmath47 , * 338 * , 939 elmhamdi a. , danziger i. j. , cappellaro e. et al .
2004 , @xmath50 , * 426 * , 963 elmhamdi a. , danziger i. j. , branch d. et al .
2006 , @xmath51&@xmath51,*450 * , 305 - 330 filippenko a. v. 1997 , @xmath52&@xmath51 , * 35 * , 309 hamuy m. , maza j. , pinto p.
2002 , @xmath49 , * 124 * , 417 hanuschik r. w. @xmath53 dachs j. 1988 , @xmath50 , * 205 * , 135 jeffery d. j. @xmath53 branch d. 1990 , in : j.c .
wheeler , t. piran , @xmath53 s. weinberg ( eds.),``supernovae '' , @xmath54 ( singapore : world scientific ) , * 149 * lucy l. b. 1991 , @xmath48 , * 383 * , 308 matheson t. , filippenko a. v. , li w. , leonard d. c. @xmath53 shields j. c. 2001 , @xmath49 , * 121 * , 1648 phillips m. m. , heathcote s. r. , hamuy m. @xmath53 navarrete m. 1988 , @xmath49 , * 95 * , 1087
|
we discuss results of analyzing a time series of selected photospheric - optical spectra of core collapse supernovae ( ccsne ) .
this is accomplished by means of the parameterized supernovae synthetic spectrum ( ssp ) code `` synow '' .
special attention is addressed to traces of hydrogen at early phases , especially for the stripped - envelope sne ( i.e. sne ib - c ) .
a thin low mass hydrogen layer extending to very high ejection velocities above the helium shell , is found to be the most likely scenario for type ib sne .
address = osservatorio astronomico di trieste , via g.b.tiepolo 11 - i-34131 trieste - italy address=<common address for author1 and author2 > address= department of physics and astronomy , university of oklahoma , norman , ok 73019 , usa address=
european southern observatory , karl - schwarzschild - strasse 2 .
d-85478 , graching - germany
|
the main goal of electronic structure methods is to solve the schrdinger equation for the electrons in a molecule or solid , to evaluate the resulting total energies , forces , response functions and other quantities of interest . in this paper
we review the projector augmented wave ( paw ) method @xcite , an electronic structure method for ab - initio molecular dynamics with full wave functions .
the main goal of this paper is not to provide a particularly complete or detailed account of the methodology , but rather to lay out the underlying ideas .
a more rigorous description can be found in the original paper @xcite . density functional theory @xcite maps a description for interacting electrons onto one of non - interacting electrons in an effective potential .
the remaining one - electron schrdinger equation still poses substantial numerical difficulties : ( 1 ) in the atomic region near the nucleus , the kinetic energy of the electrons is large , resulting in rapid oscillations of the wave function that require fine grids for an accurate numerical representation . on the other hand ,
the large kinetic energy makes the schrdinger equation stiff , so that a change of the chemical environment has little effect on the shape of the wave function .
therefore , the wave function in the atomic region can be represented well already by a small basis set . ( 2 ) in the bonding region between the atoms the situation is opposite .
the kinetic energy is small and the wave function is smooth .
however , the wave function is flexible and responds strongly to the environment .
this requires large and nearly complete basis sets .
combining these different requirements is non - trivial and various strategies have been developed .
* most appealing to quantum chemists has been the atomic point of view .
basis functions are chosen that resemble atomic orbitals .
they exploit that the wave function in the atomic region can be described by a few basis functions , while the bonding is described by the overlapping tails of these atomic orbitals .
most techniques in this class are a compromise of a well adapted basis set with complex matrix elements on the one hand and on the other hand numerically convenient basis functions such as gaussians , where the inadequacies are compensated by larger basis sets .
* pseudopotentials regard an atom as a perturbation of the free electron gas .
the most natural basis functions are plane waves .
plane waves are complete and well adapted to sufficiently smooth wave functions .
the disadvantage of the large basis sets required is offset by the extreme simplicity to evaluate matrix elements .
finite plane wave expansions are , however , absolutely inadequate to describe the strong oscillations of the wave functions near the nucleus . in the pseudopotential approach the pauli repulsion of the core electrons is therefore described by an effective potential that expels the valence electrons from the core region .
the resulting wave functions are smooth and can be represented well by plane waves .
the price to pay is that all information on the charge density and wave functions near the nucleus is lost . *
augmented wave methods compose their basis functions out of atom - like partial waves in the atomic regions and a set of functions , called envelope functions , appropriate for the bonding in between .
space is divided accordingly into atom - centered spheres , defining the atomic region , and an interstitial region for the bonds .
the partial solutions of the different regions , are matched at the interface between atomic and interstitial regions .
the projector augmented wave method is an extension of augmented wave methods and the pseudopotential approach , which combines their traditions into a unified electronic structure method . after describing the underlying ideas of the various methods let us briefly review the history of augmented wave methods and the pseudopotential approach .
we do not discuss the atomic - orbital based methods , because our focus is the paw method and its ancestors .
the augmented wave methods have been introduced in 1937 by slater @xcite and later modified by korringa @xcite , kohn and rostokker @xcite .
they approached the electronic structure as a scattered electron problem .
consider an electron beam , represented by a plane wave , traveling through a solid .
it undergoes multiple scattering at the atoms .
if for some energy , the outgoing scattered waves interfere destructively , a bound state has been determined .
this approach can be translated into a basis set method with energy dependent and potential dependent basis functions . in order to make the scattered wave problem tractable ,
a model potential had to be chosen : the so - called muffin - tin potential approximates the potential by a constant in the interstitial region and by a spherically symmetric potential in the atomic region .
the pseudopotential approach traces back to 1940 when c. herring invented the orthogonalized plane wave method @xcite .
later , phillips @xcite and antoncik @xcite replaced the orthogonality condition by an effective potential , that compensates the electrostatic attraction by the nucleus . in practice ,
the potential was modified for example by cutting off the singular potential of the nucleus at a certain value .
this was done with a few parameters that have been adjusted to reproduce the measured electronic band structure of the corresponding solid . augmented wave and
pseudopotential methods reached adulthood in the 1970s : at first o.k .
andersen @xcite showed that the energy dependent basis set of slater s apw method can be mapped onto one with energy independent basis functions , by linearizing the partial waves for the atomic regions in energy . in the original apw approach
the zeros of an energy dependent matrix had to be determined , which is problematic , if many states lie in a small energy region as for complex systems . with the new energy
independent basis functions , however , the problem is reduced to the much simpler generalized eigenvalue problem , which can be solved using efficient numerical techniques .
furthermore , the introduction of well defined basis sets paved the way for full - potential calculations . in that case the muffin - tin approximation is used solely to define the basis set .
the matrix elements of the hamiltonian are evaluated with the full potential .
hamann , schlter and chiang @xcite showed in 1979 how pseudopotentials can be constructed in such a way , that their scattering properties are identical to that of an atom to first order in energy .
these first - principles pseudopotentials relieved the calculations from the restrictions of empirical parameters .
highly accurate calculations have become possible .
a main disadvantage of these pseudopotentials has been the large basis set size required especially for first - row and transition metal atoms . in 1985 r.
car and m. parrinello published the ab - initio molecular dynamics method @xcite .
simulations of the atomic motion have become possible on the basis of state - of - the - art electronic structure methods . besides making dynamical phenomena and finite temperature effects accessible to electronic structure calculations ,
the ab - initio molecular dynamics method also introduced a radically new way of thinking into electronic structure methods .
diagonalization of a hamilton matrix has been replaced by classical equations of motion for the wave function coefficients .
if one applies friction , the system is quenched to the ground state . without friction truly dynamical simulations of the atomic structure are performed .
electronic wave functions and atomic positions are treated on equal footing .
the car - parrinello method had been implemented first for the pseudopotential approach .
there seemed to be unsurmountable barriers against combining the new technique with augmented wave methods .
the main problem was related to the potential dependent basis set used sofar : the car - parrinello method requires a well defined and unique total energy functional of atomic positions and basis set coefficients .
therefore , it was one of the main goals of the paw method to introduce energy and potential independent basis sets that were as accurate and numerically efficient as the previously used augmented basis sets .
other requirements have been : ( 1 ) the method should match the efficiency of the pseudopotential approach for car - parrinello simulations .
( 2 ) it should become an exact theory when converged and ( 3 ) its convergence should be easily controlled .
we believe that these criteria have been met , which explains why the paw method becomes increasingly wide spread today .
we would like to point out that most of these seemingly singular developments did not come out of the blue , but the ideas seemed to have evolved in the community .
in the case of the paw method , similar ideas have been developed by vanderbilt @xcite in the context of ultra - soft pseudopotentials .
the first dynamical simulations using a semiempirical electronic structure method have been performed by wang and karplus @xcite in 1973 .
the first ab - initio pseudopotentials have been published by zunger @xcite one year before hamann , bachelet and schlter @xcite .
at the root of the paw method lies a transformation , that maps the true wave functions with their complete nodal structure onto auxiliary wave functions , that are numerically convenient .
we aim for smooth auxiliary wave functions , which have a rapidly convergent plane wave expansion .
with such a transformation we can expand the auxiliary wave functions into a convenient basis set , and evaluate all physical properties after reconstructing the related physical ( true ) wave functions .
let us denote the physical one - particle wave functions as @xmath0 and the auxiliary wave functions as @xmath1 .
note that the tilde refers to the representation of smooth auxiliary wave functions .
@xmath2 is the label for a one - particle state and contains a band index , a @xmath3-point and a spin index .
the transformation from the auxiliary to the physical wave functions is @xmath4 .
@xmath5 corresponds to a ket @xmath0 , the complex conjugate wave function @xmath6 corresponds to a bra @xmath7 , and a scalar product @xmath8 is written as @xmath9 .
vectors in the 3-d coordinate space are indicated by boldfaced symbols .
the electronic ground state is determined by minimizing a total energy functional @xmath10 $ ] of the density functional theory .
the one - particle wave functions have to be orthogonal .
this constraint is implemented with the method of lagrange multipliers .
we obtain the ground state wave functions from the extremum condition for @xmath11,\lambda_{m , n})=e[\psi_n ] -\sum_{n , m}[\langle\psi_n|\psi_m\rangle-\delta_{n , m}]\lambda_{n , m}\end{aligned}\ ] ] with respect to the wave functions and the lagrange multipliers @xmath12 .
the extremum condition for the wave functions has the form @xmath13 where the @xmath14 are the occupation numbers and @xmath15 is the effective one - particle hamilton operator .
after a unitary transformation that diagonalizes the matrix of lagrange multipliers @xmath16 , we obtain the kohn - sham equations .
@xmath17 the one - particle energies @xmath18 are the eigenvalues of @xmath19 .
now we express the functional @xmath20 in terms of our auxiliary wave functions @xmath21,\lambda_{m , n } ) = e[{\cal t}\tilde\psi_n ] -\sum_{n , m}[\langle\tilde\psi_n|{\cal t}^\dagger{\cal t}|\tilde\psi_m\rangle -\delta_{n , m}]\lambda_{n , m}\end{aligned}\ ] ] the variational principle with respect to the auxiliary wave functions yields @xmath22 again we obtain a schrdinger - like equation , but now the hamilton operator has a different form , @xmath23 , an overlap operator @xmath24 occurs , and the resulting auxiliary wave functions are smooth . when we evaluate physical quantities we need to evaluate expectation values of an operator @xmath25 , which can be expressed in terms of either the true or the auxiliary wave functions . @xmath26 in the representation of auxiliary wave functions we need to use transformed operators @xmath27 .
as it is , this equation only holds for the valence electrons .
the core electrons are treated differently as will be shown below .
the transformation takes us conceptionally from the world of pseudopotentials to that of augmented wave methods , which deal with the full wave functions .
we will see that our auxiliary wave functions , which are simply the plane wave parts of the full wave functions , translate into the wave functions of the pseudopotential approach . in the paw method
the auxiliary wave functions are used to construct the true wave functions and the total energy functional is evaluated from the latter .
thus it provides the missing link between augmented wave methods and the pseudopotential method , which can be derived as a well - defined approximation of the paw method . in the original paper @xcite , the auxiliary wave functions have been termed pseudo wave functions and the true wave functions have been termed all - electron wave functions , in order to make the connection more evident .
i avoid this notation here , because it resulted in confusion in cases , where the correspondence is not clear cut .
sofar , we have described how we can determine the auxiliary wave functions of the ground state and how to obtain physical information from them . what is missing , is a definition of the transformation operator @xmath4 .
the operator @xmath4 has to modify the smooth auxiliary wave function in each atomic region , so that the resulting wave function has the correct nodal structure
. therefore , it makes sense to write the transformation as identity plus a sum of atomic contributions @xmath28
@xmath29 for every atom , @xmath28 adds the difference between the true and the auxiliary wave function .
the index @xmath30 is a label for an atomic site .
the local terms @xmath28 are defined in terms of solutions @xmath31 of the schrdinger equation for the isolated atoms .
this set of partial waves @xmath31 will serve as a basis set so that , near the nucleus , all relevant valence wave functions can be expressed as superposition of the partial waves with yet unknown coefficients . @xmath32
the index @xmath33 refers to a site index @xmath30 , the angular momentum indices @xmath34 and an additional index that differentiates partial waves with same angular momentum quantum numbers on the same site .
with @xmath35 we indicate those partial waves that belong to site @xmath30 .
@xmath36 is the position of the nucleus of site @xmath30 .
note that the partial waves are not necessarily bound states and are therefore not normalizable , unless we truncate them beyond a certain radius @xmath37 .
the paw method is formulated such that the final results do not depend on the location where the partial waves are truncated , as long as this is not done too close to the nucleus .
since the core wave functions do not spread out into the neighboring atoms , we will treat them differently .
currently we use the frozen - core approximation so that density and energy of the core electrons are identical to those of the corresponding isolated atoms .
the transformation @xmath4 shall produce only wave functions orthogonal to the core electrons , while the core electrons are treated separately .
therefore , the set of atomic partial waves @xmath38 includes only valence states that are orthogonal to the core wave functions of the atom . for each of the partial waves we choose an auxiliary partial wave @xmath39 .
the identity @xmath40 defines the local contribution @xmath28 to the transformation operator .
since @xmath41 shall change the wave function only locally , we require that the partial waves @xmath38 and their auxiliary counter parts @xmath42 are pairwise identical beyond a certain radius @xmath43 .
@xmath44 in order to be able to apply the transformation operator to an arbitrary auxiliary wave function , we need to be able to expand the auxiliary wave function locally into the auxiliary partial waves .
@xmath45 which defines the projector functions @xmath46 . the projector functions probe the local character of the auxiliary wave function in the atomic region .
examples of projector functions are shown in fig .
[ fig:2 ] . from eq .
[ eq : ps1center ] we can derive @xmath47 , which is valid within @xmath43 . it can be shown by insertion , that the identity eq . [ eq : ps1center ] holds for any auxiliary wave function @xmath48 that can be expanded locally into auxiliary partial waves @xmath39 , if @xmath49 note that neither the projector functions nor the partial waves need to be orthogonal among themselves . + + by combining eq .
[ eq : sr1 ] and eq . [ eq : ps1center ]
, we can apply @xmath50 to any auxiliary wave function .
hence the transformation operator is @xmath52 where the sum runs over all partial waves of all atoms .
the true wave function can be expressed as @xmath53 with @xmath54 in fig .
[ fig:1 ] the decomposition of eq .
[ eq : aewave ] is shown for the example of the bonding p-@xmath55 state of the cl@xmath56 molecule .
+ orbital of the cl@xmath56 molecule and its decomposition of the wave function into auxiliary wave function and the two one - center expansions .
top - left : true and auxiliary wave function ; top - right : auxiliary wave function and its partial wave expansion ; bottom - left : the two partial wave expansions ; bottom - right : true wave function and its partial wave expansion.,title="fig:",width=113 ] orbital of the cl@xmath56 molecule and its decomposition of the wave function into auxiliary wave function and the two one - center expansions .
top - left : true and auxiliary wave function ; top - right : auxiliary wave function and its partial wave expansion ; bottom - left : the two partial wave expansions ; bottom - right : true wave function and its partial wave expansion.,title="fig:",width=113 ] + orbital of the cl@xmath56 molecule and its decomposition of the wave function into auxiliary wave function and the two one - center expansions .
top - left : true and auxiliary wave function ; top - right : auxiliary wave function and its partial wave expansion ; bottom - left : the two partial wave expansions ; bottom - right : true wave function and its partial wave expansion.,title="fig:",width=113 ] orbital of the cl@xmath56 molecule and its decomposition of the wave function into auxiliary wave function and the two one - center expansions .
top - left : true and auxiliary wave function ; top - right : auxiliary wave function and its partial wave expansion ; bottom - left : the two partial wave expansions ; bottom - right : true wave function and its partial wave expansion.,title="fig:",width=113 ] to understand the expression for the true wave function , eq .
[ eq : aewave ] , let us concentrate on different regions in space .
( 1 ) far from the atoms , the partial waves are , according to eq .
[ eq : equaloutside ] , pairwise identical so that the auxiliary wave function is identical to the true wave function @xmath57 .
( 2 ) close to an atom , however , the true wave function @xmath58 is built up from partial waves that contain the proper nodal structure , because the auxiliary wave function and its partial wave expansion are equal according to eq .
[ eq : ps1center ] . in practice
the partial wave expansions are truncated .
therefore , the identity of eq .
[ eq : ps1center ] does not hold strictly . as a result the plane waves also contribute to the true wave function inside the atomic region .
this has the advantage that the missing terms in a truncated partial wave expansion are partly accounted for by plane waves , which explains the rapid convergence of the partial wave expansions .
frequently , the question comes up , whether the transformation eq .
[ eq : transf ] of the auxiliary wave functions indeed provides the true wave function .
the transformation should be considered merely as a change of representation analogous to a coordinate transform .
if the total energy functional is transformed consistently , its minimum will yield an auxiliary wave function that produces a correct wave function @xmath59 .
expectation values can be obtained either from the reconstructed true wave functions or directly from the auxiliary wave functions @xmath60 where @xmath14 are the occupations of the valence states and @xmath61 is the number of core states .
the first sum runs over the valence states , and second over the core states @xmath62 .
now we can decompose the matrix elements into their individual contributions according to eq .
[ eq : aewave ] .
@xmath63 only the first part of eq .
[ eq : expect ] , is evaluated explicitly , while the second and third parts of eq .
[ eq : expect ] are neglected , because they vanish for sufficiently local operators as long as the partial wave expansion is converged : the function @xmath64 vanishes per construction beyond some augmentation region , because the partial waves are pairwise identical beyond that region .
the function @xmath65 vanishes inside the augmentation region , if the partial wave expansion is sufficiently converged . in no region of space
both functions @xmath64 and @xmath66 are simultaneously nonzero
. similarly the functions @xmath64 from different sites are never non - zero in the same region in space .
hence , the second and third parts of eq .
[ eq : expect ] vanish for operators such as the kinetic energy @xmath67 and the real space projection operator @xmath68 , which produces the electron density . for truly nonlocal operators the second and third parts of eq .
[ eq : expect ] would have to be considered explicitly .
the expression , eq .
[ eq : expect ] , for the expectation value can therefore be written as @xmath69 where @xmath70 is the one - center density matrix defined as @xmath71 the auxiliary core states , @xmath72 allow to incorporate the tails of the core wave function into the plane wave part , and therefore assure , that the integrations of partial wave contributions cancel strictly beyond @xmath43 .
they are identical to the true core states in the tails , but are a smooth continuation inside the atomic sphere .
it is not required that the auxiliary wave functions are normalized .
for example , the electron density is given by @xmath73 where @xmath74 is the core density of the corresponding atom and @xmath75 is the auxiliary core density that is identical to @xmath74 outside the atomic region and a smooth continuation inside .
before we continue , let us discuss a special point : the matrix element of a general operator with the auxiliary wave functions may be slowly converging with the plane wave expansion , because the operator @xmath25 may not be well behaved .
an example for such an operator is the singular electrostatic potential of a nucleus .
this problem can be alleviated by adding an intelligent zero : if an operator @xmath76 is purely localized within an atomic region , we can use the identity between the auxiliary wave function and its own partial wave expansion @xmath77 now we choose an operator @xmath76 so that it cancels the problematic behavior of the operator @xmath25 , but is localized in a single atomic region . by adding @xmath76 to the plane wave part and the matrix elements with its one - center expansions , the plane wave convergence can be improved without affecting the converged result .
like wave functions and expectation values also the total energy can be divided into three parts . @xmath78,r_i)&=&\tilde{e}+\sum_r\bigl(e^1_r-\tilde{e}^1_r\bigr)\end{aligned}\ ] ] the plane - wave part @xmath79 involves only smooth functions and is evaluated on equi - spaced grids in real and reciprocal space .
this part is computationally most demanding , and is similar to the expressions in the pseudopotential approach . @xmath80
[ \tilde{n}(\mathbf{r}')+\tilde{z}(\mathbf{r}')]}{|\mathbf{r}-\mathbf{r}'| } \nonumber\\ & + & \int d^3r \tilde{n}(\mathbf{r})\epsilon_{xc}(\mathbf{r},[\tilde{n } ] ) + \int d^3r \bar{v}(\mathbf{r})\tilde{n}(\mathbf{r } ) , \label{eq : psetot}\end{aligned}\ ] ] where @xmath81 is an angular dependent core - like density that will be described in detail below .
the remaining parts can be evaluated on radial grids in a spherical harmonics expansion .
the nodal structure of the wave functions can be properly described on a logarithmic radial grid that becomes very fine near nucleus , @xmath82[n^1(\mathbf{r}')+z(\mathbf{r}')]}{|\mathbf{r}-\mathbf{r}'| } \nonumber\\ & + & \int d^3r n^1(\mathbf{r})\epsilon_{xc}(\mathbf{r},[n^1 ] ) \\
\tilde{e}^1_r&=&\sum_{i , j\in r } d_{i , j } \langle\tilde\phi_j|\frac{-\hbar^2}{2m_e}\nabla^2|\tilde\phi_i\rangle \nonumber\\ & + & \frac{e^2}{8\pi\epsilon_0}\int d^3r\int d^3r ' \frac{[\tilde{n}^1(\mathbf{r})+\tilde{z}(\mathbf{r})][\tilde{n}^1(\mathbf{r}')+\tilde{z}(\mathbf{r}')]}{|\mathbf{r}-\mathbf{r}'| } \nonumber\\ & + & \int d^3r \tilde{n}^1(\mathbf{r})\epsilon_{xc}(\mathbf{r},[\tilde{n}^1 ] ) + \int d^3r \bar{v}(\mathbf{r})\tilde{n}^1(\mathbf{r } ) \label{eq : ps1etot}\end{aligned}\ ] ] the nuclear charge density @xmath83 is defined as a sum of @xmath84-functions on the nuclear sites , @xmath85 , with the atomic numbers @xmath86 .
note that the self energy of a point charge is infinite and must be subtracted out . the compensation density @xmath87 is given as a sum of angular momentum dependent gauss functions , which have an analytical fourier transform .
a similar term occurs also in the pseudopotential approach .
in contrast to the norm - conserving pseudopotential approach however , the compensation charge is non - spherical and it is constantly adapting to the instantaneous environment .
it is constructed such that the augmentation charge densities @xmath88 have vanishing electrostatic multi - pole moments for each atomic site . as a result
the sum of all one - center contributions from one atom does not produce an electrostatic potential outside their own atomic region .
this is the reason that the electrostatic interaction of the one - center parts between different sites vanish . the compensation charge density as given here is still localized within the atomic regions , but a technique similar to an ewald summation allows to replace it by a very extended charge density .
thus we can achieve , that all functions in @xmath79 converge as fast as the auxiliary density itself .
the potential @xmath89 , which occurs in eqs .
[ eq : psetot ] and [ eq : ps1etot ] enters the total energy in the form of a zero described in eq .
[ eq : zeroop ] @xmath90 the main reason for introducing this potential is that the self - consistent potential resulting from the plane wave part is not necessarily optimally smooth .
the potential @xmath89 allows to influence the plane wave convergence beneficially , without changing the converged result .
@xmath89 must be localized within the augmentation region , where equation [ eq : ps1center ] holds .
once the total energy functional provided in the previous section has been defined , everything else follows : forces are partial derivatives with respect to atomic positions .
the potential is the derivative of the potential energy with respect to the density , and the hamiltonian follows from derivatives with respect to wave functions .
the fictitious lagrangian approach of car and parrinello @xcite does not allow any freedom in the way these derivatives are obtained .
anything else than analytic derivatives will violate energy conservation in a dynamical simulation .
since the expressions are straightforward , even though rather involved , we will not discuss them here .
all approximations are incorporated already in the total energy functional of the paw method .
what are those approximations ?
* firstly we use the frozen core approximation . in principle
this approximation can be overcome .
* the plane wave expansion for the auxiliary wave functions must be complete .
the plane wave expansion is controlled easily by increasing the plane wave cutoff defined as @xmath91 .
typically we use a plane wave cutoff of 30 ry . *
the partial wave expansions must be converged .
typically we use one or two partial waves per angular momentum @xmath34 and site .
it should be noted that the partial wave expansion is not variational , because the partial wave expansion changes the total energy functional and not only the basis set .
we do not discuss here numerical approximations such as the choice of the radial grid , since those are easily controlled .
we mentioned earlier that the pseudopotential approach can be derived as a well defined approximation from the paw method : the augmentation part @xmath92 is a functional of the one - center density matrix @xmath70 defined in eq .
[ eq:1cdenmat ] . the pseudopotential approach can be recovered if we truncate a taylor expansion of @xmath93 about the atomic density matrix after the linear term .
the term linear to @xmath70 is the energy related to the nonlocal pseudopotential .
@xmath94 thus we can look at the paw method also as a pseudopotential method with a pseudopotential that adapts to the instantaneous electronic environment , because the explicit nonlinear dependence of the total energy on the one - center density matrix is properly taken into account .
what are the main advantages of the paw method compared to the pseudopotential approach ?
firstly all errors can be systematically controlled so that there are no transferability errors . as shown by watson @xcite and kresse @xcite
, most pseudopotentials fail for high spin atoms such as cr . while it is probably true that pseudopotentials can be constructed that cope even with this situation , a failure can not be known beforehand , so that some empiricism remains in practice : a pseudopotential constructed from an isolated atom is not guaranteed to be accurate for a molecule .
in contrast , the converged results of the paw method do not depend on a reference system such as an isolated atom , because it uses the full density and potential .
the paw method provides access to the full charge and spin density , which is relevant for hyperfine parameters .
hyperfine parameters are sensitive probes of the electron density near the nucleus . in many situations
they are the only information available that allows to deduce atomic structure and chemical environment of an atom .
there are reconstruction techniques for the pseudopotential approach , which however , are poor man s versions @xcite of the paw method .
the plane wave convergence is more rapid than in norm - conserving pseudopotentials and should in principle be equivalent to that of ultra - soft pseudopotentials @xcite . compared to the ultra - soft pseudopotentials , however , the paw method has the advantage that the total energy expression is less complex and therefore is expected to be more efficient .
the construction of pseudopotentials requires to determine a number of parameters .
as they influence the results , their choice is critical .
also the paw methods provides some flexibility in the choice of auxiliary partial waves .
however , this choice does not influence the converged results .
since the first implementation of the paw method in the cp - paw code , a number of groups have adopted the paw method .
the second implementation was done by the group of holzwarth @xcite .
the resulting pwpaw code is freely available @xcite .
this code is also used as a basis for the paw implementation in the abinit project @xcite .
an independent paw code has been developed by valiev and weare @xcite .
recently the paw method has been implemented into the vasp code @xcite .
the paw method has also been implemented by w. kromen into the estcompp code of blgel and schrder @xcite .
another branch of methods uses the reconstruction of the paw method , without taking into account the full wave functions in the self - consistency .
following chemist notation this approach could be termed `` post - pseudopotential paw '' .
this development began with the evaluation for hyperfine parameters from a pseudopotential calculation using the paw reconstruction operator @xcite and is now used in the pseudopotenital approach to calculate properties that require the correct wave functions the implementation by kresse and joubert @xcite has been particularly useful as they had an implementation of paw in the same code as the ultra - soft pseudopotentials , so that they could critically compare the two approaches with each other and lapw calculations .
their conclusion is that both methods compare well in most cases , but they found that magnetic energies are seriously by a factor two in error in the pseudopotential approach , while the results of the paw method were in line with other all - electron calculations using the linear augmented plane wave method . as a short note , kresse and joubert
incorrectly claim that their implementation is superior as it includes a term that is analogous to the non - linear core correction of pseudopotentials @xcite : this term however is already included in the original version in the form of the pseudized core density . several extensions of the paw have been done in the recent years : for applications in chemistry truly isolated systems are often of great interest .
as any plane - wave based method introduces periodic images , the electrostatic interaction between these images can cause serious errors .
the problem has been solved by mapping the charge density onto a point charge model , so that the electrostatic interaction could be subtracted out in a self - consistent manner @xcite . in order to include the influence of the environment , the latter was simulated by simpler force fields using the molecular - mechanics - quantum - mechanics ( qm - mm ) approach @xcite . in order to overcome the limitations of the density functional theory
several extensions have been performed .
bengone @xcite implemented the lda+u approach @xcite into the cp - paw code .
soon after this , arnaud @xcite accomplished the implementation of the gw approximation into the cp - paw code .
the vasp - version of paw @xcite and the cp - paw code have now been extended to include a noncollinear description of the magnetic moments . in a non - collinear description
the schrdinger equation is replaced by the pauli equation with two - component spinor wave functions the paw method has proven useful to evaluate electric field gradients @xcite and magnetic hyperfine parameters with high accuracy @xcite .
invaluable will be the prediction of nmr chemical shifts using the gipaw method of pickard and mauri @xcite , which is based on their earlier work @xcite .
while the gipaw is implemented in a post - pseudopotential manner , the extension to a self - consistent paw calculation should be straightforward .
an post - pseudopotential approach has also been used to evaluate core level spectra @xcite and momentum matrix elements @xcite .
we are grateful for carefully reading the manuscript to dr.j.noffke , dr.r.schmidt and p.walther and to prof.k.schwarz for his continuous support .
this work has benefited from the collaborations within the esf programme on electronic structure calculations for elucidating the complex atomistic behavior of solids and surfaces. aaaa http://www.mapr.ucl.ac.be/abinit/ andersen o.k . , 1975 ,
rev . b**12 * * , 3060 .
anisimov v.i .
, zaanen j. and andersen o.k . , 1991 , * 44 * , 943 .
antoncik e. , 1959 , j. phys . chem .
solids * 10 * , 314 .
arnaud b. and alouani m. , 2000 , phys .
b. * 62 * , 4464 .
bengone o. , alouani m. , blchl p.e . and hugel j. , 2000 , phys . rev .
b62 , 16392 .
blchl p.e . , 1994 ,
b * 50 * , 17953 .
bchl p.e . , 1995 , j. chem .
phys.*103 * , 7422 .
bchl p.e . ,
2000 , phys , rev .
b**62 * * 6158 .
blgel s. , kromen w. und schroeder k. , 2001 , spring meeting of the german physical society 2001 , hamburg , germany , abstract http://dpg.rz.uni-ulm.de/prog/html/o_33.html car r. and parrinello m. , 1985 , phys .
lett . * 55 * , 2471 .
hamann d.r .
, schlter m. and chiang c. , 1979 , phys .
. lett . * 43 * , 1494 .
hobbs d. , kresse g. , hafner j. , 2000 , phys .
b 62 , 11556 .
holzwarth n.a.w . ,
mathews g.e . , dunning r.b .
, tackett a.r . , and zheng y. , 1997 , phys .
b * 55*,2005 . herring c. , 1940 , phys
. rev . * 57 * , 1169 .
hohenberg p. and kohn w. , 1964 , phys .
* 136 * , b864 .
jayawardane d.n .
, pickard c.j . , brown l.m . , and payne m.c . , 2001 ,
phys . rev .
b 64 , 115107 .
kageshima h. and shiraishi k. , 1997 , phys .
rev . b**56 * * , 143985 .
kohn w. and rostocker j. , 1954 , phys . rev . * 94 * , 1111 .
kohn w. and sham l.j .
, 1965 , phys . rev . *
140 * , a1133 .
korringa j. , 1947 , physica(utrecht ) * 13 * , 392 .
kresse g. and joubert j. , 1999 , phys .
b * 59 * , 1758 .
louie s.g .
, froyen s. and cohen m.l . , 1982 ,
rev . b**26 * * , 1738 .
mauri f. , pfrommer b.g . and louie s.g .
, 1996 , phys .
lett * 77 * , 5300 .
petrilli h.m .
, blchl p.e . , blaha p. , and schwarz k. , 1998 , phys .
rev . b**57 * * , 14690 .
phillips j.c . and
kleinman l. , 1959 , phys . rev . * 116 * , 287 . pickard c.j . and mauri f. , 2001 , phys . rev . b. * 63 * , 245101 .
slater j.c . , 1937 ,
rev.*51 * , 846 .
tackett a.r .
, holzwarth n.a.w . and matthews g.e . , 2001 ,
computer physics communications 135 329 - 347 , ibid .
348 - 376 .
valiev m. and weare j.h . , 1999 ,
a , 103 , 10588 .
vanderbilt d. , 1990 , phys .
b * 41 * , 7892 .
van de walle c.g . and blchl p.e . , 1993 ,
b * 47 * , 4244 . wang i.s.y and karplus m. , 1973 , j. amer .
soc . * 94 * , 8160 .
watson s.c . and carter e.a . , 1998 ,
b * 58 * , r13309 .
woo t.k .
, margl p.m. , blchl p.e . and ziegler t. , 1997 , j. phys .
b**101**,7877 ; woo t.k .
, blchl p.e . and
ziegler t. , 2000 , j. mol .
struct.(theochem ) * 506 * , 313 .
zunger a. and cohen m. , 1978 , phys .
b * 18 * , 5449 .
|
a brief introduction to the projector augmented wave method is given and recent developments are reviewed .
the projector augmented wave method is an all - electron method for efficient ab - initio molecular dynamics simulations with full wave functions .
it extends and combines the traditions of existing augmented wave methods and the pseudopotential approach . without sacrificing efficiency ,
the paw method avoids transferability problems of the pseudopotential approach and it has been valuable to predict properties that depend on the full wave functions .
|
the paper presents the model of an elementary particle and particularly the model of the neutral nucleon and its excited states .
since the model is atomic - like i will give a concise reminder both of the history of the models of an atom and of a neutron which still overlap until the present day .
possibly the first model of an atom in the 20-th century was kelvin s static one @xcite of a sphere of uniformly distributed positive electricity with embedded discrete electrons , followed closely by the thomson s model of `` the pudding with raisin muffins in it '' @xcite .
models also appeared at the end of the 19-th century , stimulated by empirical facts suggesting the complexity of the structure of the atom and its composition from smaller components @xcite .
as a result , both the theoretical intuition and the empirical indications prompted the sense of the electron to thomson .
he disagreed with the idea of the planetary `` point planets '' model of the atom but probably was influenced by maxwell s considerations on saturn s rings @xcite .
hence , contrary to kelvin s model , the one by thomson @xcite is not static and one of its characteristics is the motion of a ring of _ n _ negatively charged particles ( electrons @xcite , as they were called by stoney @xcite ) , which are set within an uniformly charged sphere ( with the radius of the order of @xmath0 m ) having a number of particles which vary from ring to ring .
the attractive force of the positively charged sphere and the repulsion of the negatively charged electrons arranged in a series of parallel rings were to guarantee the stability of the system .
the strange characteristic of the model is that if the number of the negative particles is larger than 5 , some of them must be set in the central position of the sphere yet not necessarily strictly in the center , or else the movement in the atom becomes unstable .
this problem still arouses interest until the present day @xcite .
+ thomson s atom has two types of vibrations : the first one connected with the movement of electrons along their orbits and the second one developing out of the distortion of a ring from its equilibrium circular shape in a particular spherical shell .
although the model does not explain the characteristic regularities of the observed spectral lines and the deflection of an electron when it passes through matter , it was regarded as a sufficient one by many physicists until the end of the first decade of the 20th century .
this was due to the fact that he found a close relation between the occupation of the electrons inside rings formed in the concentric shells and the regularities inside the mendeleev s periodic table .
however , at the same time some scientists considered that the planetary model of the atom was more appropriate .
+ in 1920 rutherford considered his model of the neutron in relation to the compressed hydrogen atom in the core of a star and presented his hypothesis on the structure of the neutron @xcite .
he perceived the neutron as an electron pressed into the interior of a proton hence his model of the neutron was more in line with thomson s idea of the atom than of the planetary kind . in 1932
chadwick experimentally confirmed rutherford s hypothesis on the existence of the neutron .
+ in the period from 1925 to 1928 there emerged three formulations of quantum mechanics : heisenberg s @xcite , dirac s @xcite and schrdinger s @xcite .
they were developed nearly at the same time .
the schrdinger equation provides a direct representation of the mechanical system using the notion of the wave function for a particular state . here
the eigenstate functions form the orthogonal reference frame in the hilbert space in which the wave function and its time evolution can be analyzed .
the specific schrdinger s substantial approach to the square modulus of the wave function gave the name wave mechanics to his interpretation of quantum mechanics which is in line with de broglie s theory of phase waves @xcite .
+ two questions about the main characteristics of the type of the models of an atom or nucleon built in these formalisms arise .
firstly , are they thomson s with deep overlapping of constituent fields or planetary - like , for which the constituent fields do not overlap so deeply ? on the surface , all of them carry some planetary features , especially if the hamiltonians of the particular models are analyzed .
this may be easily understood because in this respect they are classical mechanics descendants , yet in their main idea connected with the nonlocality of the wave function they are thomson oriented , regardless of the fact whether the interpretation of the wave function is schrdinger s or copenhagen s .
indeed , in the atom the electron wave function is deeply embedded in the nucleus field .
secondly , what is the geometry of the manifold ( space or space - time ) which forms the background arena for the disputed configurations of fields ?
the common feature of thomson s and schrdinger s like models of the atom is the homogeneity of their space - time structure .
we will come back to this problem in the passage below .
+ now , let us return to the neutron .
since it seemed that quantum mechanics did not permit a consistent representation of the neutron as a bound state of the proton and electron , the before mentioned rutherford s conception of the structure of the neutron has been rejected by the majority of physicists . yet , until now attempts have been made to restore this model .
for instance such attempts are connected with the construction of a covering of quantum mechanics known as hadronic mechanics @xcite for the specific objective of achieving a consistent quantitative treatment of nonlocal , nonlinear and nonpotential effects in deep wave - overlappings of particles at short distances .
as far as the quantum models of the neutron are considered , the most popular is the quantum field theory version of the bag model .
setting aside the difficulties of this model with the determination of the spin of the nucleon we give some literature references @xcite only citing after @xcite : `` ... all spin parts @xmath1 $ ] have to add to @xmath2 which is incredible in the light of the present day experiments .
this may indicate that some underlying symmetries , unknown at present , are playing a role in forming the various contributing parts such that the final sum rule gives the fermion @xmath2 value '' .
+ another story is partly connected with the name of riemann . in his lecture
entitled `` on the hypotheses that lie at the foundations of geometry '' @xcite , given in 1854 , he emphasized that the truth about space has to be discovered from physical experience .
the clue is that he noted that the geometry of space could be highly irregular at very small distances , yet appearing smooth at the observed ones .
he wrote : `` space [ in the large ] if one ascribes to it a constant curvature , is necessarily finite , provided only that this curvature has a positive value , however small ... .
it is quite conceivable that the geometry of space in the very small does not satisfy the axioms of [ euclidean ] geometry ... '' . + but the story of the atom and nucleon may be told and retold ( at least up to now ) in many ways @xcite .
not that all of them are right .
the latest attractive model of the ( hydrogen ) atom , based on the isomorphism between maxwell and dirac formalisms ( or rather on optics - mechanics isomorphism called hamilton s analogy ) , is connected with the name of sallhofer @xcite .
he worked out the formal mathematical strong similarity ( i would not call it identity ) , in the minkowski space , between electrodynamics and wave mechanics by means of which he proved that the hydrogen atom might be seen as a pair of mutually refracting electromagnetic waves .
previously this similarity was pointed out with amusement by sakurai @xcite . below the model of a neutral nucleon
will be presented which is more planetary ( fields do not overlap very deeply ) and incorporates the riemann s idea of space .
the metric of the underlying space - time was previously obtained in @xcite where a six - dimensional kaluza
klein type model at the classical level is considered . in them
the static spherically symmetric solution to the coupled six - dimensional einstein and klein - gordon equations was derived in the presence of the basic massless dilaton field @xmath3 which forms a kind of ground field for the self - field of gravitation , the notions of which were discussed elsewhere @xcite .
the solutions presented in @xcite are parameterized by parameter @xmath4 which has similar dynamical consequences as mass @xmath5 @xcite , i.e. its existence would be perceived by an observer in the same way as invisible mass which could be the extended `` center '' of a particle . because the solution is horizon free
hence it is fundamentally different from the four - dimensional schwarzschild one .
the idea of the six - dimensional space - time re - enters the physics occasionally @xcite-@xcite . yet the motivations for choosing the six - dimensional models were diverse .
foe instance in @xcite nishino , salam and sezgin suggested that one may obtain the fermion spectrum in four - dimensions within the framework of d=6 , @xmath6 kaluza - klein supergravity .
a six - dimensional model of the kaluza - klein theory was also previously investigated by maka and syska @xcite , and by ivashchuk , melnikov and bronnikov @xcite .
recently sparling @xcite has followed the concept of @xmath7 dimensional structure of the space - time @xcite .
unfortunately , models described in @xcite contain the time component with two additional time dimensions .
however there is a way to extend the four - dimensional space - time to the multi - dimensional one via the fisherian statistical analysis .
few years ago frieden proved @xcite that the fisher information channel capacity is the statistical ancestor of the kinematical part of the well known field theory models . using the channel capacity notion , which is inherently connected with the internal parametric space of the sample ( collected by the particle alone ) , and two informational principles , the structural and the variational one , frieden obtained the klein - gordon equation of motion with the proper relativistic dispersion relation @xcite .
the structural information principle has been recently proven in @xcite .
the point is that the channel capacity which has turned out to be the kinematical part of the action of any field theory model is additive both in the minkowskian @xmath8 and the sample @xmath9 ( parametric ) indices @xcite .
this enables the partition of the channel capacity @xmath10 into two parts , one for the four - dimensional space - time and the other for the inner parametric space of the sample whose dimension depends on the spinorial representation of the field @xcite .
when we note by @xmath11 the added fluctuations of the data @xmath12 ( collected by the system alone ) from the expectation positions @xmath13 @xcite , then it can be proven that it takes the following form : @xmath14 where @xmath15 are the original field amplitudes in the sample and multiplication , the likelihood of the sample is calculated . ]
@xcite and @xmath16 are the original fisherian variables .
the amplitudes are factorized as follows @xmath17 and the metric tensor @xmath18 ( @xmath19 ) is obtained effectively from this procedure @xcite .
factor @xmath20 in the kinematical form of @xmath10 signifies the fisher information origin of the action , yet it does not enter into the equation of motion @xcite as it is also factored out from the structural information @xcite . + for the complex scalar field the dimension @xmath21 of the extra parametric space is equal to @xmath22 . in this way the six - dimensionality of the space having the mixed space - time and parametric character , is chosen .
hence the model with the internal space geometry of the two - dimensional torus parameterized by two angles will be presented .
this provides us with geometry of the internal space , which in sections 3 and 4 enables the description of the spinorial field . in @xcite biesiada , maka and
syska showed that the six - dimensionality of this space - time enables the self - consistency of the background solution of the coupled einstein and klein - gordon equations with internal space , compactified in a non - homogeneous manner .
for the sake of clarity , the summary of @xcite is presented below in section 2 . in sections 3 and 4
the model will raise a new issue on the wave - mechanical analysis of the possible general structure of a particle . in section 4 its application to the description of the neutron and its excited states
is presented . in general , statistical models which one can obtain in this way
are of the kaluza - klein type with four - dimensional space - time and @xmath21-dimensional internal ( parametric ) space .
in @xcite a @xmath23 six - dimensional field theory has been considered ( see also @xcite ) which comprises the gravitational self field described by a metric tensor , @xmath18 , and a real massless `` basic '' scalar ( dilatonic ) field , @xmath3 .
we have decomposed the action into two parts : @xmath24 where @xmath25 is the einstein hilbert action and @xmath26 is the action for a real massless scalar ( dilatonic ) field with the lagrangian density equal to @xmath27 . in eq.([row_dzialanie - eh - fi ] ) @xmath28 denotes the determinant of the metric tensor , @xmath29 is the curvature scalar of the six - dimensional ( in general curved ) space - time , and @xmath30 denotes the coupling constant of the six - dimensional theory , analogous to the familiar newtonian gravity constant .
+ by extremalizing the action given by eq.([row_dzialanie - eh - fi ] ) we obtain the einstein equations @xmath31 where @xmath32 is the einstein tensor , @xmath33 is the six - dimensional ricci tensor , @xmath29 is the curvature scalar and @xmath34 is the energy - momentum tensor of a real scalar ( dilatonic ) field @xmath3 which is given by @xmath35 variation of the total action @xmath36 with respect to the field @xmath3 gives the klein - gordon equation @xmath37 where @xmath38 is the tensor dual to @xmath18 .
now consider the six - dimensional space - time which is a topological product of the curved four - dimensional physical space - time ( with the metric @xmath39 ) and the internal compactified space ( with the metric @xmath40 ) .
therefore the metric tensor can be factorized as @xmath41 the four - dimensional diagonal part is assumed to be that of a spherically symmetric geometry @xmath42 where @xmath43 and @xmath44 are ( at this stage ) two arbitrary functions .
analogously , we take the two - dimensional internal part to be @xmath45 the six - dimensional coordinates @xmath46 are denoted by @xmath47 @xmath48 where @xmath49 is the usual time coordinate , @xmath50 @xmath51 $ ] and @xmath52 are the familiar three - dimensional spherical coordinates in the macroscopic space ; @xmath53 and @xmath54 are coordinates in the internal two - dimensional parametric space and @xmath55 is the `` radius '' of this internal space .
we assume that @xmath56 is the function of the radius @xmath57 in the external three - dimensional space .
the internal space is a 2-dimensional topological torus with @xmath57-dependent parameter @xmath58 . using eqs.([row_7])-([row_8 ] )
, we can calculate the components of the ricci tensor .
they are given in appendix a. now we seek for a solution of the einstein equations ( see eq.([row_3 ] ) ) with the ricci tensor given in appendix a by eqs.([row_10])-([row_14 ] ) , with @xmath59 , and with the following boundary conditions @xmath60 which at the spatial infinity reproduces the flat external four - dimensional minkowski space - time and the static internal space of `` radius '' @xmath61 which is of the order calculated in section 4 . we also suppose that the scalar field @xmath3 is the function of the radius @xmath57 alone , @xmath62 , and we impose the following boundary condition for the scalar field @xmath3 : @xmath63 which supplements boundary conditions ( [ row_16 ] ) for the metric components . by virtue of eqs.([row_4 ] ) and ( [ row_dzialanie - eh - fi ] ) it is easy to see that the only nonvanishing components of the energy - momentum tensor are @xmath64 consequently , it is easy to verify that the self - consistent solution of the coupled einstein ( [ row_3 ] ) and klein - gordon ( [ row_5 ] ) equations is @xmath65 @xmath66 hence it results that the only nonzero component of the ricci tensor ( see eqs.([row_10])-([row_14 ] ) in appendix a ) is @xmath67 .
so the curvature scalar @xmath29 is equal to @xmath68 where @xmath4 is the real constant , with the dimensionality of length , whose value is to be taken from the observation of each particular system ( but see section 4 ) .
we notice that all of the six diagonal einstein equations ( [ row_3 ] ) have shrunk to just one @xmath69 now we can rewrite the metric tensor in the form @xmath70 with its determinant equal to @xmath71 to summarize we notice that the real massless `` basic '' free scalar field @xmath72 ( see eq.([row_22 ] ) ) can be the source of the nonzero metric tensor as in eq.([row_26 ] ) .
only when the constant @xmath4 is equal to zero , the solutions ( [ row_20 ] ) ( [ row_22 ] ) become trivial and the six - dimensional space - time is ricci flat .
it is worth noting that since ( with the solutions given by eqs.([row_20 ] ) and ( [ row_22 ] ) ) the components @xmath73 and @xmath74 of the ricci tensor ( appendix a ) are equal to zero for all values of @xmath4 , the internal space is always ricci flat .
however , we must not neglect the internal space because its `` radius '' @xmath75 is a function of @xmath57 and the two spaces , external and internal , are therefore `` coupled '' .
only when @xmath76 are these two spaces `` decoupled '' , and the four - dimensional space - time becomes minkowski flat is not equal to zero , our four - dimensional external space - time is curved .
its scalar curvature @xmath77 can be given by eq.([row_24 ] ) , i.e. @xmath78 . ] .
the self consistent solution given by eqs.([row_22 ] ) and ( [ row_26 ] ) of the coupled einstein and klein - gordon equations is unique . to answer the question on the stability of this self - consistent gravity - dilatonic configuration
let us calculate its energy : @xmath79 where using eqs.([row_19])-([row_27 ] ) we obtain : @xmath80 with @xmath81 the integral of @xmath82 does not converge on @xmath83 , hence the self consistent solution ( [ row_22 ] ) and ( [ row_26 ] ) can not be obtained in the limit @xmath84 , from the convergent solutions around @xmath85 corresponding to @xmath76 .
+ if we depart from the self consistent solution then a similar result can be obtained from the information analysis in the following way . from the notion of the fisher information @xcite
, we notice that @xmath61 is ( for @xmath86 , see eq.([row_20 ] ) ) the characteristic dimension of the inner parametric space , i.e. the radius of the expected value of the position of the system . in the frieden analysis @xmath57
is the added fluctuation to @xmath61 @xcite .
hence if we assume that the fluctuation @xmath57 is not smaller than the expected value @xmath61 then the physical limit for the fluctuation @xmath57 is @xmath87 . in eq.([q ] ) the cutoff @xmath88 can be taken which leads to @xmath89 .
therefore @xmath90 is finite and its value could be even small in contradistinction to the infinite value obtained in eq.([q ] ) for @xmath91 in the self consistent case .
hence , we see that the unique self consistent solution given by eqs.([row_22 ] ) and ( [ row_26 ] ) can not be destabilized to yield any other .
+ let us also notice that in eq.([config energy ] ) the partition of @xmath92 into two parts can be obtained by using the fisher information formalism developed for physical models by frieden @xcite . according to @xcite @xmath93
is connected with the fisherian kinematical degrees of freedom of the gravitational configuration whereas @xmath94 with its structural degrees of freedom .
+ to conclude this section it is worth noting that the metric given by eq.([row_26 ] ) , with the dilatonic field given by eq.([row_22 ] ) , serves , under further conditions discussed below , as a background fields configuration .
this metric appears in the equation of motions for all new fields which weakly enter the system .
until now the calculations were fully self - consistent . the model presented below
adds a new scalar field to the system . yet , since it procures big analytical complications the model will stop being fully self - consistent .
this means among others that we will decline from making the self - consistent correction of the metric ( [ row_26 ] ) received in section 2 .
therefore the following calculations are made totally on the basis of the wave - mechanical approach .
let us investigate the klein
gordon relativistic wave equation for the motion of a scalar particle with a wave function @xmath95 and mass @xmath96 ( for @xmath97 it would be a @xmath98-dimensional tachion ) , @xmath99 in the six - dimensional space - time given by the central gravitational field described by the metric , @xmath18 , given by eq.([row_26 ] ) with its determinant given by eq.([row_27 ] ) .
this metric tensor is used in the above klein gordon equation as the background metric only .
hence we neglect the modification of the einstein equations . therefore a particle described by the @xmath95 wave function moves without changing the metric @xmath18 in reciprocal action .
this is called the wave - mechanical approach . _ but to make our calculations self - consistent by the inclusion of the einstein equations coupled to klein - gordon equation , we should follow the pattern of the previous section for all scalar fields ( particles ) which have the mass @xmath96 and are inherently tied up by strong interaction ( compare section [ bounding_energy ] )
. this procedure should give the self - consistent change of the metric tensor @xmath18 .
_ we incorporate the so - called `` natural interpretation '' of the wave equation then we do not have any problems with possible negative values of probability density which is the inherent property of the probabilistic interpretation of quantum mechanics for the klein
gordon equation @xcite . ]
@xcite with a particle as an oscillating substance described by the wave function @xmath95 which is deformable according to eq.([rownanie ] ) .
+ using eqs.([row_26 ] ) and ( [ row_27 ] ) , we can rewrite the klein - gordon equation as follows : @xmath100 in order to isolate the time dependence , the standard procedure of the separation of variables is performed with the following factorization of @xmath95 @xmath101 in this way , we obtain the set of stationary states . as the result of further separation in the three - dimensional spherical coordinates @xmath102 of the macroscopic space and in @xmath103 coordinates of the internal two - dimensional space described by the factorization @xmath104
we obtain the `` almost '' familiar radial klein - gordon wave equation @xcite @xmath105 with the angular equation in the macroscopic space which defines states with a definite three dimensional angular momentum @xmath106 y_{lm } \nonumber \\ & = & l ( l + 1 ) \hbar^{2 } y_{lm } \ ; , \end{aligned}\ ] ] and the angular equation in the inner space with a definite internal angular momentum @xmath107 where @xmath108 is the main internal angular quantum number . in eq.([6 ] ) functions @xmath109 are the spherical harmonics ( normalized angular momentum eigenfunctions ) with @xmath110 and @xmath111 .
similarly @xmath112 are the periodic functions of the inner angles @xmath113 and @xmath114 . as the result of further separation of eq.([7 ] ) in the internal coordinates described by the factorization @xmath115 we obtain @xmath116 where @xmath117 is the internal @xmath113-angle quantum number of the torus geometry ( see section [ further ] ) .
it is not difficult to verify that for @xmath118 and @xmath120 the only solution of the stationary klein - gordon equation eq.([5 ] ) alone , vanishing in infinity ( although not normalizable ) , exists for @xmath121 and @xmath122 . for the radial part of @xmath95
this solution has the following form ( see eq.([row_22 ] ) for @xmath72 ) : @xmath123 where @xmath124 is a constant .
in this way we formally obtain @xcite up to @xmath124 the same solution which follows from the classical klein - gordon equation for the dilaton field ( see eq.([row_5 ] ) ) . yet
now the field is a scalar one , hence the energy of the configuration ( [ pcd7 ] ) is positive and equals : @xmath125 where @xmath126 is the @xmath127 component of the energy - momentum tensor of the scalar field @xmath95 .
integral @xmath128 does not converge and is given by eq.([q ] ) ) , the integral of @xmath82 in @xmath128 does not converge on @xmath83 .
hence the solution ( [ pcd7 ] ) obtained for @xmath121 , @xmath119 , @xmath118 and @xmath122 can not be obtained from the solutions around @xmath121 , @xmath119 ( i.e. in the limit @xmath129 , @xmath130 ) for which the lhs of eq.([t00 ] ) converges . ] . yet , fortunately here , we may recall the self - consistent calculations for fields which were previously introduced , i.e. the dilaton @xmath3 and metric @xmath18 ones , with the now added scalar field @xmath95 , which we for the time being assume to be given by the configuration ( [ pcd7 ] ) .
the coupled einstein and klein - gordon field equations for all these fields have to be solved .
now , the case with @xmath118 , @xmath119 , @xmath131 is the only one for which the value of @xmath132 in eq.([5 ] ) is irrelevant as then the rhs of eq.([5 ] ) becomes equal to zero .
therefore in this case the energy - momentum tensor for the massless scalar field enters into the rhs of einstein equations ( [ row_3 ] ) in the same way as the dilaton field , differing only in sign .
hence the rhs of einstein equations ( [ row_3 ] ) becomes equal to zero leading to ricci flat solution of einstein equations .
therefore the discussed new configuration of all fields , @xmath18 , @xmath3 and @xmath95 , does not describe the gravitational bound state , contrary to the case of the metric - dilaton field configuration given by eqs.([row_22 ] ) and ( [ row_26 ] ) alone .
hence the configuration given by eq.([pcd7 ] ) for @xmath133 and @xmath134 does not appear ) of eq.([5 ] ) is in this case as `` macroscopic '' as the dilatonic configuration ( [ row_22 ] ) leading to the zeroing on the rhs of the einstein equations ( [ row_3 ] ) . in other words : the only thomson like solution of the scalar configuration ( [ pcd7 ] ) , which by definition can not be treated as a small perturbation of the metric - dilaton configuration , does not appear at all . ] .
+ yet , a similar conclusion would not be true for @xmath135 even for @xmath119 .
indeed , in all these cases the planck constant @xmath132 in eq.([5 ] ) becomes important and leads to the appearance of the energy - momentum tensor for the scalar field which is of the order of @xmath136 , i.e. @xmath137 , hence constituting a very tiny quantity in comparison to the energy - momentum tensor for the dilaton field , @xmath138 ( see eqs.([row_19 ] ) and ( [ row_22 ] ) ) , which appears on the rhs of einstein equations ( [ row_3 ] ) .
this means that in the first approximation we may treat solutions of eq.([5 ] ) for @xmath139 as leading to small perturbations of the background metric ( [ row_26 ] ) , ignoring its self - consistent modification . in other words ,
the backreaction of @xmath95 is negligible in this case .
if the internal @xmath113-angle quantum number of the torus geometry is set to @xmath141 then @xmath142 , @xmath143 , are the lagrange polynomians which solve eq.([9 ] ) and hence the second factor on the rhs of eq.([8 ] ) satisfies the following equation ( see eq.([7 ] ) ) @xmath144 h_{\tilde{\mathrm{m } } } \ , .\end{aligned}\ ] ] the condition of continuity of functions @xmath145 for @xmath146 together with @xmath147 give @xmath148 where @xmath149 is the internal magnetic quantum number .
now , from the above equation we can notice that @xmath108 is a quantum number which chooses the ladder of values : @xmath150 hence @xmath151 .
for example @xmath152 in two cases : ( @xmath153 and @xmath154 ) and ( @xmath153 and @xmath155 ) .
let us also notice that under the assumption of @xmath119 the case with @xmath118 was previously excluded ( section 3.1 ) satisfies the two physical requirements , the first one of stability and the second one of the proper value of the total spin of the nucleon .
hence eventually only @xmath135 will be possible . ] .
let us focus on the case of the weak limit which means that we are far away ( @xmath156 ) from the central core of the dylatonic field given by eq.([row_22 ] ) .
at first let us rewrite eq.([5 ] ) as follows @xmath157 now , asymptotically ( i.e. @xmath158 ) the first and third term of the lhs of the above equation could be omitted as they are of smaller order then the others .
yet , the third term , with the orbital angular momentum @xmath159 , will be preserved as it is important in the angular momentum analysis . from the point of view of the energies of the excited states
( section 4 ) it would be better in the limit @xmath86 to omit consistently all terms which are proportional to @xmath160 since leaving only the one with @xmath159 leads , in light of the omission of the first one , to the appearance of nonphysical states . + nevertheless let us omit the first term in eq.([klein - gordon2 ] ) only which then , after introducing the variable @xmath161 , could be rewritten as follows : @xmath162 now , let us expand the coefficients in eq.([x ] ) at point @xmath163 ( i.e. @xmath164 ) to the second order in @xmath165 .
we obtain @xmath166 \nonumber \\ & = & \biggl\ { e^2 - \left ( \mu_{0 } + \mu_{1 } x + \mu_{2 } x^2 \right ) m^2 c^4 - \left ( \lambda_{0 } + \lambda_{1 } x + \lambda_{2 } x^2 \right ) \lambda_{d } \left ( \frac{\hbar c}{d } \right)^{2 } \biggr\ } u_{r } \ ; , \end{aligned}\ ] ] where the coefficients of the expansions and their limits for @xmath167 ( @xmath164 ) are given in appendix b. + finally , after coming back to variable @xmath57 we obtain for @xmath156 the radial part of the kline - gordon - like equation , written in the spherical coordinates as follows @xmath168 where @xmath169 @xmath170 and @xmath171 @xmath172 here @xmath173 can be interpreted as the square of the total angular momentum and @xmath174 as the four - dimensional squared mass of the scalar field @xmath95 in the flat minkowskian limit ( see @xcite ) .
if the six - dimensional mass is zero ( @xmath119 ) then the four - dimensional mass at spatial infinity would be solely of kinematical origin .
it is also easy to notice that neglecting the first term of eq.([klein - gordon2 ] ) in the limit @xmath175 is at this stage indistinguishable from zeroing the first derivative of the metric tensor also means that we neglect in eq.([klein - gordon2 ] ) a `` radial force '' proportional to @xmath176 . ] in @xmath177 which has its origin in the first term at the lhs of eq.([5 ] ) .
now , it is interesting to notice ( see eq.([14 ] ) ) that the radial motion looks like the one dimensional motion of a particle with a unit mass in the potential @xmath178 @xmath179 \ ; .\end{aligned}\ ] ] the first component on the right hand side can be interpreted as a modified centrifugal potential energy , and the second one as a potential energy related to some central attractive force .
the existence of the second term is connected with the appearance of the effective coupling constant which is proportional to @xmath180 the value of @xmath181 depends both on the parameters and the quantum numbers of the model . + after introducing @xmath182
we see @xcite that eq.([14 ] ) can be rewritten in the following form : @xmath183 u_{r } = 0 \ ; , \end{aligned}\ ] ] where @xmath184 @xmath185 with @xmath173 given by eq.([j - weak ] ) .
+ eq.([16 ] ) has nearly the same formal shape as the radial equation obtained in the schrdinger s model of the atom @xcite .
now , converting eq.([18 ] ) we obtain @xmath186 from the analysis of eq.([16 ] ) it can be shown that the final solutions for @xmath187 and @xmath188 exist only when @xmath189 where @xmath190 is zero or a natural number and @xmath191 is a nonnegative solution of the equation @xmath192 which is equal to @xmath193^{\frac{1}{2 } } \ ; .\end{aligned}\ ] ] finally , using eq.([19 ] ) we can eliminate @xmath194 from eq.([20 ] ) and rewrite it in the following form : @xmath195 , \end{aligned}\ ] ] where @xmath196^{\frac{1}{2 } } \,\end{aligned}\ ] ] and in the last equality in eq.([e2-weak ] ) the expression for @xmath181 given by eq.([stala - weak2 ] ) has been used . at the end of this section
let us notice that the original quantum klein - gordon equation given by eq.([k - g - m ] ) possesses the property of invariance according to which only the ratio @xmath197 matters .
we will discuss the relevant symmetry below . using this scale invariance
, we could move from the microscopic to the astrophysical scale . in this context
the calculations point to the similarity of both the neutron and a galaxy dark matter structure ( see also @xcite ) . at this point
the question concerning the value / s of @xmath4 arises .
previously @xcite we considered the model given by eq.([row_dzialanie - eh - fi ] ) which has the following invariance : @xmath198 where @xmath199 is the parameter of the transformation
. this would mean that it is not the change of @xmath4 but rather of @xmath200 that matters .
this is the invariance of the coupled einstein ( [ row_3 ] ) and klein - gordon ( [ row_5 ] ) equations but it is not the invariance of their solution given by eqs.([row_22 ] ) and ( [ row_26 ] ) .
this conclusion can also be drawn from eq.([row_24 ] ) .
hence we notice that the massless dilatonic field @xmath3 is the goldstone field .
what is more , using eqs.([row_19 ] ) and ( [ row_24 ] ) in eq.([row_25 ] ) we obtain @xmath201 we notice a similarity between this equation and its electromagnetic analog , where @xmath202 is the electromagnetic vector potential . ] .
eg.([screen - in - gravity - affect ] ) is the ( anti)screening current condition in gravitation , analogous to the screening current condition found in electromagnetism or in the electroweak sector in the self - consistent approach @xcite .
+ on the level of the equations of motion ( [ row_3 ] ) and ( [ row_5 ] ) the theory is scale invariant unless mass @xmath96 enters . it would seem that invariance ( [ a invariance ] ) leads to the invariance @xmath203 of the klein - gordon equation ( [ k - g - m ] ) .
yet the inclusion of any term with mass @xmath96 into the model given by eq.([row_dzialanie - eh - fi ] ) will inevitably lead to the gravitational coupling of the new matter with the metric tensor part of the lagrangian , causing the breaking of the original invariance ( [ a invariance ] ) . +
hence the conclusion emerges that there appears a relation between the four - dimensional mass @xmath204 and the parameter @xmath4 .
it results from the einstein equations which fix the value of @xmath4 after the self - consistent inclusion of the field @xmath95 with mass @xmath96 into the fields configuration .
this means that the symmetry connected with the scaling ( [ a invariance ] ) of @xmath4 is broken . hence in eq.([j - weak ] ) the dependance of @xmath173 on @xmath4 for given @xmath204 is established , where @xmath4 is not a free parameter . in view of the lack of a full ( self - consistent ) analysis ,
@xmath4 should be estimated somehow and as we shall see this will appear possible in the weak limit approximation . as we have proven that the global solution with @xmath205 and @xmath133 does not exist ( section [ global ] ) hence its weak limit does not exist also .
therefore from now on we will investigate the cases with @xmath139 only .
the reason is that , as it has been discussed in section [ global ] , they only slightly perturb the metric - dilaton configuration given by eqs.([row_22 ] ) and ( [ row_26 ] ) . at the moment
let us also suppose that @xmath133 , a value which will be dynamically chosen in the next section .
we began with the klein - gordon equation ( [ rownanie ] ) in which @xmath95 is the wave function , which in quantum mechanics is usually identified as the cause of angular momentum generators commutation relations . in our model
the square of the total angular momentum @xmath206 appears according to eq.([j - weak ] ) as the combination of terms @xmath207 and @xmath208 which originate from the internal two - dimensional and orbital three - dimensional angular momenta , respectively . yet , in the weak approximation regime these are @xmath191 and @xmath159 which take quantum mechanical values , therefore this means that @xmath209 chooses quantum mechanical values as a result of the proper addition of @xmath191 and @xmath159 only .
furthermore , since @xmath191 , @xmath159 and @xmath108 have quantum values hence @xmath210 also has this kind of values .
the strange thing is that term @xmath210 appears as a result of the pure ( field theory ) classical interaction witch originates in the background metric and now has its part in @xmath173 .
next , as @xmath211 manifests itself on the arena of the tree - dimensional space as the total angular momentum , we impose on it the usual quantum mechanical rules which lead to different representations for different values of @xmath173 .
+ this leads us to the indication of the place of wave mechanics ( quantum mechanics ) which from the point of view of the self - consistent field theory is discussed in @xcite .
wave mechanics appears as a theory for building a model of a system ( e.g. an atomic one ) with the lost of self - consistency , yet it partly works by describing its main characteristics . by `` partly ''
i mean that to receive a correction to the wave mechanics prediction the self - consistency has to be reestablished . yet
this is the realm of quantum field theory or the self - consistent field theory @xcite . to receive both the main characteristics and the corrections the gravitational interactions
should also be included in a self - consistent manner where the linear part of field fluctuations would be described by the quantum field theory . for the detailed discussion of the place of the quantum field theory see @xcite .
although the ( internal ) spin wave function could not have the @xmath212 space representation for the physical half - spin state , yet it does not eliminate motion as the origin of the spin in some kind of the classical space at all .
besides unknown , mainly three possibilities could be taken into account .
the first one states that the parameters of the spin representations are `` seen '' in the interactions only without further inner explanation . the second one exists by extending the space by the grassmann coordinates @xcite .
the third one , considered in this paper , constructs the spin of the particle from the motion in the geometry of the inner space .
the half - integer spin is then , as any other , just the numerical result .
+ the other spin representations of fields could have the similar origin ( see section 1.1 ) .
yet there is a reason to consider them as having the different dimensions of their inner spaces .
the prove goes via the fisherian statistical analysis performed by frieden @xcite and the conclusion is that the dimension of the inner space depends on the species of the field of the rank @xmath21 which is the dimension of the sample collected by the field alone during its sampling of the space - time . for the complex scalar field
the dimension of the extra parametric space is equal to @xmath22 .
for the other fields this dimension is bigger , e.g @xmath213 for the electromagnetic field and @xmath214 for the dirac one .
hence , as it has been mentioned previously , the four - dimensional space - time could be extended simultaneously to the variety of the multi - dimensional ones .
let us make some numerical adjustments for the parameters of the presented model ( @xmath22 ) .
for this purpose we begin with the state which possesses the following value of the internal quantum number ( see sections 3.1 - 2 ) : @xmath215 choosing the following value of the four dimensional mass : @xmath216 which is a little bit bigger than the observed mass of the neutron @xmath217
. solving eq([m4-weak ] ) we obtain @xmath218 as the limit value , i.e. from eq([m4-weak ] ) we see that for @xmath219 we obtain @xmath220 and for @xmath221 we obtain @xmath97 which is a tachion .
for the limit value @xmath222 the second term of @xmath223 at the rhs of eq.([j - weak ] ) is equal to @xmath224 now , it is interesting to notice that the rhs of eq.([j - weak ] ) might be interpreted as @xmath225 ( see eq.([sj ] ) ) , which means that @xmath226 after taking @xmath227 and @xmath228 for the neutron the equation on the relative value of @xmath4 and @xmath61 is obtained from the above equation .
moreover , we see that eq.([j - row - weak-2 ] ) gives us @xmath229 now , according to eqs.([e2-weak ] ) and ( [ stala - weak2 ] ) we obtain for @xmath133 and @xmath210 as in eq.([a - value - weak ] ) the following equation @xmath230 = m_{4}^2 c^{4 } \left [ 1 - \frac{1}{36 } ( \frac{3/2}{\iota})^2 \right]\end{aligned}\ ] ] and using in this equation @xmath204 from eq.([m4_first ] ) we find that the energy of the neutron in the ground state ( @xmath231 , @xmath228 and @xmath232 , see eq.([iotas ] ) ) is equal to @xmath233 which is the observed mass of the neutron . from the second term of eq.([e - weak ] )
it results that the gravitational correction to the energy of the nucleon in the ground state ( @xmath231 and @xmath234 and @xmath232 ) , i.e. the gravitational bounding energy , is equal to @xmath235 finally , let us recall that the solution for @xmath152 has according to eqs.([11 ] ) and ( [ 12 ] ) the double degeneration : ( @xmath153 and @xmath154 - which we interpret as the spin up solution ) or ( @xmath153 and @xmath155 - which we interpret as the spin down solution ) .
this assignment is very natural but by its own it does not lead to the conclusion that the particle is the neutron .
the calculation of the energy of the ground state of a field having the `` quantum '' numbers characteristic for the neutron entitles only to the conclusion that the formalism could describe the neutron .
hence we see that the model points to the origin of the spin of the nucleon as connected with the geometry of the internal manifold and with motion inside its two compactified parametric dimensions .
moreover , it indicates that gravitational effects give a visible change of the spectrum of the levels in the proposed model of the nucleon . + as it has been shown
, the background gravity - dilatonic configuration given by eqs.([row_22 ] ) and ( [ row_26 ] ) is stable ( see section 2.1 ) .
also the ground state solution obtained in section [ implication ] for the scalar field @xmath95 is the stationary one .
moreover , it perturbs the background configuration slightly only ( see section 3.1 ) and in the discussed weak approximation we could argue that the whole configuration of the fields , @xmath18 , @xmath3 and @xmath95 , is stable .
finally , for a particle with another value of its spin the dimension @xmath21 of the inner parametric space is different , yet the conclusion , if the relevant calculations are performed , would be similar .
let us still consider one particular state ( e.g. neutron ) with @xmath174 _ fixed _
, i.e. @xmath96 and @xmath61 may vary leaving @xmath174 unchanged .
now , with the use of eq.([m4-weak ] ) we notice that eq.([j - row - weak-2 ] ) can be rewritten as follows @xmath236 hence we see that @xmath223 for @xmath139 is not proportional to @xmath237 ( no matter what the integer @xmath159 is ) and it is worth noting that to keep @xmath223 equal to @xmath238 for @xmath191 equal to the multiplicity of @xmath239 , it is necessary to have @xmath133 ( hence @xmath240 ) without any variation from this particular value of @xmath96 . as in section [ neutron ]
the value of @xmath204 for the neutron has been chosen to fit the value of the observed energy @xmath241 of the neutron hence it is important to calculate the predictions for values of the energy of its excited states .
let us make the following assumption on the relation between the total angular momentum number and the orbital one : @xmath242 for the mass @xmath119 we may notice from eq.([j - row - weak-2 ] ) that the characteristic radius @xmath4 has to change with @xmath191 , @xmath159 and @xmath108 in the following way @xmath243 where @xmath61 is kept equal to @xmath244 @xmath245 $ ] as in eq.([dl ] ) . then , with @xmath246 given for each particular state by eq.([m4-weak ] ) , the energy @xmath241 is calculated according to eq.([e2-weak ] ) .
the numerical values of the energy @xmath241 for the ground state ( @xmath247 , @xmath248 ) and excited states of the neutron are given in the table [ table ] below .
+ except for the ground state ( which has been chosen to fit the neutron observed energy ) all other energies are not ideally equal to the observed ones @xcite .
nevertheless this might not be a big shortcoming of the model since we use the weak approximation only .
this statement appears to be sensible , especially in the light of the agreement of the order of the discrepancies between energies of the states belonging to the same column and different rows , which are of the order found in the experiment .
as the spacial configurations for the first column in the table 1 are the most simple ones ( @xmath227 ) hence the feeling is that only they should lie in the neighborhoods of the true states ( see also the discussion after eq.([klein - gordon2 ] ) ) .
.values of energies @xmath241 mev for the ground state and excited states of the neutron as the function of the total angular momentum number @xmath191 and internal quantum number @xmath108 . for the neutron itself
@xmath247 and @xmath248 .
the last two columns present the gravitational bounding energy @xmath249 mev and the ratio @xmath250 for the neutron and its excited states with @xmath251 and @xmath227 .
all states in the table 1 have @xmath133 and @xmath252 fm ( see the text ) . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] additionally , in the two last columns of the table 1 , the values of the gravitational bounding energy @xmath253 \ ; , \end{aligned}\ ] ] for the states with @xmath247 and @xmath227 are given . here , according to eqs.([stala - weak2 ] ) and ( [ m4-weak ] ) the effective coupling constant @xmath181 and the the four - dimensional squared mass @xmath174 are for @xmath133 equal to @xmath254 where @xmath4 is given by eq.([a ] ) .
+ the important result is that for fixed @xmath191 and @xmath159 , the increase of the energy @xmath241 of the configuration causes the system to become more tightened ( as the ratio @xmath250 decreases ) and simultaneously its gravitational bounding energy bigger . as an example , for the neutron and its excited states ( with @xmath251 and @xmath227 ) this result has been presented in the first and in the two last columns of the table 1 ( in bold ) .
+ the model could be applied to the other known particles also , yet the present calculations show that the best agreement is achieved for the neutron and its excited states .
for example , for the neutral meson @xmath255 taken as the ground state and the first five of its excited states , for all of which we choose @xmath256 , @xmath231 , @xmath257 and @xmath227 , we obtain the energies equal to @xmath258 , @xmath259 , @xmath260 , @xmath261 , @xmath262 , @xmath263 mev for @xmath108 equal to 1,2,3,4,6,7 , respectively .
so , in this case we notice the worse matching with the experimental values @xcite than in the case of the neutron . the explanation of this fact might be that the gravitational factor in the structure of the neutron , which is the lowest known barionic neutral ground state , is most decisive than in the case of other neutral barions .
the principal problem seems to be connected with the ladder of the energy state values in each particular row since in each row few of them or may be only one are physical .
this problem might be connected with the zeroing of the term with the first derivative of the metric tensor and hence with the removal of the first derivative of the radial wave function @xmath264 on passing from eq.([klein - gordon2 ] ) to eq.([x ] ) .
unfortunately , after this zeroing , the weak limit approximation was invented , leading effectively to the introduction of an extra force term ( which is equal to the neglected one with the opposite sign ) connected with the mathematics of the weak approximation only .
the addition of this unphysical extra force into the system results in appearance of row ladders of the unphysical states
. yet the unphysical differences between states of the ladder in one particular row with the established value of @xmath108 are two order of magnitude smaller than the physical ones from the columns .
this agrees with the fact that the differences in one row appear as the result of the change in the value of @xmath159 .
now , according to section [ weak interaction limit ] the orbital angular momentum @xmath159 survived with the third term @xmath265 of eq.([klein - gordon2 ] ) .
yet it survived inconsistently because the first term @xmath266 of this equation , also proportional to @xmath267 , has been neglected .
since the third term of eq.([klein - gordon2 ] ) is , for @xmath175 , of the order smaller than the ones which remained together with it in eq.([x ] ) ( and consequently in eq.([14 ] ) ) therefore the energy differences in a particular row are also smaller than in a particular column .
the analysis of eqs.([14])-([atr - weak ] ) is in agrement with the remarks above , i.e. the change of @xmath159 in @xmath268 is connected with the small energy differences in a particular row whereas the change of @xmath108 in @xmath269 is connected with the big energy differences in a particular column of the table [ table ] .
as we noticed in section [ weak interaction limit ] the term with @xmath159 has been left for the sake of the angular momentum considerations .
till now there does not exist established experimental evidence for multi - dimensionality of the world and/or our understanding of potential manifestations of higher dimensions is too poor . yet at the same time there exist in the literature multiple references indicating that the pursued effects of extra dimensions could have already been noticed earlier , both in the astrophysical @xcite , @xcite and in elementary particle scale @xcite , @xcite , contrary to the standard expectations that extremely high energies are necessary to probe higher dimensions .
the renewed interest in the kaluza - klein theories , to which the presented six - dimensional model belongs , stems from the fact that multidimensional analogues of general relativity are able , among others , to generate the four dimensional mass out of the interaction which proceeds both from the four - dimensional world and from the internal space dimensions ( here two ) with currents of the matter which modify masses in the four - dimensional world ( see also @xcite ) .
our approach to the modelling of a system , e.g. of one particle ( here neutron ) , stems from the hope that its structure might be described by the dynamics in the six - dimensional space - time , and especially that its spin and , at least partly its inertia have their origin in the internal two - dimensional space .
this is the reason for perceiving six - dimensional ( and more generally , multi - dimensional , the space - time would be eight - dimensional . ] ) theories as still attractive for understanding both fundamental interactions and the more sophisticated considerations which are related to the problem of mach s principle . to fulfil this aim
the present paper examines the implementation of the geometrical properties of the previously @xcite worked out six - dimensional kaluza - klein type model which manifest themselves in possible observational consequences in the realm of one particle physics , e.g. having an impact on the structure of neutron and its excited states .
the obtained ground state solution , here neutron , appeared to have two spinorial degrees of freedom as the total angular momentum with @xmath251 goes ( for the main internal angular quantum number @xmath248 ) with two internal magnetic quantum number possibilities , @xmath154 or @xmath155 . in section 1.1 , the statistical fisherian reasons for the six - dimensionality of the space - time are given .
the fundamentals of the model are presented in section [ geometry ] where the static spherically symmetric `` reference '' back - ground solution of the six - dimensional einstein equations coupled with the klein - gordon equation with the dilatonic field as the basic one and the gravitational field as the self - field , is recalled @xcite . a more detailed discussion of its properties is presented in @xcite .
the metric tensor part of the `` reference '' solution is , in a sense , analogous to the familiar four - dimensional schwarzschild solution but yet fundamentally different since it is horizon free . in section [ stability ] the stability of the `` reference '' self - consistent solution
section [ implication ] is devoted to the derivation of the weak limit approximation of the general wave mechanical ( quantum mechanical ) approach which is defined in the context of losing ( here gravitational ) self - consistency .
so , a new scalar field has been added to the original `` reference '' configuration of fields and the non self - consistent case for its klein - gordon equation is defined . here
the main purpose is to find the solution of this klein - gordon equation and the reason for doing this is twofold . on one hand
it is connected with the derivation of the energy of states . on the other hand with the indication of the spin origin of the configuration as the manifestation of both the geometry of the internal two - dimensional space and the kinematics of the fields inside it . in section
[ global ] the problem of the departure from the ( gravitational ) self - consistent calculations of the metric and other fields of the configuration has been discussed .
the lost of the self - consistency is connected with neglecting the perturbation term ( of the total energy - momentum tensor ) which is proportional to @xmath136 . in section [ numerical example ]
the implementation of the model to the description of neutron and its excited states has been performed , including the derivation of their energy .
finally , the discussion of the consequences of zeroing of the first derivative term in the radial klein - gordon equation has been given in section [ excited ] .
the question arises on the status of the presented model of the neutron . as discussed previously in sections [ implication ] and
[ numerical example ] it is interpreted as the wave mechanical departure from the fully ( i.e. including gravitation ) self - consistent field theory understood in accordance with the schrdinger s substantial interpretation of the wave function @xcite .
the wave mechanical approach means that the self - consistency is lost .
yet there is hope that a fully self - consistent model which incorporates all the necessary fields and interactions will result not only in the refinement of the proposed model of the neutron and its excited states but also in the construction of better models of all other elementary matter particles . yet
this is the difficult task to perform @xcite .
+ in today s physics mainly the wave mechanical ( weak approximation case ) and quantum field theory ( linear ) regimes are perceived . yet
self - consistent exceptions are also quoted , developed only for the description of the ( mainly linear ) fluctuations of the matter wave function @xcite .
the concept which lies behind this approach is not the quantum theory of interactions with quantum gravity included @xcite with the second quantization procedure for a set of aggregated quanta .
but , it is rather of a relativistic self - consistent field theory origin which includes all interactions , with gravity in this number .
the consistency of this approach is guaranteed on the theoretical basis which is stronger than the mean - field theory concept as both wave mechanics and classical field theories can be understood as having the same , fisher - information origin with finite @xmath21 @xcite .
+ finally , according to the calculation of this paper , mach s principle has to be modified .
since the mass of a particle comes both from external and internal dimensions ( where the internal part of this mass is purely of the kinetic origin ) hence it is essential that the final mass of the particle should be calculated in the self - consistent formalism .
it might be done in accord with a broadened version of thrirring and einstein analysis of einstein geometrodynamics or in accord with an effective gravity theory of the logunov type @xcite . in both cases
the internal contribution to the mass should appear as the result of the self - consistent calculations .
interestingly the same conclusions were drowned by frieden @xcite from the statistical method of estimating the physical models and just recently proved in @xcite , where the entanglement between the kinematical and structural degrees of freedom has been rigorously established .
now , because the mass has its origin in the dynamics and the kinetic motion of the ( extended ) field of every elementary particle inside the internal space , we obtain an elementary property according to which every particle is a continuously extended object .
+ yet , the presented model is obviously not the final one .
the reason is twofold .
at first , from its beginning it was formulated for the neutral particle only . the inclusion of e.g. the electric or effective weak charges , which would be necessary , changes the results . at second , the first particle in every chosen ladder of states should really be the ground state of the configuration of all fields which are taken into account to describe self - consistently both the ground state and excited states .
hence , further calculations should incorporate more realistic shapes of both the main charge densities and their fluctuations for the extended basic matter sources to which proper self - fields are coupled .
these shapes should follow both from the einstein equations coupled to e.g. klein - gordon - maxwell ( effective yang - mills ) or dirac - maxwell ( effective yang - mills ) set of equations for charge densities and fluctuations simultaneously , as it is required for the self - consistent models @xcite .
it means that from the mathematical point of view , a matter particle seems to be a self - consistent solution of field equations for the basic fields ( including fluctuations ) and their self - fields which are involved in the description of this particle . the presented model indicates the place where the weak , wave mechanical limit of the theory lies and is a step towards the construction of the self - consistent formalism of the classical theory of one , continuously extended , elementary particle .
this work has been supported by l.j.ch .. + special thanks to barbara mierzyska for the correction of the text .
this work has been also supported by the department of field theory and particle physics , institute of physics , university of silesia , the project `` _ teoretyczne i fenomenologiczne badanie modelu standardowego i jego uog ' olnie _ '' , no : bs-03 - 0507 - 022 - 07 and by the junta de andaluc ' ia project fqm 437 .
using eqs.([row_7])-([row_8 ] ) , we can calculate the components of the ricci tensor .
the nonvanishing components are @xcite @xmath270 @xmath271 @xmath272 @xmath273
the coefficients of expansions of the radial equation ( [ 13 ] ) and their limits for @xmath167 ( @xmath164 ) are equal to @xmath274 @xmath275 @xmath276 111 kelvin , lord , _ aepinus revisited _ , philos . mag .
vi * 3 * , ( 1902 ) , 257 - 283 ; _ on the weights of atoms _ , philos . mag .
vi * 4 * , 177 - 198 , ( 1902 ) .
thomson , philos . mag . * 7 * , 237 , ( 1904 ) .
chemical rubber co. , _ handbook of chemistry and physics _ ,
crc , 68th edn . ,
pp.f-157 , ( 1987 ) .
maxwell , _ on theories of the constitution of saturn s rings _ , edinburgh r. soc
* iv * , 99 - 101 , ( 1862 ) .
thomson , _ cathode rays _ , philos
. mag . * 44 * , 293 - 316 , ( 1897 ) .
stoney , _ on the electron or atom of electricity _ , letter in philos . mag . * 38 * , 418 - 420 , ( 1894 ) .
j. de luca , s.b .
rodrigues , y. levin , _ electromagnetic instability of the thomson problem _ , europhys .
lett . * 71 * , 84 , ( 2005 ) , cond - mat/0506616 .
e. rutherford , proc . of the royal society , vol .
a * 97 * , 394 , ( 1920 ) .
w. heisenberg , _ ber quantenmechanische umdeutung kinematischer und mechanischer beziehungen _ , z. phys .
* 33 * , 879 - 893 , ( 1925 ) .
dirac , _ the fundamental equations of quantum mechanics _ , proc .
london , series a * 109 * , 642 - 653 , ( 1925 ) ; _ quantum mechanics and a preliminary invastigation of the hydrogen atom _ , proc .
london , series a * 110 * , 561 - 579 , ( 1926 ) ; _ the quantum theory of the electron _ , proc .
london , series a * 117 * , 610 - 624 , ( 1928 ) ; _ the quantum theory of the electron _ , proc .
london , series a * 118 * , 351 - 361 , ( 1928 ) .
e. schrdinger , _ quantisierung als eigenwertproblem .
( erste mitteilung ) _ , ann . phys . * 79 * , 361 - 376 , ( 1926 ) ; _ quantisierung als eigenwertproblem .
( zweite mitteilung ) _ , ann .
phys . * 79 * , 489 - 527 , ( 1926 ) ; _ quantisierung als eigenwertproblem .
( dritte mitteilung : storungstheorie mit anwendung auf den starkeffekt der balmerlinien ) _ , ann . phys . * 80 * , 437 - 490 , ( 1926 ) ; _ quantisierung als eigenwertproblem .
( vierte mitteilung ) _ , ann . phys .
* 81 * , 109 - 139 , ( 1926 ) .
l. de broglie , _ recherches sur la theorie des quanta _ , ph.d .
dissertation , paris , ( 1924 ) .
espinosa , _ physical properties of de broglie s phase waves _
, am . j. phys .
* 50 * , 357 - 362 , ( 1982 ) .
albert , trans .
64 * , ( 1948 ) , 552 .
santilli , hadronic j. * 1 * , 224 , ( 1978 ) ; hadronic j. * 1 * , 574 and 1267 , ( 1978 ) ; communication of the jinr , dubna , russia , @xmath277 e4 - 93 - 352 , ( 1993 ) ; chinese j.syst.ing . & electr . *
6 * , 177 , ( 1996 ) .
s. lie , _ over en classe geometriske transformationer _ ,
english translation by e. trell , algebras groups and geometries * 15 * , 395 , ( 1998 ) .
u. stiegler , phys . rep . *
277 * , 1 , ( 1996 ) .
+ x. zheng et al.,_(the jefferson lab hall a collaboration ) _ , phys .
lett . * 92 * , 012004 , ( 2004 ) , _ precision measurement of the neutron spin asymmetry @xmath278 and spin - flavor decomposition in the valence quark region _ , nucl - ex/0308011 .
k. heyde , _ basic ideas and concepts in nuclear physics _ , 3rd ed .
, iop publishing ltd , 577 , ( 2004 ) .
riemann , _ bei die hypothesen , welche der geometrie zur grunde liegen _
, in _ gesammelte mathematische werke _ , ( 1866 ) ; reprint of 2d ed . ,
h.weber ( dover , new york ) , ( 1953 ) ; see also the translation by w.k.clifford , nature * 8 * , 14 - 17 , ( 1973 ) .
a. lakhtakia ( ed . ) , _ models and modelers of hydrogen _ , ( world scientific ) , ( 1996 ) .
h. sallhofer , _ hydrogen in electrodynamics .
vi . the general solution _ , z. naturforsch . *
45a * , 1361 - 1366 , ( 1990 ) .
sakurai , _ advanced quantum mechanics _
, ( addison - wesley publishing company ) , 169 , ( 1967 ) .
m. biesiada , r. maka and j. syska , inter .
journal of modern physics * d * , vol.9 , nu.1 , ( 2000 ) ; r. maka and j. syska , _ nonhomogeneous six - dimensional kaluza - klein compactification _ , preprint ul - tp-95/03 , university of silesia , ( 1995 ) .
j. syska , _ self - consistent classical fields in gauge field theories _ , ( university of silesia ) , phd thesis ( unpublished ) , ( 1995/99 ) .
barut , j. kraus , _ nonperturbative quantum electrodynamics : the lamb shift _ , found .
* 13 * , 189 - 194 , doi:10.1007/bf01889480 , ( 1983 ) .
barut , j.f .
van huele , _ quantum electrodynamics based on self - energy : lamb shift and spontaneous emission without field quantization _ , phys .
rev . a * 32 * , 3187 - 3195 , doi:10.1103/physreva.32.3187 , ( 1985 ) .
barut , j.p .
dowling , _ quantum electrodynamics based on self - energy : spontaneous emission in cavities _ , phys
a * 36 * , 649 - 654 , doi:10.1103/physreva.36.649 , ( 1987 ) .
barut , _ combining relativity and quantum mechanics : schrdinger s interpretation of @xmath279 _ , found .
phys . * 18 * , 95 - 105 , doi:10.1007/bf01882875 , ( 1988 ) ; _ schrdinger s interpretation of psi as a continuous charge distribution _ ,
( leipzig ) , * 45 * , 31 - 36 , ( 1988 ) ; _ the revival of schrdinger s interpretation of quantum mechanics _ , found .
* 1 * , 47 - 56 , doi:10.1007/bf00661316 , ( 1988 ) .
barut , _ quantum - electrodynamics based on self - energy _ , phys .
t * 21 * , 18 - 21 , ( 1988 ) ; a.o .
barut , y.i .
salamin , _ relativistic theory of spontaneous emission _ ,
a * 37 * , 2284 - 2296 , doi:10.1103/physreva.37.2284 ( 1988 ) .
a.o . barut and n. nal , _ an exactly soluble relativistic quantum two - fermion problem _ , j. math .
physics * 27 * , 3055 , doi:10.1063/1.527235 , ( 1986 ) ; a.o .
barut and n. nal , physica a * 142 * , _ a new approach to bound - state quantum electrodynamics : i. theory _ , pp.467 - 487 , doi:10.1016/0378 - 4371(87)90036 - 7 , and _ a new approach to bound - state quantum electrodynamics : ii .
spectra of positronium , muonium and hydrogen _ , pp.488 - 497 , doi:10.1016/0378 - 4371(87)90037 - 9 , ( 1987 ) .
j. syska , _ remarks on self - consistent models of a particle _ , in _ trends in boson research _ ,
ling , ( nova science publishers ) , 163 - 181 , ( 2006 ) .
j. syska , _ boson ground state fields in the classical counterpart of the electroweak theory with non - zero charge densities _ , in _ frontiers in field theory _ , ed
. o. kovras , ( nova science publishers ) , ( nowy jork 2005 ) , chapter 6 , 125 - 154 and in _ focus on boson research _ , ( nova publishers ) , 233 - 258 , ( 2006 ) .
ivashchuk and v.n .
melnikov , class .
quantum grav . * 11 * , 1793 , ( 1994 ) .
v.d . ivashchuk and v.n .
melnikov , grav . & cosmol .
* 1 * no2 , 133 , ( 1995 ) .
ivashchuk , phd dissertation , ( vnicpv , moscow , 1989 ) .
k.a . bronnikov and v.n .
melnikov , annals of physics ( n.y . )
* 239 * , 40 , ( 1995 ) .
sparling , proc.r.soc.a doi:10.1098/rspa.2007 , 1839 , ( 2007 ) .
p. demers , can.j.phys . *
43 * , 1687 , ( 1975 ) .
e.a.b . cole , nuovo cimento a * 40 * , 171 , ( 1977 ) .
patty , jr . , l.l .
smalley , phys.rev.d * 32 * , nu.4 , 891 - 897 , ( 1985 ) .
frieden , found .
phys.*16 * , 883 , ( 1986 ) .
frieden , _ fisher information , disorder , and the equilibrium distributions of physics _ , phys .
rev . a * 41 * , 4265 , doi:10.1103/physreva.41.4265 , ( 1990 ) .
frieden , b.h .
soffer , _ lagrangians of physics and the game of fisher - information transfer _ , phys .
e * 52 * , no.3 , 2274 - 2286 , ( 1995 ) .
frieden , _ relations between parameters of a decoherent system and fisher information _ , phys .
rev . a * 66 * , 022107 , doi:10.1103/physreva.66.022107 , ( 2002 ) .
frieden , a. plastino , a.r .
plastino and b.h .
schrdinger link between nonequilibrium thermodynamics and fisher information _ , phys .
e * 66 * , 046128 , doi : 10.1103/physreve.66.046128 , ( 2002 ) ; phys . lett . a * 304 * , 73 , ( 2002 ) .
frieden , science from fisher information : a unification , ( cambridge univ . press , 2004 ) .
frieden , b.h .
soffer , _ information - theoretic significance of the wigner distribution _
a * 74 * , 052108 , doi:10.1103/physreva.74.052108 , ( 2006 ) .
j. syska , _ fisher information and quantum - classical field theory : classical statistics similarity _ , phys .
sol.(b ) 244 , no.7 , 2531 - 2537 , ( 2007)/doi 10.1002/pssb.200674646 , arxiv : physics/0811.3600 .
piotrowski , j. sadkowski , j. syska , s. zajac , _ the method of the likelihood and the fisher information in the construction of physical models _
, accepted for publication in phys .
sol.(b ) , volume 246 , issue 5 ( 2009 ) , arxiv : physics/0811.3554 . l.i .
schiff , _ quantum mechanics _ , 3rd ed . ,
( 1968 ) .
m. biesiada , k. rudnicki and j. syska , _ an alternative picture of the structure of galaxies _ , in _ gravitation , electromagnetism and cosmlogy : toward a new synthesis _ ,
k. rudnicki , ( apeiron , montreal ) , 31 , ( 2001 ) .
sakurai , _ modern quantum mechanics _ , ed .
san fu tuan , ( addison - wesley publishing company ) , ( 1994 ) .
jaynes , _ coherence and quantum optics _ , ed . by l. mandel and e. wolf , ( plenum , new york ) , 495 - 509 , ( 1978 ) . + p.w .
milonni , _ foundations of radiation theory and quantum electrodynamics _ , ed .
barut , ( plenum , new york ) , 15 , ( 1980 ) .
( particle data group ) , j. phys .
g * 33 * , ( 2006 ) .
wesson , apj * 394 * , 19 , ( 1992 ) .
+ j. m. overduin , p. s. wesson , _ kaluza - klein gravity _ ,
. rept . * 283 * , 303 - 380 , ( 1997 ) , gr - qc/9805018 . + p.s .
wesson , _ the cosmological constant problem and kaluza - klein theory _ , int.j . mod .
d * 10 * , 905 - 912 , ( 2001 ) , gr - qc/0104045 ; _ is mass quantized ? _ , mod .
a * 19 * , 19952000 , ( 2004 ) , gr - qc/0309100 ; _ the equivalence principle as a probe for higher dimensions _
d * 14 * , 2315 - 2318 , ( 2005 ) , gr - qc/0601065 ; _ wave mechanics and general relativity : a rapprochement _ , gen.rel.grav . * 38 * , 937 - 944 , ( 2006 ) , gr - qc/0601059 .
overduin , _ solar system tests of the equivalence principle and constraints on higher - dimensional gravity _ , phys .
rev . d * 62 * , 102001 , ( 2000 ) , gr - qc/0007047 .
m. biesiada , b. malec , _ white dwarf constraints on exotic physics _ , phys .
d * 65 * , 043008 , ( 2002 ) , astro - ph/0109545 ; astrophysics and space science * 283 * , no.4 , ( 2003 ) .
m. besanon , _ looking for extra - dimensions at the weak scale : experimental search for kaluza - klein states signatures at the e+e- linear collider _ , hep - ph/9909364 ; _ experimental signatures of extra - dimensions at colliders _ , in xxix international conference of theoretical physics _ matter to the deepest _
, acta phys . pol .
b * 36 * , no.11 , 3487 , ( 2005 ) ( with numerous references ) .
b. grzadkowski , j. wudka , _ light excitatios in 5-dimensional gauge theories _ , in xxix international conference of theoretical physics _
matter to the deepest _ , acta phys .
b * 36 * , no.11 , 3523 , ( 2005 ) .
j. wudka , _ a short course in effective lagrangians _ , lectures delivered at the vii mexican workshop on particles and fields , merida , yucatan , mexico , 10 - 17
november , ( 1999 ) , aip conf.proc . *
531 * , 81 - 101 , ( 2000 ) , hep - ph/0002180 .
|
the paper examines the geometrical properties of a six - dimensional kaluza - klein type model .
they may have an impact on the model of the structure of a neutron and its excited states in the realm of one particle physics .
the statistical reason for the six - dimensionality and the stability of the solution is given .
the derivation of the weak limit approximation of the general wave mechanical ( quantum mechanical ) approach , defined in the context of losing its self - consistency ( here gravitational ) , is presented . the non self - consistent case for the klein - gordon equation
is defined .
the derivation of the energy of states and the analysis of the spin origin of the analyzed fields configuration is presented as the manifestation of both the geometry of the internal two - dimensional space and kinematics of fields inside it .
the problem of the departure from the ( gravitational ) self - consistent calculations of the metric tensor and of other fields of the configuration is discussed .
the implementation of the model for the description of a neutron and its excited states , including their spins and energies , is given .
the informational reason for the existence of the internal extra space dimensions is proposed .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore
|
in 2007 the belle collaboration claimed the discovery of the z(4430 ) .
this state attracted considerable attention because it is charged and couples to charmonium , implying that the most economical interpretation of its quark content is @xmath0 .
the recent high statistics confirmation of the @xmath1 by the lhcb collaboration , and the startling demonstration of phase motion , has brought sharp focus on exotic hadronic matter .
indeed , the @xmath2 joins a long list of other putative exotic states , several of which have been reported within the past year : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath3 multiquarks : : + x(3872 ) , z@xmath4(3900 ) , y(3940 ) , z@xmath4(4020 ) , z@xmath5(4050 ) , z@xmath6(4250 ) , y(4140 ) @xmath7 multiquarks : : + z@xmath8(10610 ) , z@xmath8(10650 ) other unusual states : : + d@xmath9(2317 ) , h dibaryon , y(2175 ) , y(4260 ) , y(4660 ) , y@xmath8(10888 ) , @xmath10 , @xmath11 , @xmath12 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ although hadronic exotics such as glueballs , hybrids , and multiquark states have been long expected , the understanding of these states is primarily at the level of conjecture .
certainly , if the confirmation of the @xmath2 marks the beginning of the exploration of a new sector of matter , the current phenomenology concerning quark interactions will need to be radically overhauled .
a compelling and unified understanding of the new states has not yet emerged , and the gap between theory and experiment remains a major deficiency in our current level of understanding of elementary particle physics .
this gap has its roots in the famously difficult problem of solving qcd in its many - body , strongly interacting , relativistic regime .
current effective field theories are inoperable in the excited spectrum , lattice field theory has difficulties with weakly bound diffuse coupled - channel systems , and extant phenomenological models are insufficiently well constrained to be confidently applied to exotic states .
even the lack of knowledge of relatively simple dynamics , such as interactions in the @xmath13 system , can affect the analysis of data concerning the new states . the flood of information initiated by @xmath14 factories ( cleo , babar , belle ) , @xmath15-charm facilities ( cleo - c , besiii ) , and hadron machines ( cdf , d0 , lhcb , atlas , cms ) is not expected to abate soon .
lhcb will continue to deliver new results in heavy quark spectroscopy for at least a decade . at the same time , besiii at the beijing electron positron collider will continue its program to collect and analyze @xmath16 data in the energy region of the putative exotic states of charmonium .
furthermore , the gluex experiment is due to start taking data in 2015 .
this experiment , situated at hall d at jlab , is designed to discover and explore the properties of light hybrid mesons .
the compass experiment at cern has been , and will continue to be , very active in hadron spectroscopy .
the panda experiment at fair is expected to start taking data in 2019 ; amongst its goals is the exploration of charmonium hybrids and other exotic states . in view of this situation ,
a workshop was convened at the institute of nuclear theory , seattle , with the aim of assessing the status of the field and drawing up a short list of questions that have the potential to move the field forward .
this document is the outcome .
we stress that this is not meant as a review , for which the reader is directed to refs .
furthermore , the topics contained herein are not meant to be comprehensive , but are offered in the hope that progress will be spurred in various directions .
the next three sections provide specific queries in the areas of lattice field theory , experiment , and theory .
the lattice method has been singled out because it has advanced to the stage where modelling issues are minimal , but where results are sufficiently complex that experimental methods must sometimes be invoked to interpret them .
finally , the interface between theory and experiment is addressed in section [ thy - int ] . here
the emphasis is on smoothing the interaction between theorists and experimental collaborations with the hope of drawing on the strengths of both communities .
1 . * compute quantities as a function of light quark mass . * + a better determination of the contribution of ( virtual ) two heavy - light meson loops in @xmath17 states below threshold is needed .
coupled channel phenomenological models suggest that for @xmath17 states near threshold these contributions are significant .
after renormalization of the bare model parameters , one finds modest shifts in the leading nonrelativistic mass spectrum .
spin splittings between ground state heavy - light mesons induce spin - dependent effects in the spectrum and allow hadronic transitions that violate the heavy quark spin symmetry expectations .
furthermore , the mass splittings between the @xmath18 , @xmath19 and @xmath20 ground state heavy - light mesons allows small isospin breaking and considerable su(3 ) breaking effects .
in particular , this may be evident in the large @xmath21 and @xmath22 transition rates . + in lattice qcd calculations , as the light quark masses
are varied down from infinite ( quenched approximation ) to the scale of the momentum in the @xmath17 system the dominant effect of light quark loops is to modify the running of the qcd coupling ( @xmath23 ) .
but as the quark masses are varied below this scale and below @xmath24 , light quark loops become spatially extended and probe the effects of coupled channels in the hadronic basis of states .
initial effort in exploring these effects are described in refs . @xcite .
+ we suggest that detailed lattice studies of the @xmath25 mass spectrum as a function of light quark masses ( @xmath26 ) for masses in a range between their physical values and @xmath27 will give much insight into the effects of coupling to decay channels in a model independent way . furthermore , a calculation of hadronic transition rates as a function of light quark masses would be very illuminating .
2 . * develop and implement coupled - channel scattering formalism .
* + the recent publication @xcite of the first determinations of _ coupled - channel _ scattering amplitudes from lattice qcd offers promise that this first - principles approach to qcd might shed light on the exotic behavior being observed in charmonium . for the case of coupled two - body scattering , resonant singularities in the amplitudes can be explored using parameterizations of the @xmath28-matrix , where the parameters are tuned to describe the finite - volume spectra calculated in lattice qcd . from the pole positions and residues , masses , widths , and branching fractions of resonances
can be determined the distribution of poles across unphysical riemann sheets may offer a discriminator for the internal structure of the resonances .
+ there has been no application of these coupled - channel techniques to meson systems featuring charm quarks , and only limited studies of elastic scattering , which is a situation in need of remedy .
early targets will be charmed systems near threshold like @xmath29 and @xmath30 as well as exotic isospin and strangeness channels@xcite .
double charmed channels like @xmath31 are also relatively simple .
hidden charm channels are challenging , because while all the tools are in place to deal with the coupled @xmath32 system , the opening of three - body channels like @xmath33 and @xmath34 occurs at rather low energies .
no complete formalism to relate finite - volume spectra to three - body scattering amplitudes yet exists
such a formalism will be required to study such systems in detail .
+ calculations of meson - baryon scattering are at a less advanced stage than those for meson - meson scattering .
current stochastic methods for dealing with quark propagation make the calculation of the @xmath35 scattering amplitudes straightforward with a modest increase in cost over meson - meson amplitudes , even for large volumes@xcite .
3 . * investigate static quark interactions .
* + recent improvements to the set of techniques available for computing light quark propagation on the lattice should encourage practitioners to revisit the problem of computing potentials between static color sources and their excitation spectra @xcite . for related work see refs .
these calculations have a long history and the static potential in the @xmath36 yang - mills theory was amongst the first lattice monte carlo computations . revisiting the potentials in the presence of light dynamical quarks @xcite will give useful insight into the nature of the @xmath37 and pentaquark experimental signals .
in particular , the bottom quark sector could be modelled very effectively with this data while exotic mesons in the charm sector are more sensitive to finite mass corrections .
phenomenological models of the exotic hadrons based on the born - oppenheimer picture would use these potentials as input .
+ with a static color and anti - color source , separated along an axis at distance @xmath38 , the eigenstates of the hamiltonian are irreducible representations of the little group of symmetries that preserve this axis .
the energy of these states as a function of distance defines the potential , @xmath39 .
the residual symmetry means these potentials are labelled by @xmath40 , where there are two spin - zero potentials since a mirror symmetry is also a good symmetry for this case .
the half - integer spin potentials do not appear in a theory of gluons alone but would be present in qcd . with two flavors of light quarks ,
qcd energy eigenstates are classified with an extra quantum number , light isospin , and this property would be inherited by states built from static sources . there
would thus be a new multiplicity of spectra with isospin @xmath41 .
the isospin 0 and 1 spectra would be the relevant ones for studies of hidden charm or bottom tetraquarks and in particular , since the @xmath42 states are charged , the isospin 1 spectrum is of particular interest .
this spectrum has not been computed in lattice qcd to date .
+ for pentaquarks containing hidden charm @xmath43 or bottom @xmath44 the isospin @xmath45 and @xmath46 potentials are relevant for modelling these states .
again , there is no counterpart for this potential in the theory of strongly - coupled gluons alone .
another possible potential that might usefully be investigated and which has no counterpart in the pure gauge theory are those associated with two color sources , @xmath47 @xcite . in order to neutralize this color charge , at least one light quark field
must be included in the creation operator .
these potentials would help model doubly - charmed or doubly - bottom baryons .
4 . * compute form factors relevant to exotic states . * + the determination of the elastic and inelastic form factors of the @xmath37 resonances directly from lattice qcd would have three major impacts .
first , it would lead to the theoretical reproduction of experimentally observed production or decay rates in a model - independent way .
second , it will give access to poorly constrained quantities that would elucidate the nature and structure of these exotic states .
examples of such quantities include the radii and electromagnetic moments of tentative molecular states .
third , it will guide future experimental searches of exotics .
although the studies of resonant electromagnetic processes are presently at their early stages , there have been a great deal of theoretical @xcite and numerical @xcite that demonstrate that they are in fact accessible from lattice qcd .
this progress resulted in the first calculation of a radiative transition of a hadronic resonance @xcite .
this calculation was performed in the light sector for @xmath48 .
+ having determined the @xmath48 amplitude for a range of values of the center of mass energy of the final @xmath49 state , the authors of ref .
@xcite were able to analytically continue the amplitude onto the @xmath50-pole and determine the @xmath51 form factor .
+ the same technology will be applicable for future calculations in the heavy quark sector .
compute decay constants for exotic states . *
+ the decay constants of the vector resonances determine their rate of production in @xmath52 and radiative transitions to lighter states offer a way to produce states of other @xmath53 .
the rigorously correct way to determine these in lattice qcd is to first determine the scattering amplitudes and their resonant content as described above , and to then introduce an external vector current . by extrapolating the calculated vector - current matrix elements to the resonance poles , off in the complex energy plane , the decay constants and radiative transition rates for resonances can be obtained .
this procedure closely resembles that for the determination of form factors discussed above .
the techniques necessary for implementing this have been previously developed in refs .
the first calculations of this type has been of @xmath49-electroproduction @xmath54 @xcite . + a slightly less rigorous approach , which may be acceptable for narrow resonances , is to ignore the hadronic decay of the states by excluding meson - meson - like operators from the basis used to determine the spectrum of states a first round of calculations of this simplified type may be justified to aid our phenomenological intuition of the vector spectrum , extending the limited calculations presented in @xcite , using the excited state technology presented in @xcite .
* in - medium hadron properties .
* + several model calculations predict that charmonium - nucleus exotic bound states should exist @xcite .
two independent , equally important binding mechanisms have been identified : multigluon exchanges in the form of color van der waals forces @xcite , and @xmath55 meson loop contributions to charmonium self - energy with medium - modified masses @xcite . a first , recent lattice calculation @xcite confirms model calculation expectations , finding relatively deeply bound states of @xmath56 and @xmath57 to several light nuclei .
another interesting class of charmed - hadrons nuclear bound states are @xmath58-mesic nuclei @xcite , which are an important source of information on chiral symmetry restoration in - medium @xcite . a lattice calculation of the @xmath59meson interaction with nucleons and of @xmath59nuclei binding energies would be of great importance for constraining models , given the present lack of experimental information on the @xmath59nucleon interaction .
1 . * publish upper limits for negative searches . *
+ candidates for exotic hadrons have been observed in many channels . while it is clearly important to find new decay channels for these states , it is also important to limit their decays to other channels when these searches are negative .
this is a reminder to experimentalists to publish upper limits on cross sections and branching fractions in a wide range of channels .
2 . * confirm marginal states .
* + a variety of signals have been observed that require confirmation .
there is some urgency in achieving this because attempts to understand the data can be seriously misled by the acceptance of spurious signals as hadronic states . alternatively , many signals are statistically significant but contain unknown systematic errors due to assumptions in modelling ( for example using interfering breit - wigner amplitudes to obtain asymmetric line shapes ) .
additional and more varied amplitude analysis is required in these cases . amongst states requiring confirmation
are @xmath60 , @xmath61 , @xmath62 , @xmath63 , @xmath64 , and @xmath65 .
3 . * unravel the excited @xmath66 spectrum .
* + the masses of the charmonium @xmath67 states are expected to be around 3.84.0 gev/@xmath68 @xcite , while @xmath69 and @xmath70 are well above the @xmath71 threshold but below @xmath72 threshold ; they are expected to be wide .
if the mass of @xmath73 is high enough , @xmath74 will be its dominant decay mode .
the @xmath70 may decay into @xmath75 as well .
+ so far the @xmath76 observed in @xmath77 @xcite is regarded as the @xmath70 state , and the @xmath78 observed in @xmath79 @xcite is supposed to be a @xmath69 candidate , although its mass is a bit too close to @xmath70 , and it was not observed in @xmath77 .
+ further study on @xmath80 decaying to @xmath71 and @xmath75 should be performed to identify @xmath69 and @xmath73 and to confirm the @xmath70 .
+ with more data collected in @xmath16 annihilation at the @xmath81 and @xmath82 peaks , the e1 transitions @xmath83 and @xmath84 should be searched for ; e1 transitions of @xmath85 are also expected to be large compared with @xmath86 and @xmath87 .
+ hadronic transitions @xmath88 should be searched for , and the reaction @xmath89 may occur if the mass difference between @xmath80 and @xmath90 is large enough .
the spin - parity of the @xmath69 candidate , @xmath78 , needs to be measured and its production and decay patterns should be examined carefully to see if it is the @xmath69 .
* measure @xmath16 cross sections . *
+ the region at center - of - mass energies above the open charm threshold is of great interest due the plethora of vector charmonium states : the @xmath91 , @xmath81 , @xmath82 , and @xmath92 observed in the inclusive hadronic cross section , and the vector charmonium - like states , the @xmath61 , @xmath93 , @xmath94 , @xmath95 , and @xmath96 observed in exclusive hadronic modes .
these states were discovered in one specific mode and are not observed in other modes .
searches for these states in all possible final states are desired .
this suggests high precision measurements of as many as possible exclusive @xmath16 annihilation modes , including multi - body open charm modes , hadronic transitions , radiative transitions , and even exclusive light hadron final states .
[ xsections ] shows an example of measured cross sections of two - body open charm final states and two- or three - body hadronic transition modes .
common features of the distributions are a richness of structures and a lack of precision . with more data from open charm threshold
up to about 5 gev and improved precision , better theoretical models will likely be needed to describe the line shapes of all the final states simultaneously . in this way
better knowledge on the excited @xmath97 and the @xmath98 states can be extracted .
this may result in an understanding of the nature of these states and reveal if any are charmonium hybrids .
+ + with the existing data samples , besiii can already improve precision of the open charm cross sections significantly @xcite , considering the besiii experiment will continue run for another 6 - 8 years , better measurements at more energy points are expected .
belle - ii @xcite will start data taking in 2018 with a data sample expected to be fifty times larger than belle s .
thus the precision of all the measurements with initial state radiation will be improved .
+ the cross sections of @xmath16 annihilation into open bottom and bottomonium final states should also be measured to understand the excited bottomonium and bottomonium - like states .
this can only be done at the belle - ii experiment @xcite .
* search for flavor analog exotic states . *
+ the majority of recently discovered exotic states are placed firmly in the charmonium spectrum .
flavor - independence of gluon exchange implies that flavor analog states should exist .
for example , the @xmath99 and @xmath100 are evidently hidden bottom partners to the @xmath101 and @xmath102 multiquark candidate states .
it is possible that the @xmath103 is the hidden strange partner of the @xmath93 .
finding flavor - analog states will yield valuable information on the dynamics underlying the new states and will probe the robustness of putative models .
+ the case of a bottomonium analog of the @xmath104 is interesting , both because of the novelty of the @xmath105 and because of differences that may arise .
for example , if the @xmath105 is a weakly bound @xmath106 system then the @xmath107 would be expected at mass of 10604 mev . however , some models @xcite rely on the proximity of the hidden charm @xmath108 and @xmath109 channels to stabilize the @xmath105 .
this coincidence is _ not _ repeated in the case of the @xmath107 , where the @xmath110 threshold lies 370 mev away .
this also implies that the novel isospin - breaking features of the @xmath105 will not be repeated in the @xmath107 ( isospin symmetry breaking is related to the hidden flavor mixing and to the splitting between charged and neutral @xmath106 channels neither of which is repeated in the case of the @xmath107 ) .
finally the proximity of the @xmath111 to the @xmath104 is likely to be important .
again , this numerical coincidence is not repeated in the case of the @xmath107 , where nearby @xmath112 states are at 10255 mev ( 1p ) , 10512 mev ( 2p ) , or 10788 mev ( 3p@xcite ) . 6 . * search for flavor analogs of the @xmath113 . * + the recent evidence for the resonant @xmath114 structures in @xmath115 in the @xmath116 decays found by the lhcb experiment has renewed the interest of the experimental and theoretical communities in pentaquark states . further experimental work is critical for clarification of the nature of these structures .
+ most established and candidate exotic hadron states contain hidden heavy flavor , @xmath17 .
this is mainly due to experimental constraints for production and detection .
however , other sectors of flavor deserve to be investigated .
let us give two examples .
+ the isospin partner @xmath117 and the strangeness partners such as @xmath118 should be searched for , along with their @xmath119 analogs .
one should not restrict to hidden heavy flavor .
pentaquark states @xmath120 , where @xmath121 denotes @xmath122 , @xmath123 or @xmath124 were predicted in 1987 on the basis of a chromomagnetic mechanism very similar to the one leading to speculations about the @xmath125 dibaryon .
this flavored pentaquark has been searched for at fermilab and hera .
searches with higher statistics are desirable , especially if more hidden - flavor states such as @xmath126 are found @xcite . + exotic mesons with double heavy flavor ,
@xmath127 have been predicted with many methods such as potential models , qcd sum rules , lattice qcd and the meson - meson molecular picture .
it has also been stressed that more effort should be put on double - charm and other doubly - heavy baryons .
we thus suggest a search of doubly - heavy hadrons besides @xmath128 : double - charm baryons , double - charm mesons and double - charm dibaryons , and in the future , their analogs with charm and beauty or double beauty@xcite .
* search for quantum number partners of the @xmath93 .
* + if the @xmath93 is a hybrid state it represents the first example of what is expected to be a large array of novel hadrons .
specifically , a spin multiplet analogous to those in the conventional spectrum is expected .
a lattice computation of the lightest charmonium hybrid multiplets is shown in fig .
[ fig - cch ] . a clear structure with quantum numbers @xmath129 and @xmath130 is seen .
this multiplet can be conveniently interpreted as arising due to an effective constituent gluon with quantum numbers @xmath131 mixing with conventional quark - antiquark degrees of freedom@xcite .
+ + given this information one can expect the spectrum shown in table [ tab - cch ] .
the increasing luminosity expected at the colliders raises interesting possibilities for detecting these states .
for example , the @xmath132 and @xmath133 lie below the @xmath93 and therefore should be accessible in radiative decays in p - wave .
the hybrids can then be detected in decay modes such as @xmath134 or @xmath135 ( see table [ tab - decays ] ) .
+ .expected hybrid multiplet @xcite .
[ cols="<,<",options="header " , ] + [ tab - decays ] 8 .
* pursue properties of the @xmath104 . *
+ although properties of the @xmath104 are reasonably well known , additional experimental effort can greatly assist in improving the understanding of this state .
for example , the rate for decays to light hadrons , such as @xmath136 , can be compared to those for @xmath137 states in an effort to determine the expected mixing of the @xmath105 with the bare @xmath73 . + analog hidden charm states are predicted in some models and can be searched for .
for example a @xmath138 @xmath139 state is expected at 4019 mev in pion - exchange models@xcite .
+ intriguing analog flavor - exotic states are also expected in qcd . in particular
, it has been argued that @xmath140 states must exist in the limit where the heavy quark mass goes to infinity@xcite .
the phenomenology of such states is discussed by tornqvist@xcite and was anticipated long ago@xcite .
specific possibilities include isoscalar @xmath141 , @xmath142 , @xmath143 states with @xmath144 and @xmath145 , @xmath146 , @xmath147 , etc .
nevertheless , flavor exotic vector - vector bound states are unlikely , except possibly in the doubly charged bottom sector@xcite .
* measure additional channels to investigate the @xmath113 . *
+ the interpretation of the lhcb pentaquark signal remains open .
tightly bound pentaquarks , molecular states , and rescattering effects have been proposed .
more accurate determination of the quantum numbers of the pentaquark candidates would greatly help their interpretation .
even before more data is accumulated by the lhcb , improving parameterizations in the amplitude fits to the existing data may help this end .
for example , models of @xmath148 excitations , which dominate the data via decays to @xmath149 , need to be checked for completeness since the previous experiments may have not discerned all the relevant states , especially at high masses .
non - resonant terms with slowly varying magnitude and phase can also be significant .
alternative approaches to the isobar model may be helpful , like , for example , the recently published coupled - channel model by fernandez - ramirez et al.@xcite .
see also refs .
+ it is important to confirm the @xmath113 via other channels .
there are already suggestions@xcite such as @xmath150 , or @xmath151 , which are cabibbo suppressed , or @xmath152 .
the predictions of rescattering models for the @xmath114 amplitudes can be tested by fitting them directly to the data .
the presence ( or lack thereof ) of the same structures in the other channels , like @xmath150 or @xmath153 is of great importance . + rescattering models predict presence of structures related to the @xmath114 peaks induced by the analyticity in the coupled channels , like @xmath154 , @xmath155 and @xmath156 . ideally , simultaneous coupled channel analysis of the related final states should be performed .
the investigation of possible structures , which may include depressions rather than peaks , is a good start .
even total relative rates between the different channels would be interesting .
negative searches have theoretical implications and should be published .
+ bound - state models for the observed @xmath114 states predict other pentaquark states built by the same binding mechanisms .
the same @xmath114 states may be observable in the other decay modes too .
thus , every accessible decay mode of @xmath157 with @xmath158 and @xmath159 quarks among the final state hadrons should be examined , e.g. @xmath160 .
other final states can be accessible from @xmath161 to charmonium decays .
+ different production mechanisms for the @xmath114 states , or their siblings , should be investigated .
examples are prompt production at lhc or photo - production at jlab
. 10 . * test ideas for meson - nuclear interactions . *
+ presently there is a complete lack of experimental information on the low - energy interactions of charmed mesons and charmonium with nucleons and nuclei .
we look forward to several forthcoming experimental programs in this area : the near - threshold experiments by the athena collaboration @xcite as part of the 12 gev program at jefferson lab , the proton - antiproton experiments by the panda collaboration at fair @xcite , and the experiments with 50 gev high - intensity proton beams at the j - parc complex @xcite .
we also envisage opportunities for finding exotic charmonium - nucleon and charmonium - nucleus bound states with the ongoing heavy - ion experiments at rhic and lhc . in particular , we suggest studies on the formation of such exotic bound states by coalescence in the late - stage evolution of heavy - ion collisions , as their production yields should be of comparable magnitude to those of anti - nuclei and anti - hypernuclei recently observed at rhic @xcite and lhc @xcite . 11 . *
improve meson classification scheme . *
+ there is a wide range in signal robustness in the spectrum of new states . because this can lead to confusion amongst theorists who are attempting phenomenological descriptions of these new states , we recommend that a star system for mesons be implemented by the particle data group for use in the _ review of particle properties_. furthermore , the current exotic particle naming scheme
is somewhat confused and is applied inconsistently ; we therefore recommend that a consistent and flexible nomenclature be implemented
. 12 . * search for @xmath162 in decays at lhc for panda . *
+ heavy - flavor physics will benefit from experiments with medium - energy antiproton beams .
in the past , a precursor signal of the @xmath163 was seen at the cern isr , and many properties of the @xmath164 and other charmonium states were obtained from the @xmath165 experiment at the fermilab accumulator .
+ to assist future experiments , it is desirable to get information on the coupling of heavy hadronic systems to proton - antiproton pairs by detecting @xmath166 production at @xmath14-factories and at the lhc .
this is already under way , and this should be accompanied by more theoretical studies .
for instance , it remains rather mysterious that @xmath167 decays more often to @xmath168 than suggested by simple perturbative qcd , while @xmath169 is more weakly coupled to that channel .
* study exclusive @xmath16 cross sections using better coupled - channel formalism . * + the identification of possible new resonances implied by the @xmath37 phenomena requires studies of analytical amplitudes that describe the relevant production and decay characteristics@xcite .
for example , in the case of the @xmath101 that is observed in the @xmath170 spectrum in decays of the @xmath93 to @xmath171 , the relevant direct channels involve the nearby open - charm , @xmath172 states and need to be included in a coupled channel formalism .
the open charm resonances in the production channel , e.g. , the @xmath173 can produce rapid variations of the direct channel partial waves near the @xmath174 signal and need to be taken into account in production . the singularity structure of partial wave amplitudes is constrained by unitarity , therefore a comprehensive analysis requires implementation of unitarity constraints in all relevant channels .
this requires simultaneous studies of quasi two - to - two scattering amplitudes of open flavor and heavy quarkonia , e.g. @xmath175 , and eventually a study of three - to - three scattering , i.e. @xmath176 amplitudes .
2 . * develop tests for the dynamical diquark picture .
* + in an alternate proposal for the structure of the heavy quarkonium - like exotics , both for the tetraquarks @xcite and pentaquarks @xcite , the states are composed of compact diquark - antidiquark ( -antitriquark ) pairs rapidly separating and hence ultimately achieving large ( @xmath177 fm or greater ) separation before decay .
this picture has features in common with the diquark models previously mentioned @xcite , but differs in that the states are extended , dynamical rather than compact , static objects , and therefore does not necessarily admit a hamiltonian description .
nevertheless , in the limit of small separation , the two pictures should coincide .
the first priority in this case is therefore the development of a formalism in which the spectrum can reliably be predicted .
a first attempt in the pentaquark sector @xcite , still using a hamiltonian formalism , gives a natural explanation for a broad @xmath178 lying just below a narrow @xmath179 , consistent with the lhcb findings @xcite , but also predicts a large number of undiscovered states .
a lattice calculation of the potential corresponding to a well - separated static diquark - antidiquark pair may provide valuable information on the possible spectrum . since the exotics are so prominent in the @xmath3 and @xmath7 channels , some hints of the same mechanism with @xmath180 ( hidden - strangeness pentaquarks ) should appear in processes such as @xmath181 @xcite or @xmath182 photoproduction @xcite .
a primary benefit of the dynamical picture is its natural explanation of strong overlaps with spatially larger states , so a precision measurement of the ratio @xmath183 vs. @xmath90 and to other states will be illuminating .
the dynamical and compact diquark models share an expected enhancement of @xmath184 compared to the corresponding rate in molecular models @xcite .
the extended structure of the state may also offer interesting opportunities for the production of unusual final - state particle correlations .
the multiparticle nature of states produced via , say , electroproduction or @xmath185 annihilation ( at jlab or panda , respectively ) can be probed by means of constituent counting rules @xcite , and can help to distinguish whether compact multiquark components are produced .
* develop experimental tests for tetraquarks . * + compact tetraquark configurations ,
in which all four quarks participate in strong mutual interactions , can be distinguished from the hadron molecular picture or threshold effects through a variety of experiments .
the most well developed tetraquark models are of the diquark - antidiquark class @xcite , and rely on hamiltonians with spin - spin interactions between the quark pairs . a comparison of the expected spectra in this tetraquark model versus hadronic molecular models ( and also hadrocharmonium )
@xcite indicates that many more states should arise if tetraquarks are the dominant exotic component ; for example , the @xmath104 should have isotriplet charged partners of the same @xmath186 parity [ and opposite that of the @xmath187 and @xmath188 . due to the proximity of thresholds , such states might exist only as very broad yet - undiscovered resonances .
large prompt production cross sections at colliders @xcite argue against @xmath104 being a @xmath189 molecule forming through coalescence ; indeed , an extrapolation @xcite of data from alice shows that production of loosely bound hadronic molecules such as @xmath190 and @xmath191he at high @xmath192 will be quite suppressed , unlike current indications for @xmath104 , an effect that can be decisively checked in future alice and lhcb experiments .
the molecular and diquark pictures also differ radically in the ratios of their branching fractions of @xmath184 vs. @xmath193 or @xmath194 @xcite , the former being dozens of times less frequent in molecular models .
loosely bound molecules also must obey well - known universal relations ( independent of the potential ) between binding energy and width , and precision measurements of the resonance widths and constituent masses can help determine whether these constraints are satisfied @xcite .
4 . * develop techniques for @xmath195 and @xmath196 systems * + potential models provide some guidance for qcd calculations .
two - body calculations are obvious once an explicit potential is given .
three - body and four - body computational methods new yield accurate spectra , although they require more delicate tools .
the case of five - body and six - body systems are still debated .
for instance , with similar ansatze for the interaction , the @xmath197 can be found to be either stable or unbound .
we suggest to publish a set of benchmark calculations to remove the ambiguities .
pursue the born - oppenheimer method ( adiabatic surface mixing ) . *
+ the presence of heavy charm or bottom quarks in the new putative tetraquark mesons suggests that they may be successfully studied using the _ born - oppenheimer expansion_. this approach was introduced by born and oppenheimer in 1920 @xcite to understand the binding of atoms into molecules by exploiting the large ratio between the mass of an atomic nucleus and an electron , which implies that the time scale for the motion of electrons is orders of magnitude faster than that for the motion of the nuclei .
the energies of stationary states of the electrons in the presence of fixed nuclei can be calculated as functions of the separation of the nuclei .
the resulting functions are called _ born - oppenheimer potentials_. in the born - oppenheimer approximation , these functions are used as potential energies in the schrdinger equation for the nuclei , under the assumption that the electrons respond very rapidly to the motion of the nuclei .
the born - oppenheimer _ expansion _ involves taking the large mass ratio into account more systematically by incorporating non - adiabatic couplings between different stationary states of the electrons .
this results in coupled - channel schrdinger equations that systematically improves the description of a molecule . +
the born - oppenheimer expansion was applied to mesons containing a heavy quark ( @xmath198 ) and antiquark ( @xmath199 ) in 1999@xcite , exploiting the fact that , since the mass of the heavy quark is much larger than the typical energies of the gluons and light quarks , the time scale for the evolution of the gluon and light - quark fields is much faster than that for the motion of the @xmath198 and @xmath199 . in ref .
@xcite , lattice qcd was used to calculate the born - oppenheimer potentials defined by the energies of the gluons in the presence of fixed @xmath198 and @xmath199 as functions of the @xmath25 separation .
these energies were then used as the potential energies in the schrdinger equation for the @xmath198 and @xmath199 .
the bound states in the born - oppenheimer potentials were interpreted as meson resonances .
these bound - state energies were compared with corresponding meson masses computed directly using lattice qcd and agreement in the level spacings to within 10% was found , strongly supporting the validity of the born - oppenheimer expansion for such systems . + the approach used in ref .
@xcite should be extended to apply to the @xmath37 mesons and to include nonadiabatic effects that can be incorporated through coupled - channel schrdinger equations .
the born - oppenheimer potentials for heavy tetraquark mesons and the nonadiabatic couplings between the potentials could be calculated using lattice qcd .
the heavy quark and antiquark would be treated as static , and the energies of the gluons and light quarks could then be computed as a function of the separation between the quark and the antiquark .
the resulting coupled - channel schrdinger equations could then be solved to determine the energies and widths of resonances , which can be compared with the observed @xmath37 mesons , and possibly to predict new tetraquark mesons .
* revisit conventional meson models . *
+ while the successes of the constituent quark model are well - known in the heavy quark sector , the efficacy of the model is not expected to survive higher in the spectrum , where gluonic and coupled channel effects become important . of course
, it should be possible to extend constituent models to include these additional degrees of freedom , but experimental and theoretical guidance will be required .
+ even the simple problem of assessing the accuracy of the constituent quark model above threshold has difficulties .
for example , there are eight charmonium states below @xmath200 threshold that are all well - described by models .
alternatively , the situation above threshold is considerably more confused ; of the approximately twenty claimed states , most of them are not understood , and even well - known states such as the @xmath91 lie 50 mev below the prediction of the godfrey - isgur model . in the bottom sector
the 14 states that lie below @xmath201 are well - described . in this case
there are only six states above threshold , but , again , the experimental and theoretical situation is confused .
+ since the new experimental data lie firmly in the continuum region , it is very likely that more sophisticated versions of the quark model that respect unitarity must be developed .
of course , this has been known in the community for many decades , and much work has been done@xcite .
there are daunting issues to be overcome , including determining the form of the nonperturbative gluonic transition operator and evaluating the ( divergent ) sum over infinitely many virtual channels@xcite .
nevertheless it is difficult to imagine progress being made without a successful outcome to this effort .
alternative approaches exist of course : lattice gauge theory is rapidly making progress in working in the coupled channel regime , and one hopes that effective field theory approaches will be developed that can accommodate the extra scales present . 7 .
* develop the dyson - schwinger formalism .
* + the dyson - schwinger equations ( dses ) of qcd , together with various many - body equations for bound states ( bethe - salpeter equations for the two - body problem , faddeev and faddeev - yakubovsky equations for the three- and four - body problem ) have the potential to reveal the connections between the physics in different sectors of qcd .
the equations encode the running of qcd green s functions ( for example , the quark mass function ) and therefore connect the perturbative current quark region with the non - perturbative constituent quark domain .
furthermore , they connect the heavy quark regime , where nrqcd or potential models are applicable , with the light quark sector , where the concept of a potential is not very well defined . + the explanatory power of the dse framework with respect to exotic hadrons is still in its early exploration stage .
so far , light scalar mesons have been treated as tetraquarks in an approach that takes into account two - body correlations within the bound state equation for two quarks and two antiquarks@xcite . the resulting bethe - salpeter amplitude for scalar tetraquarks
is dominated by pseudoscalar meson - meson correlations . for the lightest state ,
the @xmath202 , this explains its large decay width into two pions , whereas the @xmath203 and @xmath204 are dominated by their @xmath205 components .
in general , it turns out that all two - body correlations inside the tetraquark ( i.e. , ( anti-)diquarks or mesons ) contribute to the wave function and it is a question of the internal dynamics which is the dominant cluster . for the light scalar mesons
this is the ` meson molecule ' configuration , but other results are in principle possible for other quantum numbers and different quark flavors and masses .
+ whether this mechanism has the potential to shed some light on the question of the internal structure of the tetraquarks among the @xmath37-states , in particular their ( anti-)diquark , molecular or hadrocharmonium nature , needs to explored . to this end , non - scalar quantum numbers need to be studied and the framework needs to be extended toward heavy - light systems .
furthermore , more quantitative precision is needed to confirm the prediction of an all - charm tetraquark in the 5.0 - 6.5 gev mass region @xcite . + complementary ongoing projects within the dse framework concern the glueball spectrum @xcite and the question whether states with exotic quantum numbers can be accounted for with relativistic quark - antiquark systems ( in contrast to the non - relativistic quark model ) @xcite .
the status of large @xmath206 considerations . * + one striking thing about modern exotics the @xmath37 states is that they all involve the physics of heavy quarks .
this raises an interesting issue : are heavy quarks necessary for the formation of exotics or do exotics exist for light quarks systems .
the experimental data on this is murky .
the large @xmath206 limit may provide a bit of insight .
the subtle point is that the large @xmath206 and heavy quark limits may not commute so that generic large @xmath206 arguments based on scaling arguments really apply for light quark systems . the standard version of the large @xmath206 limit with quarks in the fundamental representation of su(@xmath206 )
can be shown not to have narrow tetraquarks at large @xmath206@xcite , apparently supporting the proposition that the heavy quarks are necessary for the existence of tetraquark states .
however , there is a variant of the large @xmath206 limit where quarks are in the two - index anti - symmetric representation of su(@xmath206 ) in which it can be shown that states with exotic tetraquark quantum numbers must exist as narrow resonances ( i.e. states whose widths go to zero as @xmath206 goes to infinity ) regardless of the mass of the quarks@xcite .
minimally this shows that qcd - like gauge theories are not excluded from having tetraquark states even if the quarks are light .
the following are a few suggestions that could help facilitate collaboration between theory and experiment . 1 .
* improve parameterizations of the data . *
+ one of the challenges in many of the experimental studies of the @xmath37 states is to develop correct parameterizations of the data .
for example , amplitude analyses often find a need to introduce non - resonant terms . at present ,
very little theoretical guidance is provided except for the amplitude formulations based on the k - matrix approach . however , the latter is not always practical .
+ to improve this situation , we have two recommendations .
+ first , we encourage that , when appropriate and beneficial , experimentalists and theorists directly work together on the analysis of data .
this could be accommodated by theorists becoming co - authors on specific experimental papers they substantially contributed to , or joint submission of experimental and theoretical papers cross - referencing each other .
the experiments are encouraged to formalize procedures making such collaboration possible , and theorists are encouraged to approach the experiments when they think they might directly aid specific data analysis topics .
further progress could be made by more persistent forms of collaboration , including direct involvement of theorists in the data analysis process within the established procedures of the experimental collaborations .
+ second , we encourage theorists , when possible , to publish complete functional forms ( amplitudes , etc . ) that could be used in the fitting of data .
one example of this is in the parameterization of rescattering amplitudes .
the current theoretical calculations are dependent only on the center of mass energy @xcite , whereas amplitudes used in fitting require a flexible parameterization involving the angular information of the decay .
if theorists develop more complete rescattering amplitudes , experimentalists could use them in analyses .
this would most likely involve some collaborative effort in understanding both how the experimental analysis is performed and what are the theoretical requirements for such amplitudes .
another example is in resonance parameterizations : it would be useful for experimentalists if a number of alternate resonance parameterizations were available that could be used in systematic studies .
+ 2 . *
make experimental results more accessible for subsequent interpretation . *
+ the analysis of data from many modern experiments often necessitates complying with internal rules designed to provide collaborative controls over the quality of statistical methods used and the proper evaluation of systematic uncertainties .
therefore it is unrealistic to expect that all data will be made available for analyses outside of this collaborative setting .
a correct analysis of data would benefit from the types of closer interaction between experiment and theory discussed in the previous point . +
a different issue is how published data ( for example , dalitz plots ) should be subsequently interpreted .
it often occurs that experimental results are made public in a manner that does not easily allow for subsequent interpretation .
+ one example is the discovery of the @xmath101 decaying to @xmath170 in the process @xmath207 @xcite .
the data presented in the discovery papers include several effects that are difficult to take into account when performing theoretical fits .
first , the besiii data @xcite was taken at @xmath208 gev , while the belle data @xcite includes a range of energies around the @xmath93 peak .
this makes the belle data , in particular , hard to subsequently fit , since any changes ( beyond the size of available phase space ) in the @xmath170 mass spectrum as a function of @xmath209 mass are unknown .
second , the two experiments have different experimental efficiencies over the dalitz plot due to differing detectors and kinematics .
these effects are not quantified in the publications .
the importance of these two effects can be seen when comparing the @xmath210 plots from besiii and belle , which differ substantially , especially in the low - mass region ( fig .
[ zcshapes ] ) .
it is therefore not clear how one could correctly analyze the published data with various new parameterizations to test , for example , differences between cusp and resonant models of the @xmath101 .
+ when deemed appropriate , experiments are therefore encouraged to make efficiency - corrected data available for external analyses . or , when possible or desired , experiments could make published plots publicly available along with efficiency curves and instructions for how to use the plots for subsequent analysis .
this could be provided as supplemental information to a publication .
this may be easy for simple three - body final states ( like @xmath211 , where the dalitz plot could be provided ) , but impractical for more complicated final states .
+ another suggestion , especially when amplitude analyses have been performed , is to publish a complete parameterization of the data , including both the formulas and the numerical values for each fit parameter .
+ making data public in these ways could help facilitate ongoing efforts to test and build models .
it could also permit combined fits of data from different experiments or different channels , thus helping a more global picture to emerge .
* preview upcoming analysis results .
* + it may be useful , in some circumstances , for experimental collaborations to provide a list of upcoming experimental results .
this might be a fruitful way to elicit new theoretical predictions or ideas .
for example , if the community knows a certain measurement is being performed , there is then a chance to make predictions prior to the publication of experimental results .
it also permits setting priorities in theoretical computations and enhances the possibility of arranging collaborative effort . hosting such a list on a common platform may prove useful and should be explored .
4 . * create a list of publications in @xmath37 physics . *
+ it would be helpful to have a centralized running bibliography of publications relevant for @xmath37 physics .
we are grateful to david kaplan and the int for financial and logistics support .
the authors thank the following agencies for financial support : the u.s .
department of energy ( cohen ) ; the institute of modern physics and chinese academy of sciences under contract y104160yq0 and agreement no .
2015-bh-02 ( coito ) ; the u.s .
department of energy , for grant de - ac05 - 06or23177 , under which jefferson science associates , llc , manages and operates jefferson laboratory and de - sc0006765 , early career award ( dudek ) ; fermilab , operated by the fermi research alliance under contract number deac02 - 07ch11359 with the u.s .
department of energy ( eichten ) ; bmbf , under contract no .
06gi7121 , and the daad under contract no .
56889822 and by the helmholtz international center for fair within the loewe program of the state of hesse ( fischer ) ; the german research foundation dfg under contract number collaborative research centre crc-1044 ( gradl ) ; the conselho nacional de desenvolvimento cientfico e tecnolgico - cnpq , grant no .
305894/2009 - 9 and fundao de amparo pesquisa do estado de so paulo - fapesp , grant no . 2013/01907 - 0 ( krein ) ; u.s . national science foundation , under grants phy-1068286 and phy-1403891 ( lebed ) ; the brazilian national council for scientific and technological development under grant cnpq / capes-208188/2014 - 2 ( machado ) ; u.s .
department of energy under grant de - fg02 - 05er41374 ( mitchell ) ; u.s .
national science foundation under grant phy-1306805 ( morningstar ) ; u.s .
department of energy , supported by jefferson science associates , llc under contract no .
de - ac05 - 06or23177 ( pennington ) ; the national natural science foundation of china ( nsfc ) under contract no . 11575017 , the ministry of science and technology of china under contract no .
2015cb856701 ( shen ) ; u.s .
department of energy , under grant de - fg02 - 05er41374 ( shepherd ) ; u.s
. national science foundation under grant phy-1507572 ( skwarnicki ) ; u.s .
department of energy , under contract de - ac05 - 06or23177 and grant de - fg0287er40365 ( szczepaniak ) ; the national natural science foundation of china ( nsfc ) under contract numbers 11235011 and 11475187 ( yuan ) .
e. s. swanson , phys .
rept . * 429 * , 243 ( 2006 ) [ hep - ph/0601110 ] .
s. godfrey and s. l. olsen , ann .
nucl . part .
* 58 * , 51 ( 2008 ) [ arxiv:0801.3867 [ hep - ph ] ] .
n. brambilla _ et al .
_ , eur .
j. c * 71 * , 1534 ( 2011 ) [ arxiv:1010.5827 [ hep - ph ] ]
. s. eidelman , b. k. heltsley , j. j. hernandez - rey , s. navas and c. patrignani , arxiv:1205.4189 [ hep - ex ] . g. t. bodwin , e. braaten , e. eichten , s. l. olsen , t. k. pedlar and j. russ , arxiv:1307.7425 .
s. prelovsek , l. leskovec , phys .
111 * 192001 ( 2013 ) ; c.b .
lang , l. leskovec , d. mohler , s. prelovsek jhep * 1509 * , 089 ( 2015 ) .
j. j. dudek _ et al .
_ [ hadron spectrum collaboration ] , phys .
lett . * 113 * , no .
18 , 182001 ( 2014 ) [ arxiv:1406.4158 [ hep - ph ] ] .
d. j. wilson , j. j. dudek , r. g. edwards and c. e. thomas , phys .
d * 91 * , no . 5 , 054008 ( 2015 ) [ arxiv:1411.2004 [ hep - ph ] ] .
d. j. wilson , r. a. briceo , j. j. dudek , r. g. edwards and c. e. thomas , phys .
d * 92 * , no .
9 , 094502 ( 2015 ) [ arxiv:1507.02599 [ hep - ph ] ] .
d. mohler , c.b .
lang , l. leskovec , s. prelovsek , r.m .
woloshyn , phys .
* 111 * , 222001 ( 2013 ) ; c.b .
lang , l. leskovec , d. mohler , s. prelovsek , r.m .
woloshyn , phys .
* d90 * , 034510 ( 2014 ) .
s. perantonis and c. michael , nucl .
b * 347 * , 854 ( 1990 ) .
k. j. juge , j. kuti and c. morningstar , phys .
lett . * 90 * , 161601 ( 2003 ) [ hep - lat/0207004 ] .
e. braaten , c. langmack and d. h. smith , phys .
d * 90 * , no . 1 , 014044 ( 2014 ) [ arxiv:1402.0438 [ hep - ph ] ] . m. berwein , n. brambilla , j. t. castella and a. vairo , arxiv:1510.04299 [ hep - ph ] .
a. p. szczepaniak and p. krupinski , phys .
d * 73 * , 034022 ( 2006 ) e. s. swanson and a. p. szczepaniak , phys .
d * 59 * , 014035 ( 1999 ) g. s. bali _ et al .
_ [ sesam collaboration ] , phys . rev .
d * 71 * , 114513 ( 2005 ) [ hep - lat/0505012 ] .
p. bicudo , k. cichy , a. peters , b. wagenbach and m. wagner , phys .
d * 92 * , no . 1 , 014507 ( 2015 ) [ arxiv:1505.00613 [ hep - lat ] ] .
p. bicudo , k. cichy , a. peters and m. wagner , arxiv:1510.03441 [ hep - lat ] . z. s. brown and k. orginos , phys .
d * 86 * , 114506 ( 2012 ) [ arxiv:1210.1953 [ hep - lat ] ] .
l. lellouch and m. luscher , commun .
phys . * 219 * , 31 ( 2001 ) [ hep - lat/0003023 ] .
r. a. briceo , m. t. hansen and a. walker - loud , phys .
d * 91 * , no . 3 , 034501 ( 2015 ) [ arxiv:1406.5965 [ hep - lat ] ] .
a. agadjanov , v. bernard , u. g. meiner and a. rusetsky , nucl .
b * 886 * , 1199 ( 2014 ) [ arxiv:1405.3476 [ hep - lat ] ] .
r. a. briceo and m. t. hansen , phys .
d * 92 * , no . 7 , 074509 ( 2015 ) [ arxiv:1502.04314 [ hep - lat ] ] .
r. a. briceo and m. t. hansen , arxiv:1509.08507 [ hep - lat ] .
j. j. dudek , r. edwards and c. e. thomas , phys .
d * 79 * , 094504 ( 2009 ) [ arxiv:0902.2241 [ hep - ph ] ] .
c. j. shultz , j. j. dudek and r. g. edwards , phys . rev .
d * 91 * , no .
11 , 114501 ( 2015 ) [ arxiv:1501.07457 [ hep - lat ] ] .
r. a. briceo , j. j. dudek , r. g. edwards , c. j. shultz , c. e. thomas and d. j. wilson , arxiv:1507.06622 [ hep - ph ] .
h. b. meyer , phys .
lett . * 107 * , 072002 ( 2011 ) [ arxiv:1105.1892 [ hep - lat ] ] .
x. feng , s. aoki , s. hashimoto and t. kaneko , phys .
d * 91 * , no . 5 , 054504 ( 2015 ) [ arxiv:1412.6319 [ hep - lat ] ] .
j. bulava , b. hrz , b. fahy , k. j. juge , c. morningstar and c. h. wong , arxiv:1511.02351 [ hep - lat ] . j. j. dudek , r. g. edwards and d. g. richards , phys . rev .
d * 73 * , 074507 ( 2006 ) [ hep - ph/0601137 ] .
j. j. dudek and r. g. edwards , phys .
lett . *
97 * , 172001 ( 2006 ) [ hep - ph/0607140 ] .
d. becirevic , m. kruse and f. sanfilippo , jhep * 1505 * , 014 ( 2015 ) [ arxiv:1411.6426 [ hep - lat ] ] .
d. becirevic and f. sanfilippo , jhep * 1301 * , 028 ( 2013 ) [ arxiv:1206.1445 [ hep - lat ] ] .
y. chen _ et al .
_ , phys .
d * 84 * , 034503 ( 2011 ) [ arxiv:1104.2655 [ hep - lat ] ] .
s. j. brodsky , i. schmidt , and g. de teramond , phys .
lett . * 64 * , 1011 ( 1990 ) .
t. barnes , s. godfrey and e. s. swanson , phys .
d * 72 * , 054026 ( 2005 ) [ hep - ph/0505002 ] .
e. j. eichten , k. lane and c. quigg , phys .
lett . * 89 * , 162002 ( 2002 ) [ hep - ph/0206018 ] .
s. uehara _ et al .
_ [ belle collaboration ] , phys .
lett . * 96 * , 082003 ( 2006 ) [ hep - ex/0512035 ] .
s. uehara _ et al . _ [ belle collaboration ] , phys .
lett . * 104 * , 092001 ( 2010 ) [ arxiv:0912.4451 [ hep - ex ] ] . c. z. yuan , front .
china * 10 * , no . 6 , 101401 ( 2015 ) [ arxiv:1509.06850 [ hep - ex ] ] .
t. abe _ et al . _ [ belle - ii collaboration ] , `` belle ii technical design report , '' arxiv:1011.0352 [ physics.ins-det ] .
e. s. swanson , phys .
b * 588 * , 189 ( 2004 ) [ hep - ph/0311229 ] .
s. godfrey and k. moats , phys .
d * 92 * , no . 5 , 054034 ( 2015 ) [ arxiv:1507.00024 [ hep - ph ] ] .
c. gignoux , b. silvestre - brac and j. m. richard , phys .
b * 193 * , 323 ( 1987 ) .
h. j. lipkin , phys .
b * 195 * , 484 ( 1987 ) .
j. p. ader , j. m. richard and p. taxil , phys .
d * 25 * , 2370 ( 1982 ) .
a. v. manohar and m. b. wise , nucl .
b * 399 * , 17 ( 1993 ) [ hep - ph/9212236 ] .
dudek , phys .
d 84*,074023 ( 2011 ) .
c. a. meyer and e. s. swanson , prog . part .
phys . * 82 * , 21 ( 2015 ) [ arxiv:1502.07276 [ hep - ph ] ] .
l. liu _ et al . _
[ hadron spectrum collaboration ] , jhep * 1207 * , 126 ( 2012 ) [ arxiv:1204.5425 [ hep - ph ] ] .
n. a. tornqvist , z. phys .
c * 61 * , 525 ( 1994 ) [ hep - ph/9310247 ] . c. fernandez - ramirez , i. v. danilkin , d. m. manley , v. mathieu and a. p. szczepaniak , arxiv:1510.07065 [ hep - ph ] .
t. j. burns , arxiv:1509.02460 [ hep - ph ] .
f. k. guo , u. g. meiner , w. wang and z. yang , phys .
d * 92 * , no . 7 , 071502 ( 2015 ) [ arxiv:1507.04950 [ hep - ph ] ] .
m. mikhasenko , arxiv:1507.06552 [ hep - ph ] . .
b. i. abelev [ star collaboration ] , science * 328 * , 58 ( 2010 ) [ arxiv:1003.2030 [ nucl - ex ] ] .
j. adam _ et al .
_ [ alice collaboration ] , arxiv:1506.08951 [ nucl - ex ] .
a. p. szczepaniak , [ arxiv:1510.01789 [ hep - ph ] ] .
a. p. szczepaniak , phys .
b * 747 * , 410 ( 2015 ) [ arxiv:1501.01691 [ hep - ph ] ] .
brodsky , d.s .
hwang , and r.f .
lebed , phys .
lett . * 113 * , 112001 ( 2014 ) [ arxiv:1406.7281 [ hep - ph ] ] .
lebed , phys .
b * 749 * , 454 ( 2015 ) [ arxiv : 1507.05867 [ hep - ph ] ] .
l. maiani , f. piccinini , a.d .
polosa , and v. riquer , phys .
d * 71 * , 014028 ( 2005 ) [ hep - ph/0412098 ] .
l. maiani , f. piccinini , a.d .
polosa , and v. riquer , phys .
d * 89 * , 114010 ( 2014 ) [ arxiv:1405.1551 [ hep - ph ] ] .
r. zhu and c.f .
qiao , arxiv:1510.08693 [ hep - ph ] .
r. aaij _ et al . _
[ lhcb collaboration ] , phys .
* 115 * , 072001 ( 2015 ) [ arxiv:1507.03414 [ hep - ex ] ] .
lebed , arxiv:1510.06648 [ hep - ph ] .
lebed , arxiv:1510.01412 [ hep - ph ] .
a. esposito , a.l .
guerrieri and a. pilloni , phys .
b * 746 * , 194 ( 2015 ) [ arxiv:1409.3551 [ hep - ph ] ] .
s.j . brodsky and r.f .
lebed , phys .
d * 91 * , 114025 ( 2015 ) [ arxiv:1505.00803 [ hep - ph ] ] .
m. cleven , f.k .
guo , c. hanhart , q. wang , and q. zhao , phys .
d * 92 * , 014005 ( 2015 ) [ arxiv:1505.01771 [ hep - ph ] ] .
s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1304 * , 154 ( 2013 ) [ arxiv:1302.3968 [ hep - ex ] ] .
a. esposito , a.l .
guerrieri , l. maiani , f. piccinini , a. pilloni , a.d .
polosa and v. riquer , phys .
d * 92 * , 034028 ( 2015 ) [ arxiv:1508.00295 [ hep - ph ] ] .
g. li , x.h .
liu and z. zhou , phys .
d * 90 * , 054006 ( 2014 ) [ arxiv:1409.0754 [ hep - ph ] ] .
polosa , phys .
b * 746 * , 248 ( 2015 ) [ arxiv:1505.03083 [ hep - ph ] ] .
m. born and j.r .
oppenheimer , _ on the quantum theory of molecules _ , annalen der physik * 389 * , 457 ( 1927 ) .
e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. m. yan , phys .
lett . * 36 * , 500 ( 1976 ) .
e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. m. yan , phys .
d * 17 * , 3090 ( 1978 ) [ erratum - ibid .
d * 21 * , 313 ( 1980 ) ] .
e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. m. yan , phys .
d * 21 * , 203 ( 1980 ) .
tornqvist , ann . phys . * 123 * , 1 ( 1979 ) ; acta . phys . polon .
* b16 * , 503 ( 1985 ) .
s. ono and n. a. tornqvist , z. phys .
c * 23 * , 59 ( 1984 ) .
k. heikkila , s. ono and n. a. tornqvist , phys .
d * 29 * , 110 ( 1984 ) [ erratum - ibid .
d * 29 * , 2136 ( 1984 ) ] .
n. a. tornqvist and p. enczykowski , phys .
d * 29 * , 2139 ( 1984 ) .
p. enczykowski , annals phys .
* 169 * , 453 ( 1986 ) .
p. geiger and n. isgur , phys .
d * 41 * , 1595 ( 1990 ) .
p. geiger and n. isgur , phys .
d * 44 * , 799 ( 1991 ) .
p. geiger and n. isgur , phys .
d * 47 * , 5050 ( 1993 ) .
d. morel and s. capstick , `` baryon meson loop effects on the spectrum of non strange baryons , '' arxiv : nucl - th/0204014 . e. van beveren and g. rupp , arxiv : hep - ph/0312078 .
e. van beveren and g. rupp , mod .
a * 19 * , 1949 ( 2004 ) [ arxiv : hep - ph/0406242 ] .
e. van beveren and g. rupp , phys .
lett . * 93 * , 202001 ( 2004 ) [ arxiv : hep - ph/0407281 ] . c. amsler and n. a. tornqvist , phys .
rept . *
389 * , 61 ( 2004 ) .
s. kalashnikova , phys .
d * 72 * , 034010 ( 2005 ) [ arxiv : hep - ph/0506270 ] .
c. hanhart , yu .
s. kalashnikova , a. e. kudryavtsev and a. v. nefediev , arxiv:0704.0605 [ hep - ph ] .
s. kalashnikova , aip conf .
proc . * 892 * , 318 ( 2007 ) .
e. j. eichten , k. lane and c. quigg , phys .
d * 69 * , 094019 ( 2004 ) [ arxiv : hep - ph/0401210 ] .
t. barnes and e. s. swanson , phys .
c * 77 * , 055206 ( 2008 ) [ arxiv:0711.2080 [ hep - ph ] ] .
m. r. pennington and d. j. wilson , arxiv:0704.3384 [ hep - ph ] .
g. rupp , e. van beveren and s. coito , acta phys .
* 8 * , no . 1 , 139 ( 2015 ) [ arxiv:1502.05250 [ hep - ph ] ] .
e. s. swanson , j. phys .
g * 31 * , 845 ( 2005 ) [ hep - ph/0504097 ] .
w. heupel , g. eichmann and c. s. fischer , phys .
b * 718 * ( 2013 ) 545 [ arxiv:1206.5129 [ hep - ph ] ] .
g. eichmann , c. s. fischer and w. heupel , arxiv:1508.07178 [ hep - ph ] .
j. meyers and e. s. swanson , phys . rev .
d * 87 * ( 2013 ) 3 , 036009 [ arxiv:1211.4648 [ hep - ph ] ] .
t. hilger , m. gomez - rocha and a. krassnigg , phys .
d * 91 * ( 2015 ) 11 , 114004 [ arxiv:1503.08697 [ hep - ph ] ] .
t. d. cohen and r. f. lebed , phys .
d * 90 * , no . 1 , 016001 ( 2014 ) [ arxiv:1403.8090 [ hep - ph ] ] .
t. d. cohen and r. f. lebed , phys .
d * 89 * , no . 5 , 054018 ( 2014 ) [ arxiv:1401.1815 [ hep - ph ] ] . m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 110 * , 252001 ( 2013 ) [ arxiv:1303.5949 [ hep - ex ] ] . z. q. liu _ et al . _ [ belle collaboration ] , phys .
lett . * 110 * , 252002 ( 2013 ) [ arxiv:1304.0121 [ hep - ex ] ] .
|
the last few years have been witness to a proliferation of new results concerning heavy exotic hadrons .
experimentally , many new signals have been discovered that could be pointing towards the existence of tetraquarks , pentaquarks , and other exotic configurations of quarks and gluons .
theoretically , advances in lattice field theory techniques place us at the cusp of understanding complex coupled - channel phenomena , modelling grows more sophisticated , and effective field theories are being applied to an ever greater range of situations .
it is thus an opportune time to evaluate the status of the field . in the following
, a series of high priority experimental and theoretical issues concerning heavy exotic hadrons is presented .
|
it is now well - established that the kinetic feedback from massive stars creates shell structures in the interstellar medium ( ism ) .
indeed , studies of these objects in the magellanic clouds provide the best empirical understanding of this phenomenon .
these types of shells can be roughly classified into three categories ( chu 1996 ; meaburn 1980 ) , which are summarized in table 1 : `` bubbles '' and supernova remnants ( snrs ) , which result from stellar winds and supernovae ( sne ) of individual massive stars ; `` superbubbles , '' which result from the action of a few , to hundreds of , ob star winds and sne clustered in an ob association ; and `` supergiant shells , '' which have sizes of order @xmath0 kpc or more .
if the supergiant shells are analogously created by massive star feedback , they must be associated with starburst phenomena having total wind and sn energies of @xmath1 erg .
this review will focus on recent ( 1990 s ) studies of superbubbles in the magellanic clouds , with also a brief look at supergiant shells .
.ob star shells . [ cols="<,<,^,^ " , ] how do these shells evolve ?
depending on the dominant physics , there are three self - similar relations that are usually used to describe the shell evolution . in the case of a single point energy deposition ,
as would apply to a sn , we have the traditional sedov ( 1959 ) blastwave : @xmath2 where @xmath3 is the shell radius , @xmath4 is the input mechanical energy , @xmath5 is a uniform ambient density , and @xmath6 is elapsed time . for a constant mechanical power @xmath7 over an extended period , there is a standard , adiabatic evolution ( pikelner 1968 ) given by , @xmath8 this model has a double - shock structure , in which the inner shock , close to the energy source , heats the wind ejecta to temperatures of order @xmath9 k , driving the shell growth by the thermal pressure behind the outer shock at the shell .
if the hot interior is able to radiate away its energy , the inner shock will collapse upon the outer one , and the shell may assume a momentum - conserving model ( steigman 1975 ) : @xmath10 where @xmath11 is the wind terminal velocity .
this model describes a growth rate intermediate between those of equations [ sedov ] and [ ad ] .
there are other variations of these models ; for example , an extensive review of astrophysical blastwaves is given by ostriker & mckee ( 1988 ) .
assuming coeval star formation in ob associations , the growth of superbubbles quickly becomes dominated by sn activity , although most of those found in are extremely young objects that are still wind - dominated , @xmath12 myr old .
most empirical , recent studies of superbubbles attempt to test the applicability of the standard , adiabatic model .
the lmc in particular , offers a fine selection of objects , which have been used to evaluate the shell dynamics and search for enclosed hot gas .
recent dynamical studies ( oey & massey 1995 ; oey 1996 ; mac low _ et al . _ 1998 ; oey & smedley 1998 ) of individual lmc superbubbles have found a class of objects whose observed @xmath13 is far too high for self - similar growth , in view of the stellar population present ( @xmath14 is the shell expansion velocity ) .
based on excess observed x - ray luminosities , chu & mac low ( 1990 ) and wang & helfand ( 1991a ) suggest that most of these objects suffered recent snr impacts on the shell walls .
complex velocity structures in several objects ( rosado 1990 ; ambrocio - cruz 1997 ) are consistent with this interpretation .
however , oey & smedley ( 1998 ) and mac low ( 1998 ) demonstrate that blowout from a high - density into a low - density region can perfectly mimic the shell kinematics .
although many optical superbubbles show this high @xmath13 evidence of disruption , it is important to note that other objects have been found whose dynamics are fully consistent with self - similar growth ( oey 1996 ) .
however , all of the objects , along with the high @xmath13 objects , appear to indicate a growth rate discrepancy from the standard model , where the assumed @xmath15 is overestimated by up to an order of magnitude .
it is unclear whether this is caused by an overestimate in stellar wind power , or underestimate in ambient @xmath5 , or both .
the search for hot gas within the superbubbles is an important test of the standard model .
several studies ( chu & mac low 1990 ; wang & helfand 1991a ; magnier 1996 ; mac low 1998 ) have found objects with x - ray emission in excess of predictions .
as mentioned above , these may be caused by snr impacts .
reassuringly , there are other superbubbles that do not show x - ray enhancements ( chu 1995 ) , and detection limits for these remain consistent with the predicted x - ray emission .
furthermore , in all superbubbles that have been examined for c iv and si iv absorption , these high - ionization species have been found ( chu 1994 ) , and can be attributed to an interface layer between the shell wall and hot interior .
supergiant shells ( sgs s ) are more poorly understood than superbubbles .
it is possible that many of these do not originate from star formation events at all .
alternative mechanisms include impacts by high - velocity clouds ( tenorio - tagle 1987 ) and gamma - ray bursters ( efremov 1998 ; loeb & perna 1998 ) . at any rate , the evolution of sgs s is likely to differ substantially from that of superbubbles because the size scales are similar to those of galactic parameters like scale height and length .
the sgs s in the lmc have been catalogued by meaburn ( 1980 ) , and only two of these have been examined in any detail .
the most well - known example is lmc-4 , associated with the constellation iii stellar region .
this region is a remarkably well - defined shell , @xmath16 pc in radius , seen in both and as a ring of optical regions .
there have been numerous studies of lmc-4 over the last two decades ; the most recent work is by domgoergen ( 1995 ) and bomans ( 1996 ) studying its gaseous properties , and by olsen ( 1997 ) and braun ( 1997 ) on the stellar content .
the other lmc sgs that has been examined is lmc-2 , just east of 30 dor .
its gas properties were studied by wang & helfand ( 1991b ) and caulet & newell ( 1996 ) .
superbubbles , as a direct manifestation of massive star feedback , are of critical importance to the global properties of the ism . these include ism structure and kinematics , relationship between the different gas phases , and interstellar processes like star formation . to gain insight on the ism structure and kinematics caused by superbubbles , it is possible to use equation [ ad ] to derive a size and velocity distribution for superbubbles in the ism .
this also serves as an additional test of the standard , adiabatic evolution over the full lifetimes of the superbubbles .
oey & clarke ( 1997 ) computed the size distribution @xmath17 for the smc , using the region luminosity function and assuming a constant creation rate for ob associations .
the observed slope @xmath18 for superbubbles ( staveley - smith 1997 ) is in remarkable agreement with the predicted slope of @xmath19 , suggesting that superbubble activity can fully explain the structure of the smc .
this can be compared to @xmath20 for holes in a fractal ism structure ( elmegreen 1997 ) , which is being quantified for the smc by stanimirovi ( 1998 ) .
they find a volume fractal dimension @xmath21 , translating directly to @xmath22 .
it is therefore too difficult to distinguish between a superbubble and fractal ism model , using only @xmath23 .
indeed , the similarity in observed @xmath23 for both cases in the smc , might be indicative of a physical relation between the two structural models .
it is also possible to test for superbubble structure in the ism by examining the morphology of the holes , of which @xmath24 were catalogued as shells , based on elliptical shape ( staveley - smith 1997 ) .
in addition , the velocity distribution @xmath25 computed by oey & clarke ( 1998 ) is also consistent with the observed shell velocities for the smc .
a superbubble velocity structure could be a dominant input to the kinematics and turbulence of the ism .
as superbubbles expand and age , their surface brightness quickly diminishes .
hence , they could be an important component of the diffuse , warm , ionized medium ( wim ) . hunter ( 1994 )
obtained emission - line spectra of superbubbles in the lmc , which do show that some of the line ratios are intermediate between those for classical regions and wim detected in other galaxies .
kennicutt ( 1995 ) examined the morphological structure of the wim in the magellanic clouds , and also found a substantial contribution from faint superbubbles and sgs s .
finally , it is widely thought that superbubbles can trigger renewed star formation in their shells of swept - up cool gas .
several candidates have been examined in the lmc .
the clearest example is dem 34 / n11 ( walborn & parker 1992 ; rosado 1996 ) , and additional candidates are dem 152 / n44 ( oey & massey 1995 ) and dem 192 / n51 d ( oey & smedley 1998 ) .
as mentioned above , the sgs lmc-4 has a long history of investigation as a triggered star formation candidate ( efremov 1998 ) .
ambrocio - cruz , p. , laval , a. , marcelin , m. , & amram , p. 1997
, , 319 , 973 bomans , d. j. , de boer , k. s. , koornneef , j. , & grebel , e. k. 1996 , , 313 , 101 braun , j. m. , bomans , d. j. , will , j .-
, & de boer , k. s. 1997 , , 328 , 167 caulet , a. & newell , r. 1996 , , 465 , 205 chu , y .- h .
1996 , in _ the interplay between massive star formation , the ism , and galaxy evolution , _ eds .
d. kunth , b. guiderdoni , m. heydari - malayeri , & t. x. thuan , gif - sur - yvette : editions frontires , 201 chu , y .- h . &
mac low , m .-
1990 , , 365 , 510 chu , y .- h . , chang , h .- w .
, su , y .-
, & mac low , m .- m .
1995 , , 450 , 157 chu , y .- h . , wakker , b. , mac low , m .- m . , garca - segura , g. 1994 , , 108 , 1696 domgoergen , h. , bomans , d. j. , & de boer , k. s. 1995 , , 296 , 523 efremov , y. n. 1998 , this volume efremov , y. n. , elmegreen , b. g. , & hodge , p. w. 1998 , , 501 , l163 elmegreen , b. g. 1997 , , 477 , 196 hunter , d. a. 1994 , , 107 , 565 kennicutt , r. c. , bresolin , f. , bomans , d. j. , bothun , g. d. , & thompson , i. b. 1995 , , 109 , 594 loeb , a. & perna , r. 1998 , , 503 , l35 mac low , m .-
chang , t. h. , chu , y .- h . , points , s. d. , smith , r. c. , & wakker , b. p. 1998
, , 493 , 260 magnier , e. a. , chu , y .- h . , points , s. d. , hwang , u. , & smith , r. c. 1996 , , 464 , 829 meaburn , j. 1980 , , 192 , 365 oey , m. s. 1996 , , 467 , 666 oey , m. s. & clarke , c. j. 1997 , , 289 , 570 oey , m. s. & clarke , c. j. 1998 , in _ interstellar turbulence , _ eds . j. franco & a. carraminana , cambridge : cambridge univ .
press , in press oey , m. s. & massey , p. 1995
, , 452 , 210 oey , m. s. & smedley , s. a. 1998 , , in press olsen , k. a. g. , hodge , p. w. , wilcots , e. m. , & pastwick , l. 1997 , , 475 , 545 ostriker , j. p. & mckee , c. f. 1988 , rev .
phys , 60 , 1 pikelner , s. b. 1968 , astrophys .
lett . , 2 , 97 rosado , m. 1990 , , 238 , 315 rosado , m. 1996 , , 308 , 588 sedov , l. i. 1959 , _ similarity and dimensional methods in mechanics _
, new york : academic stanimirovi , s. , staveley - smith , l. , dickey , j. m. , sault , r. j. , & snowden , s. l. 1998 , this volume staveley - smith , l. , sault , r. j. , hatzidimitriou , d. , kesteven , m. , & mcconnell , d. 1997 , , 289 , 225 steigman , g. , strittmatter , p. a. , & williams , r. e. 1975 , , 198 , 575 tenorio - tagle , g. , franco , j. , bodenheimer , p. , & ryczka , m. 1987 , , 179 , 219 walborn , n. r. & parker , j. w. 1992 , , 399 , l87 wang , q. & helfand , d. j. 1991a , , 373 , 497 wang , q. & helfand , d. j. 1991b , , 379 , 327
|
superbubbles that result from the stellar winds and supernovae of ob associations probably play a fundamental role in the structure and energetics of the ism in star - forming galaxies .
their influence may also dominate the relationship between the different interstellar gas phases .
how do superbubbles form and evolve ? how do they affect the local and global ism ?
the magellanic clouds provide a superior opportunity to study this shell - forming activity , since both stellar content and gaseous structure can be examined in detail . here , the results of recent studies of superbubbles in the magellanic clouds are reviewed .
# 1to 0pt#1
|
superconducting correlations in graphene - based structures have been the focus of intensive theoretical and experimental research even before graphene became one of the most important topics in condensed matter physics .
following last decade experimental evidences reporting hints of superconductivity behavior in graphite @xcite and graphite intercalated compounds @xcite , a considerable amount of theoretical studies have been devoted to provide a clear understanding about possible mechanisms that could induce intrinsic superconducting states in single and multilayer graphene @xcite .
more recently experimental investigations have reported intriguing traces of high - temperature superconducting behavior in highly oriented pyrolytic graphite ( hopg ) samples @xcite , feeding speculations about the existence of intrinsic superconducting correlations in graphite and graphite - based compounds . despite the fact that most of these experimental evidence suggests that superconductivity in graphite compounds appears due to external causes , several theoretical studies reveal the possibility of inducing superconductivity in graphite by considering unconventional symmetry of the order parameter @xcite .
however , these calculations show that superconductivity becomes stable after considering disorder @xcite or high doping in the graphene layers @xcite while surface superconductivity appears to be stable in clean rhombohedral graphite in the absence of external doping @xcite .
considering that most of these calculations are based on a reduced hamiltonian or are performed within two - dimensional models , by ignoring the interplanar hopping , a numerical description of the superconducting correlation in multilayer graphene is urgently needed .
in view of this , we provide in the following a numerical description of intrinsic superconductivity in multilayer graphene at the tight - binding level . following ref . , we consider a simple s - wave pairing symmetry in a abc ( or rhombohedral ) stacking multilayered graphene structure .
calculations are also performed for aba ( or bernal ) stacking where its quadratic low - energy band structure @xcite shows a remarkable difference from the @xmath0 momentum dependent band structure seen in abc multilayer @xcite structures .
in particular we are interested in the limit of large number of layers ( @xmath1 ) where the lower - energy band in the rhombohedral case is flat over a large region in k - space signaling the suppression of the kinetic energy and therefore resulting in strong effect from interactions .
in addition , because of the sensitive stacking dependent band structure in multilayer graphene , we also considered hybrid stacking cases .
it is known that exfoliated few - layer graphite samples are usually found to exhibit very stable bernal stacking but often also display rhombohedral structures in part of the sample @xcite .
it is worth mentioning that during the preparation of the manuscript , new experimental results revealed the existence of superconducting correlations at two - dimensional interfaces that appear when angle misalignments about the _
c_-axis exist in hopg @xcite . by using highly efficient graphics card ( gpu ) simulations of the tight - binding bogoliubov - de - gennes equations , we are able to solve self - consistently for the pair correlation in multilayer graphene by considering both planar and inter - planar coupling between nearest neighbors .
translational invariance is assumed along the 2-dimensional direction within the graphene sheets . in this way
, an adequate description for the profile of superconducting correlations along the direction perpendicular to the graphene layers is achieved .
our results confirm the main features of recent analytical approaches for abc rhombohedral graphite where an enhancement of surface superconductivity , with respect to its bulk analog , was recently predicted @xcite .
we find that the opposite behavior is present in aba stacking where bulk dominates over surface superconductivity .
this fact requires that a strong pairing potential is needed in order to obtain a non - zero pair correlation for the aba case .
in contrast , in the abc case , a lower pairing potential is sufficient to induce a large pair amplitude at the surface .
in addition , we show how doping influences surface superconductivity , i.e. it strongly suppresses it .
the paper is organized as follows . in sec .
[ model ] we introduce the model and the numerical approach that we use . in sec .
[ results ] we present and discuss the results of our numerical calculations .
finally , we briefly summarize our findings in sec .
[ conclusions ] .
schematic layout of the side view of a multilayer graphene structure stacked in two different configurations : rhombohedral or abc ( left ) and bernal or aba ( right ) .
sublattices a and b are coupled by @xmath2 within the same layer , while interlayer coupling is described by @xmath3 .
integer coordinate @xmath4 correspond to the index layer . ]
superconducting correlations in multilayered graphene structures are described by the following mean - field single particle hamiltonian written in nambu space : @xmath5 where the summation , @xmath6 , is done over nearest neighbors within each layer while the summation , @xmath7 , is done for adjacent layers .
the non - diagonal elements @xmath8 correspond to the s - wave superconducting order parameter at the atomic site @xmath9 in layer @xmath10 while the diagonal elements @xmath11 are the normal state components of the hamiltonian : @xmath12\delta_{l , m}-t(\delta_{l , m+1}+\delta_{l , m-1 } ) \;\;\;\end{aligned}\ ] ] where @xmath13 is the chemical potential and , according to the layout of fig .
[ layout ] , nearest - neighbors sublattices a and b are coupled within the layers by the hopping parameter @xmath142.8ev while @xmath15 describes the hopping which couples a sites with the nearest b sites in the adjacent ( upper and lower ) layers .
bernal and rhombohedral stacking are defined according to the vertical symmetry along the z - axis as shown in fig .
[ layout ] . due to this symmetry ,
rhombohedral stacking allows us to reduce the description of the superconducting parameter to one of the sublattices , whereas the other sublattice can be deduced from a mirror reflection transformation . for bernal
stacking , we consider a more practical way of sorting sites inside the supercell as dimmer sites , which correspond to the sublattices coupled by the interlayer hopping @xmath3 , and no - dimmer sites or sublattices which does not participate in the coupling between adjacent graphene sheets . as we previously pointed out in a recent work @xcite , the pair correlation behaves differently in these inequivalent sites because at dimmer sites the density of states vanishes around the dirac point while being finite at no - dimmer sites where the formation of andreev states is more feasible .
we will not specify the origin of superconductivity in the multilayer graphene structure , but rather assume @xmath16 to be a conventional s - wave symmetric order parameter and the pairing potential @xmath17 is fixed and homogeneous in the whole structure . under this assumption and
considering translational invariance along the transversal directions , we solve self - consistently for the amplitude of the pair correlation @xmath18 for the sublattice a ( or both in the case of n - bernal stacked layers for n odd ) , in the @xmath4 direction .
we have considered graphene multi - layer supercells of size 42nm@xmath1925nm@xmath20 such that the order parameter is converged and no additional momentum summations in the parallel direction are needed . instead of a direct diagonalization of the hamiltonian we performed the self - consistent mean - field calculation through a numerical approximation of the gorkov green s function by using the chebyshev - bogoliubov - de - gennes method @xcite .
both , the normal and anomalous gorkov green s function , can be approximated by a superposition of a finite number of chebyshev polynomials as follows : @xmath21,\end{aligned}\ ] ] where the expansion coefficients for the diagonal , or normal ( @xmath22 ) , and the off - diagonal , or anomalous ( @xmath23 ) , components of the 2@xmath192 green function are defined respectively as @xcite : @xmath24 where @xmath25 $ ] is the chebyshev polynomial of order @xmath26 , which satisfies the following recurrence relation : @xmath27 .
once the hamiltonian has been normalized according to @xmath28 , where the rescaling factors are @xmath29 and @xmath30 , with @xmath31 being a small number , the expansion coefficients can be obtained through the dot product @xmath32 , where @xmath33 are the vectors @xmath34 and @xmath35 and the recursive vector @xmath36 .
we refer the reader to ref . for more details about the numerical procedure .
since this iterative procedure involves mainly a successive application of the hamiltonian matrix ( [ ham ] ) on iterative vectors @xmath37 , most of the computational effort corresponds to sparse matrix - vector multiplication which can be performed with high - efficiency by implementing parallel computations on gpus by using cuda nvidia .
we are therefore able to solve efficiently multilayered graphene structures containing typically several hundreds of thousand of atoms , which is half of the size of the hamiltonian matrix ( [ ham ] ) .
physical quantities , like the local density of states and the pair correlation function , can be easily determined once the green s functions are known : @xmath38 and @xmath39d\omega$ ] .
( color online ) ( a ) order parameter profile ( @xmath40 ) along the z - direction perpendicular to the graphene sheets ( see fig . [ layout ] ) for various values of the s - wave attractive pairing potential @xmath17 in abc stacked multilayer graphene with @xmath41 layers .
the inset shows the corresponding @xmath40 profile for the aba case where dimmer and non - dimmer sites follow different curves for higher values of @xmath17 .
@xmath17-dependence of the maximum ( b ) and minimum ( c ) value of @xmath40 for both abc and aba stacking configuration .
@xmath42 nm is the inter - layer distance . ]
we solved self - consistently for the order parameter along the z - direction , and show in fig .
[ upro](a ) the different profiles of the order parameter ( @xmath40 ) for different values of the pairing potential ( @xmath43 ) in abc stacked graphene with @xmath1=20 layers .
the order parameter is only shown for the a sublattice , while the b sublattice value is achieved by a mirror reflexion symmetry along z , as can be deduced from fig .
[ layout ] .
analog results are shown in the inset of fig .
[ upro](a ) for aba stacking and different values of the pairing potential .
we notice in fig .
[ upro](a ) , that the superconducting order parameter at the outermost layers is larger than the vanishing pair correlation in the bulk for all the u - values considered here .
the same surface enhancement is observed when we decrease the pairing potential such that the penetration of the superconducting order parameter into the bulk becomes strongly suppressed . in the limit of very low pairing potential good agreement could be found with the analytical result previously reported in ref . .
on the other hand , an opposite surface - bulk superconducting ratio is found for the bernal stacking configuration ( see inset in fig [ upro](a ) ) .
self - consistent calculations performed for aba show that the bulk value of the order parameter is dominant while surface superconducting correlations are suppressed . also , we can observed a sublattice polarization in the @xmath40 profile where pair correlation is found to be higher in non - dimmer sites as compared to dimmer sites with an energy difference which becomes smaller as the pairing potential is decreased .
we return to this issue in a later discussion about the density of states .
a direct comparison between the surface and the bulk value of the pair correlation is given in figs .
[ upro](b ) and [ upro](c ) for both stacking configurations .
[ upro](b ) shows the maximum values of the superconducting correlation for both abc and aba cases for different values of @xmath17 . according to the profiles observed in fig .
[ upro](a ) , the maximum value of the order parameter , max@xmath44 , corresponds to the surface for abc stacking while for the aba stacking the maximum value corresponds to the bulk non - dimmer sites .
in contrast , the log - linear representation presented in fig . [ upro](c )
shows the u - dependence of the minimum value of the superconducting order parameter , min@xmath44 , which corresponds to the bulk and surface locations for abc and aba , respectively .
two different regimes can be inferred from figs .
[ upro](b ) and [ upro](c ) , depending whether the pairing potential u is larger or smaller than a critical value , @xmath45 .
this value is also very close to the critical pairing for the aba stacking , below which superconducting correlations vanish . as @xmath46 decreases , but is still larger than @xmath47 , the order parameter decays exponentially for both abc and aba stacking , as the main contribution comes from the bulk . when @xmath48 , for abc stacking , the bulk order parameter vanishes exponentially but the surface one is still finite . even more interesting is the fact that the surface order parameter decays only linearly and is non - zero for all the values of @xmath46 that we considered in the simulation , down to @xmath49 giving a surface order parameter @xmath50 .
the self - consistent calculation becomes increasingly intensive as the order parameter decreases since more chebyshev moments are needed to resolve the green s function at higher and higher resolutions .
the bulk behavior resembles the superconducting critical point reported for graphene , where in the undoped case it was found that the critical temperature vanishes below a finite value of the s - wave pairing interaction @xcite .
in contrast , the @xmath46 dependence of the surface order parameter in the abc stacking configuration suggests that the surface states , which form a flat band with suppressed kinetic energy , are strongly influenced by any exponentially small interaction .
+ ( color online ) ( a ) order parameter profile ( @xmath40 ) for a rhombohedral multilayer graphene consisting of various total number of layers @xmath1=10 , 12 , 16 and 20 .
( b ) order parameter profile for n=10 and 20 for different values of the interlayer hopping @xmath51=0.1 , 0.092 , 0.084 , 0.076 and 0.068 , in decreasing order as this is indicated by the arrows .
( c ) dependence of the surface pair correlation , max@xmath44 , as a function of @xmath3 .
here we used @xmath52 . ] since the pair correlation at the surface is always enhanced for abc stacking and survives even for lower values of the pairing potential , we will further only also report numerical results for this stacking configuration .
the dependence of the order parameter on the total number of graphene layers , @xmath1 , and the interlayer coupling , @xmath3 , is shown in fig .
[ npro ] for @xmath17=-1.76@xmath2 . an asymptotic enhancement of surface superconductivity is observed in fig .
[ npro](a ) as the fermi surface size , defined by the flat band localized at the surfaces , increases with the number of layers . on the other hand , the decrease of
the interlayer coupling leads to the suppression of the order parameter as this is seen in fig .
[ npro](b ) .
the evolution of the surface pair correlation as a function of @xmath3 is shown in fig .
[ npro](c ) and indicates an almost linear suppression as a consequence of the linear dependence of the fermi surface size on @xmath3 .
@xcite in order to provide a better understanding of the peculiar behavior of the order parameter profile observed in fig .
[ upro ] for the rhombohedral case , we next present the local density of states ( ldos ) .
[ ldos ] shows the ldos in different layers for both sublattices , a and b , within a small energy interval near the fermi energy .
the left panel of fig .
[ ldos ] represents the ldos at the surface where the superconducting order parameter is enhanced for sublattice a. there , the normal state ldos shows a sharp peak due to the existence of flat bands with dispersion @xmath53 .
the corresponding wave function of these states is localized at the surface and only on sublattice a. a zoomed - out view of the ldos , over a wider range of energies , is shown in the inset of fig .
this is very different for the b sublattice where the density of states vanishes around the fermi energy and the superconducting coherence peaks are not visible .
despite this , a non - zero solution for the order parameter is obtained for atomic sites belonging to this sublattice as we can see in figs .
[ upro](a ) and [ npro ] .
this non - zero solution appears as a consequence of the proximity effect between the intra - layer neighbors a and b sites at the surface .
we also observe less pronounced coherence peaks appearing in the ldos of sublattice a in the layer adjacent to the surface ( see central panel of fig . [ ldos ] ) while the ldos vanishes for both sublattices in the bulk ( see right panel of fig .
[ ldos ] ) . according to this behavior of the ldos for rhombohedral
stacking , we expect that superconducting correlation will be more stable on the surface and on few adjacent layers rather than in the bulk where the ldos vanishes around the fermi energy .
+ ( color online ) local density of states ( ldos ) showing the formation of the s - wave superconducting gap for @xmath54 at both sublattices a and b in different layers around the fermi energy . left and center panel show the ldos at the surface and its adjacent layer while right panels show the ldos at the bulk .
the dashed line represent the corresponding normal state ldos which shows the localized flat band at the outermost layers .
the inset of the central panel shows the surface ldos in abc stacked graphene over a wider range of energies . ] with respect to the ldos in the bernal case , it is well known that a sublattice polarization appears around the fermi energy .
a finite density of states is found for non - dimmer sites while the density of states vanishes at the dimmer sites @xcite .
such a polarization allows a finite order parameter to be induced only by large s - wave pairing potentials and therefore bulk superconductivity will not be stable for values of u for which surface superconductivity in abc is still finite ( see fig . [ upro](b ) ) . in this way , the suppression of surface superconductivity seen in fig .
[ upro](c ) for aba appears as a consequence of the lower density of states for surface non - dimmer sites when compared to its bulk value .
( color online ) effect of homogeneous doping on the intrinsic s - wave superconducting order parameter in a multilayer rhombohedral graphene with @xmath54 .
we shown here half of the profile for different doping values , @xmath55=0 , 0.8 , 1.6 , 2.4 and 3.1 , in decreasing order indicates by the arrow .
inset : evolution of surface superconductivity ( max@xmath44 ) as a function of doping ( @xmath56 ) showing that an increase in @xmath56 leads to a strong suppression of the order parameter at the surface . ]
+ ( color online ) ldos at the surface , on the a sublattice , for @xmath54 and three different dopings considered before in fig .
[ dopef ] . the normal state represented by the dashed line
has been included for the corresponding values of doping @xmath56 .
@xmath57 corresponds to max@xmath58 . ] in order to see the effect of external factors we have considered homogeneous doping , as well as an inhomogeneous field - induced charge distribution , along the z - direction .
the first case corresponds to graphite - intercalated compounds where dopant atoms are placed between the graphene layers while the inhomogeneous case can be easily realized in an experimental set - up where top and bottom electrodes have opposite gate voltages @xcite .
[ dopef ] shows the evolution of the order parameter profile as a function of the homogeneous doping @xmath59 for a fixed value of the pairing potential @xmath54 .
notice that , surface superconductivity becomes strongly suppressed as the doping shifts the dirac point away from the fermi energy .
total suppression occurs for doping lower than the value of the order parameter at the surface .
looking at the ldos we found that this critical doping coincides with the extinction of the coherence peaks and the rising of the single peak away form the fermi energy . despite the fact that we do not find any relevant difference between homogeneous doping and surface only doping in our self - consistent calculation ,
the value of the critical doping is slightly higher than the one reported by the analytical results @xcite where the critical doping was found to be @xmath60=@xmath61 .
in addition to the homogeneous case we have also considered inhomogeneous doping , achieved when an electric field is applied perpendicular to the graphene sheets . following ref . where the potential distribution due to an electric field was self - consistently calculated by taking into account screening effects
, we consider the z - dependent doping as described in the inset of fig .
[ dopel ] . for comparison purposes
we have considered the same pairing potential @xmath54 , as we did in fig .
[ dopef ] , and three different cases for which the doping at the surfaces is strongly suppressing the superconducting correlations in the homogeneous case .
surprisingly , in contrast to homogeneous doping , we found that the field - induced doping suppresses only slightly the pair correlation at the surface .
according to this result , the electric - field induced gap in the inhomogeneous case is much lower than the value of the doping at the surface @xcite .
therefore , even considering the same surface doping in both cases , superconducting correlations are weakly affected by an inhomogeneous doping configuration . fig .
[ ldoel ] shows how coherence peaks still survive for this doping configuration even if the value of the doping at the surface is on the order of the surface order parameter ( as obtained for zero - doping ) .
this is in contrast with the effect observed in fig .
[ ldop ] where for a similar level of doping causes a strong suppression of the superconducting gap .
+ ( color online ) effect of inhomogeneous field - induced doping along the z - direction on the order parameter in a rhombohedral multilayer graphene .
some cases previously shown in fig .
[ dopef ] are included for comparison .
inset : doping profile along the z - direction perpendicular to the graphene sheets . ]
( color online ) ldos showing the effect of field - induced doping along the z - direction ( see inset in fig .
[ dopel ] ) on surface superconductivity , obtained for @xmath54 in abc multilayer graphene .
dashed curves represent the corresponding normal state ldos . ] finally , we consider multilayer graphene with hybrid stacking . in order to shift
the high - temperature surface superconductivity , observed in abc , to the bulk of the structure , we propose intercalated hybrid stacked graphite with few layers .
numerical calculations considering stacking faults in multilayer graphene were recently reported @xcite , showing that surface superconductivity survives at the interface between abc and aba stacking .
however , superconductivity in the bulk is still seen being suppressed when the thickness of the hybrid layers is large . by considering a few - layer structure where the external layers have the abc stacking configuration while the inner ones have aba stacking , we expected a slight suppression of surface superconductivity and an enhancement of its bulk value .
[ hyb ] shows self - consistent results obtained for the hybrid structure depicted in the top part of the figure . while surface superconductivity appears suppressed as compared to the non - hybrid abc multilayer , bulk correlation are also enhanced .
on the basis of these results we suggest that this kind of hybrid stacked multilayer structure , or more complex combinations , could support high temperature superconductivity due to the interplay between surface superconductivity present in abc stacking and the bulk superconductivity preserved in the aba case .
( color online ) self - consistent order parameter profile for a multilayer graphene considering the stacking configuration shown in the top part .
different point types correspond to different values of the pairing potential , @xmath62 2.16 , 2.12 and 2.08 , where @xmath17 decrease in the direction pointed by the arrow .
open points represent the corresponding cases for an ordered abc stacked multilayer graphene . ]
by using a highly efficient gpu - based numerical procedure we solved self - consistently for the s - wave order parameter within a mean - field approach for a tight - binding model of aba and abc - stacked multilayer graphene .
main findings show that a surface superconducting state appear when the multilayer is in the abc stacking configuration .
opposite behavior is seen for the aba stacking where bulk superconductivity is predominant but unstable below a certain critical pairing potential .
the ldos for surface sites shows peculiar coherence peaks and sublattice polarization , i.e. large ldos in one sublattice and zero in the other .
we showed that under homogeneous doping this state is quickly suppressed .
we extracted a critical doping which is slightly higher than the one reported previously based on an approximate analytical study@xcite .
in addition , we considered a field - induced inhomogeneous doping and showed that the superconducting correlations survive in this case for higher values of @xmath63 .
finally we pointed out the importance of hybrid stacking structures where surface superconductivity related to abc stacking could be preserved even in the bulk of the structure , suggesting a possible path for the survival of high temperature superconductivity .
this work was supported by the flemish science foundation ( fwo - vl ) and the methusalem funding of the flemish government .
28ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase 10.1023/a:1004637814008 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.147001 [ * * , ( ) ] link:\doibase 10.1016/j.physc.2004.02.039 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.134519 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.107008 [ * * , ( ) ] link:\doibase 10.1103/physrevb.63.134421 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.146801 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.134512 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.035429 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.085431 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.220503 [ * * , ( ) ] link:\doibase 10.1038/nphys2181 [ * * , ( ) ] link:\doibase 10.1038/nphys2208 [ * * , ( ) ] link:\doibase 10.1088/0953 - 2048/24/9/092001 [ * * , ( ) ] link:\doibase 10.1002/adma.201202219 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/15/2/023024 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.075404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.245426 [ * * , ( ) ] link:\doibase 10.1021/nl1032827 [ * * , ( ) ] link:\doibase 10.1021/nl301932v [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.184505 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.167006 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.241401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.246808 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.195401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.115432 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.125304 [ * * , ( ) ] http://arxiv.org/abs/1210.7075 [ ( ) ]
|
using highly efficient gpu - based simulations of the tight - binding bogoliubov - de gennes equations we solve self - consistently for the pair correlation in rhombohedral ( abc ) and bernal ( aba ) multilayer graphene by considering a finite intrinsic s - wave pairing potential .
we find that the two different stacking configurations have opposite bulk / surface behavior for the order parameter .
surface superconductivity is robust for abc stacked multilayer graphene even at very low pairing potentials for which the bulk order parameter vanishes , in agreement with a recent analytical approach . in contrast , for bernal stacked multilayer graphene , we find that the order parameter is always suppressed at the surface and that there exists a critical value for the pairing potential below which no superconducting order is achieved .
we considered different doping scenarios and find that homogeneous doping strongly suppresses surface superconductivity while non - homogeneous field - induced doping has a much weaker effect on the superconducting order parameter . for multilayer structures with hybrid stacking ( abc and aba )
we find that when the thickness of each region is small ( few layers ) , high - temperature surface superconductivity survives throughout the bulk due to the proximity effect between abc / aba interfaces where the order parameter is enhanced .
|
laser control of molecular rotation , alignment and orientation has received significant attention in recent years ( for a review , see e. g. @xcite ) .
interest in the field has increased , mainly due to the improved capabilities to manipulate characteristics of the laser pulses ( such as time duration and temporal shape ) , which in turn leads to potential applications offered by controlling the angular distribution of molecules . since the typical rotational motion is slow ( @xmath0 ) with respect to the typical short pulse ( @xmath1 ) , effective rotational control and manipulation are in reach . during the last decade ,
temporal rotational dynamics of pulse - excited molecules was studied @xcite , and multiple pulse sequences giving rise to the enhanced alignment were suggested @xcite and realized @xcite .
further manipulations such as optical molecular centrifuge and alignment - dependent strong field ionization of molecules were demonstrated @xcite .
selective rotational excitation in bimolecular mixtures was suggested and demonstrated in the mixtures of molecular isotopes @xcite and molecular spin isomers @xcite .
recently , several groups suggested a method for exciting field - free unidirectional molecular rotation , in which the angular momentum along the laser beam propagation direction is provided by two properly delayed ultrashort laser pulses that are linearly polarized at 45 degrees to each other @xcite .
the first experimental demonstration of the effect was done in @xcite for benzene molecules .
the mechanism behind this double - pulse scheme is rather clear , and is most easily explained in the case of linear molecules @xcite .
the first ultrashort laser pulse , linearly polarized along the @xmath2-axis , impulsively induces coherent molecular rotation that continues after the end of the pulse .
the molecules rotate under field - free conditions until they reach an aligned state , in which the molecular axis with the highest polarizability is confined in a narrow cone around the polarization direction of the first pulse .
the second short laser pulse is applied at the moment of the best alignment , and at an angle with respect to the first pulse . as a result ,
the aligned molecular ensemble experiences a torque causing molecular rotation in the plane defined by the two polarization vectors .
the torque acting on a linear molecule is maximal when the laser pulse is polarized at 45 degrees with respect to the molecular axis of the highest polarizability .
this defines the optimal angle between the laser pulses .
the direction of the excited rotation ( clockwise or counter - clockwise ) is determined by the sign of the relative angle ( @xmath3 degrees ) between the first and the second pulse in the polarization plane . this double pulse scheme
( see fig .
[ propeller ] ) was termed `` molecular propeller '' @xcite , as it resembles the action ( side kick ) needed to ignite a rotation of a plane propeller .
the heuristic arguments presented above are , in fact , essentially classical , although it is clear that a detailed quantum analysis may be needed to consider the long - time evolution of the excited molecules , or the operation of this scheme at relatively weak pulses when only low - lying rotational states are involved .
that is why all the previous treatments of the problem @xcite presented full - scale quantum - mechanical studies , combining both analytical and numerical methods . in the present paper ,
we provide a detailed comparison between the classical and the quantum - mechanical analysis of the double pulse scheme for exciting unidirectional molecular rotation , both for linear molecules and for benzene ( that was the object of the experiment in @xcite )
. we will demonstrate the predictive power of the classical treatment that correctly prescribes the optimal delay between the pulses , and the angle between them , and will find the advantages and limitations of such an approach in quantitative description of the effect at experimental conditions .
we believe that by comparing the complementary classical and fully quantum mechanical views on the same phenomenon , the reader will get a comprehensive idea about the underlying physics .
recent paper @xcite studied quantum - mechanically the `` molecular propeller '' scheme in the case of linear molecules .
dependence of the induced angular momentum on the pulses intensity , and the delay between them , was analyzed .
moreover , the angular distribution of linear molecules subject to unidirectional rotation was considered , and it was shown to be confined in the plane defined by the two polarization vectors of the pulses .
this anisotropic angular distribution was characterized by the observable @xmath4 , which was referred to as the azimuthal factor . in the present paper
, we describe the same scheme using a classical description of the double - pulse molecular excitation , followed by a monte carlo averaging over the thermal ensemble .
we compare the time dependence of the azimuthal factor according to the classical analysis , to the quantum mechanical one , and show that both results coincide for initial times .
in addition , we introduce a long - time - averaged angular distribution function , which conveniently depicts the molecular orientation following the excitation .
this distribution function seems to behave similarly in the classical and quantum cases .
finally , we analyze classically the excitation of the unidirectional rotation for more complicated oblate symmetric top molecules , like benzene , and compare this analysis with the full quantum - mechanical treatment of the problem .
the outline of the paper is as following . in sec .
ii we study classically the rotational motion of a linear molecule kicked by a pair of short nonresonant laser pulses that are linearly polarized in arbitrary directions . in sec .
iii , we describe the response of a thermal ensemble of linear molecules to such an excitation by using the monte carlo method . angular distribution function and its time - averaged behavior are studied in sec .
iv , and the results are compared to the full quantum mechanical treatment . in sec .
v , we analyze ( both classically and quantum mechanically ) the problem of inducing unidirectional rotation of benzene molecules . sec .
vi summarizes and concludes our study .
here we consider classical dynamics of a model diatomic ( or , more generally , linear ) polarizable molecule that interacts with an ultrashort laser pulse .
the molecule is treated as a linear rigid rotor .
we assume that the central frequency of the pulse is far detuned from any molecular transition .
the potential energy of the pulse interaction with the induced molecular polarization is given by @xcite : @xmath5 where @xmath6 is the difference between the polarizability along the molecular axis and the one perpendicular to it , @xmath7 is the envelope of the electric field of the _ linearly polarized _ laser pulse , and @xmath8 is the angle between the molecular axis and the direction of polarization of the pulse . here
@xmath9 are the polar and the azimuthal angles characterizing the direction of the molecular axis , respectively .
we assume the pulse duration is very short compared to the rotational period , so that the pulse can be described in the impulsive ( @xmath10-kick ) approximation .
we define the dimensionless interaction strength @xmath11 , which characterizes the pulse , as @xmath12 let us remind again the pulse sequence producing unidirectionally rotating molecules @xcite : ( 1 ) at time @xmath13 molecules are kicked by a pulse linearly polarized along the @xmath2-axis ; ( 2 ) at the time of the maximal molecular alignment ( @xmath14 , which is the moment when the alignment factor @xmath15 reaches the maximum value ) , a second pulse is applied , polarized at 45 degrees to the @xmath2-axis ( we choose the polarization vector to be in the @xmath16-@xmath2 plane ) .
we represent the molecule by a unit vector pointing from the center of mass towards one of the atoms ( the center of mass is defined as the coordinates origin ) .
the tip of the vector defines a point on the surface of the unit sphere .
the orientation of the unit vector is given by @xmath17 the rotational velocity of the unit vector @xmath18 can be described by two orthogonal angular velocities @xmath19 and @xmath20 ( the dot denotes a derivative with respect to time ) , or , in vector notation , @xmath21 where the spherical unit vectors are @xmath22 this rotational velocity is referred to as simply `` velocity '' in the following .
the rotational kinetic energy in this notation is given by @xmath23 where @xmath24 is the moment of inertia of the molecule .
as the first step , we find the change in the molecular velocity due to interaction with a single short laser pulse .
for a @xmath2-polarized pulse , angle @xmath25 is equal to the polar angle @xmath26 , and we can use the potential energy from eq .
( [ v ] ) to find the torque @xmath27 applied to the molecule in spherical coordinates : @xmath28 newton s second law gives @xmath29 and @xmath30 .
we assume that the molecule does not change its orientation during the short pulse ( impulsive approximation ) . by integrating the newton s equations of motion over the short duration of the pulse
, we obtain the change of the velocity due to the laser pulse action : @xmath31 integrating again , and assuming that the molecule was initially at rest , we get @xmath32 here @xmath33 and @xmath34 describe the initial orientation of the molecule . every molecule starts to rotate with a constant velocity ( depending on @xmath33 and @xmath34 ) , and @xmath18 is tracing a circle passing through the north and the south poles of the unit sphere . as the next step , we consider the action of a pulse linearly polarized along some arbitrary unit vector @xmath35 , and determine the vector of the resulting velocity change @xmath36 for a molecule oriented along some direction @xmath37 .
the norm @xmath38 can be found similarly to the previous case of a @xmath2-polarized pulse , and it is equal to @xmath39 , where @xmath40 is the angle between the polarization direction of the pulse @xmath35 and the initial orientation direction of the molecule @xmath37 ) .
the direction of @xmath36 can be defined from the vector diagram in fig .
[ kick_picture ] .
notice that @xmath36 is always perpendicular to @xmath37 .
also , @xmath36 is directed parallel or antiparallel to the vector component of @xmath35 perpendicular to @xmath37 , which is equal to @xmath41 . as a result
, we arrive at : @xmath42 we can easily check that ( [ delta_v_vec ] ) gives the correct result for @xmath2-polarized pulse , for which @xmath43 , by transforming ( [ delta_v ] ) to cartesian coordinates using eq .
( [ v_spher2cart ] ) . , is kicked by a laser pulse linearly polarized along an arbitrary direction @xmath35 . as a result , it receives a velocity @xmath36.,scaledwidth=50.0% ] finally , we provide a formula for the time dependence of the position @xmath18 and velocity @xmath44 of the point representing molecular orientation on the unit sphere during the field - free rotation , given the initial orientation @xmath37 and initial velocity @xmath45 . as @xmath44 is always perpendicular to @xmath18 , the tip of the @xmath18 vector traces a circle on the sphere . the orientation of the molecule at time @xmath46 is @xmath47 where @xmath48 . taking the derivative
, we find the molecular velocity : @xmath49
we are interested in the response of a thermal molecular ensemble to the double - pulse excitation , and in the calculation of ensemble averaged values of the alignment factor and other observable quantities . one way to calculate
the time dependence of an observable averaged over the ensemble is to find its value at time @xmath46 using ( [ theta_phi ] ) , or ( [ r_of_t ] ) , and to integrate it over the initial angular distribution function @xcite .
for example , at _ zero temperature _ , the molecules are initially at rest and isotropically distributed in all directions .
therefore , for any observable @xmath50 , we have to substitute the time - dependent expressions ( [ theta_phi ] ) , ( [ r_of_t ] ) for @xmath26 and @xmath51 to obtain @xmath52 , and to integrate over the isotropic distribution of @xmath33 and @xmath34 : @xmath53 here the brackets denote averaging over the ensemble . for a _ nonzero temperature _ , an averaging over the initial velocity distribution should be performed as well . in a thermal ensemble at temperature @xmath54 , the velocities @xmath55 and @xmath56
are distributed according to the boltzmann distribution : @xmath57 } dv'_{\theta,0 } \nonumber \\ \label{boltzmann } & \times \frac{1}{\sqrt{2\pi\sigma_{\text{th}}^2 } } \exp{\left[-\frac{\left(v'_{\varphi,0}\right)^2}{2\sigma_{\text{th}}^2}\right ] } dv'_{\varphi,0}~,\end{aligned}\ ] ] where the dimensionless width of the thermal distribution @xmath58 is given by @xmath59 . here
@xmath60 is the rotational constant of the molecule , and @xmath61 is the dimensionless velocity @xmath62 . instead of ( [ avg_over_ens_0temp ] ) , we have @xmath63 these integrals have fast oscillating integrands and are difficult ( although possible ) to calculate numerically . a somewhat simpler way of finding the time dependence of observables averaged over the molecular ensemble is by means of a monte carlo simulation .
we start with molecules isotropically oriented in space , with rotational velocities distributed according to eq .
( [ boltzmann ] ) .
these initial values are obtained using a numerical random number generator . in order to randomly distribute points on the surface of the sphere
, one uses the probability function @xmath64 the realization of @xmath65 is straightforward if one has a _ uniform _ random number generator . if @xmath66 denotes a random number from a uniform distribution in the interval @xmath67 $ ] , then @xmath68 . on the other hand , in order to realize @xmath69 one should use the transformation method @xcite . starting with the equality of probabilities @xmath70
, we obtain @xmath71 . as a function of time after the second pulse exciting unidirectional molecular rotation .
it is calculated by the classical monte carlo method with @xmath72 molecules ( dashed black ) , and by the exact quantum ( finite - difference time - domain ) simulation ( solid red ) .
the polarization vectors are @xmath73 and @xmath74 , and @xmath75 for both pulses .
the calculation was done for nitrogen molecules at the rotational temperature of @xmath76 . the time is measured in units of the quantum revival time @xmath77.,scaledwidth=100.0% ] in order to realize the velocity distribution ( [ boltzmann ] ) , one uses the _ normal _ random number generator . using the transformation method again
, we obtain @xmath78 and @xmath79 , where @xmath80 denotes a random number from a normal distribution with zero mean and standard deviation of unity .
now we have all the ingredients needed to simulate the scheme for exciting unidirectionally rotating molecules , as explained at the beginning of sec .
randomly oriented molecules are `` kicked '' by a pulse polarized along the @xmath2-axis , and get velocity additions according to eq .
( [ delta_v_vec ] ) , which should be added to their initial thermal velocities .
the molecules then evolve freely according to eqs . ( [ r_of_t ] ) and ( [ v_of_t ] ) . at each time step , we calculate the ensemble averaged alignment factor @xmath15 . at the time of the maximal alignment , the 45-degrees rotated pulse is applied , velocities are changed again according to eq .
( [ delta_v_vec ] ) , and another free propagation follows . as the first application of this approach , we investigate the azimuthal factor @xmath4 that characterizes confinement of the molecular angular distribution to the @xmath16-@xmath2 plane .
this factor takes the unity value when all the molecules are confined to this plane , and it is equal to @xmath81 for the isotropic ensemble . in figure [ cos2phi ] , we plot the classically calculated azimuthal factor @xmath4 as a function of time following the second pulse , along with the exact quantum result ( for details of the quantum calculation see @xcite ) .
the calculation here ( and also at figs .
[ z_kick_dist ] and [ z_then45_dist ] below ) is done for nitrogen molecules with the rotational constant of @xmath82 .
the rotational temperature is @xmath83 , or @xmath84 .
it is clearly seen that the classical calculation ( monte carlo realization using @xmath72 molecules ) perfectly reproduces the first oscillation of the azimuthal factor .
obviously , later fractional and full revivals are quantum features , which are not reproduced by the classical treatment .
it is also evident that the time averaged azimuthal factor ( that is around 0.52 here ) coincides for the classical and the quantum calculation .
this value is higher than the isotropic value of 0.5 , hence the molecular distribution is squeezed towards the plane defined by the polarization directions of the first and the second pulses .
we see that a relatively simple classical model reproduces the important features of this problem .
in this section we estimate classically the angular distribution function of an ensemble of linear molecules excited by two pulses , and compare it with the results of full quantum - mechanical simulation @xcite . to achieve this goal
, we use the kernel estimator method @xcite to reconstruct the distribution function from a set of @xmath85 points on the unit sphere , which are participating in the monte carlo simulation . according to this method , each one of the @xmath86 points
is surrounded by a narrow symmetric gaussian function on the surface of the sphere .
the sum of all gaussians represents the angular distribution function .
its time dependence is determined numerically by propagating independently @xmath86 rotating molecules , as explained in the previous sections . a simple way to build a symmetric gaussian function around a point @xmath87 on a sphere (
@xmath88 denotes the molecule number in the monte carlo realization ) is by referring to the angle @xmath89 between the unit vector to this point , and the unit vector to a general point @xmath18 in its neighborhood : @xmath90 .
the corresponding ( unnormalized ) distribution function is @xmath91 , where @xmath92 is the distribution width ( a free parameter ) .
because only small width values are used ( @xmath93 ) , the quantity @xmath94 can be replaced by the expression @xmath95 , which is equivalent to it for @xmath96 , while for @xmath97 the gaussian distribution function is negligible any way .
normalizing the two dimensional surface gaussian function , we finally obtain the following estimate for the time dependent distribution function : @xmath98 an important object showing the symmetry and anisotropy of our system is the time - averaged angular distribution function . a straightforward , but somewhat cumbersome way to generate it is by numerical averaging the distribution function ( [ td_distribution ] ) over a long time period . a much simpler approach to time averaging is described below . after the end of the pulses ,
each molecule moves with a constant speed on some circle on the surface of the unit sphere . within a sufficiently long time period
, each molecule spends the same amount of time at each length element of the circle ( apart from very rare cases in which a molecule stays strictly at rest ) .
thus , the long - time - averaged angular distribution function for each molecule is a `` belt '' ( a circle with a width ) on the unit sphere .
its orientation depends on the initial position @xmath37 and velocity @xmath99 on the unit sphere , and its width is a free parameter @xmath100 , which , for consistency , is taken equal to the parameter @xmath92 that was introduced before .
the total time - averaged distribution function is obtained , as before , by adding the `` belts '' for all the molecules participating in the monte carlo simulation .
the orientation of the `` belt '' for the @xmath88-th molecule is given by the angular momentum unit vector @xmath101 , which is perpendicular to the plane defining the circle of rotation .
thus , the gaussian `` belt '' for a single molecule is located close to the points @xmath18 on the unit sphere , for which the dot product @xmath102 is nearly zero . inserting the correct normalization factors , we finally obtain for the time - averaged angular distribution function , @xmath103 : @xmath104 where the overline denotes a long time average . .
the curve plots and inset ( a ) are for molecules initially at zero rotational temperature .
the dashed green curve presents the exact quantum result , and the dash - dotted red line displays the analytical classical distribution according to eq .
( [ dist_analytic ] ) . the blue solid curve and inset ( a )
are calculated by the monte carlo method according to eq .
( [ time_avg_distribution ] ) , with @xmath105 and @xmath106 .
inset ( b ) shows monte carlo results for the initial rotational temperature of @xmath83 .
, scaledwidth=100.0% ] before considering the double pulse scheme , we first analyze the time - averaged angular distribution function for molecules subject to a single pulse polarized along the @xmath2-axis .
the result for zero temperature ( all molecules are initially at rest ) is shown in fig .
[ z_kick_dist ] ( a ) and duplicated at the plot by the solid line . in fig .
[ z_kick_dist ] ( b ) , the distribution function is plotted for the same system , but at finite temperature . as expected , the distribution is `` cigar''-shaped , and azimuthally symmetric . at zero temperature , an analytical result can be obtained for the time - averaged distribution function in the classical limit . in this case , all the trajectories of the kicked molecules are circles passing through the poles of the unit sphere .
these circles are uniformly distributed in the azimuthal direction , and the distribution function is @xmath51-independent .
therefore , at any given value of @xmath26 we find the same total number of molecules . as a result , the time - averaged distribution function multiplied by the geometrical factor
@xmath107 should be a constant .
it is seen from here , that the distribution function averaged over long time should be proportional to @xmath108 , or , after normalization : @xmath109 this equation was verified also by evaluating numerically eq .
( 9 ) of reference @xcite , including summation over all branches . the function of eq .
( [ dist_analytic ] ) is plotted in fig .
[ z_kick_dist ] by a dash - dotted line .
our numerical monte carlo result in fig .
[ z_kick_dist ] ( solid line ) agrees well with the analytical result , but it is not singular at the north and the south poles , because a finite width @xmath100 was given to every molecule trajectory , and the singularities are smeared out .
finally , we compare the classical result to the quantum angular distribution function averaged over a single revival period .
the quantum result is given in fig .
[ z_kick_dist ] by a dashed line .
two features should be stressed here .
first , the singular classical behavior disappears in the quantum description .
second , there is a fine oscillatory structure in the quantum distribution , which is absent in the classical distribution . besides these features ,
the distributions are similar .
we also compare the values of @xmath110 , @xmath111 and @xmath112 , calculated both classically and quantum mechanically at the conditions shown in fig .
[ z_kick_dist ] ( b ) . here
the brackets denote the ensemble averaging and the overline denotes long - time averaging .
the quantities @xmath113 are the cartesian components of the unit vector @xmath18 . in both cases , the same values of @xmath114 and @xmath115 are obtained .
this should be compared to the `` isotropic '' values of @xmath116 . for the double pulse scheme ,
leading to the unidirectional rotation of the molecules , we present the classical time - averaged angular distribution in fig .
[ z_then45_dist ] ( a ) , which is very similar , except of fine details , to the corresponding quantum distribution in fig .
[ z_then45_dist ] ( b ) , presented also in @xcite .
the distribution is squeezed to the @xmath117-plane ( the plane defined by the polarization directions of the two pulses ) .
we expected @xmath111 to be smaller than @xmath110 and @xmath112 , and we indeed obtain numerically @xmath118 , @xmath119 and @xmath120 classically , as well as quantum mechanically .
these results demonstrate very good agreement between the classical and quantum calculations even for relatively small values of the interaction strength @xmath11 . .
the calculation was done using the classical monte carlo method , eq .
( [ time_avg_distribution ] ) , in ( a ) , and using the quantum finite - difference time - domain simulation followed by the averaging over a revival period in ( b ) .
both pulses had the kick - strength of @xmath75 .
the following parameters were used in ( a ) : @xmath105 , and @xmath106.,scaledwidth=100.0% ]
now we address a more difficult problem of the unidirectional rotation of benzene molecules .
these molecules , considered as rigid bodies , behave as oblate symmetric tops .
the classical treatment of the free rotation of a symmetric top is analytically simple ( although more complicated than the rotation of a linear rotor ) , and is well documented in literature ( see , e. g. @xcite ) .
there are two moments of inertia characterizing a symmetric top , @xmath121 and @xmath122 , which are defined using the body fixed coordinate system @xmath123 , shown in fig .
[ symmetric_top ] , where the molecular symmetry axis is directed along the @xmath124 axis .
the figure is taken from @xcite ; the @xmath125 axis is perpendicular to the plane of the figure and is not shown . for benzene molecule , which is an oblate symmetric top , and has the form of a planar ring , the @xmath124 axis is perpendicular to the plane of the ring and passes through its center .
hence , for benzene , @xmath126 @xcite .
when the molecule is freely rotating , the angular momentum , denoted by @xmath127 , is constant in magnitude and direction .
the rotational kinetic energy is also a constant of motion and can be written as : @xmath128 where @xmath129 , @xmath130 and @xmath131 are the components of angular momentum along the body - fixed coordinate axes @xmath132 ,
@xmath125 and @xmath124 , respectively , and @xmath133 .
the free rotation is such that the angular momentum @xmath127 , the angular velocity @xmath134 and the molecular axis @xmath124 all lie in the same plane .
the molecular axis @xmath124 ( as well as the vector @xmath134 ) performs precession around the angular momentum vector @xmath127 with an angular velocity @xmath135 , which is shown in fig .
[ symmetric_top ] .
its magnitude is : @xmath136 the molecular axis @xmath124 moves on the surface of a cone and forms an angle @xmath137 with the cone axis ( which is along the vector @xmath127 ) .
in addition to the precession motion , the top rotates around its @xmath124 axis , but this motion is not important in our problem .
the reason is that this last rotation does not change the molecular orientation ( defined by the molecular axis ) , and that the laser pulse does not influence this motion , as will be explained later on .
we describe the orientation of the molecular axis by a unit vector @xmath18 in the direction of @xmath124 .
its value at @xmath138 is @xmath37 .
the velocity of this vector is given by @xmath139 we now define two more unit vectors .
one is @xmath140 , a unit vector in the direction of @xmath127 .
another is in the plane of the circle traced by the tip of @xmath18 : @xmath141 using the notation defined , it is easy to see that the time dependence of @xmath18 is given by : @xmath142 for the velocity we obtain : @xmath143 it is important to stress , that the tip of the unit vector @xmath18 does not necessarily trace a circle of a unit radius on the surface of the unit sphere , as for the linear molecule .
it can trace any circle , which can even shrink to a point , in the case of @xmath144 and @xmath145 .
a unit circle is traced only if @xmath146 , or @xmath147 and @xmath148 .
although we are dealing here with a more complicated molecule , the interaction with an ultrashort laser pulse is described by the same formula as in eq .
( [ v ] ) .
the reason is that the polarizability tensor of the symmetric top has the same symmetry as the one of the linear molecule .
the interaction depends , as before , on the electric field envelope squared , on @xmath149 , which is negative in the case of benzene , and on the angle between the pulse polarization direction @xmath35 and the molecular axis orientation @xmath37 .
notice that a linearly polarized laser pulse does not produce a torque _ parallel _ to the molecular axis .
the torque exerted by the pulse is , therefore : @xmath150 where @xmath151 is a unit vector in the direction of the cross product @xmath152 .
this equation shows that the laser pulse `` kicks '' the symmetry axis of the benzene molecule towards the plane , which is perpendicular to the polarization vector @xmath35 . using newton s second law and integrating over the short time of the pulse , we obtain for the change in the angular momentum : @xmath153 and two similar equations for the @xmath154- and the @xmath2-components .
angle @xmath40 can be found using @xmath155 .
now , after the new angular momentum vector is known , and assuming the molecule did not move during the interaction with the pulse , the molecular rotation is completely defined by eq . ( [ r_benz ] ) . naturally , new values of @xmath137 , @xmath140 , @xmath156 , @xmath45 and @xmath157 should be found using the new value of @xmath127 .
specifically , @xmath137 is found using the relation @xmath158 , @xmath156 is then defined using eq .
( [ r0_par ] ) , @xmath157 by eq .
( [ omega_pr ] ) and @xmath45 using eq .
( [ v_benz_def ] ) .
the monte carlo simulation for benzene molecules has several differences compared with the one used for linear molecules . here , the kinetic energy is more easily separable when expressed using the angular momentum components along the body - fixed coordinate system , as in eq .
( [ erot_sym_top ] ) .
we start with random generation of the angular momentum components @xmath159 and @xmath131 , according to the thermal distribution : @xmath160 } l'_{\parallel}\ , dl'_{\parallel } \nonumber \\ ~&\times & \frac{1}{\sqrt{2\pi\sigma_{\text{th},3}^2 } } \exp{\left[-\frac{\left(l'_3\right)^2}{2\sigma_{\text{th},3}^2}\right ] } dl'_3~,\nonumber\end{aligned}\ ] ] where we introduce @xmath161 and @xmath162 , similar to eq .
( [ boltzmann ] ) , and where @xmath163 represents the dimensionless angular momentum @xmath164 . employing the transformation method , and using the notation introduced after eq .
( [ rand_r_sphere ] ) , we obtain explicitly for the random components : @xmath165}}$ ] and @xmath166 . knowing these components
, we deduce the _ magnitude _ of the angular momentum @xmath167 , but its _ direction _ is yet to be defined .
we obtain also the precession velocity using eq .
( [ omega_pr ] ) , and @xmath137 from the relation @xmath168 .
next , we determine the _ direction _ of the angular momentum in the laboratory coordinate system , which is given by the vector @xmath140 .
we do this by randomly generating two spherical angles @xmath169 and @xmath170 .
these two angles are distributed the same as @xmath26 and @xmath51 in eq .
( [ rand_r_sphere ] ) , because the tip of the vector @xmath140 is distributed uniformly on the surface of the unit sphere .
finally , by randomly generating an angle @xmath51 , uniformly distributed between @xmath171 and @xmath172 , we obtain the initial orientation of the molecular axis on the cone around the vector @xmath127 .
now , the coordinates of @xmath37 are completely defined in the frame in which the @xmath2-axis is fixed along the vector @xmath127 .
these coordinates are given by @xmath173 .
using two rotational matrices , the coordinates of @xmath37 in the laboratory frame are then found .
the above monte carlo procedure was used to generate data that are discussed later in sec . c. here we outline the quantum mechanical description of unidirectional rotation of benzene molecules by two delayed ultrashort laser pulses with crossed linear polarizations .
details will be presented in @xcite .
the molecular hamiltonian , @xmath174 , in the field - free conditions is obtained simply by replacing the components , @xmath175 @xmath176 , of the classical angular momentum by the corresponding quantum mechanical operators , @xmath177 , in eq .
( [ erot_sym_top ] ) .
the rotational eigenstates and eigenenergies are given by the time - independent schr@xmath178dinger equation , @xmath179 where @xmath180 stands for an index to identify the eigenstates , explicitly expressed by a set @xmath181 , with @xmath182 being the total angular momentum , and @xmath183 and @xmath184 being its projection onto the molecular symmetry axis and the space - fixed axis , respectively .
at finite temperature , an initial ensemble is a mixture of molecules in many different eigenstates , but we first take a single rotational level as an initial state , @xmath185 . the nonadiabatic interaction with the nonresonant ultrafast laser field , given by eq .
( [ v ] ) , converts the stationary state to a rotational wave packet , @xmath186 , which is expanded as @xmath187 here , @xmath188 is the time - evolution operator from time @xmath189 to @xmath190 and @xmath191 .
the expansion coefficients , @xmath192 , vary during the interaction with the laser pulses , and subsequently become constant after the laser field vanishes .
they can be derived by solving the time - dependent schr@xmath178dinger equation ( tdse ) with the initial condition , @xmath193 , @xmath194 \vert \psi_{r_{i } } ( t ) \rangle ~,\end{aligned}\ ] ] where @xmath195 is given in eq .
( [ v ] ) . substituting eq .
( [ psi ] ) , this tdse is recast into the following coupled differential equations for the expansion coefficients : @xmath196 with @xmath197 .
these coupled equations can be solved numerically to determine the complex expansion coefficients , once the matrix elements for @xmath198 are specified .
if another laser pulse is irradiated onto the molecule at @xmath199 , the wave packet created by the first pulse , represented in eq .
( [ psi ] ) , is further modified by the interaction with the second pulse .
the resultant wave packet at @xmath46 after the second pulse is expanded as @xmath200 where the prime on @xmath201 indicates the properties of the second pulse .
the quantity @xmath202 is the transition amplitude from the initial @xmath203 state to @xmath204 by the interaction with the two pulses , and is represented by @xmath205 to evaluate the transition amplitude given in eq .
( [ b ] ) , it is convenient to adopt the space - fixed axis system defined differently from the previous one . here , we set the @xmath2 axis along the laser propagation direction and the polarization of the first pulse parallel to the @xmath16 axis .
then , the interaction given in eq . ( [ v ] ) is recast as @xcite @xmath206 \cos^2{\theta } \right .
\label{v_new } ~&-&\left .
\delta\alpha \cos(2 \varphi ) - ( \alpha_{\parallel } + \alpha_{\perp } ) \right\ } ~,\end{aligned}\ ] ] or , re - expressed as @xmath207 \right\ } ~,\end{aligned}\ ] ] where @xmath208 is the rotational matrix @xcite . with this expression ,
the evaluation of the matrix elements for a symmetric - top basis set is straightforward by applying the spherical tensor algebra , and the following selection rules are derived : @xmath209 , @xmath210 ( then @xmath211 for @xmath212 and @xmath213 for @xmath214 ) , or @xmath215 ( then @xmath211 for @xmath216 and @xmath213 for @xmath217 ) .
we set the polarization of the second pulse tilted against that of the first one by angle @xmath218 .
when the new axis system is defined so as the @xmath219 axis is set parallel to the polarization of the second pulse , the new and old angular coordinates are related as : @xmath220 .
consequently , the symmetric - top wave function transforms as @xmath221 by taking the transformation into account , @xmath202 in eq .
( [ b ] ) is rewritten , if the second pulse is a replica of the first one , as @xmath222 ~,\end{aligned}\ ] ] with @xmath223 standing for a set of @xmath224 .
the transition amplitude for the two pulses can be evaluated for an arbitrary @xmath225 at once , if the set of the coefficients , @xmath226 , are derived by solving the coupled equations in ( [ cde ] ) .
the expectation value of a given observable @xmath227 for a certain initial condition with finite temperature is derived as the following ensemble average , @xmath228 where @xmath229 is the boltzmann factor for @xmath230 with appropriate degeneracy factor due to nuclear - spin statistics @xcite . for the discussion of the molecular alignment and orientation , we take @xmath231 , @xmath232 , and @xmath233 ( corresponding to the classical @xmath234 ) , as @xmath227 . in this section
we analyze the process of exciting unidirectional rotation of benzene , following the setup of @xcite .
the scheme operation is practically the same as described after eq .
( [ p ] ) , except of several details specific for the benzene molecules . shortly after the first ultrashort pulse , polarized along @xmath73
, the molecules experience a transient anti - alignment , as opposed to alignment in the case of linear molecules .
this happens because @xmath149 is negative for benzene , as opposed to linear molecules , such as nitrogen , and the direction of the angular momentum transferred to the molecule by the pulse is reversed .
therefore , shortly after the first pulse the molecular axis angular distribution is squeezed to the plane _ perpendicular _ to @xmath235 .
quantitatively , after the first pulse , the ensemble average of @xmath15 decreases below the isotropic value of @xmath236 .
the second delayed and tilted laser pulse brings this anisotropic ensemble to the rotation with a preferred sense . in the following
, we shall consider the time development of the alignment factor @xmath15 after the first pulse , and the dependence of the average value of the induced angular momentum as the function of the delay of the second pulse .
we show first the classical results . the upper part of fig .
[ benzene_jy_cos2_class ] presents the alignment factor @xmath15 as a function of time passed after the first pulse ( polarized along @xmath73 ) , for three values of the pulse strength @xmath11 . here and
below we use the following parameters for the benzene molecule @xcite : rotational constants @xmath237 , @xmath238 and the polarizabilities @xmath239 , @xmath240 .
all plots are for rotational temperature of @xmath241 ( as in experiment @xcite ) , which corresponds to the dimensionless parameters @xmath242 and @xmath243 . for benzene molecules kicked by a single pulse , plotted as a function of time after the pulse .
the pulse has a strength of @xmath244 ( black , blue and red lines online , respectively ) and polarization direction @xmath73 .
the lower panel displays the ( normalized ) value of the oriented angular momentum @xmath245 induced by a second ultrashort pulse , shown as a function of the delay of this pulse .
the polarization direction of the second pulse is @xmath246 ( at @xmath247 degrees to @xmath235 ) , and the value of @xmath248 is the same as that of @xmath249 , with the same color code used for the curves .
the initial rotational temperature is @xmath241 .
the calculations were done using the monte carlo method with @xmath250 realizations .
, scaledwidth=100.0% ] are plotted here .
they were calculated using the quantum mechanical treatment .
half and full rotational revivals are seen , which are not present in fig .
[ benzene_jy_cos2_class].,scaledwidth=100.0% ] ( solid lines ) , and quantum mechanical results of fig . [ benzene_jy_cos2_quant ] ( dashed lines ) at short times .
the values of @xmath11 are @xmath251 for black , blue and red lines , respectively .
, scaledwidth=100.0% ] the plots for @xmath15 show that the anti - alignment becomes more pronounced and occurs at earlier times as the pulse intensity increases , as is naturally expected .
similar to the case of linear diatomic molecules , we anticipate that the second pulse should be fired at the moment of the maximal squeezing of the molecular angular distribution , i. e. when the alignment factor takes the minimal value . in this case , the maximal average torque is applied to the molecules , facilitating the most emphasized unidirectional rotation .
this is indeed seen from the lower part of fig .
[ benzene_jy_cos2_class ] . here
the quantity @xmath245 ( calculated after the second pulse ) is plotted as a function of the time delay between the first pulse ( with polarization direction @xmath235 ) and the second pulse with polarization direction @xmath246 , where the angle between @xmath235 and @xmath252 is @xmath247 degrees .
the maximal modulus of the transferred angular momentum happens indeed when the second pulse is applied at the moment of the maximal anti - alignment .
we also notice that the optimal delay becomes shorter with the strength of the pulses .
figure [ benzene_jy_cos2_quant ] presents the same quantities calculated fully quantum mechanically , following the procedure outlined in section v - b . on the long - time scale
, it demonstrates quantum rotational revivals , such as half and full revivals , that are absent in fig .
[ benzene_jy_cos2_class ] for obvious reasons . for a detailed comparison with the classical results at shorter times ( corresponding to the conditions of the experiment @xcite ) , fig .
[ benzene_jy_cos2_compar ] displays the classical and the quantum graphs overlapped on the same plot .
the results seem to be practically identical at short times , and the agreement is improving with the increase of @xmath253 , as more rotational states are involved in the dynamics of the wavepacket .
parameters of the experiment @xcite correspond to the curves for @xmath254 in fig .
[ benzene_jy_cos2_compar ] , with the delay time chosen near the minima of the curves . as follows from fig .
[ benzene_jy_cos2_compar ] , the rotational dynamics in this range is adequately described by the classical treatment .
we studied the problem of exciting unidirectional molecular rotation with the help of two delayed cross - polarized laser pulses using classical mechanics , and compared this approach with the results of fully quantum mechanical treatment .
the problem was analyzed both for linear molecules and for symmetric tops , like benzene .
although the comprehensive description of the fine details of rotational dynamics undeniably requires the full scale quantum formulation , we found a very good agreement between the classical and quantum approaches in a wide range of experimentally relevant parameters .
the classical model identifies the origin of the mechanism , defines the optimal delay between the pulses , and the best angle between them .
moreover , it accurately predicts the magnitude of the induced oriented angular momentum , and correctly reproduces the behavior of the optimal delay with pulse strength , as it was observed in @xcite . at the same time , a fully quantum treatment is required to describe quantitatively the operation of the double - pulse scheme for relatively weak laser pulses , when only low rotational states are involved . in this case , considering quantum interference of excitation pathways by the first and the second pulses @xcite provides an instructive complementary view on the origin of the phenomenon .
moreover , operation of the scheme at long delays between the pulses ( comparable with the period of rotational revivals ) needs quantum description as well . at the same time ,
quantum and classical results agree well again if the long - time - averaged observable quantities are considered .
we illustrated this by comparing classical and quantum time - averaged molecular angular distributions that show anisotropy and confinement to the plane spanned by two polarization vectors of the laser pulses .
we believe that this comparative study provides the reader a comprehensive view on the physics behind the double pulse scheme for ultrafast preparation of molecular ensembles with oriented rotational angular momentum and controlled sense of rotation .
the authors are thankful to kenji ohmori for stimulating this collaborative work .
yk and ia appreciate many fruitful discussions with yehiam prior , sharly fleischer and erez gershnabel
. financial support of this research by the israel science foundation is gratefully acknowledged .
additional financial supports by a grant - in - aid from mext japan , c - phost , and extreme photonics " are also appreciated .
ia is an incumbent of the patricia elman bildner professorial chair .
this research is made possible in part by the historic generosity of the harold perlman family .
h. stapelfeldt and t. seideman , rev .
phys . * 75 * , 543 ( 2003 ) .
t. seideman and e. hamilton , adv . at .
. phys . * 52 * , 289 ( 2006 ) .
j. ortigoso , m. rodriguez , m. gupta and b. friedrich , j. chem . phys . * 110 * , 3870 ( 1999 ) .
f. rosca - pruna and m. j. j. vrakking , phys .
lett . * 87 * , 153902 ( 2001 ) .
averbukh and r. arvieu , phys .
87 * , 163601 ( 2001 ) ; m. leibscher , i. sh .
averbukh and h. rabitz , phys
lett . * 90 * , 213001 ( 2003 ) .
m. leibscher , i. sh .
averbukh and h. rabitz , phys .
a * 69 * , 013402 ( 2004 ) .
m. renard , e. hertz , s. gurin , h. r. jauslin , b. lavorel and o. faucher , phys .
a * 72 * , 025401 ( 2005 ) . c. z. bisgaard , m. d. poulsen , e. pronne , s. s. viftrup and h. stapelfeldt , phys .
* 92 * , 173004 ( 2004 ) .
k. f. lee , i. v. litvinyuk , p. w. dooley , m. spanner , d. m. villeneuve and p. b. corkum , j.phys .
b : at . , mol . ,
opt . phys . * 37 * , l43 ( 2004 ) . c. z. bisgaard , s. s. viftrup and h. stapelfeldt , phys .
a * 73 * , 053410 ( 2006 ) .
d. pinkham and r. r. jones , phys .
a * 72 * , 023418 ( 2005 ) .
j. karczmarek , j. wright , p. corkum and m. ivanov , phys .
lett . * 82 * , 3420 ( 1999 ) .
i. v. litvinyuk , kevin f. lee , p. w. dooley , d. m. rayner , d. m. villeneuve and p. b. corkum , phys .
lett . * 90 * , 233003 ( 2003 ) .
s. fleischer , i. sh .
averbukh and y. prior , phys .
a * 74 * , 041403 ( 2006 ) .
m. renard , e. hertz , b. lavorel , and o. faucher , phys . rev . a * 69
* , 043401 ( 2004 ) .
s. fleischer , i. sh .
averbukh and y. prior , phys .
99 * , 093002 ( 2007 ) .
s. fleischer , i. sh .
averbukh , and y. prior , in ultrafast phenomena xvi , proceedings of the 16th international conference , p.75 - 77 , june 2008 , stressa , italy , ( springer , berlin , 2008 ) ; http://www.ultraphenomena.org/programme/tue3.pdf s. fleischer , y. khodorkovsky , i. sh .
averbukh , and y. prior , in _ conference on lasers and electro - optics / international quantum electronics conference _ , ( baltimore , maryland , may 2009 ) osa technical digest ( cd ) ( optical society of america , 2009 ) , paper ituc5 .
s. fleischer , y. khodorkovsky , y. prior and i. sh .
averbukh , new j. phys . * 11 * , 105039 ( 2009 ) .
a. g. york , opt . exp . * 17 * , 13671 ( 2009 ) .
k. kitano , h. hasegawa , and y. ohshima , phys .
. lett . * 103 * , 223002 ( 2009 ) .
b. w. silverman , _
density estimation for statistics and data analysis _ , chapman & hall / crc ( 1998 ) .
m. leibsher , i. sh .
averbukh , p. rozmej and r. arvieu , phys .
a * 69 * , 032102 ( 2004 ) .
l. d. landau and e. m. lifshitz , _ mechanics _ , third edition , pergamon press ( 1976 ) .
k. kitano , h. hasegawa , and y. ohshima , to be published .
y. ohshima and h. hasegawa , int .
phys . chem .
* 29 * , 619 ( 2010 ) .
r. n. zare , _ angular momentum _ , wiley - interscience , new york ( 1988 ) . s. m. beck , m. g. liverman , d. l. monts , and r. e. smalley , j. chem . phys . *
70 * , 232 ( 1979 ) .
h. hasegawa and y. ohshima , chem .
lett . * 454 * , 148 ( 2008 ) .
|
recently , it was predicted theoretically and verified experimentally that a pair of delayed and cross - polarized short laser pulses can create molecular ensembles with a well defined sense of rotation ( clockwise or counterclockwise ) .
here we provide a comparative study of the classical and quantum aspects of the underlying mechanism for linear molecules and for symmetric tops , like benzene molecules , that were used for the first experimental demonstration of the effect .
very good quantitative agreement is found between the classical description of the process and the rigorous quantum mechanical analysis at the relevant experimental conditions .
both approaches predict the same optimal values for the delay between pulses and the angle between them , and deliver the same magnitude of the induced oriented angular momentum of the molecular ensemble .
as expected , quantum and classical analysis substantially deviate when the delay between pulses is comparable with the period of quantum rotational revivals .
however , time - averaged characteristics of the excited molecular ensemble are equally good described by the these two approaches .
this is illustrated by calculating the anisotropic time - averaged angular distribution of the double - pulse excited molecules , which reflects persistent confinement of the molecular axes to the rotation plane defined by two polarization vectors of the pulses .
|
the origin of the isotropic diffuse emission ( sreekumar et al .
1998 ) in energies between @xmath0 and @xmath1 , detected by the _
energetic gamma - ray experiment telescope _ ( egret ) aboard the _ compton gamma - ray observatory _
, remains one of the great unknowns of gev - energy astrophysics .
there are two major questions that still remain unanswered
how much of the diffuse emission detected by egret is , in fact , extragalactic , and what is the spectrum of this extragalactic background ? and 2 .
what fraction of the extragalactic emission can be attributed to each of the observationally established classes of gamma - ray emitters ? despite the associated uncertainties , these two issues are critical in any attempt to use gamma - ray observations to constrain exotic high - energy physics and yet - undetected classes of theorized gamma - ray emitters . to answer the first question , a good understanding of the galactic diffuse emission in the egret energy range is required . in order to obtain the intensity and spectrum of the extragalactic emission from the egret sky maps , the galactic emission needs to be modeled and subtracted .
this is made complicated by the discrepancy between the observed milky way spectrum in energies of @xmath2 and theoretical expectations ( hunter et al .
the observed spectrum is more shallow than model predictions based on the local demodulated cosmic ray spectrum .
this deviation is known as the `` gev excess '' , and although various explanations have been proposed to account for part or all of the discrepancy , its origin remains a matter of debate ( e.g. , pohl et al .
1997 ; bsching et al . 2001 ; strong et al . 2004b ; kamae et al . 2005 ; de boer et al . 2006
; strong 2006 ; stecker et al . 2007 ) . as a result
, determinations of the gamma - ray background using different galactic emission models yield very different answers , both in intensity and in spectrum , despite being based on the same set of observations ( e.g. , sreekumar et al . 1998 ; strong et al .
2004a ) .
attempts to answer the second question have been plagued by uncertainties in the cosmic density and evolution of the two established classes of extragalactic gamma - ray emitters : normal galaxies and blazars .
our observational knowledge of the gamma - ray properties of normal galaxies is very limited , as the sample of normal galaxies which have been observed in gamma rays consists of only the milky way and a marginal detection of the large magellanic cloud ( sreekumar et al .
1992 ; hunter et al .
1997 ; hartman et al .
for this reason , the accuracy of theoretical estimates of the contribution of normal galaxies to the gamma - ray background is unavoidably at the order - of - magnitude level ( e.g. , lichti et al .
1978 ; pavlidou & fields 2002 ) .
but even in the case of blazars , which are by far the most numerous and best studied class of identified gamma - ray emitters , estimates of their contribution to the gamma - ray background vary from a few percent to @xmath3 of the background originally reported by the egret team ( e.g. , padovani et al .
1993 ; stecker & salamon 1996a ; kazanas & perlman 1997 ; mukherjee & chiang 1999 ; mcke & pohl 2000 ; narumoto & totani 2006 ; dermer 2007 ) .
the issue is further complicated by the existence of 171 sources which , at the time of publication of the 3rd egret catalog ( hereafter 3eg ; hartman et al .
1999 ) , had not been positively or potentially associated with a lower - energy counterpart .
these sources are collectively known as the unidentified egret sources , and they are more numerous than any established group of gamma - ray emitters .
the distribution of these sources on the sky is such that a galactic feature can be clearly distinguished - however a large number of sources are located away from the galactic plane and the galactic center .
no more than a handful of sources can be associated with the milky way halo if the milky way is not many times brighter in gamma - rays than similar galaxies such as m31 ( siegal - gaskins et al .
hence , it is almost certain that the egret unidentified sources include a significant extragalactic component .
although the nature of these extragalactic sources remains unknown , it is reasonable to believe that there is a large number of fainter , unresolved objects of the same class , which are guaranteed to have _ some _ contribution to the extragalactic gamma - ray background ( egrb ) .
if these sources represent yet unidentified members of some known class of gamma - ray emitters ( e.g. blazars ) , then excluding them from any calculation of the contribution of the parent class to the diffuse background would lead to a significantly underestimated result due to an incorrect normalization of the bright - end of the gamma - ray luminosity function .
if they represent an unknown class of gamma - ray emitters , then the contribution of their unresolved counterparts to the diffuse emission would significantly limit the diffuse flux left to be attributed to known classes , exotic processes , and truly diffuse emission .
hence , some contribution of unresolved unidentified sources to the egrb is certain .
it is therefore clear that until we either answer the question of the nature of unidentified sources or derive some strong constraint indicating that a possible contribution of such unresolved objects to the egrb would indeed be minor , we can not hope to fully understand the origin of the egrb .
detailed predictions for the level of the unidentified source contribution to the egrb involve important uncertainties : since no low - energy counterparts have been identified , we have no estimates of distance , and therefore no estimates of the gamma - ray luminosities of these sources . as a result
, very few constraints can be placed on their cosmic distribution and evolution .
however , very simple estimates can offer some guidance on whether ignoring this egrb component may be a safe assumption to make .
for example , we can use the number of unidentified sources , the minimum flux resolvable by egret , and the observed intensity to the extragalactic gamma - ray background to place rough limits on the distance scales associated with resolved and unresolved unidentified sources so that unresolved sources do not overproduce the background .
a population of unbeamed , non - evolving , single - luminosity sources uniformly distributed in euclidian space are resolvable out to a distance @xmath4 by an instrument of number flux sensitivity @xmath5 .
the relation between @xmath4 , @xmath5 , and number luminosity @xmath6 in this case is simply @xmath7 .
if the instrument detects @xmath8 such sources , their number density @xmath9 can be estimated to be @xmath10 .
if the same distribution of sources continues out to a distance @xmath11 , the isotropic intensity ( photons per unit area per unit time per unit solid angle ) from the _ unresolved _ members of this population will be @xmath12 , where @xmath13 is the contribution from sources within a spherical shell located at a distance @xmath14 from the observer .
substituting our results for @xmath9 and @xmath6 above , and performing the integral , we obtain @xmath15 if we require that the unresolved emission from this population does not exceed the egrb observed by egret ( @xmath16 ) , we obtain @xmath17 . substituting @xmath18 for the population of extragalactic unidentified sources
( see discussion in [ samples ] ) , @xmath19 for the sensitivity of egret , and @xmath20 ( sreekumar et al . 1998 ) , we obtain @xmath21 .
this result implies that the largest distance out to which such a distribution of objects persists can not be larger than a few times the distance out to which these objects are currently resolved , since in any other case these sources would overproduce the egrb .
it is therefore conceivable that the unidentified source contribution to the egrb is significant , if not dominant . in this work ,
we approach the problem from a purely empirical point of view . instead of attempting to _ predict _ the level of a diffuse component due to unresolved objects of the same class as unidentified egret sources , we try to assess whether there are any empirical indications that this component is , in fact , minor .
we construct samples of unidentified sources which , based on their sky distribution , are likely to consist mostly of extragalactic objects . under the assumption that the majority of these sources can be treated as members of a single class of gamma - ray emitters ,
we seek to answer the following three questions : \(1 ) is it likely that unresolved objects of the same class could have a significant contribution to the egrb at least in some energy range ?
\(2 ) how would the collective spectrum of their emission compare to the measured spectrum of the egrb deduced from egret observations ?
\(3 ) how are glast observations expected to improve our understanding of the nature of unidentified sources , based on the insight gained from our analysis ?
this paper is structured as follows . in
[ samples ] we discuss the samples of resolved unidentified sources used in our analysis .
our formalism for constructing the collective emission spectrum of unresolved unidentified sources is presented in [ formalism ] .
inputs from egret data used in our analysis are described in [ inputs ] . in [ results ] we describe our results , and in [ sec : glast ] we discuss how these results are expected to improve once glast observations become available .
finally , we summarize our conclusions in [ sec : disc ] .
the observational input constraining our calculations will be the fluxes of the unidentified sources and the spectral index of each source .
the fluxes we use for our analysis are the p1234 photon fluxes quoted in 3eg .
these fluxes were determined from the combined maps from all viewing periods during which any particular object was within the egret field of view .
the spectral index @xmath22 of each object is also taken from 3eg , and was derived through a fit assuming that the photon flux of each object obeys @xmath23 where @xmath24 is the photon energy .
we seek to constrain a possible contribution to the egrb from objects of the same class as unidentified sources .
for this reason , we would like to base our calculations on a _ resolved _ source sample which is as representative as possible of the population of _ extragalactic _ unidentified egret sources .
however , the population of sources which , at the time of publication of 3eg , did not have confident associations with known astrophysical objects , clearly exhibits a galactic feature , with many of the sources concentrated around the galactic plane and at small galactic longitudes ( close to the galactic center ) .
thus , the original population of egret unidentified sources is not an appropriate sample as it includes many sources which are most likely galactic .
since these sources are unidentified , there is no counterpart information which could be used to distinguish between galactic and extragalactic objects .
additionally , inferences based on their spectral indices and variability are also very hard to make due to the large error bars associated with measurements of these quantities ( see siegal - gaskins et al . 2007 ) .
finally , both galactic and extragalactic identified objects exhibit a large range of luminosities , so that inferences about the location of unidentified sources based on their fluxes are not possible either .
the only relatively robust indication concerning the location of these sources is their position on the sky . a galactic population associated with the disk and bulge
will most likely be concentrated around low galactic latitudes and small galactic longitudes . on the other hand , a halo population , which may appear more uniformly distributed on the sky ,
can not have too many members among the resolved egret unidentified sources , based on the gamma - ray properties of galaxies similar to the milky way ( siegal - gaskins et al .
thus , excluding objects around the galactic plane and the galactic center will most likely minimize any galactic contamination of the sample of unidentified sources .
for this reason , we have excluded from our studies all sources with galactic latitude @xmath25 for galactic longitudes @xmath26 and galactic latitude @xmath27 for galactic longitudes @xmath28 .
the source exclusion map we have applied is plotted in fig .
[ mask ] , and is identical to the mask applied by sreekumar et al .
( 1998 ) in their selection of egret diffuse emission data to use for the derivation of the egrb .
the solid angle associated with this mask is @xmath29 , leaving @xmath30 of allowed solid angle .
for this reason , and under the assumption that an extragalactic population will be uniformly distributed on the sky , we have renormalized all derived flux distributions by a factor of @xmath31 .
additionally , we have excluded from all samples objects j0516 + 2320 , j0852 - 1216 , j1424 + 3734 , and j2241 - 6736 because no p1234 flux was quoted for them in 3eg , and objects j0824 - 4610 , j0827 - 4247 , j0828 - 4954 , j0841 - 4356 , j0848 - 4429 , and j0859 - 4257 because they were marked as possible or likely artifacts in 3eg .
we have applied further selection criteria which result in three distinct samples of resolved unidentified sources .
the use of different samples will allow us to test the sensitivity of our results to the details of the included members .
sample 1 includes all sources which were unidentified at the time 3eg was published _ and
_ all sources which were characterized as `` possible blazars '' ( identification code `` a '' in 3eg ) .
these latter objects are routinely not included in studies of the blazar luminosity function , so it is reasonable to examine the potential contribution of this class to the egrb by including it in our analysis .
sample 1 includes 84 sources .
sample 2 only includes sources which were unidentified at the time 3eg was published . in this case , we do not consider `` possible blazars '' to belong to this sample .
sample 2 includes 63 sources , and is a subset of sample 1 .
sample 3 is the most conservative of our samples , in that it contains the smallest number of objects . in this case , not only have we excluded all `` probable blazars '' , but also any object for which even a tentative or potential identification has been suggested since the publication of 3eg .
sample 3 includes 53 sources , and is a subset of sample 2 .
sources which are present in sample 2 but not in sample 3 are the following : j0010 + 7309 , j0724 - 4713 , j1212 + 2304 , j1825 + 2854 , j1835 + 5918 , j0329 + 2149 , j0725 - 5140 , j1249 - 8330 , j1621 + 8203 , and j1959 + 6342 .
the locations of all the sources in our samples are shown in fig .
a quantity which , under the assumption of isotropy , fully describes the egrb is the differential photon intensity @xmath32 ( photons per unit area - time - energy - solid angle ) . in this section ,
we calculate @xmath32 for a population of unresolved sources .
we make the following assumptions for the population of unresolved gamma - ray sources we consider . 1 .
each of the sources has an energy spectrum which , in the egret and glast energy ranges , can be well described as a single power law of spectral index @xmath22 , and a total photon flux @xmath33 at energies @xmath34 ( where @xmath33 is the photon flux in units of @xmath35 ) .
2 . the cumulative flux distribution , @xmath36 ( number of sources with flux greater than @xmath33 versus @xmath33 ) , of this population can be adequately described , between fluxes @xmath37 and @xmath38 , as a power law , @xmath39 3 .
the probability distribution of spectral indices of the population @xmath40 is independent of source luminosity and does not evolve with redshift .
the differential flux of each object of this population , @xmath41 , is related to @xmath33 through @xmath42 provided @xmath43 .
each unresolved source with a flux @xmath33 at energies @xmath44 , has a contribution @xmath45 to the diffuse emission which is given by @xmath46 where @xmath22 is the spectral index of the source , and @xmath47 in the case of 3eg fluxes .
the @xmath48 normalization factor comes from assuming an isotropic distribution of sources , the collective emission of which is uniformly distributed over the celestial sphere .
the differential flux distribution of the population is @xmath49 the collective diffuse emission due to this population of unresolved sources with fluxes between @xmath50 and @xmath51 will be : @xmath52_{f_{\rm 8 , min}}^{f_{\rm 8 , max } } \!\!\!\!\!\!\times \int_{\alpha = -\infty}^{\infty } \!\!\!\!\!\!d\alpha \ , p(\alpha ) \left(\frac{e}{e_0}\right)^{-\alpha } \!\!\!\!\!\ !
( \alpha-1)\ , , \nonumber \\\end{aligned}\ ] ] where @xmath53 .
physically , the upper - limit flux @xmath54 represents the sensitivity limit of the telescope ( objects with flux higher than this are resolved and do not contribute to the background ) , while the lower - limit flux @xmath50 represents the flux below which the approximation of eq .
( [ fit1 ] ) for the cumulative flux distribution breaks down .
the features of the spectral shape of this emission in the simple case when @xmath40 is a gaussian are discussed in appendix [ shape ] .
equation ( [ contr ] ) can provide a first simple estimate for the possible contribution of unresolved sources of the same class as extragalactic unidentified egret sources to the isotropic diffuse gamma - ray background .
the required inputs are the cumulative flux function of eq .
( [ fit1 ] ) , and the spectral index distribution of the source class , @xmath40 . in this section ,
we derive these inputs from the three source samples discussed in [ samples ] .
possible source variability has not been accounted for in this analysis .
a discussion of possible effects of source variability , as well as our reasoning for not considering them for the purposes of our study , are presented in appendix [ var ] .
we use data from the resolved objects in our three samples to construct the cumulative flux distribution of egret unidentified sources .
the cumulative flux distributions of the three samples are shown in fig .
( [ lognlogs ] ) . due to non - uniformities in egret exposure ,
the lowest resolvable flux is not a hard - set number , but rather the efficiency in resolving low - flux objects decreases gradually as the flux decreases , which is one of the major factors in the slow turnover of the cumulative distribution function .
for this reason , we only use data within the range @xmath55 to obtain a fit to the cumulative flux distribution .
we take @xmath56 , and assume that all sources with fluxes above this limit have been resolved by egret .
we define @xmath57 to be the flux for which , in each sample , 5 objects with fluxes higher than or equal to it have been resolved .
the best - fit values of @xmath58 and @xmath59 for each sample are shown in table [ sometable ] . as we can see in fig .
( [ lognlogs ] ) as well as from the parameter values in table [ sometable ] , the slopes of the cumulative flux distribution in the regime where most objects are expected to have been resolved by egret are fully consistent with each other .
this in turn implies that the data are consistent with the hypothesis that most members of all samples have been drawn from a single population of extragalactic emitters . however , we note that the flux distribution slope , close to @xmath602.5 in all cases , is steeper than the slope that would be expected from a single - luminosity population of sources with a uniform distribution in a flat cosmology ( which is equal to @xmath61 , e.g. dermer 2007 ) , and is also steeper than the slope of the flux distribution of the confidently identified blazar population
. this result may reflect the cosmological distribution and evolution properties of the population of extragalactic unidentified sources .
however , it may also originate in a selection effect in the identification of gamma - ray sources : the positional identification of brighter sources is more frequent as more photons are detected from these sources which allows for a more accurate pinpointing of their location .
in addition , multiwavelength campaigns for the identification of gamma - ray sources naturally target the brightest objects first . for these reasons , the high - flux end of the flux distribution
may be preferentially depleted , which would lead to an apparent steepening of the distribution of the sources that remain unidentified .
finally , the steepness of the slope could be an effect of viewing a distribution with real curvature in a very small dynamical range in flux .
a simple example quantifying such a possible effect is the following : if we were to treat all sources in sample 1 and all confidently identified blazars which are located outside our exclusion mask as a single population , then the cumulative flux distribution power - law fit in the flux range @xmath62 would have a slope @xmath63 , much closer to @xmath61 .
finally , we should add a cautionary note on fits to the cumulative , rather than differential , flux distribution function , such as the one presented here .
our fits purposfully do not account for uncertainties in each bin since , from a statistical point of view , such uncertainties are almost perfectly correlated ( each bin contains the same data as its adjacent ones plus / minus one data point ) . on the other hand ,
the poor dynamical range in flux and the dependence of uncertainties of the _ differential _ flux function on the size of the bin make constructing and fitting the differential flux function problematic as well . for these reasons , it is important to not over - interpret these results , which should be viewed as broad , order - of - magnitude assessments of the behavior of the flux distribution , rather than robust statistical evaluations and strict constraints .
the second important observational input in our calculation is the distribution of spectral indices of unidentified objects @xmath40 . in the limit that the spectral index of sources is independent of source redshift and luminosity
, the spectral index distribution uniquely determines the spectral shape of the collective emission from the unresolved population .
the spectral shape of the unresolved emission provides , in turn , an additional tool to assess the possibility that unidentified sources constitute a dominant contribution to the egrb through _ shape comparison _ between the observed and the predicted spectra of diffuse extragalactic emission . in our analysis
we adopt the assumption that the spectral index distribution of unresolved unidentified sources is the same as that of the resolved unidentified sources .
the latter can be deduced from measurements of the spectral index @xmath22 for sources in each one of our samples , following the method presented in venters & pavlidou ( 2007 ) .
the details of this calculation , properly accounting for measurement uncertainties in individual spectral indices of resolved sources , are presented in appendix [ spindex ] .
.fit parameters of the cumulative flux distribution [ sometable ] [ cols="<,<,<,<",options="header " , ]
certain classes of sources contributing to the gamma - ray background may exhibit significant time variability ( blazars are a prominent example of such sources ) .
this fact has generated some concern that the objects resolved by egret ( especially the fainter ones ) are preferentially objects that happened to be in their flaring state when observed , and hence the luminosity function as determined by the resolved objects is not an adequate description of all the unresolved sources contributing to the gamma - ray background .
stecker & salamon ( 1996a ) have addressed that uncertainty in the case of blazars using an explicit treatment of source variability .
this concern is less pronounced if the fluxes used for normalizing the luminosity function of a class of sources are averaged over exposure times reasonably large compared to the source duty cycle . to minimize the possible impact of variability
, we use in this study the p1234 fluxes quoted in 3eg .
sources which have been detected in individual viewing periods but which have no p1234 flux ( which implies that their detection significance in the cumulative map was not high enough ) are omitted from our analysis .
we have not made any further attempt to treat variability in this work , because the unknown parameters in the variability properties of sources are severe enough that any simple treatment of variability is likely to introduce at least as much uncertainty as it attempts to remove , while any uncertainty in the flux distribution of the source samples due to source variability is minor compared to the uncertainties introduced due to our ignorance concerning the nature of unidentified sources .
we have simply used the time - averaged flux of resolved unidentified sources as representative of the emission properties of such sources .
the spectral shape of the cumulative emission from a population of unresolved sources with individual power - law photon spectra @xmath64 and a gaussian spectra index distribution @xmath40 is given by eq .
( [ contr ] ) . if @xmath40 is a gaussian of the form of eq .
( [ spindexes ] ) , as assumed here , then eq . ( [ contr ] )
can be integrated analytically .
defining @xmath65_{f_{8 , \rm min}}^{f_{8 , \rm max}}$ ] , we can write @xmath66\exp\left[-\alpha\ln \frac{e}{e_0}\right ] \nonumber \\ & = & \frac{\tilde{i_0}}{\sqrt{2\pi}\sigma_0 } \exp\left[-\frac{b_0}{2\sigma_0 ^ 2}\right ] \int_{-\infty}^{\infty } d\alpha ( \alpha-1 ) \exp\left[-\frac{(\alpha - a_0)^2}{2\sigma_0 ^ 2}\right ] = \tilde{i}_0(a_0 - 1)\exp[-\frac{b_0}{2\sigma_0 ^ 2}]\,,\end{aligned}\ ] ] where @xmath67 and @xmath68 ^ 2 $ ] . the usual way of plotting the isotropic background spectrum is @xmath69 vs @xmath24 in logarithmic scale ( as in fig .
[ ress ] ) . defining @xmath70 and @xmath71 ,
we can rewrite eq .
( [ apeq ] ) as @xmath72 where @xmath73 .
note that the limits of integration of eq .
[ apeq ] should strictly extend from @xmath74 ( instead of @xmath75 ) to @xmath76 , to ensure that all @xmath77 are positive and finite .
the result of eq .
( [ apeq ] ) is a good approximation as long as @xmath78 is high enough and @xmath79 is low enough so that @xmath40 has no appreciable power at @xmath80 .
equation ( [ anres ] ) consists of a parabolic convex part and a logarithmic concave part .
the parabolic part dominates at low energies and is responsible for the characteristic convex shape of the unresolved emission from a population of sources with power - law spectra and a finite - spread spectral index distribution ( see , e.g. , brecher & burbidge 1972 ; stecker & salamon 1996 a , b ; pohl et al .
the spectrum remains convex ( @xmath81 is positive ) for @xmath82 .
the curvature of the spectrum depends on the spread of the spectral index distribution , @xmath79 ( the second - order term coefficient in eq .
[ anres ] is @xmath83 ) .
consequently , an overestimation of the spread of the population spectral index distribution will lead to an overestimation of the curvature of the corresponding unresolved emission spectrum ( as pointed out by pohl et al .
1997 , pavlidou et al .
2006 , and venters & pavlidou 2007 ) . in the limit @xmath84 all sources have identical spectra of index @xmath78 , and eq .
( [ anres ] ) yields a power - law , @xmath85 or , equivalently , @xmath86 , as expected . for @xmath87 , eq .
( [ anres ] ) has a local minimum , at @xmath88 , which corresponds to the energy at which the population has the smallest energy flux contribution to the ( unabsorbed ) isotropic background .
|
the large majority of egret point sources remain without an identified low - energy counterpart , and a large fraction of these sources are most likely extragalactic .
whatever the nature of the extragalactic egret unidentified sources , faint unresolved objects of the same class must have a contribution to the diffuse extragalactic gamma - ray background ( egrb ) .
understanding this component of the egrb , along with other guaranteed contributions from known sources , is essential if we are to use this emission to constrain exotic high - energy physics . here , we follow an empirical approach to estimate whether a potential contribution of unidentified sources to the egrb is likely to be important , and we find that it is .
additionally , we show how upcoming glast observations of egret unidentified sources , as well as of their fainter counterparts , can be combined with glast observations of the galactic and extragalactic diffuse backgrounds to shed light on the nature of the egret unidentified sources even without any positional association of such sources with low - energy counterparts .
|
transition radiation detectors ( trds ) have a long tradition in direct cosmic - ray measurements .
they have been utilized as threshold detectors already by e.g. @xcite and @xcite , recently by heat @xcite , and now by the ams-02 @xcite experiment .
they have been used to measure the energy of highly relativistic cosmic - ray nuclei in several other experiments : crn @xcite , tracer @xcite , and cream @xcite .
the light weight per area and signal response at very high energies make transition radiation detectors valuable for the direct observation of cosmic radiation on balloon - borne or space - based experiments . transition radiation ( tr )
is emitted as x - ray photons @xcite by particles with lorentz factors @xmath1 above about 1000 , when they traverse boundaries of different dielectric constants . above the threshold the tr yield increases with lorentz factor , until it reaches saturation usually above @xmath4 .
the exact onset of tr , the shape of the response curve , and its saturation depend on the radiator materials and the detector configuration .
it is an advantage of trds that they can be calibrated with light particles over the whole range of their response , see e.g. @xcite .
about 100 boundary transitions are needed on average for a singly charged particle to emit one tr photon . in practice
this means that a lot of boundaries are needed in a trd ( realized with foam radiators ) and that an energy measurement is usually only possible for heavier nuclei that produce more signal ( the tr yield scales with @xmath5 ) . for more information on trds
see @xcite , @xcite , and references therein . in this paper
we investigate the possibility to use the trd of the ams-02 experiment for energy measurements of heavy cosmic - ray nuclei .
first , the ams-02 experiment is described .
then the response curve of the transition radiation detector is determined and the energy resolution calculated .
] the ams-02 experiment is a multi - purpose cosmic - ray detector installed on the international space station ( iss ) since may 16th 2011 .
its active detector elements consist of a silicon tracker within a strong magnetic field , a time - of - flight detector ( tof ) , a ring imaging cerenkov detector , an electromagnetic calorimeter ( ecal ) , and a transition radiation detector ( trd ) .
a detailed description of the detector is available by @xcite and references therein . a schematic view of the detector is shown in figure [ fig01 ] . for cosmic - ray ions a measurement of rigidity
is provided by the tracker .
the tracker consists of nine layers of silicon sensors , of which seven layers ( 2 to 8) are the inner tracker located within the magnetic field . two outer layers ( 1 and 9 )
complete the full tracker @xcite .
its limiting figure of merit is the maximum detectable rigidity ( mdr ) at which the uncertainty on the measured rigidity equals 100% . for protons
the mdr is about 1.8 tv @xcite , for helium it is about twice as high ( 3.6 tv ) . because of the long exposure time of ams-02 the measurement of cosmic - ray ions will be limited by this systematic rather than by statistics .
therefore it is worthwhile to investigate the possibility to extend the energy range with the trd .
the trd of ams-02 is described in detail by @xcite , @xcite , and @xcite .
it consists of 5248 proportional tubes of 6 mm diameter with a maximum length of 2 m. the tubes are filled with a xe / co@xmath6 gas mixture ( 90:10 ) .
they are arranged side by side in 328 modules that are mounted in 20 layers .
each layer is interleaved with a 22 mm thick fiber fleece radiator ( lrp375 ) with a density of @xmath7 g/@xmath8 . in each layer
the trd measures the sum of the specific ionization of the particle traversing the gas and the produced transition radiation , `` de / dx+tr '' . in the energy range of the positron fraction measurement @xcite , light particles , like electrons ,
emit transition radiation , in contrast to heavy particles , like protons .
this leads to a discrimination power between electrons and protons of up to @xmath9 for the trd alone . ]
figure [ fig02 ] shows the pulse height spectra of a single straw for low lorentz factor ( protons ) and high lorentz factor ( electrons ) particles with @xmath10 . at low lorentz factor
the signal in the trd is purely due to specific ionization and distributed according to a landau distribution . at high lorentz factor
the signal is superimposed with a tr signal .
the resulting signal distribution is greatly enhanced at high signal values compared with the landau distribution due to specific ionization only .
this difference allows the separation of protons and electrons .
however , when the particle s charge is known , the signal also encodes the energy of the nucleus , as the transition radiation yield depends on the lorentz factor @xmath1 .
the trd of ams-02 samples the signal distribution 20 times for every event .
the mean signal is then clearly larger for particles with high lorentz factor than for particles with low lorentz factor .
in order to exploit the tr emission to measure the energy of cosmic ray ions , the response curve of trd has to be precisely known . for this purpose ,
the trd was cross - calibrated with protons and electrons measured by the ams-02 tracker and ecal on the iss , to determine the onset of the tr , the increase rate of the trd signal with the lorentz factor , and its saturation .
a similar strategy has been used by the cream experiment @xcite , although it could not observe particles in the tr region . for the cross - calibration only events
are considered that pass strict quality cuts , ensuring that well reconstructed events with good energy determination are used .
singly - charged particles are selected with the tof and tracker .
the tracker also determines the charge sign of every particle . among the positively charged particles , protons
are by far the most dominant species and additional cuts based on the ecal reject a possible residual contamination of positrons . for negatively charged events
, the ecal is used to reject the small background of anti - protons from the sample in order to identify electrons . in this study
protons are used with a measured rigidity from 1 to 1000 gv in the tracker .
they thus span a range of about 1 to 1100 in lorentz factor @xmath1 .
electrons are used with a measured energy in the ecal above 0.7 gev , which corresponds to @xmath11 .
^.[fig03 ] ] the average trd signal for protons and electrons is shown in figure [ fig03 ] as a function of lorentz factor .
the signal , + `` de / dx+tr '' , is determined in each layer from the deposited energy and the corresponding track length `` dx '' determined from the precise tracking of the silicon tracker .
the proton data follow an expected trend for specific ionization in gas with a minimum around @xmath12 and a log - linear relativistic rise at higher energies . at around @xmath13
the onset of tr is already evident .
the increasing signal can then be traced using the electron data .
the response of the trd saturates above about @xmath14 .
the saturated , maximum trd signal is about 730 adc / cm compared to 200 adc / cm at the energy of minimum ionizing .
the response curve has been found to be stable in time and stable with respect to small changes in gas mixture that occur during the operation of the detector .
also shown in fig .
[ fig03 ] is a parameterization of the response curve .
the data have been fit by a log - linear function to which a modified fermi function was added .
the fermi function with variable width accurately describes the onset , increase and saturation of the tr signal .
the saturation of the response curve sets in beyond the tracker mdr for protons and heliumfor highly relativistic particles holds @xmath15 , in which the factor of two in mdr between protons and helium is canceled by the factor of two in @xmath16 .
thus , the mdr for both species is the same in terms of @xmath17 . ] .
it is therefore possible to extend the energy range of the ams-02 experiment with the trd beyond the limitations of the tracker , if the signal fluctuations are small enough to utilize the trd for an energy measurement .
for an individual measurement with up to 20 layers in the trd . the solid line is a fit to the data according to the expected trend ( see text).[fig04 ] ] the intensity of transition radiation scales with the square of the traversing particle s charge .
thus , the relative fluctuations decrease for heavier nuclei proportional to @xmath18 and the mean trd signal for one individual event can be more precisely determined for heavier nuclei .
this behavior can be seen in figure [ fig04 ] , in which the relative signal fluctuations are shown as a function of charge for nuclei with lorentz factors between 10 and 100 .
the trd signal at this energy range includes no transition radiation but is only due to specific ionization .
this allows the analysis to include nuclei up to boron ( @xmath3 ) in the dynamic range of the trd .
the fluctuations of the average trd signal are about 31% for @xmath10 , but decrease quickly to the level of 10% for heavier nuclei . the solid line in fig
. [ fig04 ] represents a fit to the fluctuations @xmath19 using ( see also @xcite ) : @xmath20 with the constants @xmath21 and @xmath22 . the expected asymptotic relative signal fluctuation for heavy nuclei is therefore about 9% and is due to uncertainties in tracking , gain calibration , and geometrical alignment of the trd . for protons and helium
the signal fluctuations have been found to be very stable over this energy range of @xmath1 between 10 and 100 . at much higher lorentz factors , where transition radiation is present
, the signal uncertainty for electrons is slightly smaller compared to the uncertainty of protons at low @xmath1 .
this behavior has previously been discussed by @xcite . nevertheless , the measured fluctuations of figure [ fig04 ] can here be used as an estimate of the detector performance . ] the signal fluctuations together with the response curve determine the energy resolution that can be reached with the trd . the calculated resolution for a measurement of the lorentz factor @xmath1 with the trd of ams-02
is shown in figure [ fig05 ] for helium and boron . for comparison
the resolution of the tracker for helium has been calculated in terms of lorentz factor and is also shown . for helium the trd performs better than the tracker above @xmath23 .
an energy measurement with the trd for helium seems possible from @xmath24 to 2500 , above which the uncertainty reaches 100% at about @xmath25 .
this could extend the energy range by a factor of 1.5 compared to the tracker . for boron ,
the trd reaches a best resolution of about 20% and a measurement may be possible in the range from @xmath24 to 5000 ( about 5 tev / nucleon in energy ) . the tracker resolution for cosmic - ray species heavier than helium expressed in terms of lorentz factor
is expected not to exceed the resolution for helium due to saturation effects in the silicon layers .
therefore , the tracker resolution for helium shown in fig .
[ fig05 ] may serve as a approximate comparison even for boron .
because of the worsening trd resolution towards lower energies a rejection of low - energy particles is needed .
this can be achieved by the tracker .
straight particle tracks in the central seven layers of the tracker ensure a rigidity larger than a couple hundred gv .
this can sufficiently reject background for the trd analysis from low energies that would otherwise spill into the high energy region and bias the measurement of cosmic - ray spectra . for helium , for which the signal fluctuations are largest and the signal distribution still has a significant landau tail towards higher signals , this rejection of low - energy background is even more important . in this case
the full tracker with all nine layers can be used to identify all low - energy spill over .
the response curve of the trd of ams-02 determined with proton and electron data shows that a measurement of lorentz factor @xmath1 is possible above about @xmath24 . for helium
the accessible energy range extends to about @xmath26 or 1.5 times the mdr of the tracker .
thus , a trd energy measurement may serve as an independent cross - check of a conventional tracker analysis . for heavier elements like boron or carbon the resolution of the trd improves , so that a measurement up to @xmath27 may be possible , greatly extending the systematic limitations of the silicon tracker . for a successful measurement of elemental spectra with the trd
a powerful rejection of low energy background events is imperative because of the poor energy resolution before the onset of tr .
this can be achieved with the central seven layers of the tracker .
compared to a conventional analysis based on the full tracker ( all nine layers ) a trd - based analysis will utilize a ten times larger geometrical acceptance , only using the trd , tof , and inner tracker .
another important aspect of an energy measurement with the trd is the dynamic range of the detector .
the dynamic range of the electronics is 12 bit and at nominal operations the most probable value ( mpv , not the mean value ) of the signal of a minimum ionizing @xmath10 particle is set to 60 adc counts ( this corresponds to 100 adc / cm and a mean value of about 200 adc / cm due to the asymmetric signal distribution ) . in this configuration
the saturated tr signals of high - energy helium ( @xmath28 ) is still well within the dynamic range . for a measurement of elements up to carbon ( @xmath29 ) the operation of the trd would have to be changed so that the mpv of the signal of minimum ionizing helium corresponds to 40 adc counts .
such a change in amplification of the initial ionization signal will change the overall normalization of the trd response , but its shape is expected to remain stable . ]
the potential of a trd - based measurement of heavy nuclei is illustrated in figure [ fig06 ] .
it shows the b / c ratio as measured by various experiments with the highest data points at about 1 tev / nucleon . the expected result with statistical uncertainties of a trd measurement with ams-02
is added for an exposure time of one year .
the measurement could reach an energy of 5 tev / nucleon with unprecedented statistical accuracy .
the expected result is based on published data on the carbon energy spectrum @xcite and the b / c ratio . a realistic live time and efficiency of the ams-02 detector
was assumed for the calculation .
it should be noted that an actual measurement of cosmic - ray fluxes will of course also suffer from systematic uncertainties , for example due to corrections for interactions in the detector and unfolding of the energy scale .
this project is supported by the german space agency dlr under contract number 50oo1109 .
the ams experiment is grateful for the support of nasa and doe .
m. aguilar et al . , first result from the alpha magnetic spectrometer on the international space station : precision measurement of the positron fraction in primary cosmic rays of 0.5 - 350 gev , phys . rev .
110 , 1041102 , 2013 p. maestro et al . , energy cross - calibration from the first cream flight : transition radiation detector versus calorimeter , proceedings of the 30th international cosmic ray conference , mexico , vol . 2 ( og part 1 ) , 333 - 336 , 2008 a. d. panov et al . , relative abundances of cosmic - ray nuclei b - c - n - o in the energy region from 10 gev /
n to 300 gev / n .
results from atic-2 ( the science flight of atic ) , proceedings of the 30th international cosmic ray conference , mexico , vol .
2 ( og part 1 ) , 3 - 6 , 2008
|
since may 2011 the ams-02 experiment is installed on the international space station and is observing cosmic radiation .
it consists of several state - of - the - art sub - detectors , which redundantly measure charge and energy of traversing particles . due to the long exposure time of ams-02 of many years the measurement of momentum for protons and ions
is limited systematically by the spatial resolution and magnetic field strength of the silicon tracker .
the maximum detectable rigidity for protons is about 1.8 tv , for helium about 3.6 tv .
we investigate the possibility to extend the range of the energy measurement for heavy nuclei ( @xmath0 ) with the transition radiation detector ( trd ) .
the response function of the trd shows a steep increase in signal from the level of ionization at a lorentz factor @xmath1 of about 500 to @xmath2 , where the transition radiation signal saturates . for heavy ions the signal fluctuations in the trd are sufficiently small to allow an energy measurement with the trd beyond the limitations of the tracker .
the energy resolution of the trd is determined and reaches a level of about 20% for boron ( @xmath3 ) . after adjusting the operational parameters of the trd
a measurement of boron and carbon could be possible up to 5 tev / nucleon .
transition radiation , proportional tubes , space - borne detector , cosmic rays 29.40.cs , 98.70.sa , 07.87.+v = 0.5 cm
|
3c 391 ( g31.9 + 0.0 ) , a supernova remnant ( snr ) with irregular morphology , has been observed in several electromagnetic bands .
radio observations with the vla reveal an elongated structure extending from the northwest ( nw ) to the southeast ( se ) , surrounded by a shell except on the se border , where it appears that the snr has broken out of a dense region into an adjacent region of lower density ( reynolds & moffett 1993 ) . in x - rays , both _ einstein _
( wang & seward 1984 ) and _ rosat _ ( rho & petre 1996 ) observations show that the centroid of the soft x - ray emission sits in the se region .
the _ asca _
observation reveals a hard x - ray enhancement in the nw region , and confirms the decrease in hydrogen column density across the remnant from nw to se ( chen & slane 2001 ) .
two hydroxyl radical 1720 mhz maser spots are found along the radio shell ( frail et al .
the masers and the strong enhancement in [ oi ] @xmath1 m emission near the northwestern edge ( reach & rho 1996 ) both provide evidence for shock interactions with a molecular cloud .
co and other molecular line observations place the location of the remnant at the southwestern ( sw ) edge of a molecular cloud ( wilner , reynolds , & moffett 1998 ; reach & rho 1999 ) .
3c 391 is an example of the `` thermal composite '' or `` mixed morphology '' category of supernova remnants , which also includes w28 , w44 , ic 443 , g349.7 + 0.2 , and others ( rho & petre 1998 ) .
they generate bright _ thermal _ x - ray emission interior to their radio shells , and have faint x - ray rims .
they are usually found to interact with adjacent molecular clouds , characterized by the hydroxyl radical maser emission ( green et al .
1997 , yusef - zadeh et al .
non - thermal diffuse x - ray emission inside snrs is widely believed to be synchrotron radiation from pulsar wind nebulae , while the nature of the internal thermal x - ray emission seen in mixed morphology remnants is still uncertain .
at least four distinct scenarios have been proposed to explain centrally - brightened x - ray morphology .
the first scenario is radiative cooling of the rim gas . under this hypothesis ,
the gas at the rim has been cooled down in the radiative stage of evolution , with a temperature so low ( @xmath2 ) that its x - ray emission is very weak , while the gas in the inner volume is still hot enough to emit strong x - rays ( e.g. harrus et al .
1997 , rho & petre 1998).the second mechanism invokes thermal conduction .
it is suggested that thermal conduction in the remnant can smooth out the temperature gradient from the hot interior to the cooler shell , and increase the central density in response to the associated change in pressure conditions .
this results in luminous x - rays in the interior , and is dominant in the radiative stage ( cox et al . 1999 , shelton et al .
the third scenario invokes cloudlet evaporation in the snr interior .
when a snr expands in an inhomogeneous interstellar medium ( ism ) whose mass is mostly contained in small clouds , the clouds engulfed by the blast wave can be evaporated to slowly increase the density of the interior gas ; as a result , the snr appears internally x - ray brightened ( white & long 1991 ) .
the fourth suggestion is that the mixed morphology is a projection effect . for shell - like snrs that evolve in a density gradient such as at the edge of a molecular cloud , if the line of sight is essentially aligned with the density gradient as well as the magnetic field
, the snrs will appear as thermal composites ( petruk 2001 ) . to date
, x - ray studies have not yet provided strong evidence favoring any one of these mechanisms .
here we report on high resolution _
chandra _ observations of 3c 391 aimed at addressing the nature of the centrally filled x - ray emission in this snr .
we observed snr 3c 391 with the advanced ccd imaging spectrometer ( acis ) on board the _ chandra _ x - ray observatory on 03 august 2002 ( obs .
i d 2786 ) for an exposure of 62 ks .
the target center ( @xmath3 , @xmath4 ) was placed @xmath5 off the telescope aim - point for an optimal coverage of the supernova remnant on the s3 ( # 7 ) ccd chip ( using very faint mode ) .
the level 1 raw event data were reprocessed to generate a new level 2 event file to capitalize on the chandra interactive analysis of observations ( ciao ) software package ( version 2.3 ) . in the reprocessing ,
bad grades have been filtered out and good time intervals have been reserved .
we also removed time intervals with significant background flares ( peak count rates @xmath6 and/or a factor of @xmath7 off the mean background level of the observation ) .
this cleaning , together with a correction for the dead time of the observation , resulted a net 60.7 ks exposure for subsequent analysis .
furthermore , a correction was made for the time - dependence of the acis gain . for imaging analysis , we produced exposure maps in 0.3 - 1.5 , 1.5 - 3.0 , and 3.0 - 7.0 kev bands .
these maps were used for flat - fielding , accounting for bad pixel removal as well as correcting for telescope vignetting and the quantum efficiency variation across the detector .
the images in the three bands are adaptively smoothed ( using ciao tool _ csmooth _ with signal - to - noise ratio of 3 ) and exposure - corrected . the tri - color x - ray image ( 0.3 - 1.5 kev in red , 1.5 - 3.0 kev in green , and 3.0 - 7.0 kev in blue ) , overlaid with the 1.5 ghz radio contours ( from moffett & reynolds 1994 ) ,
is shown in fig.1a . another tri - color image with 1.5 ghz emission in red , 0.3 - 3.0 kev in green , and 3.0 - 7.0 kev in blue
is shown in fig.1b .
a broad band ( 0.3 - 7.0 kev ) image overlaid with the radio contours are shown in fig.1c .
the two oh maser points ( frail et al . 1996 ) are labeled in these three maps .
we searched for point - like sources in three broad bands , 0.3 - 1.5 kev ( s ) , 1.5 - 7.0 kev ( h ) , and 0.3 - 7.0 kev ( b ) .
a combination of source detection algorithms were applied : wavelet , sliding - box , and maximum likelihood centroid fitting ( see wang , chaves , & irwin 2003 for details of the approach ) .
the estimation of the count rate of a source was based on the data within the 90% energy - encircled radius ( eer ) determined with the calibrated point spread function of the instrument .
the information obtained for the point - like sources detected this way on the s3 chip are summarized in table 1 . in order to produce a map of diffuse emission
, we removed the point - like sources with each source region as a circle of twice the 90% eer . after filling the source - removed `` holes '' ( in a swiss - cheese - like map ) by interpolating the intensity from the surrounding regions
, we produced a smoothed ( with count - to - noise ratio of 6 ) , exposure - corrected diffuse x - ray map in 0.3 - 7 kev ( fig.1d ) .
also shown in fig.1d are the regions used for spectral fitting ( see 2.2 ) .
these x - ray images display the se - nw elongated morphology , similar to that observed in the earlier radio ( moffett , & reynolds 1994 ) and x - ray ( rho & petre 1996 , chen & slane 2001 ) observations , but with considerable small - scale structure revealed by the superb angular resolution of _
chandra_. the x - ray maps reveal a highly clumpy structure of the remnant , with clumps or knots located in both the se and nw parts . a bright hard enhancement is peaked at ( @xmath8 ) near the the nw border ( see fig.1a ) ; this corresponds to region # 3 in fig.1d .
a complex mixture of knots is seen in the se part of the remnant interior , including at least four bright enhancements in all colors centered at @xmath9 ( region # 6 ) , ( @xmath10 , @xmath11 ( region # 7 ) , ( @xmath12 , @xmath13 ( region # 8) , and ( @xmath14 , @xmath15 ( region # 9 ) .
similar clumpy structures are also seen in the interior of snr n49b ( park et al .
several remarkable , very bright knotty features appear on the east and se border of 3c 391 .
also , the eastern knot seems to coincide with the location of the infrared source iras 18470 - 0100 ( whose fluxes are 1.32 , 4.66 , 23.8 , and 236 jy at 12 , 25 , 60 , and 100 @xmath16 m , respectively ) .
these bright knots on the border may be small clouds that have recently been shocked by the supernova blast wave . in the rgb 3-color map
fig.1a , yellow and light - green colors indicate strong soft emission in the se part , while fainter soft emission appears brown in color and is seen in the middle section of the elongated body of the remnant .
a faint arc appears in the rim of the northern brown patch .
this arc / shell seems to connect with another shell like structure at the northeastern ( ne ) rim , which is evident in radio ( fig.1b ) and discernible in x - rays , and harbors one of the previously detected oh masers ( frail et al .
1996 ) . on the opposite side ,
no shell - like x - ray structure can be discerned along the southern and sw border , but faint diffuse x - ray emitting gas seems to spill outside the radio border in the southwest . in fig.1a
, the green color representing the middle - band emitting gas almost pervades the whole postshock region of the nw part .
the pervasive green color in the nw , mixed with a small blue patch , shows hardening of the x - ray spectrum .
this is consistent with gradual increase in hydrogen column density across the remnant ( chen & slane 2001 ; see also 3.2 below ) .
the radio contour overlays make it very apparent that there is a shock interaction with a molecular cloud on the nw border because of the flat morphology of the radio emission . at the west rim ,
a soft x - ray brightened slab appears to be very close to ( just slightly behind ) the radio peak emission ( region # 12 ) , and may be related to a small dense region there .
a few point sources are seen in the remnant s geometric center and the south and west portions ( fig.1 and table 1 ) .
the small number of counts collected make it difficult to determine their nature .
a bright hard enhancement ( @xmath17 ) appears as an unresolved source on the nw border ( source # 17 in table 1 ; region # 5 in fig.1d ) .
we discuss the spectrum of this source below ( 3.3 . ) .
taking advantage of _ chandra _ s superb angular resolution , we have performed the spatially - resolved spectral analysis of 3c 391 . prior to the extraction of spectra , we removed the point - like sources detected on the s3 chip ( see table 1 ) except for weak sources with @xmath18 ( see table 1 for the detailed definition of the false detection probability @xmath19 ) and those with a high possibility of being clumps of diffuse emission ( judged by eye ) .
we defined 13 small - scale regions , 3 medium - scale regions , and the region of the entire remnant ( as diagramed in fig.1d ) for spectral investigation .
most of the small - scale regions are chosen to include the small features of x - ray enhancement such as the knots and the faint shell like structures .
two medium - size circles are used to compare the mean spectral variation between the inner se and nw portions .
the remaining area on the s3 chip outside region # 17 for the entire remnant was used for background . for each chosen region
, we use the ciao script _
acisspec _ to extract the spectrum and obtain the weighted response matrix .
the net count rates of these extracted spectra are tabulated in table 2 and table 3 .
corrections for the continuous degradation in the acis quantum efficiency have been applied to auxiliary response files .
all the spectra mentioned above were regrouped to contain at least 25 net counts per bin .
there are three distinct line features , mg he@xmath20 ( @xmath21 ) , si he@xmath20 ( @xmath22 ) , and s he@xmath20 ( @xmath23 ) in these spectra except the spectrum of a nw unresolved source ( region # 5 ) , confirming their thermal origin ( see fig . 2 and fig . 3 )
the fe l complex could have significant contribution in the range @xmath24-@xmath25 .
the emission diminishes rapidly above the photon energy @xmath26 and no fe k@xmath20 emission is observed . using the morrison & mccammon ( 1983 ) interstellar absorption
, we test to fit the thermal spectra with various collisional equilibrium ionization and non - equilibrium ionization ( nei ) models for the optically thin plasma of snrs available in the xspec code .
we find that our spectra can be best described by the vnei model with nei spectral version 2.0 , which uses the astrophysical plasma emission database ( aped ) to calculate spectrum .
considering the evident line emission from mg , si , s , and possibly fe l , we treat the abundances of these element species as free parameters . in most cases , however , thawing the abundances did not significantly improve the spectral fit relative to that using solar abundances .
for a few cases , especially for the emission from medium - scale regions and from the entire remnant , thawing the abundances do indeed improve the fitting .
the spectral fit results are tabulated in table 2 .
the spectrum of the nw unresolved source ( region # 5 ) can be best fitted with an absorbed power - law emission ( see the spectral fit parameters summarized in table 3 ) and can not be fitted with a blackbody emission .
the spectral fits show that the diffuse emission from various regions have ionization parameters ( @xmath27 ) close to or higher than @xmath28 .
this implies that the hot plasma in the snr is very close to , or is basically in , the ionization equilibrium .
the spectral fits also show that the diffuse emission from small - scale regions can be well fitted with solar abundances , and the emissions from larger regions can be fitted with abundances very close to solar values .
the temperature of the gas interior to the snr [email protected]@xmath30 , with only small fluctuations .
for instance , the lowest temperature ( @xmath31 ) is found at the faint se rim ( region # 13 ) , and the highest temperature ( @xmath32 ) is seen at the ne rim ( region # 10 ) .
the absorption column density is found to generally increase across the remnant from se to nw .
this is consistent with the conclusion of the _ rosat _ ( rho & petre 1996 ) and _ asca _ studies ( slane & chen 2001 ) , and is also consistent with the emission hardening shown in the previous section .
the absorption in the se part is about @xmath33 ( region # 14 ) . the highest absorption ( @xmath34 ) is found in a nw region ( region # 4 ) which is in blue in the tri - color image ( fig.1a ) .
this hardest region was also seen in the _ asca _ study ( see fig.3 in chen & slane 2001 ) . a very high absorption ( @xmath35 ) is also found at the westmost rim ( region # 12 ) .
the difference of mean @xmath36 between the central se region ( region # 14 ) and the central nw region ( region # 15 ) is @xmath37 .
such a variation in @xmath36 is consistent with the existence of a molecular cloud to the nw ( wilner et al . 1998 ;
reach & rho 1999 ) .
similar to the hydrogen column density distribution across the remnant obtained in previous x - ray studies , the @xmath36 values obtained in the spectral analysis are confirmed to increase from the se toward the nw .
the spatial analysis also shows an obvious spectral hardening in the nw part .
this could naturally be explained by the scenario that the nw part of the remnant is embedded in the molecular cloud ( reynolds & moffett 1993 ) .
the mean difference of @xmath36 between the se and nw interior portions , @xmath37 , is consistent with that found in the _ asca _ study , and implies a mean density of the ambient molecular gas @xmath38 of order @xmath39-@xmath40 on the assumption that the snr line - of - sight depth in the cloud is similar to the remnant radius ( chen & slane 2001 ) .
we also adopt the distance @xmath41 estimated there . according to the nei fit to the spectrum of the entire remnant ( region # 17 ) , the x - ray luminosity in 0.510 kev of the remnant is @xmath42 . from the volume emission measure @xmath43 , where @xmath44 is the filling factor of the x - ray emitting gas , the mean number density of the gas interior to the remnant is @xmath45 ( here @xmath46 has been used and a cylinder - like volume of a diameter @xmath47 and a height @xmath48 has been assumed ) .
the remnant volume is @xmath49 .
the mass of the x - ray emitting gas is @xmath50 .
raw estimates for the gas density of the regions used for spectral fit ( apart from # 17 , the entire remnant ) can also be given in terms of the volume emission measures .
we assume that the circular regions correspond to spheres , the elliptical regions to oblate spheroids , and the rectangular regions to shell structures .
the gas densities thus obtained are listed in the last column of table 2 .
the densest x - ray emitting gas is located in the se and eastern regions # 1 and # 2 that contain very bright knots ( with density @xmath51 ) .
the other bright knots ( regions # 3 in the nw and # 6 , # 7 , # 8 , and # 9 in the inner central se portion ) have a gas density @xmath52@xmath53 .
the regions along the remnant border ( # 10 , # 11 , # 12 , # 13 , and # 16 ) have a density @xmath54@xmath55 .
as mentioned in the introduction , four main mechanisms have been proposed to explain the x - ray morphology of thermal composite snrs : projection effect , radiative shells , thermal conduction , and cloudlet evaporation . here
we compare them with the properties found from our spatially - resolved spectral analysis .
the projection effect does not match the properties found in 3c 391 .
as mentioned above , the molecular cloud is located in the nw but the x - ray emission is not only enhanced in the nw half , but also in the se half .
moreover , the hydrogen column density increases along the se - nw elongation direction , so the density gradient of the ambient medium seems to be close to the projection plane , and hence will make the projection effect weak according to the petruk ( 2001 ) model .
though there is no report in the available literature of the presence of an hi shell around the remnant like that in w44 , another thermal composite snr ( koo & heiles 1995 ) , the diffuse filamentary near - infrared [ fe ii ] and the mid - infrared 12 - 18 @xmath16 m [ ne ii ] and [ ne iii ] emission along the nw radio shell provide some evidence for a radiative cooling at the rim ( reach , rho , & jarrett 2001 ) .
as viewed in fig.1 , the volume of the remnant as delineated by the radio shell ( including the regions of the near- and mid - infrared emission ) is basically filled with x - ray emitting gas . several examined portions of x - ray emission on the borders , such as the ne rim ( region # 10 ) , the northern rim ( region # 11 ) , the western rim ( region # 12 ) ,
the sw faint rim ( region # 16 ) , and the se rim ( region # 13 ) , which should be dominated by the postshock gas of the supernova blast wave ( and are indeed located at the edge of the radio shell ) , are found to be at temperatures higher than @xmath56 .
this indicates that a considerable amount of gas at the blast shock has not yet suffered heavy radiative cooling , and is capable of generating considerable x - ray emission along the remnant s border . if the gas density of these regions , @xmath54@xmath55 , is approximated as the postshock density , then the preshock intercloud medium ( icm ) has a density about @xmath57@xmath58 .
according to the infrared spectroscopic study , the ambient environment in the nw part has complicated physical properties , with at least three typical preshock components ( reach & rho 2000 ) . for the atomic component implied by the [ oiii ] and [ niii ] emission , the preshock density @xmath59 and the shock velocity @xmath60 ; for the moderate - density molecular component revealed by the [ oi ] , [ siii ] , and [ feii ] emission , the gas density @xmath61 and the shock velocity @xmath62 ; and for the dense clumps giving rise to h@xmath63 , oh , co , cs , and h@xmath63o molecular emission , the gas density @xmath64 and the shock velocity @xmath65 .
thus there may be two possibilities about the coexistence of the radiative filaments and the x - ray emitting gas along the nw rim : 1 ) the cooling shell may be in the early stage of its formation ; or 2 ) the radiative filaments may be caused by the propagation of the shock waves into some regions of higher density in the molecular cloud so that these regions of postshock gas have cooled down severely while most of the gas at the rim is still hot . in the first case , we note that the bright [ feii ] emission coincides precisely with the brightest radio bar along the western border , which is regarded as the interface between the remnant and the molecular cloud ( reach , rho , & jarrett 2001 ) , and that , as we find here , the soft x - ray brightened slab along the west border ( in region # 12 ) appears to be very close to , but slightly behind , the radio bar .
the shell formation time can be estimated according to @xmath66 ( cox et al .
if the mean preshock hydrogen nucleus density is @xmath67 ( twice the molecular density @xmath38 @xmath39-@xmath40 given above ) , @xmath68 ; if the preshock density is @xmath69 for the moderately dense cloud , @xmath70 . the remnant s age would not be much larger than these estimates . in the second case , the blast shock propagates in an inhomogeneous medium .
after a blast shock impacts a cloud , a transmitted shock ( at velocity @xmath71 ) is expected to move into the cloud .
the pressure balance between the shocked cloud and the shocked icm gives @xmath72 , where factor @xmath73 for a shock interaction with small clouds ( mckee & cowie 1975 ) and @xmath74 and 6 for a strong shock hitting a rigid cloud plane for @xmath75 and @xmath76 , respectively ( zeldovich & raizer 1967 ) .
the moderate - density ( @xmath77 ) cloud may , after it is shocked ( with @xmath78 ) , be responsible for the filamentary near- and mid - infrared emission and the tenuous atomic component ( @xmath79 ) may , after shocked , correspond to the postshock x - ray emitting gas detected by _
chandra_. for the western x - ray slab ( @xmath80 ) along the radiative filaments , the velocity of the blast wave in the hot gas is @xmath81 ( where the mean atomic weight @xmath82 ) , and the preshock density is @xmath83 .
thus we have @xmath84 , which suggests that the snr shock wave hits the plane surface of a dense cloud .
this is consistent with the flat morphology of the nw radio bar .
the spatially - resolved spectral analysis demonstrates that the hot gas over most interior regions of the snr is at similar temperatures of around 0.5-@xmath30 .
such a uniform distribution of the gas temperature inside the remnant is not consistent with the canonical sedov distribution in which the inner temperature is higher than that near the rim . on the other hand ,
the x - ray morphology of this remnant is not limb brightened , and this is inconsistent with the sedov solution as well .
cox et al . ( 1999 ) explain this sort of morphology ( as observed in w44 ) in terms of the effect of thermal conduction of the interior gas , which could prevent the formation of a hot vacuous cavity described by the adiabatic sedov solution and provide enough material in the center to make it bright in x - rays .
the thermal conduction scenario could apply to 3c 391 to some extent . if the suppression of conduction by magnetic fields can be ignored , the conduction timescale would be @xmath85 , where @xmath86 denotes the linear scale of the temperature gradient and the conductivity is given by @xmath87 with the coulomb logarithm @xmath88 ( spitzer 1962 ) .
if @xmath86 is taken to be of order the remnant s mean radius @xmath89 , the timescale would be @xmath90 , much larger than the remnant s inferred age .
it would , however , be much smaller than this in the early stage of the evolution in which @xmath91 would be higher and @xmath86 would be smaller .
if we look at small spatial scales of order the separation between clumps ( typified by @xmath92 ) , the timescale would be @xmath93 , comparable to the remnant s age ( see 3.1.4 . ) , and hence thermal conduction may play a role in smoothing of the interior temperature profile .
we note that x - ray clumps are also present in w44 ( shelton , kuntz , & petre 2004 ) where thermal conduction may be an efficient mechanism .
the fact that the hot gas in 3c 391 is near ionization equilibrium suggests that there may be relatively little newly shocked material and is consistent with the idea of some mechanism smoothing the properties .
more significantly , the presence of the radiative filaments along the nw border is also in favor of the thermal conduction model .
there are , however , disagreements in the distribution of physical properties between the observation and the model interpretation . in the thermal conduction scenario in which a centrally brightened x - ray morphology is reproduced ( cox et al . 1999 ,
shelton et al .
1999 ) , the central pressure is about 0.3 of that at the edge and the central density would be about much lower ( 0.13 times at radiative shell formation ) than the preshock icm density , but this is not consistent with the case of 3c 391 .
in fact , this remnant has a uniform temperature around 0.5-@xmath30 throughout and the inner mean density @xmath94 ( regions # 14 and # 15 ) is not smaller than the hot gas density @xmath54@xmath55 along the periphery . in the model
the temperature would be expected to decrease gradually away from the inner portions to the outer . unlike w44 in which the gas temperature is found to drop by a factor @xmath95 between the center and @xmath96 from the center ( shelton et al .
2004 ) , in 3c 391 our measurement does not suggest the temperature decrease even if the line - of - sight projection is taken into account . the inner temperature is not higher than the gas temperature in most portions along the rim , and highest temperatures appear at a location on the border ( region # 10 , the ne rim ) . the phenomenon of low temperature in the inner portions could result from substantial cloudlet evaporation because part of the thermal energy is deposited to the gas evaporated from the cold cloudlets .
the relatively uniform distribution of temperature ( even with slightly lower values at the center ) is actually expected by the white & long ( 1991 ) cloudlet evaporation model for model parameters @xmath97 and @xmath98 ( see fig.2 and fig.4 therein ) , where @xmath99 is the ratio of the cloud evaporation timescale to the snr s age and @xmath100 is the ratio of the mass in the cloudlets to the mass of icm .
the ratio between the mean density ( @xmath101 ) and the density along the border ( @xmath54@xmath55 ) is basically consistent to that predicted in the evaporation model . in fact , the highly clumpy structure unveiled in this observation lends support to the conjecture that the ambient molecular cloud is inhomogeneous .
the cloudlets engulfed by the supernova blast wave can act as a large reservoir of interior gas by gradual evaporation .
this could also explains why the mean density of the hot gas inside the remnant ( as mentioned above ) is much lower than that of the ambient cloud gas .
the association between the thermal composites and the oh maser emission in molecular clouds is strongly suggestive of the role played by clouds in the snrs ( yusef - zadeh et al .
2003 ) . with a little adjustment of the model parameters used in the _ asca_study of 3c 391 ( chen & slane 2001 , see table 6 therein )
, we here adopt @xmath10236 . using the temperature measured from the entire remnant ( region # 17 ) @xmath103 as the average temperature
, the white & long ( 1991 ) model would predict a postshock temperature 0.46@xmath104 , consistent with most of the postshock temperatures measured from the spectral analysis .
the velocity of the blast wave is @xmath105@xmath106 ( where the mean atomic weight @xmath82 ) . adopting a mean radius of the remnant @xmath107
, we estimate the dynamical age of the remnant : @xmath108@xmath109 . because our _ chandra _ observation can measure the postshock temperature of 3c 391 , this age estimate should be more accurate than those obtained from _ rosat _ and _ asca _
observations using the same model , and the age of remnant is somewhat smaller than previously thought . the preshock gas density ( @xmath110@xmath58 ) estimated above could be regarded as the preshock icm density .
an alternative estimate of the preshock icm density is obtained from the remnant s x - ray luminosity ( @xmath111 ) : @xmath112@xmath113 by use of eq.(21 ) in white & long s ( 1991 ) model . according to this model , the explosion energy of the supernova remnant is @xmath114({r_{s}}^5/t^2 ) \sim0.3$]@xmath115 ( where the adiabatic index @xmath116 and the ratio of thermal to kinetic energy @xmath1170.132 ) .
the application of the evaporating cloudlets model to 3c 391 may raise a few questions , which are discussed below and need further clarification .
first , the remnant ages may be systematically underestimated by the white & long model for a given explosion energy .
w44 is a typical example , where an age of @xmath118 estimated from the evaporation model ( rho et al .
1994 ) is smaller than the spindown time scale @xmath119 of the associated pulsar ( wolszczan , cordes , & dewey 1991 ) .
the age derived above for 3c 391 ( @xmath120 ) using the evaporation model may possibly be an underestimate .
if a sedov evolution is adopted , the postshock temperature @xmath121 would correspond to a similar age @xmath122 ; as a comparison , reynolds & moffett ( 1993 ) present an estimate @xmath123 for a cloud density @xmath69 .
considering the possibility of the snr being in the radiative phase , chen & slane ( 2001 ) obtain an age @xmath124 for a preshock density @xmath67 .
second , the cloudlet evaporation scenario needs a model parameter @xmath97 ( with @xmath98 ) , and thus the influence of cloud evaporation on the properties of the interior gas seems possibly negligible .
similar parameters are required to explain the x - ray morphology of another thermal composite msh11 - 61a observed with _
( slane et al . 2002 ) .
in contrast , hydrodynamic simulations ( e.g. , xu & stone 1995 , using a density contrast @xmath125 ) show that clouds tend to get disrupted soon after passing through a strong shock . in our observation of 3c 391 , however , the se bright knots ( in region # 1 ) appear compact and survive the shock passage .
the ratio of the hot gas density of these knots to the mean density of the entire remnant is about 10 ( see table 2 ) , so it would be reasonable to assume that the cloud - to - icm density contrast @xmath126 is much higher than this number or to assume a high mass contrast @xmath100 .
the other x - ray knots interior to the remnant may also be the dense regions that have not been destroyed by the blast shock .
the denser clouds could be expected to survive longer time , and still a large amount of dense matter expected to have not been evaporated from cloudlets yet .
infrared observations of 3c 391 have detected dense ( @xmath127 ) molecular clumps in the southwest `` broad molecular line '' region , each h@xmath63 clump being of size @xmath128 ( reach & rho 1999 , reach & rho 2000 , reach et al .
( as a comparison , the compact se knots in region # 1 are of size @xmath129 . )
it is not impossible that the x - ray knots contain similar dense cloudlets .
third , if the bright knots stands for the newly evaporated cloud material , it should have been seen cooler than the mean temperature . nonetheless ,
because the evaporated gas would be soon heated up to the environmental temperature about 2 times the cloud radius from the cloud ( cowie & mckee 1977 ) , the spectral properties of the cooler component could be smoothed out if the knots are much larger than the cold cloudlets ( for instance , 0.5 pc versus 0.1 pc as described above ) .
fourth , the newly evaporated gas should have a low ionization age .
the knots in regions # 2 , # 6 , # 7 , # 8 , and # 9 marginally satisfy this condition ( @xmath130 ) , especially if the evaporated gas had been pre - ionized by the ultraviolet photons and/or the transmitted shock wave , while those in region # 1 has a high ionization age ( @xmath131 ) .
the latter case does not satisfy the condition , unless , again , the low ionization age could be smeared by the surrounding gas if the cold clouds are very small . concluding this section
, we suggest that three mechanisms , radiative rim , thermal conduction , and cloudlet evaporation , all play roles in the inner brightened x - ray morphology of 3c 391 . by comparison , the cloudlet evaporation can largely explain some important properties such as the highly clumpy structure , the uniform temperature and density distribution almost all over the remnant , and the much lower x - ray emitting gas density than the mean ambient cloud density .
nonetheless , even cloudlet evaporation is driven by saturated conduction .
it is probably true in many thermal composite remnants , and in 3c 391 as well , that the morphology arises from a combination of expansion into an inhomogeneous medium and heat conduction on various scales . as in many other thermal composites , hydroxyl
radical 1720 mhz maser spots are found within 3c 391 .
the two maser spots are located on the radio shell ( frail et al .
1996 ) , one of which is here found to be harbored in the ne rim ( region # 10 ) where x - ray emission is modestly enhanced at a high plasma temperature ( @xmath32 ) and the other of which is in the southwest region ( # 16 ) where the x - ray emission is faint ( see fig.1 ) .
the oh ( 1720 mhz ) maser spots are believed to be due to c - type shock collisional excitation in the clumpy molecular clouds , with cloud temperature and density conditions @xmath132 and @xmath133 ( elitzur 1976 ; lockett , gauthier , & elitzur 1999 ) . the production of oh molecules is enhanced by the x - ray emission ( with an ionization rate @xmath134 ) behind the c - shock ( wardle 1999 ) . according to wardle s model formula , the x - ray induced ionization rate in the postshock regions of 3c 391 is about @xmath135 ( using the x - ray luminosity @xmath136 and the mean radius @xmath137 ) , well above the rate needed for the oh enhancement .
let us take the x - ray enhanced region # 10 and x - ray faint region # 16 as oblate spheroids .
thus the volume in region # 10 is @xmath138 and the mean gas density is @xmath139 .
so , the pressure in the hot intercloud gas , @xmath140 . similarly , for region # 16 , the volume is @xmath141 , the mean gas density is @xmath142 , and the gas pressure is @xmath143 .
such gas pressures are consistent with the that estimated for the maser portions , @xmath144@xmath145 .
the far - infrared h@xmath146o and oh emission lines are found around the sw maser site .
they are consistent with the passage of shock wave through dense clumps and the postshock gas is estimated to have a density @xmath147 and temperature 1001000 k ( reach & rho 1998 ) .
such density and temperature are similar to the condition for the maser production , with a pressure range @xmath148@xmath149 , comparable to the above pressure estimates for the sw maser region .
the association of 3c 391 with a dense molecular cloud makes it possible that 3c 391 is the interstellar remnant of the supernova explosion of a massive star , which may also leave behind a compact star as a result of gravitational core collapse .
it is thus important to see if there is any evidence for this stellar remnant .
we do locate numerous point - like sources within the boundary of the remnant ( for example , evident point sources # 22 and # 26 right in the center , # 24 and # 32 in the south , etc . )
( see fig.1d and table 1 ) .
however , spectral analysis for these sources can not be carried out because of the small number of counts . a bright unresolved source ( source # 17 )
appears near the nw rim and seems to be located at the north apex of the nw diffuse hard emission ( the blue patch in fig.1a and fig.1b ) .
the spectrum of this source extracted from region # 5 ( with 158 net counts ) is better described with a powerlaw than with a bremsstrahlung or a blackbody ( see fig.4 and table 3 ) .
the hydrogen column density obtained is similar to those of other portions of this remnant , so the possibility of the association of this source with the remnant can not be ruled out .
if this source is at the distance of 3c 391 , its unabsorbed x - ray luminosity is @xmath150 .
the steep photon index of the possible powerlaw model 2.5 - 4.5 is higher than the typical indices 1.2 - 2.2 of active galactic nuclei ( turner & pounds 1989 ) , but is similar to those of the some `` compact central objects '' within snrs ( e.g. pavlov et al .
if the association with the gaseous remnant is true , the stellar remnant may have moved away from its explosion site .
if we follow reynolds & moffett s assumption that the explosion site is located at the center of the nw half , then the source would have displaced 2 toward the border .
taking an age of @xmath0 , then the displacement velocity would be @xmath151 .
such a transverse velocity is quite high , but not unreasonable ; pulsar proper motion studies have provided evidence that a small population of neutron stars have velocities in excess of @xmath152 ( lai 2003 ) .
for example , the pulsar b2224 + 65 in the guitar nebular pulsar has a velocity @xmath153@xmath154 ( chatterjee & cordes 2004 ) .
since the age used may be an underestimate , a higher remnant age would allow for a lower velocity . if the unresolved source is an offset pulsar , a trailing pulsar wind nebula might be expected to be present to the south , roughly in the hard emission region ( the blue patch mentioned in 2.1 ) .
however , the spectrum of the diffuse gas of the blue patch ( region # 4 ) is well described by a thermal model ( table 2 ) .
if a power law component is added to fit the spectrum , the photon index would be @xmath155 , too steep to be physically true .
since the spectrum of region # 4 is very week below 1.3 kev ( fig.2 ) and the hydrogen column density is very high ( @xmath156 ) , this hard region seems to be the diffuse hot gas suffering heavy absorption in soft x - rays .
we observed the thermal composite supernova remnant 3c 391 for 60 ks using _ chandra _ acis - s detector and carried out a detailed spectro - spatial x - ray analysis which results in following conclusions . 1 .
the southeast - northwest elongated morphology is similar to that previously found in radio and x - ray studies .
faint shell - like x - ray structures appear along the northern and ne boundary and a faint slab appears at the west rim , while on the sw side the diffuse gas seems to expand out of the radio boundary .
the x - ray emission is hardened in the nw , consistent with the increase of hydrogen column density from se to nw , and all the images and spectra show compellingly that the intervening column density is the primary influence on the broad band x - ray appearance of the remnant .
2 . this observation unveils a highly clumpy structure in the remnant .
the spatially resolved spectral analysis for the small - scale features found that the gas is generally of normal metal abundance and has approached or basically reached ionization equilibrium .
3 . three mechanisms , radiative rim , thermal conduction , and cloudlet evaporation , may all play roles in the x - ray visage of 3c 391 as a
thermal composite " snr , but there are difficulties with each of them in explaining some physical properties . by comparison , the cloudlet evaporation model can largely explain some important properties such as the highly clumpy structure , the uniform temperature and density distribution almost all over the remnant , and the much lower x - ray emitting gas density than the mean ambient cloud density . the enhanced emission in both the se and nw halves as well as the hydrogen column density gradient does not support the projection effect model .
the directly measured postshock temperature implies a young age , @xmath0 , for the remnant .
the postshock gas pressure derived from the ne and sw rims , which harbor maser spots , is consistent with the estimate for the maser regions .
an unresolved source on the northwest border is best fitted in spectrum by a power - law , and its position and absorption column density might suggest a possibility for an association with 3c 391 .
yc and ys acknowledge support from nsfc grants 10073003 & 10221001 and grant nkbrsf - g19990754 of china ministry of science and technology , pos acknowledges support from nasa contract nas8 - 39073 and grant go2 - 3081x , and qdw acknowledges nasa ltsa grant nag5 - 8935 .
the authors thank an anonymous referee for helpful comments and advices .
chatterjee , s. & cordes , j.m .
2004 , , 600 , l51 chen , y. & slane , p.o .
2001 , , 563 , 202 cowie , l.l . &
mckee , c.f .
1977 , , 211 , 135 cox , d. p. , shelton , r. l. , maciejewski , w. , smith , r. k. , plewa , t. , pawl , a. , & rzyczka , m. 1999 , , 524 , 179 elitzur , m. 1976 , , 203 , 124 frail , d.a . ,
goss , w.m . ,
reynoso , e.m . ,
giacani , e.b . ,
green , a.j . , & otrupcek , r. 1996 , , 111 , 1651 goss , w.m . ,
winkler , p.f .
, & williams , b.f . 1997 , , 113 , 1379 green , a. j. , frail , d. a. , goss , w. m. , otrupcek , r. 1997 , , 114 , 2058 harrus , i. m. , hughes , j. p. , singh , k. p. , koyama , k. , & asaoka , i. 1997 , , 488 , 781 koo , b .- c . & heiles , c. 1995 , , 442 , 679 lai , d. 2003 , in 3d signatures of stellar explosion , a workshop honoring j.c .
wheeler s 60th birthday ( astro - ph/0312542 ) lockett , p. , gauthier , e. , & elitzur , m. , 1999 , , 511 , 235 mckee , c.f . & cowie , l.l .
1975 , , 195 , 715 moffett , d. a. , & reynolds , s. p. , 1994 , , 425 ,
668 morrison , r. , & mccammon , d. , 1983 , , 270 , 119 park , s. , hughes , j.p . , slane , p.o . ,
burrows , d.n . , warren , j.s .
, garmire , g.p .
, & nousek , j.a .
2003 , , 592 , l41 pavlov , g.g . ,
sanwal , d. , kiziltan , b. , & garmire , g.p.2001 , , 559 , l131 petruk , o. 2001 , , 371 , 267 reach , w.t . , & rho , j.h . , 1996 , , 315 , l277 , 1998 , , 507 , l93 , 1999 , , 511 , 836 , 2000 , , 544 , 858 reach , w.t . ,
rho , j.h .
, & jarrett , t.h . , 2002 , , 564 , 302 reynolds , s.p . , & moffett , d.a .
1993 , , 105 , 2226 rho , j.h . , & petre , r. 1996 , , 467 , 698 rho , j.h . , & petre , r. 1998 , , 503 , l167 rho , j.h . ,
petre , r. , schlegel , e.m .
, & hester , j.j .
1994 , , 430 , 757 shelton , r. l. , cox , d. p. , maciejewski , w. , smith , r. k. , plewa , t. , pawl , a. , & rzyczka , m. 1999 , , 524 , 192 shelton , r. l. , kuntz , k. d. , & petre , r. 2004 , , accepted , astro - ph/0407026 slane , p. , smith , r. , hughes , j.p . , & petre , r. 2002 , , 564 , 284 spitzer , l. jr . 1962 , physics of fully ionized gases ( new york : interscience ) turner , t.j . & pounds , k.a .
1989 , , 240 , 833 wang , q.d . ,
chaves , t. , & irwin , j.a .
2003 , , 598 , 969 wang , z.r .
& seward , f.d .
1984 , , 279 , 705 wardle , m. 1999 , , 525 , l101 wilner , d.j .
, reynolds , s.p . , & moffett , d.a .
1998 , , 115 , 247 white , r.l . , &
long , k.s .
1991 , , 373 , 543 wolszczan , a. , cordes , j.m . , &
dewey , r.j .
1991 , , 372 , l99 xu , j. & stone , j.m .
1995 , , 454 , 172 yusef - zadeh , f. , wardle , m. , rho , j. , & sakano , m. 2003 , , 585 , 319 zeldovich , ya .
b. & raizer , yu . m. 1967 , physics of shock waves and high - temperature hydrodynamic phenomena ( new york : academic press ) lrrrrr 1 & j184915.0 - 005651 & 0.3 & @xmath157&@xmath158 & @xmath159 + 2 & j184916.1 - 005827 & 0.3 & @xmath157&@xmath160 & + 3 & j184916.2 - 005624 & 0.4 & @xmath157&@xmath161 & @xmath162 + 4 & j184916.8 - 005841 & 0.5 & @xmath163&@xmath164 & + 5 & j184917.0 - 005701 & 0.7 & @xmath165&@xmath166 & + 6 & j184917.4 - 005710 & 0.4 & @xmath157&@xmath167 & + 7 & j184917.4 - 005625 & 0.6 & @xmath168&@xmath169 &
+ 8@xmath170 & j184917.5 - 005900 & 0.4 & @xmath171&@xmath172 & + 9 & j184918.9 - 005252 & 1.0 & @xmath165&@xmath173 & + 10 & j184919.0 - 005749 & 0.3 & @xmath157&@xmath174 &
+ 11@xmath170 & j184920.4 - 005928 & 0.5 & @xmath175&@xmath176 &
+ 12@xmath170 & j184921.1 - 005357 & 0.7 & @xmath177&@xmath178 &
+ 13@xmath170 & j184921.5 - 005413 & 0.6 & @xmath177&@xmath179 &
+ 14@xmath170 & j184921.6 - 005355 & 0.6 & @xmath180&@xmath181 &
+ 15@xmath170 & j184921.9 - 005351 & 0.6 & @xmath182&@xmath183 &
+ 16@xmath170 & j184922.0 - 005946 & 0.8 & @xmath184&@xmath185 &
+ 17@xmath170 & j184922.3 - 005334 & 0.3 & @xmath157&@xmath186 & @xmath187 + 18 & j184922.6 - 005458 & 0.5 & @xmath188&@xmath189 & @xmath190 + 19@xmath170 & j184923.1 - 005736 & 0.5 & @xmath191&@xmath192 & + 20 & j184923.7 - 005705 & 0.3 & @xmath193&@xmath194 & + 21 & j184924.5 - 005447 & 0.5 & @xmath180&@xmath195 & + 22 & j184925.0 - 005628 & 0.2 & @xmath157&@xmath196 & @xmath197 + 23 & j184925.5 - 005344 & 0.5 & @xmath157&@xmath198 & + 24 & j184925.7 - 005749 & 0.2 & @xmath157&@xmath199 & @xmath200 + 25 & j184926.1 - 010024 & 0.4 & @xmath157&@xmath201 & + 26 & j184927.0 - 005640 & 0.2 & @xmath157&@xmath202 & @xmath203 + 27@xmath170 & j184927.0 - 005410 & 0.5 & @xmath204&@xmath205 &
+ 28@xmath170 & j184927.4 - 005349 & 0.5 & @xmath206&@xmath178 & + 29 & j184927.8 - 005318 & 0.8 & @xmath207&@xmath208 &
+ 30@xmath170 & j184927.9 - 005330 & 0.7 & @xmath204&@xmath209 &
+ 31@xmath170 & j184928.3 - 005734 & 0.4 & @xmath210&@xmath194 &
+ 32 & j184929.1 - 005741 & 0.2 & @xmath157&@xmath211 & @xmath212 + 33 & j184929.4 - 005951 & 0.6 & @xmath157&@xmath192 & + 34 & j184929.5 - 005905 & 0.3 & @xmath157&@xmath213 & @xmath214 + 35@xmath170 & j184931.3 - 005924 & 0.6 & @xmath215&@xmath216 & + 36@xmath170 & j184932.7 - 005805 & 0.5 & @xmath217&@xmath218 & + 37 & j184933.9 - 005406 & 0.9 & @xmath219&@xmath220 & + 38 & j184934.4 - 005238 & 0.9 & @xmath221&@xmath222 & + 39 & j184937.0 - 005825 & 0.4 & @xmath223&@xmath224 & + 40@xmath170 & j184937.3 - 005438 & 1.3 & @xmath225&@xmath218 & + 41@xmath170 & j184938.5 - 005848 & 0.4 & @xmath226&@xmath227 & @xmath228 + 42@xmath170 & j184938.6 - 005629 & 0.5 & @xmath229&@xmath230 & @xmath231 + 43@xmath170 & j184938.6 - 005845 & 0.4 & @xmath232&@xmath233 & @xmath234 + 44@xmath170 & j184939.6 - 005820 & 0.4 & @xmath235&@xmath236 & @xmath237 + 45 & j184939.7 - 005721 & 0.6 & @xmath238&@xmath239 & @xmath240 + 46@xmath170 & j184939.7 - 005817 & 0.5 & @xmath229&@xmath241 & @xmath242 + 47@xmath170 & j184939.8 - 005849 & 0.5 & @xmath243&@xmath244 & @xmath245 + 48 & j184939.9 - 005532 & 0.4 & @xmath157&@xmath246 &
+ 49@xmath170 & j184940.0 - 005848 & 0.3 & @xmath157&@xmath247 & @xmath248 + 50@xmath170 & j184940.0 - 005643 & 0.3 & @xmath157&@xmath249 & @xmath250 + 51 & j184941.4 - 005441 & 0.9 & @xmath251&@xmath252 & @xmath253 + 52 & j184944.7 - 005428 & 0.4 & @xmath157&@xmath254 & + cc|cccccc|c 1 & @xmath255 & 90.5/65 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & 0.84 & 18 + 2 & @xmath260 & 87.0/79 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & 1.31 & 11 + 3 & @xmath265 & 68.8/48 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & 1.05 & 6.4 + 4@xmath270 & @xmath271 & 108.9/79 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & 2.95 & 4.8 + & & + 6 & @xmath276 & 65.2/58 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & 0.97 & 6.1 + 7 & @xmath281 & 57.8/58 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & 1.20 & 6.8 + 8 & @xmath286 & 72.4/52 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & 0.95 & 5.4 + 9 & @xmath291 & 58.2/43 & @xmath292 & @xmath293 & @xmath294 & @xmath295 & 0.62 & 4.5 + 10 & @xmath296 & 44.2/44 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & 0.40 & 2.3 + 11 & @xmath301 & 96.9/72 & @xmath256 & @xmath302 & @xmath303 & @xmath304 & 0.94 & 1.9 + 12 & @xmath305 & 49.8/32 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & 0.68 & 2.0 + 13 & @xmath310 & 107.0/82 & @xmath297 & @xmath311 & @xmath312 & @xmath313 & 2.27 & 2.6 + 14@xmath270 & @xmath314 & 273.6/180 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & 10.8 & 2.9 + & & + 15@xmath270 & @xmath319 & 237.8/163 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & 10.2 & 2.8 + & & + 16@xmath270 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & 2.94 & 1.1 + & & + 17@xmath270 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & 45.5 & 1.9 + & & & cc|cccccc 5 & @xmath336 & power & 4.08/4 & @xmath337 & @xmath338 & & 9.1 + & & bremsstrahlung & 4.86/4 & @xmath339 & & @xmath340 & 5.5 + & & blackbody & 6.95/4 & @xmath341 & & @xmath342 & 3.3 +
fig.1 ( f1.jpg )
|
we present a 60 ks _ chandra _ acis - s observation of the thermal composite supernova remnant 3c 391 .
the southeast - northwest elongated morphology is similar to that previously found in radio and x - ray studies .
this observation unveils a highly clumpy structure of the remnant .
detailed spatially resolved spectral analysis for the small - scale features reveals that the interior gas is generally of normal metal abundance and has approached or basically reached ionization equilibrium .
the hydrogen column density increases from southeast to northwest .
three mechanisms , radiative rim , thermal conduction , and cloudlet evaporation , may all play roles in the x - ray appearance of 3c 391 as a thermal composite " remnant , but there are difficulties with each of them in explaining some physical properties .
comparatively , the cloudlet evaporation model is favored by the main characteristics such as the highly clumpy structure and the uniform temperature and density distribution over most of the remnant .
the directly measured postshock temperature also implies a young age , @xmath0 , for the remnant . the postshock gas pressure derived from the ne and sw rims , which harbor maser spots , is consistent with the estimate for the maser regions .
an unresolved x - ray source is observed on the northwest border and its spectrum is best fitted by a power - law .
|
stars with masses @xmath0 have short kelvin times that enable them to reach the main sequence while they are still accreting @xcite .
the resulting nuclear burning leads to a huge luminosity , which produces a correspondingly large radiation pressure force on dust grains suspended in the incoming gas .
this force can exceed the star s gravitational pull , possibly halting accretion and setting an upper limit on the star s final mass .
early spherically symmetric calculations suggested that this phenomenon sets an upper limit on stellar masses of @xmath1 , @xcite for typical galactic metallicities .
more recent non - spherical calculations have loosened that constraint by considering the role of accretion disks @xcite .
disks reduce the effects of radiation pressure by concentrating the incoming matter into a smaller solid angle , increasing its ram pressure .
they also absorb stellar radiation in a thin ring and re - radiate it isotropically , casting a shadow of reduced radiation pressure .
even with a disk , however , radiation pressure can still be a significant barrier to accretion .
@xcite were able to form a 50 @xmath2 star only if it had a thin accretion disk with a radius @xmath3 au .
simulations by @xcite in 2d found limiting masses of @xmath4 before radiation pressure reversed the inflow .
however , observations show that considerably more massive stars exist , and their formation mechanism remains uncertain
. recent observations of massive protostars have added an element to this picture .
@xcite , @xcite and @xcite report interferometric measurements showing outflows from massive protostars with collimation factors of @xmath5 up to @xmath6 .
they conclude that high and low mass outflows have similar collimation factors , typically @xmath7 @xcite .
( the collimation factor is the ratio of the outflow s length to its width . )
@xcite show that the momentum of co outflows driven by massive stars scales with the bolometric luminosity of the source in the same manner as for low mass stars . from these observations ,
the natural conclusion is that low and high mass stellar outflows have a common driving mechanism and similar morphologies .
previous theoretical work on massive star formation did not include the effects of outflows , and therefore assumed that the stellar radiation was either isotropic ( e.g. jijina and adams 1996 ) or had only those anisotropies induced by the presence of a disk ( e.g. yorke and sonnhalter 2002 ) . in this letter , we calculate radiative transfer through dense envelopes accreting onto massive protostars , and we study the effects of outflow cavities in the envelopes on the radiation field . in [ models ]
we present our models for radiative transfer and for the protostellar environment .
we then present the results of our calculations in [ results ] and discuss conclusions in [ conclusions ] .
we used the monte carlo / diffusion radiative transfer code written by @xcite to find the temperature distribution of gas in a circumstellar envelope .
the opacity comes from dust grains , which are thermally well - coupled to the gas in the high density envelope @xcite .
the grain size distribution depends on the location in the flow .
the stellar spectrum is a kurucz model atmosphere . rather than discuss the code , grain , and stellar spectral models here , we refer readers to @xcite , and to [ outflow ] for a discussion of opacity in the outflow.we used @xmath8 photons , sufficient to produce high signal - to - noise spectra . to calculate the radiation force on the gas , we begin with the temperature distribution determined by the monte carlo code . from a given position , we pick ray directions based on the healpix scheme @xcite . along each ray , we solve the transfer equation for emission and absorption by dust grains to obtain @xmath9 , the specific intensity coming from direction @xmath10 .
we also perform this calculation along a ray to the star itself . once we have found the intensity along each ray , we compute the flux @xmath11 where @xmath12 is the number of rays .
we increase @xmath12 until the change in flux is less than 2.5% between iterations .
we then integrate over frequency to obtain the radiation force per unit mass , @xmath13 .
we ignore scattering when calculating the force ( but not the temperature ) .
scattering of ir photons is negligible .
uv and visible photons can scatter significantly , but for the fiducial envelope and star we use , photons at the peak of the stellar spectrum are all reprocessed into ir in the inner few au of the envelope .
( in the disk the distance is vastly smaller . ) outside this layer , neglecting scattering does not change the radiation pressure force . for the same reason , the choice of stellar spectrum model is relevant only in the inner few au of the envelope . to study the effects of outflow cavities , we choose a single fiducial model for the star , disk , and envelope , and vary the properties of the outflow .
we place an @xmath14 @xmath2 zero - age main sequence star in a 50 @xmath2 envelope .
the zams models of @xcite , predict a radius , surface temperature , and luminosity of @xmath15 @xmath16 , @xmath17 k , and @xmath18 @xmath19 .
@xcite predict a formation time of @xmath20 yr for such a star , so we adopt an accretion rate of @xmath21 @xmath2 yr@xmath22 .
the accretion luminosity is @xmath23 , negligible in comparison to the central star .
although the cores that form massive stars are probably turbulent @xcite , for simplicity we adopt a simple rotationally flattened density distribution given by @xcite @xmath24^{-1 } \\
u_r & = & -\left(\frac{2 g { m_{*}}}{r}\right)^{1/2 } \left(1+\frac{\cos\theta}{\cos\theta_0}\right)^{1/2 } \\ \label{tsc2 } \frac{{r_{\rm cen}}}{r } & = & \frac{\cos\theta_0 - \cos\theta } { \sin^2\theta_0 \cos\theta_0}.\end{aligned}\ ] ] here @xmath25 is the legendre polynomial , @xmath26 and @xmath27 give the position in the envelope in spherical coordinates , and @xmath28 is the centrifugal radius of the flow .
the gas at position @xmath29 was at angle @xmath30 when it began falling toward the star , where @xmath30 is given implicitly by ( [ tsc2 ] ) .
the centrifugal radius @xmath28 is determined by the angular momentum of material arriving at the star . when the last material in the core accretes , it is roughly @xmath31 , where @xmath32 is the ratio of the core s rotational kinetic energy to gravitational binding energy and for most cores @xmath33 @xcite . since our core is half accreted , we take @xmath34 .
we therefore adopt values of @xmath35 pc and @xmath36 au , which give the correct mass in the envelope and satisfy @xmath34 .
in addition to the envelope , our star has a disk of radius @xmath28 .
observations of disks around massive stars are limited , and we therefore adopt a disk mass of @xmath37 based on the consideration that a disk more massive of order 10% of the central object is likely to be subject to gravitational instabilities @xcite that will cause matter to accrete until marginal stability is restored .
we model the disk surface density and scale height as power laws @xmath38 , @xmath39 au .
we base the choice @xmath40 on images from simulations of massive star formation ( e.g. yorke & sonnhalter 2002 ) which show disks with little or no flaring .
we can partially describe the outflow cavity from observations . as discussed in
[ intro ] , high mass outflows have collimation factors from @xmath7 , similar to low mass stars .
this corresponds to half - angles from @xmath41 , including the uncertain inclination .
we therefore test opening angles of @xmath42 , @xmath43 and @xmath44 .
unfortunately , observations constrain only the asymptotic opening angle when the outflow is far from its parent core .
cavity walls from nearby low mass sources seen in near - ir are generally curved when viewed on sufficiently small scales ( e.g. padgett et al 1999 and reipurth et al 2000 ) .
this curvature is a natural result of outflow collimation by the rotationally - flattened envelope @xcite . for massive protostars that are still embedded in dense envelopes ,
near - ir observations are impossible due to dust extinction .
millimeter observations , even with interferometers , are unable to probe length scales comparable to the size of the protostellar disk , at which we expect the strongest collimation and curvature . following @xcite , therefore , we parameterize the uncertain shape of the cavity as @xmath45 , where @xmath46 is the vertical distance from the star , @xmath47 is the distance from the outflow axis , and @xmath48 is a constant chosen to give an opening angle of @xmath49 at the edge of the core .
we try values of @xmath50 , 1.5 , and 2.0 , coupled with a fixed wind opening angle @xmath51 , to study the effects of variations in cavity shape . as a baseline
, we also study a case with no outflow cavity .
some gas at the outflow base will be ionized by the star s uv flux ( tan & mckee , in preparation ) . whether the ionized gas remains near the star depends on the outflow s structure at its base , where the density is highest and recombinations are fastest .
if the outflow is ionized , it will be too hot to contain dust grains , and its opacity will come primarily from resonant scattering by metal ions @xcite .
this is a complex subject beyond the scope of this letter .
we simply note that the opacity produced by this scattering must be far smaller than that produced by dust grains in the envelope . in the case
that the outflow cavity is ionized , we may therefore set the opacity in the cavity to zero .
if the outflow is neutral , grains will re - form , growing in radius @xmath52 at rate ( cf .
hoyle 1946 ) @xmath53 where @xmath54 is the mass fraction of the element of which the grain is composed , @xmath55 is the thermal velocity of gaseous atoms of that element , @xmath56 is the gas density , and @xmath57 is the grain density .
growth is fastest for carbon grains , since carbon is the most abundant refractory metal in the galactic ism and has a large thermal velocity due to its low atomic mass .
@xcite estimate its mass fraction to be @xmath58 .
grains can only condense below the dust destruction temperature of @xmath59 k , so @xmath60 km s@xmath22 .
the density of carbon grains is @xmath61 g @xmath62 . for our adopted cavity shape ,
the wind density at a distance @xmath46 from the equatorial plane must be roughly @xmath63 $ ] , where @xmath64 is the wind velocity ( taken to be constant ) and @xmath65 is the fraction of mass reaching the star that is ejected into the wind . we take @xmath66 and @xmath67 km s@xmath22 @xcite , where @xmath68 is the keplerian velocity at the stellar surface .
the smallest distance at which grains can form is the dust destruction radius , which we calculate to be @xmath69 au for our fiducial model
. we may therefore integrate ( [ graingrowth1 ] ) from @xmath70 to @xmath71 to obtain the maximum size that grains can attain before escaping the core .
the largest grain sizes occur for @xmath72 and @xmath73 , which gives @xmath74 @xmath75 m in effect , grains can not grow past the size of molecules .
in contrast , the typical grain sizes in the envelope are @xmath76 @xmath75 m .
stellar and envelope radiation has wavelengths @xmath77 , so grain opacity scales as @xmath78 .
thus , even if all the c goes into grains , the opacity in the outflow is smaller than that in the envelope by at least a factor of @xmath79 .
if we included cooling of the outflow and evaluated the grain size at radii smaller than @xmath71 , the reduction would be even larger .
we conclude , therefore , that dust opacity in a neutral outflow is negligible in comparison to the opacity in the infalling envelope .
since for either a neutral or an ionized outflow the opacity is negligible , we set the opacity in the outflow cavity to zero in our calculations .
we show the gas temperature distribution for each of our models in figure [ tempmap ] .
although our core extends to 0.22 pc , we concentrate on the inner few thousand au , where the radiation pressure force is strongest .
the plot shows that , in models with a cavity , the gas is hotter on the inside edge of the outflow cavity but is cooler elsewhere . since , due to rotational flattening , most of the gas is attempting to accrete close to the equatorial plane , this means that the bulk of the accreting gas is cooler in runs with an outflow .
the effect is stronger for outflow cavities with larger opening angles , and for cavities that are wider at their base .
we show the radiation pressure force versus radius at angles of @xmath80 , @xmath81 , and @xmath82 from the pole in figure [ angplot ] .
consistent with the reduced temperature , the radiation pressure force is also substantially smaller in runs with an outflow cavity . for our `` intermediate '' model with @xmath51 , @xmath73 , the radiation pressure force is smaller than in the run with no cavity by a factor of up to 4.6 . for the other runs ,
the peak reductions in radiation force were factor of 1.7 ( @xmath83 , @xmath73 ) , 14.4 ( @xmath84 , @xmath73 ) , 4.2 ( @xmath51 , @xmath50 ) , and 7.2 ( @xmath51 , @xmath85 ) .
thus , the amount by which the wind cavity reduces radiation pressure shows moderate dependence on the wind curvature and opening angle , but for only one combination of parameters was the reduction less than a factor of 4 . comparing radiation and gravitational forces , it is clear that this reduction can mean the difference between accretion halting or continuing . to check how the results depend on properties of the envelope , which has a comparatively low surface density compared to most high - mass cores , we re - ran the intermediate and no cavity cases with identical @xmath28 and @xmath71 , but a 100 @xmath2 envelope .
the case with no wind cavity showed little change in force with envelope mass . with a cavity ,
the increased envelope mass decreased the radiation pressure force at equatorial angles so that radiation was stronger than gravity only at @xmath86 , increasing the fraction of solid angle through which accretion could occur .
thus , our results likely represent a _ lower limit _ on the reduction in radiation pressure force that outflow cavities actually produce . to determine how the effect of the cavity compares to that produced by the disk and the rotationally - flattened envelope alone
, we also considered a spherical envelope , with no cavity , no disk , and @xmath87 in ( [ tsc1])-([tsc2 ] ) .
we compared this to the run with our fiducial disk and envelope parameters and no cavity . at angles
@xmath88 , the radiation pressure force in the spherical case was a factor of a few higher , while at angles @xmath89 it was comparable or smaller . thus , our intermediate wind cavity reduces the radiation pressure force relative to the disk - and - envelope only case by about the same amount that the disk - and - envelope case reduces it relative to purely spherical .
collimation of the radiation field by the disk and envelope and collimation by the outflow cavity are about equally important , and reinforce each other .
together , they reduce the radiation pressure force in the intermediate cavity case by a factor of @xmath6 relative what one would find for an isotropic radiation field . in all our tests , the degree of radiation collimation is roughly consistent with expectations .
tan & mckee ( in preparation ) show that the fraction of radiation escaping an envelope through a path of optical density @xmath90 is proportional to @xmath91 . in our case ,
the flux fraction escaping from the core at angles between @xmath27 and @xmath92 should roughly satisfy @xmath93^{-1 } d\left[\cos(\theta)\right],\ ] ] where @xmath94 is the optical depth at angle @xmath27 .
there is considerably uncertainty in applying this to our problem because @xmath94 is frequency - dependent .
however , if we use the rosseland mean opacity at @xmath95 k , then at most angles ( [ fluxfrac ] ) predicts roughly the correct flux fraction .
we have shown that an outflow can substantially change the radiation field , and radiation pressure , around a massive protostar .
the outflow cavity provides an optically thin channel through which radiation can escape , significantly reducing the radiation pressure . with no wind cavity , in our fiducial model
the radiation pressure force is stronger than gravity essentially everywhere except inside the accretion disk , and it is therefore likely that accretion would be halted . in our intermediate model , outside the centrifugal radius radiation is weaker than gravity over about @xmath96 sr , providing a large funnel through which accretion can continue .
the calculation we have performed here is only a proof of principle , and we are currently performing 3-d radiation hydrodynamic amr simultions .
our results strongly argue , however , that the presence of outflows provides a mechanism for circumventing the radiation pressure limit to protostellar accretion .
surprisingly , outflows that drive gas out of a collapsing envelope may increase rather than decrease the size of the final , massive star .
we thank jonathan tan for useful discussions .
this work was supported by : nasa gsrp grant ngt 2 - 52278 ( mrk ) ; nsf grant ast-0098365 ( cfm ) ; nasa atp grant nag 5 - 12042 ( rik and cfm ) ; and the us department of energy at the lawrence livermore national laboratory under the auspices of contract w-7405-eng-48 ( rik ) .
grski , k. m. , hivon , e. , and wandelt , b. d. 1999 , in proceedings of the mpa / eso cosmology conference `` evolution of large - scale structure '' , eds .
banday , r.s .
sheth and l. da costa , ( printpartners : ipskamp , nl ) , pp .
37 - 42 ( also astro - ph/9812350 ) richer , j.s . ,
shepherd , d.s . , cabrit , s. , bachiller , r. , and churchwell , e. 2000 , in protostars and planets iv , eds . v. mannings , a.p .
boss , s.s .
russell , ( tucson : the university of arizona press ) , p. 867
|
we consider the effects of an outflow on radiation escaping from the infalling envelope around a massive protostar . using numerical radiative transfer calculations , we show that outflows with properties comparable to those observed around massive stars lead to significant anisotropy in the stellar radiation field , which greatly reduces the radiation pressure experienced by gas in the infalling envelope .
this means that radiation pressure is a much less significant barrier to massive star formation than has previously been thought .
|
this paper is a tribute to bruce shore s role in establishing the concept of atomic coherence and the related techniques of irreducible tensor operators @xcite .
we demonstate that this viewpoint emerges quite naturally when describing the evolution of a spinor condensate . during the past decade
condensed bose gases have energed as flexible test system to explore the rich structure of many - body physics .
a systematic introduction into cold gases in general and spinor condensates in particular may be found , e.g. , in the book by pethick and smith @xcite . with the development of purely optical traps @xcite
, we are in the position to hold atoms with their complete hyperfine submanifold , thus entering the regime of spinor physics @xcite .
this is exactly the condition under which one expects the techniques of atomic coherence @xcite to be most helpful .
there are two major approximations used in the literature when describing spinor condensates theoretically : the common mean - field approximation ( mf ) and the single mode aproximation ( sma ) . within the sma ,
the spatial dependence of the bose field operator describing the spinor gas is assumed given and decoupled from the internal dynamics , as illustrated schematically in fig .
[ schema ] .
the reason one can separate the local redistribution of particles amongst hyperfine sublevels due to collisions from the long range dynamics is clear : a typical trapped alkali atom is paramagnetic , but provides only a single bohr magneton of magnetic moment .
consequently the resulting magnetic dipolar interaction is very weak compared to the typical interactomic van - der - waals energy .
rb condensate in the single - mode approximation.,scaledwidth=50.0% ] the single mode approximation is commonly introduced on top of the mean - field approximation .
one exception is the paper by koashi and ueda @xcite where the sma is introduced directly into the many - body treatment ( see also @xcite ) .
this , and also neglecting the quadratic zeeman part of the energy , allowed the authors to construct an explicit solution to the problem including collisions . in this paper
we include the quadratic zeeman part which plays a crucial role in the dynamics of the condensate as realized with @xmath2rb or @xmath3na atoms @xcite .
consequently , we are left with an interacting three mode hamiltonian , whose dynamics will be the subject of study in this article .
it should be pointed out that our approach would not apply directly to the recently realised chromium bose - einstein condensate @xcite where the dipolar interaction is comparable to the van - der - waals energy .
even less so in the case of heteronuclear molecules pursued by many experimental groups around the world where the dipolar interaction is expected to be the dominant mechanism .
independence of the internal and spatial motion is completely lifted e.g. in the proposed einstein de haas effect @xcite .
the sma is obviously not aplicable under the conditions for condensate fragmentation and domain formation ( see @xcite and references therein ) . the article is organized as follows : in section [ secmanybody ] we review the derivation of the three mode spinor hamiltonian . for later reference
, we will also discuss the standard mean - field approximation in the language of hamiltonian mechanics in section [ secmf ] . visualizing the results of a full many - body evolution of the three mode system
is a nontrivial problem .
therefore , we introduce the concepts of irreducible tensor operators in section [ seccohev ] . finally , we discuss and compare the results of the many - body calculation in section [ secresdisc ] .
the @xmath4 bosons are described in the usual way by a field operator with the three spherical components , @xmath5 , where @xmath6 ( in this order ) labels the @xmath1 projection of the atomic spin .
the effective low energy hamiltonian of such bosons in a homogeneous magnetic field is written as : ( cf .
@xcite ) @xmath7here , @xmath8 \hat\psi_k({\mbox{\rm\boldmath$r$ } } ) \end{aligned}\]]contains the kinetic energy and the trapping potential which holds the atoms in place , and @xmath9describes the effects of the magnetic field . for weaker magnetic fields ,
this is mainly the linear zeeman effect , @xmath10 , and the quadratic zeeman , or paschen - back , effect , @xmath11 .
the latter is crucial for the purposes of our analyses .
the general breit - rabi formulas for @xmath12 may be found in ref .
@xcite .
both @xmath13 and @xmath14 describe single - body physics .
two - body collisions enter via @xmath15 and @xmath16 .
the former describes the familiar spin - independent density - density interaction , @xmath17where @xmath18 , and @xmath19 stands for the normal ordering of the field operators .
the latter describes the spin - dependent part of the collision , @xmath20where @xmath21 is the density of the angular momentum of the condensate , @xmath22_{kl}\hat\psi^{\dag } _
k({\mbox{\rm\boldmath$r$ } } ) \hat\psi _ l({\mbox{\rm\boldmath$r$ } } ) .
\end{aligned}\]]the cartesian components of the single - particle spin operator , @xmath23 , are defined as : @xcite @xmath24 the collisional interaction constants , @xmath25 and @xmath26 , are expressed in terms of the scattering lengths @xmath27 and @xmath28 , for two spin-1 atoms in the combined symmetric channels of total spins , respectively , 0 and 2 , as @xcite @xmath29 a typical value of @xmath26 is one - two orders of magnitude below @xmath25 .
this leads to a natural separation of time scales : on smaller time scales , one may neglect @xmath16 , whereas for longer time scales it may be of crucial importance .
the ultimate example of a longer time scale is the ground state of the condensate , which may be polar or ferromagnetic depending on the sign of @xmath26 @xcite .
an example of the opposite are experiments where the spinor condensate is prepared in a particular magnetic state , and the evolution of the condensate is then observed @xcite . in this case , an approximation commonly used in the literature is the so - called single mode approximation ( sma ) .
technically , sma is recovered by assuming that the spatial dependence of the field operator is fixed , @xmath30 , @xmath31 . under the sma ,
the hamiltonian of the system of @xmath32 spin - one particles becomes : ( cf .
@xcite ) @xmath33here , @xmath34 , @xmath35 , and @xmath36 .
the `` wide hat '' symbol denotes mapping of a single - body operator into the fock space , @xmath37with @xmath38 being the mode annihilation operators ; in our case , @xmath39 ( in this order ) . the `` wide hat '' should not be confused with the `` normal '' hat which just signifies that a quantity is an operator in the fock space .
in particular , @xmath40 is the full angular momentum operator , and @xmath41 is the fock - space operators of the total particle number , where @xmath42 is a 3@xmath433 unit matrix . the second line in ( [ eq : hfockspace ] ) is found using that , @xmath44 in numerical terms , constituents of ( [ eq : hfockspace ] ) are sparse matrices ; these are handled very efficiently by common high - level numerical packages .
some details on how these were implemented numerically may be found in appendix [ tofock ] .
in the mean field approximation , operators @xmath45 are replaced by complex numbers @xmath46 . however ,
this transition is not uniquely defined . using different orderings ( normal , symmetric , etc . ) of creation and annihilation operator yields different _ functional forms _ of the quantum hamiltonian . in quantum mechanics
these functional forms represent the same hamiltonian . in the mean field picture where ordering is immaterial for the complex numbers @xmath46
, @xmath47 , these different functional forms result in different classical hamiltonians . in the scalar case ,
the corresponding corrections are of the order of @xmath48 , where @xmath32 is the total number of atoms . for the spinor condensate ,
these corrections are of the order of one over the population of zeeman sublevels .
there can well be pathological cases when one level is much depleted , leading to large corrections .
it is therefore of importance to have the procedure of introducing the mean field approximation for the spinor condensate clearly defined .
there is a number of ways the mean field picture can be introduced .
the most rigorous one is via the phase - space technique ( see , e.g. , @xcite and references therein ) , where the time - dependent gross - pitaevskii equation appears on dropping quantum noises in the equations for the phase - space amplitudes .
this procedure is equivalent to using the normally - ordered form of the interaction hamiltonian , replacing the field operators by the c - numbers , and then _ postulating _ the resulting c - number function to be the classical hamiltonian . the lack of rigour is clearly seen in the fact that this procedure leaves it unclear how quantum averages are expressed by the c - number amplitudes .
some insight can be gained from the fact that if we postulate the many body state as a coherent state , @xmath49 , and then minimize the energy subject to the condition @xmath50 , the time - independent gross - pitaevskii equation is recovered .
this recipe is in fact general : a correct mean - field expression for the average of an operator @xmath51 is found as @xmath52 .
this is equivalent to replacing field operators by c - numbers in the _ normally ordered representation _ of @xmath51 . following the above recipe , we postulate the classical hamilton function to be , @xmath53 , \end{aligned}\label{eq : hamiltonfunction } \end{aligned}\]]where ( [ eq : hfockspace ] )
was used to find the explicit form of the hamilton function . to use the powerfull methods of analytical mechanics we have to know the canonical conjugated momentum @xmath54 to the coordinate @xmath55 . if , by definition , @xmath56 and @xmath57 are three - dimensional canonical conjugated variables , defining the complex variable @xmath55 @xmath58 than we can identify @xmath59 .
one can verify easily , that this is a proper canonical transform , as @xmath60 with the following conventional definition of the poisson brackets , @xmath61 so we obtain the hamilton equations of motion @xmath62 by employing complex coordinates we have doubled the dimension of the phase space making the two sets of equations of motion redundant .
explicitly , we find the three components of the time - dependent gross - pitaevskii equation @xmath63,\ \ h_1=\left [ \begin{matrix}|\zeta_0|^2+\mathcal{f}_z & \zeta_{-1}^\ast\zeta_0 & 0\\ \zeta_{-1}\zeta_0^\ast & |\zeta_{1}|^2+|\zeta_{-1}|^2 & \zeta_{1}\zeta_0^\ast\\0 & \zeta_{1}^\ast\zeta_0 & |\zeta_0|^2-\mathcal{f}_z\end{matrix}\right ] , \end{gathered}\ ] ] with @xmath64 . for the initial condition as in the experiment of the sengstock group @xcite , these nonlinear equations have been solved analytically in terms of the periodic jacobi elliptic functions , see @xcite .
consider now how an effective mean - field amplitude is recovered in the many - body techniques . were the population and coherence properties of the matter field investigated in a measurement
, all the relevant information enters through the single - body density matrix , @xmath65where @xmath66 is the initial many - body state , and @xmath67 is the heisenberg field operator , @xmath68it can be also written as an object in the schrdinger picture , @xmath69where @xmath70 . the order or matrix indices
was chosen so as to have a natural property , @xmath71for any single - body operator @xmath72 .
the `` wide hat '' symbol is defined by eq .
( [ eq : tofock ] ) . in the mean - field approach ,
the density matrix is expressed by the mean - field amplitude , @xmath73 , as : @xmath74mathematically speaking , this relation stipulates that @xmath75 has a single nonzero eigenvalue equal to @xmath32 , with @xmath76 being the corresponding eigenvector .
it is then only natural to introduce effective mean - field amplitudes as eigenvectors of @xmath77 : @xmath78 in general , we have three eigenvectors of @xmath79 , @xmath80 , and three related eigenvalues , @xmath81 , @xmath82 . for a mean - field approximation to be recovered , there must exist a dominant eigenvalue , @xmath83 .
below in section [ secresdisc ] we shall see that it is exactly the situation occuring at early stages of the condensate evolution . at later stages
all three eigenvalues are of the same order of magnitude so that the mean - field approximation breaks down completely .
the physical information contained in the correlations expressed by the single - body density matrix is _ atomic coherence _ @xcite .
a natural framework for discussing such correlations is the picture of irreducible tensor operators @xcite .
an irreducible tensor operator of angular momentum @xmath84 is a set of @xmath85 square matrices , @xmath86 , @xmath87 . under three dimensional rotations ,
they obey the same transfomation laws as spherical harmonics @xmath88 , thus forming a natural basis for expressing any physical quantity with angular momentum @xmath84 .
furthermore , irreducible tensor operators form a full orthonormal set in the space of matrices in respect of the hilbert - schmidt norm : @xmath89and for any @xmath90 matrix @xmath72 , @xmath91 for the spin-1 condensate , the single - body density matrix ( [ eq : rhott ] ) is a @xmath92 matrix and hence only contains angular momenta 0 , 1 , and 2 : @xmath93@xmath94 and @xmath95 have a very simple interpretation : @xmath94 is in essence the total population , @xmath96and @xmath95 is related to magnetisation , @xmath97where the spherical components of the angular momentum operator are defined as , @xmath98 , @xmath99 . @xmath100
characterise the quadrupole momentum ( alignment ) of the atomic state .
it is instructive to rewrite @xmath77 in the cartesian basis , ( with @xmath101 ) @xmath102then , atomic population corresponds to the trace of the density matrix , magnetisation to its antisymmetric part , and alignment to its traceless symmetric part . taking the symmetric part of an hermitian matrix
is the same as taking its real part , hence @xmath103 contains a mix of the population and alignment while @xmath104 is pure magnetisation .
explicitly , @xmath105where @xmath106 is the magnetisation vector , and @xmath107 is the fully antisymmetric tensor , @xmath108 , @xmath109 . introducing the quadrupole tensor as per ref .
@xcite , @xmath110we find the irreducible - tensor - operator expansion of @xmath77 as @xmath111 in the mean - field picture , @xmath112 , where the cartesian components of @xmath113 are found in full analogy to ( [ eq : tocart ] ) .
the population - and - alignment part of the density matrix is then , @xmath114magnetisation also follows easily , @xmath115 .
\end{aligned}\label{eq : fmeanf } \end{aligned}\]]we can assume that either @xmath116 is real , @xmath117 , or the real and imaginary parts of @xmath116 are orthogonal , @xmath118 .
indeed , since we can always find such @xmath119 that @xmath120this orthogonality can be achieved by an overall phase transformation .
then , eq . ( [ eq : reim ] )
explicitly diagonalises @xmath121 , making it evident that in the mean - field approximation @xmath121 always has a zero eigenvalue .
another two eigenvalues equal @xmath122 and @xmath123 .
if @xmath124 is real , two eigenvalues of @xmath121 are zeros , leaving the system in a maximal - alignment state along @xmath116 . in this case
, the magnetisation which is related to @xmath125 vanishes .
the inverse is also true : zero magnetisation is characteristic of real @xmath116 .
an immediate word of caution is necessary here .
orthogonality of @xmath126 and @xmath127 is not preserved by the gross - pitaevskii equation ( [ eq : gp - equation ] ) .
one can rewrite ( [ eq : gp - equation ] ) so as to maintain the said orthogonality , but the resulting equations are rather bulky and do not seem to allow for a new insight ; worse still , under certain conditions the phase evolution is discontinuous . in order to gain further insight into the evolution of the atomic coherence ,
consider the equations of motion for the single - body density matrix written in terms of the magnetisation and quadrupole matrix .
direct calculation with hamiltonian ( [ eq : hfockspace ] ) yields , @xmath128_{\text{coll . } } .
\end{aligned}\label{eq : evolfq } \end{aligned}\]]the benefits of the representation of irreducible tensor operators are now clearly seen .
collisions are spherically symmetric so that only the linear and quadratic zeeman effects may couple magnetisation and alignment .
in actuality , additional reflection symmetries leave only quadratic zeeman effect to do it .
the linear zeeman effect leads only to the overall larmor precession around the @xmath1 axis .
furthermore , collisions only contribute directly to the evolution of the quadrupole , @xmath129_{\text{coll . } } = 2 g_2 \im \sum_{\sigma } \left\langle \hat a^{\dag}_{\alpha } \hat a^{\dag}_{\beta } \hat a^{2}_{\sigma } \right\rangle .
\end{aligned}\]]in general , this contribution contains nontrivial two - body correlations and can not be simplified any further . within the mean - field approximation , @xmath130_{\text{coll . } } = 2g_2\big [ 2\big ( { \mbox{\rm\boldmath$\zeta $ } } ^{\prime } \cdot { \mbox{\rm\boldmath$\zeta $ } } ^{\prime \prime }
\big ) \big ( \zeta ' _ { \alpha } \zeta ' _ { \beta } - \zeta ' ' _ { \alpha } \zeta ' ' _ { \beta } \big ) - \big ( { \mbox{\rm\boldmath$\zeta $ } } ^{\prime 2 } - { \mbox{\rm\boldmath$\zeta $ } } ^{\prime \prime 2 } \big ) \big ( \zeta ' _ { \alpha } \zeta ' ' _ { \beta } + \zeta ' ' _ { \alpha } \zeta ' _ { \beta } \big ) \big ] .
\end{aligned}\label{eq : qcollmf } \end{aligned}\]]note that here we do not assume that @xmath118 .
equations ( [ eq : evolfq ] ) allow one to make qualitative statements about the evolution of the system s coherence independently of the details of collisional interactions .
assume that the initial state of the system is as in experiment @xcite .
then , the initial magnetization vector is along the @xmath131 axis , @xmath132 . since @xmath133 , cf .
( [ eq : evolfq ] ) , magnetization @xmath134 never leaves the @xmath135 plane . as a result , under the mean field approximation the alignment eigenvectors stay in a plane orthogonal to the @xmath135 plane , cf .
( [ eq : fmeanf ] ) .
if at a particular time magnetisation vanishes , the mean field picture predicts that the system must be axially symmetric .
hence , under the sma , absence of axial symmetry at zero magnetisation is a clear signature of many - body effects .
since thermalisation via symmetric collisions can only increase the symmetry of the system , it appears plausible that this statement should survive beyond the sma . verifying this conjecture is subject to further analysis .
the choice of the basis in the fock space is inherently linked to the angular - momentum properties of hamiltonian ( [ eq : hfockspace ] ) .
a native bookkeeping is to introduce the basis states by the populations of the single - particle states with @xmath136 . by construction ,
such basis state in the fock space is an eigenstate of the three mode population operators : @xmath137these states also happen to be the eigenstates of @xmath138 and @xmath139 : @xmath140where @xmath141 and @xmath142 .
since @xmath143 commutes with @xmath138 and @xmath139 , states with different @xmath32 s ans @xmath144 s are not mixed by free evolution .
another simplification when using this `` natural '' basis is the ease with which the many - body operators are implemented ( appendix [ tofock ] ) . in the absence of the quadratic zeeman effect ( qze ) ,
hamiltonian ( [ eq : hfockspace ] ) conserves the full angular momentum , @xmath145 , and its component , @xmath146 . in this case
the time evolution is solved @xcite by simply classifying the basis states by the eigenvalues of the three operators : @xmath147here , @xmath148 and @xmath149 .
the subscript @xmath150 distinguishes the `` angular momentum '' basis from the `` natural '' basis introduced above . in the presence of the qze
, @xmath151 is not a good quantum number any more , while @xmath144 is still conserved .
hamiltonian ( [ eq : hfockspace ] ) can then be diagonalised within subspaces of given @xmath144 of the `` natural '' basis .
conversely , rotating the states by the radio - frequency field as in ref .
@xcite conserves @xmath151 while mixing different @xmath144 s . using the `` angular momentum '' basis
may thus be of some help . in the calculations summarised in this paper ,
the natural basis was used .
exploring the advantages of the angular momentum basis remains a subject to further work .
all simulations summarised in this paper were carried out with the experiment by sengstock and co - workers in mind @xcite . in this experiment ,
the condensate is firstly prepared in a state where all atomic spins are antiparallel to the @xmath1 axis .
a radio - frequency pulse is then used to re - orient the spins antiparallel to the @xmath131 axis . in our notation , this state looks as @xmath152where @xmath153 . since the manifold of @xmath154 is non - degenerate , @xmath155 is also an eigenstate of the total angular momentum , @xmath156 . furthermore , since the evolution does not mix different @xmath144 s , @xmath155 must be an eigenstate of the hamiltonian , providing us with a test of numerics .
another test is setting @xmath157 , yielding @xmath158 , which is also an eingenstate of the hamiltonian .
having thus checked that both for @xmath159 and for @xmath157 `` nothing at all happens , '' we proceeded with a series of simulations with @xmath160 . to compare the many - body results to the mean - field approximation
, we also ran a mean - field simulation of the system using ( [ eq : gp - equation ] ) .
it easy to see that , with our choice of the many - body initial state , eqs .
( [ eq : mfeigsvd ] ) are exact at @xmath161 .
this provides us with the initial condition for the mean - field simulation .
this is certainly the same initial mean - field amplitide as used in refs .
@xcite .
simulations presented in this paper were run for the following set of parameters : @xmath162 , @xmath163 , and @xmath164 .
these differ from the parameters characteristic of experiments @xcite , @xmath165 , @xmath166 , by setting @xmath167 to zero , and enhancing nonlinearity by an order of magnitude .
it should be emphasised that observation times in these experiments are too low for the many - body effects to manifest themselves . simply extending the `` observation '' time in simulations
allows for the many body effects to be seen , but produces two many oscillations of physical quantities on the relevant time scales .
enhancing the nonlinearity allows one to produce comprehensible graphics , while making no qualitative difference to the condensate evolution .
setting @xmath167 to zero is of no consequence : in essence this means introducing a reference frame following the larmor precession .
the number of particles in our simulations was @xmath168 .
importantly , this number is large enough to clearly separate the relaxation and the revival time scales . indeed ,
relaxation is exacty that what we expect from a many body treatment , whereas revivals are artifact of the model . at the same time , leaving the atom number relatively low simplifies numerics enourmously , making the problem easily tractable on an average desktop computer . , @xmath162 , @xmath163 , @xmath169 .
only the @xmath131 component of the magnetisation is nonzero . in the mean - field approximation ,
magnetisation oscillates without relaxation.,scaledwidth=50.0% ] , @xmath162 , @xmath163 , @xmath169 . in the mean field approximation
, one eigenvalue always equals one and the other two eingenvalues are zero.,scaledwidth=50.0% ] figure [ fig : f ] shows the evolution of the condensate magnetisation , eq .
( [ eq : fmeanf ] ) , calculated both within the many - body model and under the mean - field approximation . with our choice of the inital state and parameters , only the @xmath131 component of the magnetisation
is nonzero . in the mean - field approximation , this component oscillates periodically between values @xmath170 and @xmath32 . _
exact _ periodicity follows from the analytical solution @xcite .
the many body result shows _ decaying _ oscillations with the same period .
qualitatively , the decay can be understood if considering the simplest generic model of collisional nonlinearity : the kerr oscillator @xcite .
the latter is a single - mode quantum oscillator with quartic nonlinearity , @xmath171assume the kerr oscillator is initially in a coherent state with amplitude @xmath172
. a straightforward calculation of the average position of the oscilator yields , @xmath173\big\ } \\
\approx x_{\text{mf}}(t)\,{\rm e}^{-n_{\text{eff}}g^2_{\text{eff}}t^2/2 } , \ \ \ |g_{\text{eff}}t| \ll 1 .
\]]where @xmath174 and @xmath175 \end{aligned}\]]is the mean - field approximation to @xmath176 which accounts for the frequency shift but not the dephasing .
given the obvious crudeness of the kerr model , it reproduces the evolution of the magnetisation , fig .
[ fig : f ] , with surprising accuracy . setting @xmath177 results in a good quantitative estimate of the decay time , @xmath178 .
the gaussian nature of the decay is also evident in figs .
[ fig : f ] and [ fig : eigen ] .
contrary to what could be expected , the system does not evolve towards a spherically symmetric state . a formal indication of this may be seen in fig . [
fig : eigen ] .
this figure shows scaled eigenvalues of the single - body density matrix , cf .
( [ eq : eigenv ] ) . in the mean field approximation
, one eigenvalue always equals one and the other two eingenvalues are zero .
this holds for the initial state , but for nonzero times we find that all three eigenvalues are nonzero .
most interestingly , they do not evolve towards the value of 1/3 indicative of spherical symmetry .
two of them stabilise at @xmath179 , while the third one remains close to @xmath180 .
since the magnetisation decays completely , cf .
[ fig : f ] , the persistent asymmetry must be of quadrupole nature .
plane with @xmath181 tilt . _
bottom row : _ later stages of the evolution at which the mean - field picture fails .
_ left : _ a typical picture of coherence at intermediate stages of the evolution .
_ right : _ the final state of condensate coherence .
the magnetisation and tilt are zero , and the ellipsoid is axially symmetric in respect of the quantisation axis.,title="fig:",scaledwidth=49.0% ] plane with @xmath181 tilt . _
bottom row : _ later stages of the evolution at which the mean - field picture fails . _ left : _ a typical picture of coherence at intermediate stages of the evolution .
_ right : _ the final state of condensate coherence .
the magnetisation and tilt are zero , and the ellipsoid is axially symmetric in respect of the quantisation axis.,title="fig:",scaledwidth=49.0% ] plane with @xmath181 tilt . _
bottom row : _ later stages of the evolution at which the mean - field picture fails .
_ left : _ a typical picture of coherence at intermediate stages of the evolution .
_ right : _ the final state of condensate coherence .
the magnetisation and tilt are zero , and the ellipsoid is axially symmetric in respect of the quantisation axis.,title="fig:",scaledwidth=49.0% ] plane with @xmath181 tilt . _
bottom row : _ later stages of the evolution at which the mean - field picture fails .
_ left : _ a typical picture of coherence at intermediate stages of the evolution .
_ right : _ the final state of condensate coherence .
the magnetisation and tilt are zero , and the ellipsoid is axially symmetric in respect of the quantisation axis.,title="fig:",scaledwidth=49.0% ] , @xmath162 , @xmath163 , @xmath169 . left : the alignment eigenvalues , right : the tilt angle .
the smallest eigenvalue corresponding to the alignment axis along the magnetisation is shown as solid line , the other two as dashed lines .
the modulus of magnetisation ( dotted line ) helps to visually synchronise this picture with fig .
[ fig : f ] . in the mean field approximation ,
the smallest eigenvalue is exactly zero while the other two eigenvalues and the tilt oscillate periodically without relaxation.,title="fig:",scaledwidth=49.0% ] , @xmath162 , @xmath163 , @xmath169 .
left : the alignment eigenvalues , right : the tilt angle .
the smallest eigenvalue corresponding to the alignment axis along the magnetisation is shown as solid line , the other two as dashed lines .
the modulus of magnetisation ( dotted line ) helps to visually synchronise this picture with fig .
[ fig : f ] . in the mean field approximation ,
the smallest eigenvalue is exactly zero while the other two eigenvalues and the tilt oscillate periodically without relaxation.,title="fig:",scaledwidth=49.0% ] being a mathematically rigorous way of characterising alignment , the quadrupole tensor is not the best choice so far as _ visualisation _ of the results is concerned .
a much more convenient way is thinking in terms of the _ alignment ellipsoid_. the latter is found by diagonalising the real part of the density matrix , cf .
( [ eq : rhobyref ] ) .
the eigenvectors of @xmath103 are the alignment axes .
the eigenvalues of @xmath103 are positive , their sum equals population , @xmath182 , and differences between them characterise alignment . in particular , three equal eigenvalues , @xmath183 , mean the atomic state has no alignment at all .
the system of three orthogonal eigenvectors and three positive eigenvalues is naturally visualised as a three - dimensional ellipsoid .
if we additionally shift the centre of this alignment ellipsoid by the magnetisation vector , the resulting picture contains all details of the atomic coherence , cf .
figure [ fig : visual ] .
an additional simplification is that the smallest axis of the ellipsoid was always found to be collinear with the magnetisation vector .
this also holds if @xmath184 when the whole picture is subject to the larmor precession .
thus the only parameter we need in order to fully specify the orientation of the alignment ellipsoid is the tilt angle ; the latter is defined as the angle between the largest axis of the ellipsoid and the quantisation ( @xmath1 ) axis . together with the eigenvalues of @xmath185 [ cf .
( [ eq : rhobyref ] ) ] and the magnetisation vector , this angle completes characterisation of the atomic coherence .
visualisation of atomic coherence at different stages of the evolution may be seen in fig .
[ fig : visual ] .
evolution of the alignment eigenvalues and of the tilt angle is shown in fig .
[ fig : q ] . at early stages of the evolution ,
the mean field picture holds with a very good accuracy . in the mean field approximation
, the alignment ellipsoid reduces to a flat ellips orthogonal to the magnetisation vector .
there are two characteristic moments of the mean - field evolution [ cf .
( [ eq : fmeanf ] ) and fig .
[ fig : q ] , top row ] .
if the magnetisation is maximal , the alignment axis are equal ; the alignment ellipsoid becomes a disk ( fig .
[ fig : q ] , top - left ) .
if the magnetisation is zero , only one alignment eigenvalue is nonzero ; the alignment ellipsoid reduces to a straight line at @xmath181 tilt ( fig .
[ fig : q ] , top - right ) .
the mean - field evolution of the atomic coherence is in essence periodic oscillations between these charateristic configurations .
on later stages of the evolution the mean field picture breaks down .
all three eigenvalues of @xmath103 are nonzero , so that the alignment ellipsoid acquires thickness , becoming an ellipsoid proper ( fig .
[ fig : visual ] , bottom - left ) .
the tilt oscillations are damped and finally cease ( fig .
[ fig : q ] ) .
the final state of atomic coherence is depicted in fig .
[ fig : visual ] , bottom - right .
the magnetisation is zero so that the ellipsoid is centered at the centre of coordinates .
the larger axis is along @xmath1 , the other two are equal so that the alignment ellipsoid is axially symmetric .
this state of atomic symmetry is impossible to obtain in the mean - field picture .
thus in the course of the evolution atomic coherence undergoes a complete change of symmetry .
initially , the system is prepared in an axially symmetric state in respect of the @xmath131 axis . at intermediate stages of the evolution ,
all symmetries are broken .
finally , the system again tends towards an axial symmetry .
not unexpectedly , the final symmetry coinsides with the dynamical symmetry , with the symmetry axis parallel to the magnetic field .
we have modelled the quantum many - body dynamics of a spin-1 condensate numerically within a single mode approximation , using the representation of irreducible tensor operators so as to conveniently visualise the condensate evolution .
comparing results of the numerical simulation to those obtained within the mean field approximation shows a number of characteristic features of the many - body evolution which are impossible to obtain under the mean field approximation .
in particular , the final state of the condensate coherence is of a characteristic quadrupole nature , axially symmetric in respect of the quantisation axes , whereas this kind of atomic coherence is impossible in principle in the mean field approximation .
experimentally observing this state of atomic coherence would be a distinct signature of spin - mixing collisions without the mean - field approximation .
l.p . and r.w .
acknowledge numerous stimulating discussions with k. bongs , k. sengstock , j. kronjger and p. navez .
l.p . gratefully acknowledges support of the program atomoptik of the landesstiftung baden - wrtemberg .
to be specific , consider how mode annihilation operators , @xmath186 , are implemented in a high - level numerical package .
such packages handle sparse rectangular matrices very effectively , so the first step is to replace a triple index , @xmath187 , characterising a basis state in the fock space , by a linear index .
a helpful trick is to start from introducing an extended index , @xmath188 , where @xmath189 .
the third index , @xmath190 , is found from @xmath191 .
each value of the extended index , @xmath192 , thus corresponds to a unique triple , @xmath193this includes nonphysical entries with @xmath194 which are to be filtered out later .
the advantage of using the extended indexing is that , as a rule , many - body operators written in the extended index are simple banded matrices .
for example , if @xmath195 , @xmath196that is , in the extended index , @xmath197 and @xmath198 are sparse matrices with a single nonzero side diagonal . generating such matrices numerically is straightforward ;
@xmath199 which is not found in such simple way can then be obtained by a suitable permutation of rows and columns of @xmath197 ( say ) . in the matrices
generated using ( [ eq : amatr ] ) one should retain only columns for which @xmath200 and rows for which @xmath201 and @xmath202 .
( recall that @xmath186 acts from the subspace of @xmath32 atoms into the subspace of @xmath203 atoms . )
this is equivalent to building a `` physical '' linear index , changing @xmath204 first , and @xmath205second , while watching for the condition @xmath206 ; however the property of having a single nonzero side diagonal does not hold for the physical indexing .
generalisation to higher - order spin values is straightforward .
there is a further subtlety in how products of field operators are implemented numerically .
an unwanted by - product of introducing field operators is a formal non - conservation of the particle number .
we wish to restrict ourselves to a subspace of given @xmath32 and introduce annihilation operators as acting from this subspace into the subspace of @xmath203 atoms .
then , only bilinear normally ordered operator products , @xmath207 , are immediately defined for numerical purposes . viewed literally , quartic normally ordered operators products [ cf . the first line of ( [ eq : hfockspace ] ) ] imply knowledge of the field operators acting from the subspace of @xmath203 atoms into the subspace of @xmath208 atoms and hence can not be implemented under this restriction .
the solution to this problem is obvious : use commutation relations .
so , with @xmath209 being the unit matrix in the subspace of @xmath32 atoms , @xmath210 etc .
also the second line of ( [ eq : hfockspace ] ) ] .
equations ( [ eq : commnum ] ) hold in the complete fock space , but only their right - hand sides make numerical sense under the above restriction .
bruce w. shore , _ the theory of coherent atomic excitation _ , v. 1,2 ( wiley , 1989 ) .
pethick and h. smith , _ bose - einstein condensation in dilute gases _ ( cambridge university press , 2002 ) d. m. stamper - kurn _ et al .
_ , phys .
lett . * 80 * , 2027 ( 1998 ) .
m. d. barrett , j. a. sauer , and m. s. chapman , phys .
lett . * 87 * , 010404 ( 1998 )
. h. schmaljohann _ et al .
_ , appl .
b * 79 * , 1001 ( 2004 ) .
tin - lun ho , phys .
lett . * 81 * , 742 ( 1998 ) .
t. ohmi and k. machida , j. phys .
. jpn . * 67 * , 1822 ( 1998 ) . c. k. law , h. pu and n. p. bigelow , phys .
* 81 * , 5257 ( 1998 ) .
et al . _ ,
lett . * 82 * , 2228 ( 1999 ) ; d.m .
stamper - kurn _ et al .
_ , phys .
lett . * 83 * , 661 ( 1999 ) .
a. omont , progr .
electronics * 5 * , 69 ( 1977 ) .
varshalovich , a.n .
moskalev , v.k .
khersonskii , quantum theory of angular momentum ( world scientific , 1988 ) .
masato koashi and masahito ueda , phys .
* 84 * , 1066 ( 2000 ) . ming - shien chang _
et al . _ , cond - mat/0509341 .
j. kronjger _
et al . _ , phys .
rev . a * 72 * , 063619 ( 2005 ) .
a. griesmaier _
et al . _ ,
lett . * 94 * , 160401 ( 2005 ) l. santos and t. pfau , _
e_-print cond - mat/0510534 .
yuki kawaguchi , hiroki saito , and masahito ueda , _
e_-print cond - mat/0511052 .
et al . _ ,
lett . * 95 * , 180403 ( 2005 ) wenxian zhang _
et al . _ ,
a * 72 * , 013602 ( 2005 ) .
j. vanier and c. audoin , _ the quantum physics of atomic frequency standards _ ( hilger , philadelphia , 1988 ) . m .- s .
chang _ et al .
_ , phys .
* 92 * , 140403 ( 2004 ) .
m. steel _ et al .
_ , phys .
a * 58 * , 4824 ( 1998 ) .
a. imamoglu , m. lewenstein , and l. you , phys .
lett . * 78 * , 2511 ( 1997 ) .
|
we analyse and numerically simulate the full many - body quantum dynamics of a spin-1 condensate in the single spatial mode approximation .
initially , the condensate is in a `` ferromagnetic '' state with all spins aligned along the @xmath0 axis and the magnetic field pointing along the @xmath1 axis .
in the course of evolution the spinor condensate undergoes a characteristic change of symmetry , which in a real experiment could be a signature of spin - mixing many - body interactions .
the results of our simulations are conveniently visualised within the picture of irreducible tensor operators .
cold atoms , trapped atoms , bose condensate , spinor condensate , nonequilibrium dynamics , many body theory 03.75.fi , 05.30.jp , 67.40.db , 05.70.ln
|
biological systems present us with a bewildering fauna of oscillators : cell cycles @xcite , circadian rhythms @xcite , calcium oscillations @xcite , pace maker cells @xcite , protein responses @xcite , and so on .
sometimes , however , it is hard to see what are the minimal models behind these oscillations .
typically , the models are categorized into two classes : ( i ) _ limit cycle oscillator _ :
the fixed point is linearly unstable and the oscillations are described by stable limit cycles sustained by nonlinearity of the system in the deterministic case @xcite .
noise ( e.g. molecular noise due to limited number of copy numbers ) can be added on the top of the deterministic oscillations .
( ii ) _ noise - induced oscillator _ : the fixed point is linearly stable for the system without noise and the system relaxes to the fixed point with damped oscillations when temporally perturbed .
addition of noise to such a system is known to show sustained oscillations with fluctuating amplitude @xcite . for some systems ,
both limit cycle oscillators ( i ) and noise - induced oscillators ( ii ) are proposed as a mechanism for the oscillation @xcite . here
, we propose a way to distinguish the two , by using the phenomenon of _ entrainment _ to a periodic perturbation .
it is well known that , when an periodic perturbation is added to a deterministic limit cycle , the system s oscillation frequency @xmath0 will be entrained to the external frequency @xmath1 with various rational numbers of frequencies @xmath2 for _ all _ positive integers @xmath3 and @xmath4 in a _
finite window of the external frequency _ @xmath1 , where the width of the window depends on the amplitude of the external forcing @xcite .
entrainment , also called mode - locking , has been observed in variety of physical systems during the last decades , from onset of turbulence @xcite , josephson junctions @xcite , one - dimensional conductors @xcite , semiconductors @xcite and crystals @xcite .
it has been predicted , and verified experimentally , that the mode - locking structure possesses certain universal properties @xcite . in biological systems
, entrainment has been investigated theoretically for circadian rhythms @xcite as well as in model systems for protein responses @xcite .
experimental observation of entrainment in biological systems is often rather difficult due to noisy signals , but it has been observed for circadian rhythms @xcite and synthetic genetic oscillators @xcite . in this paper , we study the difference in the entrainment behavior for the limit cycle oscillators and noise - induced oscillators . our main question is the following : can we distinguish the two cases by means of the entrainment behavior ?
we employ the famous van der pol equation with noise as an example , because there we can easily study both cases by changing parameters .
we show that entrainments to all the rational ratios are seen only in the limit cycle oscillator . in the case of the noise - induced oscillator with additive external forcing
, the oscillator can entrain only at one - to - one ratio , meaning that the entrainment to other than the one - to - one ratio is the sign of the dominance of the limit cycle mechanism .
when the external forcing is multiplicative , we find that the noise - induced oscillator with weak nonlinearity can show some entrainment ratios other than one - to - one , but not all the ratios .
to confirm the generality of the entrainment behavior for the limit cycle system under weak noise , we also study a biological example , the tnf - driven oscillating nf-@xmath5b system , and confirm that @xmath6 entrainments can be seen .
consider the following two - dimensional equation with noise : @xmath7 with @xmath8 @xmath9 here , @xmath10 , @xmath11 , and @xmath12 are parameters , and @xmath13 are uncorrelated , statistically independent gaussian white noise , satisfying @xmath14 first let us consider the deterministic case , @xmath15 .
the model has a fixed point at @xmath16 , and the eigenvalues around this fixed point are @xmath17 indicating that the system experiences a hopf bifurcation at @xmath18 .
when @xmath19 , the fixed point relaxes to the fixed point with damped oscillation with the angular frequency @xmath20 , while when @xmath21 and @xmath22 the model shows a stable limit cycle ( van der pol oscillator ) . in the stochastic case with @xmath23 , however , the system shows a sustained oscillation even in the linearly stable case , @xmath19 , because the noise keeps activating the oscillation with frequency @xmath24 .
this is the case of the linear p53 model introduced in ref .
when @xmath21 , @xmath23 adds fluctuations on top of the stable oscillation around the limit cycle .
we investigate the entrainment behavior of the model , focusing on the following three classes of parameter sets . 1 .
the _ limit cycle oscillator _ , with @xmath21 and @xmath22 .
2 . for the _ noise - induced oscillator _
, we consider the two subcategories . 1 . the _ linear system with a stable fixed point _ , with @xmath19 and @xmath25 , i.e. , the equations are linear in @xmath26 and the fixed point is stable .
the _ nonlinear system with a stable fixed point _
, with @xmath19 and @xmath22 , i.e. , the fixed point is linearly stable but the equations possess a nonlinear term .
when noise - induced oscillators are studied , normally only linear terms are considered . however , in reality , there are often nonlinear terms , which can play a role when distance from the stable fixed point @xmath27 is sufficiently large .
this is the reason why we consider both linear and nonlinear noise - induced oscillators .
when needed , numerical integration of stochastic differential equations are performed by using euler method .
figure [ noforce ] shows the typical behavior of the model in each categories .
the parameters are chosen so that the period and amplitude are in similar range . without noise ,
the limit cycle is the only case with stable oscillation ( fig . [
noforce]a ) , while linear and nonlinear systems with a stable fixed point exhibit damped oscillations relaxing to the fixed point ( fig .
[ noforce]bc ) .
when noise is added , the oscillation is perturbed for limit cycle oscillator ( fig .
[ noforce]d ) ; here the noise level is chosen so that the base oscillation is still recognizable . for linear and nonlinear noise - induced oscillators ( fig . [
noforce]ef ) , we observe oscillations with the expected angular frequency ( @xmath28 ) . in order to demonstrate the difference between the two , we apply the exact same sequence of noises in both cases .
we observe a bigger difference when linear noise - induced oscillator have large ( @xmath29 ) amplitude , because the nonlinear term becomes more important .
naturally this effect depends on the value of @xmath12 ( data not shown ) .
we study these oscillators under the following two kinds of external periodic perturbation .
[ [ additive - forcing . ] ] additive forcing .
+ + + + + + + + + + + + + + + + + the first case is an _ additive _ forcing , in the form of @xmath30 with @xmath31 [ [ multiplicative - forcing . ] ] multiplicative forcing .
+ + + + + + + + + + + + + + + + + + + + + + + the second case is an _ multiplicative _ forcing ( also called parametric forcing ) , in the form of @xmath32 where @xmath33 in the next section , we first present the behavior of the model under the additive forcing , and then show the parallel results for the multiplicative forcing .
in the case of the additive periodic forcing to a linear deterministic system , we have in general @xmath34 where @xmath35 is a coefficient matrix of the linearized equation , and @xmath36 is periodic function in time with a period @xmath37 , satisfying @xmath38 . by expressing @xmath39 , with using eigenvectors @xmath40 of the matrix @xmath41 given by @xmath42 , we can show that in the long - time limit we have @xmath43 note that @xmath44 because the fixed point @xmath45 is stable .
@xmath46 is defined by the fourier expansion of @xmath36 as @xmath47 where @xmath48 is the left eigenvector .
therefore the solution will always be a periodic function of @xmath49 with the period @xmath37 in the long time limit , and contains only the frequencies that the external forcing has .
in other words , the system will be always in a 1/1 entrained state if the perturbation is pure sine or cosine wave .
when gaussian white noise is added to eq .
( [ linearsolve ] ) , we have @xmath50 in this case , we can evaluate the auto - correlation of @xmath51 for large enough @xmath52 ( @xmath53 ) as @xmath54 namely , the response contains oscillations with the frequencies from forcing @xmath55 and from the complex part of the eigenvalue @xmath56 , and the amplitude of the latter is proportional to @xmath11 .
we now investigate numerically the entrainment behaviors for all three categories .
first we demonstrate the behavior without noise , and then show how the noise modify this behavior .
[ [ without - noise . ] ] without noise .
+ + + + + + + + + + + + + + figure [ aforceexample ] illustrates typical entrainment behaviors for additive forcing when noise is absent . with a limit cycle oscillator ( fig .
[ aforceexample ] a , d ) , the system s angular frequency can entrain to the external angular @xmath1 with various ratios , while in the linear system , one - to - one entrainment occurs ( fig .
[ aforceexample ] b , e ) .
the nonlinear system shows very similar behavior to the linear system , where we see only one - to - one entrainment ( fig .
[ aforceexample ] c , f ) . in order to define the system s angular frequency in a simple way
, we adopt the polar coordinate @xmath57 using @xmath58 as proposed in ref .
we define @xmath59 so that @xmath60 gives the winding number , i.e. , how many times the orbit went around the fixed point by time @xmath49 .
the system s angular frequency is numerically calculated from @xmath61\ ] ] for long enough @xmath37 ( typically 1000 times external forcing period ) . with this definition ,
[ aforceexample](a ) shows the entrainment of the ratio @xmath62 , while fig .
[ aforceexample](d ) gives @xmath63 .
[ [ with - noise . ] ] with noise .
+ + + + + + + + + + + the addition of noise blurs the entrainment behavior , as depicted in fig .
[ aforceexamplenoise ] . for the limit cycle oscillator ( fig .
[ aforceexamplenoise]a and d ) , we can see that the noise makes the orbit irregular , which can make the phase to slip . in the linear noise - induced oscillator for small external angular frequency
, we can clearly see that the noise induces the oscillation with angular frequency close to @xmath64 on top of one - to - one entrainment behavior ( fig .
[ aforceexamplenoise]b ) , as expected from the auto - correlation eq .
( [ ac ] ) .
when @xmath1 is larger than @xmath64 , the external angular frequency is more visible , because the noise @xmath11 is small compared to the amplitude @xmath65 for this case , though both frequencies should be present .
the nonlinear noise - induced oscillator behaves again very similar to the linear case in entrainment behavior ( fig .
[ aforceexamplenoise]c and f ) .
the visible difference is a suppression of large amplitude by the nonlinear term .
[ [ devils - staircase - and - arnolds - tongues . ] ] `` devil s staircase '' and `` arnold s tongues '' .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for deterministic limit cycles , the plot of @xmath66 vs @xmath1 for a fixed amplitude of external forcing shows a infinitely complex structure with fractal nature , known as devil s staircase @xcite . for the present system of limit cycle oscillator without noise , this is also observed as shown in fig .
[ figure - dsa](a ) ( solid line ) . as noise increases ,
the phase slips occasionally , therefore narrow entrainment regions become harder to recognize ( fig .
[ figure - dsa]a , dashed and dotted line ) .
for the systems with a stable fixed point , there is only one - to - one entrainment for the no noise case ( fig .
[ figure - dsa]b , solid line ) , while noise induced oscillation around the entrained solution will add some phase slips giving a change in the angular frequency when the entrainment is not so strong , resulting in an escape from the one - to - one ratio as shown in fig .
[ figure - dsa](b ) .
when entrainment regions for various values of @xmath66 are plotted in the @xmath65-@xmath1 plain , it gives an `` arnold s tongue '' structure for the deterministic limit cycles : the entrainment regions widen as the external forcing amplitude @xmath65 grows , resulting in tongue - like shapes of the entrainment region
when @xmath65 is large enough the tongues start to overlap @xcite .
this can be seen in the limit cycle oscillator without noise in fig .
[ figure - tonguea ] ( a ) .
when noise is added , the phase of the oscillator sometimes slips , resulting in narrower tongues ( fig .
[ figure - tonguea ] b ) . for the noise - induced oscillators ( i.e. with a stable fixed point ) ,
there exists only 1/1 entrainment without noise , and with noise 1/1 entrainment is the only case that gives the tongue - like structure , both for the linear and nonlinear cases ( fig .
[ figure - tonguea ] cd ) .
we see other ratios of entrainment regions , because for a given @xmath65 with changing @xmath1 , @xmath66 changes continuously outside of the entrainment region ( e.g. fig.[figure - dsa ] b ) .
we next consider the multiplicative forcing @xmath67 where the matrix @xmath68 satisfies @xmath69 with @xmath70 .
it is known from floquet theory @xcite that the solution matrix of this equation is expressed as @xmath71 where @xmath72 and a general solution is the linear combinations of column vectors consisting of @xmath73 .
the eigenvalues of the matrix @xmath74 , called floquet exponents , determine the stability of the solution : the solution will converge to the fixed point when the real parts of the floquet exponents are all negative , and diverges if some floquet exponent have positive real parts . therefore , no entrainment behavior will be observed for a linear noise - induced oscillator without noise under multiplicative forcing . in fig.[mplot ] , we show numerically calculated the maximum real part of the floquet exponents @xmath75 for ( [ multieq ] ) with ( [ whatism ] ) with @xmath76 and @xmath25 , as a function of amplitude of forcing @xmath77 and external frequency @xmath1 .
when @xmath78 ( dark blue region ) , @xmath27 will exponentially decays to zero , otherwise @xmath27 will diverge except for the marginal case @xmath79 . for various @xmath77 and @xmath1 , for the linear system with stable fixed point ( @xmath76 and @xmath25 ) without noise ( @xmath15).,scaledwidth=45.0% ] [ [ without - noise.-1 ] ] without noise .
+ + + + + + + + + + + + + + figure [ mforceexample ] shows the entrainment behaviors for multiplicative forcing . for the limit cycle oscillator ( fig .
[ mforceexample ] ac ) , there is no qualitative difference from the additive noise case , i.e. , the system shows entrainment with various frequency ratio @xmath2 . for the linear system with a stable fixed point without noise ,
on the other hand , the system can either decay to the fixed point ( fig .
[ mforceexample ] b ) or diverge ( fig .
[ mforceexample ] d ) , which can be predicted from the flouquet exponents ( fig.[mplot ] ) .
when nonlinear term is added , it does not prevent the decay ( fig . [ mforceexample ] c ) , but the divergent behavior is suppressed and system shows the entrainment behavior ( fig .
[ mforceexample ] f ) .
the frequency ratio @xmath66 is not necessarily 1/1 ; the example in fig .
[ mforceexample](f ) gives @xmath80 .
[ [ with - noise.-1 ] ] with noise .
+ + + + + + + + + + + when noise is added , the behavior changes drastically in the noise - induced oscillators , as shown in .
[ mforceexamplenoise ] .
the noise can induce the oscillation with the angular frequency close to @xmath64 for the case where the no - noise system would decay to the fixed point ( fig .
[ mforceexamplenoise ] b , c ) . on the other hand , in the linear noise - induced oscillator
, adding noise does not prevent divergence ( fig .
[ mforceexamplenoise ] e ) . for
the parameters where no - noise system would entrain , the noise blurs the entrainment due to occasional phase slip for both limit cycle oscillator ( fig .
[ mforceexamplenoise ] ac ) and nonlinear noise - induced oscillator ( fig .
[ mforceexamplenoise ] f ) .
[ [ devils - staircase - and - arnolds - tongue . ] ] `` devil s staircase '' and `` arnold s tongue '' .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we also study the `` devil s staircase '' for the multiplicative forcing .
for the limit cycle oscillator without noise , we again see proper devil s staircase , where noise will blur the entrainment behaviors ( fig.[figure - dsm ] a ) . for the noise - induced oscillators ,
only the nonlinear case is studied because the linear case may diverge depending on the parameter values . without noise , we see discrete finite regions of entrainment ( fig.[figure - dsm ] b squares ) , while noise induces the oscillations in the decaying region resulting in a continuous line ( fig.[figure - dsm ] b dashed line ) .
the arnold s tongue structure for the limit cycle is similar to those in the additive forcing case , as seen in fig.[figure - tonguem](a ) and ( b ) .
the arnold s tongues for all the entrainment ratios are observed without noise , and noise makes the regions smaller . for the nonlinear system with a stable fixed point without noise , there are entrainment regions for a few rational ratios , but the ones that appear are problem specific - for instance , in the present case , the @xmath81 is not observed at all in fig.[figure - tonguem](c ) . with noise ( fig.[figure - tonguem]d ) ,
the entrainment regions shrinks , but at the same time the system can occasionally pass the given ratio of @xmath66 , resulting in narrow line of `` fake '' entrainment .
in this section , we study a biological example , tnf - driven nf-@xmath5b system , to confirm the generality of the entrainment behavior for the limit cycle system under weak noise . the system has been studied for the deterministic case in ref.@xcite .
nf-@xmath5b is a transcription factor , and it has been verified experimentally that nf-@xmath5b level in the nucleus shows sharp oscillations after treatment with tumor - necrosis factor ( tnf)@xcite .
the interaction network involves a negative feedback loop between the nf-@xmath5b and an inhibitor , ikb@xmath82 , which is the main mechanism for the oscillations .
tnf modulates the state of the ikb@xmath82 and hence affects the oscillation .
tnf can be added externally to the cell , therefore it can serve as a possible probe to study the entrainment , i.e. , we can use tnf level as the external forcing term . in ref.@xcite ,
the system was modeled by 5 dimensional coupled nonlinear odes , @xmath83),\ ] ] where @xmath84(i_{tot}-x_4-x_5)-k_ix_4,\\ f_5&=&k_ix_4 -k_px_5\frac{k_{a20}}{k_{a20}+a_{20}[tnf]}.\label{x5}\end{aligned}\ ] ] the variable @xmath85 denotes the nuclear nf-@xmath5b level , and @xmath86 $ ] denotes the tnf level , which we shall change to a periodically external forcing onto the system .
the biological meaning of the variables and parameter values are summarized in table [ parameters ] .
note that @xmath86 $ ] appears twice in eq .
( [ x5 ] ) in the terms multiplied with @xmath26 , therefore this is an example of multiplicative forcing .
we study this system with adding a gaussian white noise in each term , i.e. , @xmath83)+\sigma { \ensuremath{\boldsymbol{\gamma}}}.\ ] ] .[parameters ] variables and the parameters in the tnf - driven nf - kb oscillation , from ref.@xcite . [ cols= " < , < " , ] figure [ nfkbexamples](a - c ) shows the spontaneous oscillation of nuclear nf-@xmath5b , when @xmath86 $ ] is kept constant at @xmath86=0.5 $ ] , without ( a ) and with noise ( b - c ) .
we see clear periodic oscillation with the period around @xmath87 ( minutes ) .
we then modulate the @xmath86 $ ] level around this basal level @xcite as @xmath88=0.5+m_{tnf}\sin(\omega t).\ ] ] this has been studied in the no - noise case by jensen and krishna @xcite , and it was found that the entrainments of various ratios can occur , when the frequency of the nf-@xmath5b level is determined based on the frequency of the peaks .
figure [ nfkbexamples](c ) shows an example of 1/2 entrainment , for @xmath89 and @xmath90 , in the deterministic case . with weak enough noise , the entrainment
is maintained ( fig .
[ nfkbexamples]e ) , but larger noise induces phase slips ( fig .
[ nfkbexamples]f ) , as has been seen in the van der pol system . b system .
spontaneous oscillations with [ tnf]=0.5 for ( a ) no noise ( @xmath15 ) case and ( b)@xmath91 , ( c)@xmath92 . ] in fig .
[ nfkbstairs ] , several devil s staircases are shown with and without noise . in ref . @xcite the arnold tongues have been calculated , and it has been demonstrated that general p / q entrainments occur .
a characteristic observation to this system is that the tongues overlap easier for larger external frequency ; e.g. , @xmath93 for 1/3 and 1/2 tongues to overlap , while the 2/1 and 5/2 tongues do not overlap even at @xmath94 . when the devil s staircase are calculated for overlapping region , non - smooth or irregular jumps between the steps
can be seen , and is thus dependent on initial conditions in general .
this is visible in our data in fig.[nfkbstairs ] , for large @xmath95 and larger @xmath1 .
when weak noise is added , it enables the system to jump to other overlapping tongues , which results in irregular behavior around the entrainment regions .
as the noise becomes larger , the entrainment is again smoothed away by phase slips .
b system without and with noise , with ( a)@xmath96 , ( b)@xmath97 , ( c)@xmath89 , and ( d)@xmath94 .
the entrainment regions are calculated from the frequency of the peaks in the the deterministic case . in the finite noise case
, we define the nuclear nf-@xmath5b peak as follows : we first determined the maximum value @xmath98 and the minimum value @xmath99 of @xmath85 of the steady state in the deterministic simulation for the given parameters .
we then calculate two thresholds , @xmath100 and @xmath101 .
next we perform the corresponding simulation with finite @xmath11 .
we define a switching event from the `` low '' state to `` high '' state when @xmath85 exceeds @xmath102 , while the reverse switching happens when @xmath85 becomes smaller than @xmath103 .
the number of peaks are calculated from how often the `` high '' states are reached . this way we can filter out the wiggly motion due to the noise and thus define the overall peak . ]
@l*15@l oscillator & no noise & with noise & force + limit cycle & entrainment to any @xmath6 & entrainment to any @xmath6 with phase slips & a + & entrainment to any @xmath6 & entrainment to any @xmath6 with phase slips & m + linear & one - to - one entrainment@xmath104 & one - to - one entrainment@xmath104 with phase slips & a + noise - induced & decay or diverge & noise - induced oscillation with @xmath105 & m + & & or diverge & + nonlinear & one - to - one entrainment@xmath104 & one - to - one entrainment@xmath104 with phase slips & a + noise - induced & small decay or & noise - induced oscillation with @xmath105 or & m + & some @xmath6 entrainment & some @xmath6 entrainment with phase slips & + * @xmath104 all the frequencies contained in the forcing can be observed .
our motivation behind this work was to ask : can one by applying an external periodic forcing and studying entrainment determine whether an oscillating system is driven by a linear mechanism ( noise induced oscillator ) or a non - linear mechanism ( limit cycle oscillator ) ?
our answer to this question is generally yes . our obtained results on entrainment behavior of oscillators are summarized in table [ tablesummary ] . when the forcing is additive , there is clear difference between the limit cycle oscillators and the noise - induced oscillators .
the former can entrain to any frequency ratio , while the latter shows only one - to - one entrainment . therefore ,
if one see entrainment to @xmath106 ratio , under additive forcing , it is a sign of limit cycle oscillator .
when the forcing is multiplicative , the non - linear noise - induced oscillators can also show @xmath106 ratio entrainment , but not necessarily all of the rational ratios .
if the system is noise - induced oscillator and the non - linear term is small , one might be able to capture the diverging tendency of the amplitude , because saturation happens when the amplitude is large enough to make the non - linear term relevant .
in such a case , one might see big difference in amplitude for a fixed @xmath77 with varying @xmath1 .
we thus urge experiment to be performed on oscillating biological systems .
it is well known that some proteins ( p53 , nf - kb , wnt ) can oscillate in cells under stress responses . in the case of p53 ,
both non - linear @xcite and linear models have been proposed @xcite . by applying an external time dependant signal such as dna damaging radiation or drugs which specifically perturb the p53 circuit , it might be possible to entrain the internal oscillation and draw conclusions on the basis of our results summarized in table [ tablesummary ] . in the case of nf - kb oscillations
, one might be able to entrain the internal oscillation by an externally varying cytokine ( like tnf ) signal @xcite .
potentially , it could lead to a way of controlling the dna - repair pathway .
the present research also opens for further theoretical investigations .
the @xmath106 entrainment of the winding number for a nonlinear noise - induced oscillator with multiplicative forcing is a purely numerical observation , and further research is needed to refine the condition when this can occur .
we did not study the strong noise case either , and it would be interesting to investigate in more details the active role of noise in the entrainments of limit cycles . in many biological examples , where dynamics are
molecular reaction based , additive gaussian white noise is not appropriate for large noise because it does not reflect the noise amplitude dependence on the molecule number : instead either a concentration dependent noise amplitude or a stochastic treatment of molecule numbers should be performed .
finally , we would like to briefly comment on `` noise - induced '' oscillations by mechanisms other than the linear model studied here .
it has been long known that , when noise is added to excitable system with a stable fixed point , regular oscillatory behaviour can be observed at a certain level of noise ( coherence resonance ) @xcite . since the nonlinearity plays an important role in an oscillation , such a system shows mode - locking behaviour similar to the deterministic nonlinear oscillators @xcite .
more recently , in gene network models with negative feedback , it has been shown that the noise due to finiteness of the number of molecules can modify the condition for oscillatory behaviour @xcite or enhance the oscillation @xcite .
it would also be interesting to see the entrainment behaviour in such systems .
this study was supported by the danish national research foundation through the center for models of life . nelson de , ihekwaba ae , elliott m , johnson jr , gibney ca , foreman be , nelson g , see v , horton ca , spiller dg , edwards sw , mcdowell hp , unitt jf , sullivan e , grimley r , benson n , broomhead d , kell db , white mr .
_ s__cience * 306 * , 704 - 8 .
naama geva - zatorsky , erez dekel , eric batchelor , galit lahav , and uri alon , `` fourier analysis and systems identification of the p53 feedback loop '' , pnas * 107 * , 13550 ( 2010 ) .
lang m , waldherr s , allgower f `` amplitude distribution of stochastic oscillations in biochemical networks due to intrinsic noise '' .
pmc biophys * 2*:10 , ( 2009 ) .
jensen m , bak p , and bohr t , 1983 _ _ p__hys .
50 * 1637 - 9 jensen m , bak p , and bohr t , 1984 _ _ p__hys
a * 30 * , 1960 - 9 nelson , d.e . ,
ihekwaba , a.e.c . , elliott , m. , johnson , j.r . ,
gibney , c.a . ,
foreman , b.e .
, nelson , g. , see , v. , horton , c.a . , spiller , d.g . , edwards , s.w . ,
mcdowell , h.p . ,
unitt , j.f . ,
sullivan , e. , grimley , r. , benson , n. , broomhead , d. , kell , d.b . and
white , m.r.h . oscillations in nf - kb signaling control the dynamics of gene expression .
2004 _ science _ 306 , 704 .
|
theoretical models that describe oscillations in biological systems are often either a _ limit cycle oscillator _ , where the deterministic nonlinear dynamics gives sustained periodic oscillations , or a _ noise - induced oscillator _ , where a fixed point is linearly stable with complex eigenvalues and addition of noise gives oscillations around the fixed point with fluctuating amplitude .
we investigate how each class of model behaves under the external periodic forcing , taking the well - studied van der pol equation as an example .
we find that , when the forcing is additive , the noise - induced oscillator can show only one - to - one entrainment to the external frequency , in contrast to the limit cycle oscillator which is known to entrain to any ratio .
when the external forcing is multiplicative , on the other hand , the noise - induced oscillator can show entrainment to a few ratios other than one - to - one , while the limit cycle oscillator shows entrain to any ratio .
the noise blurs the entrainment in general , but clear entrainment regions for limit cycles can be identified as long as the noise is not too strong . * biological systems present us with a wide range of oscillators , which include cell cycles , circadian rhythms , calcium oscillations , pace maker cells , and protein responses , but it is often a challenging task to identify the minimal models behind these oscillations .
the proposed models are typically categorized into two classes : ( i ) _ limit cycle oscillator _
, where fixed points are linearly unstable and the oscillations are described by stable limit cycles sustained by nonlinearities of the system which are deterministic .
noise can be added on the top of the deterministic oscillations .
( ii ) _ noise - induced oscillator _
, where the fixed point is linearly stable for the system without noise and the system relaxes to the fixed point with damped oscillations when temporally perturbed .
addition of noise to this type of system is known to show sustained oscillations with fluctuating amplitudes .
we propose a way to distinguish the two , by using the phenomenon of entrainment to a periodic perturbation .
taking the van der pol equation with noise as an example , we show that entrainments to all the rational ratios are seen only in the limit cycle oscillator . in the case of the noise - induced oscillator with additive external forcing
, the oscillator can entrain only at one - to - one ratio , meaning that the entrainment to other than the one - to - one ratio is the sign of the dominance of the limit cycle mechanism .
when the external forcing is multiplicative , we find that the noise - induced oscillator with weak nonlinearity can show some entrainment ratios other than one - to - one , but not all the ratios . *
|
since the relativistic laser - plasma interaction was first investigated by akhiezer and polovin @xcite , many nonlinear phenomena have been found , such as solitons , vortices , double layers and so on .
in particular the discovery of electromagnetic solitary waves have attracted the attention of many people due to its robust and resilient behaviors @xcite . in past decades
many works have been performed in conventional electron - ion ( ei ) plasma @xcite . on the other hand , there exists also electron - positron - ion ( epi ) plasma in most astrophysical environments and in the laboratory @xcite , for example , in the pulsar magnetosphere @xcite , the active galactic nuclei @xcite , and the early universe @xcite .
recently the studies of nonlinear waves in the intense laser field interacting with epi plasma have revealed some new features @xcite .
usually the epi plasma is very different from the conventional ei plasma and also different from the pure electron - positron ( ep ) plasma @xcite .
an obvious fact is that there is no soliton solutions for pure ep plasma because the charge density cancels each other for charge neutralized electrons and positrons .
let us recall some researches that have been done about the localized solitary wave solutions in two - component plasma , ei or / and ep , and three - component plasma , epi .
farina and bulanov @xcite studied the relativistic electromagnetic solitons in ei and ep plasma . in our previous study we obtained the parameter region of existence of solitons and the bifurcation diagram for ei plasma @xcite .
@xcite studied the relativistic solitary waves in cold ep plasma in an external magnetic field .
they also found the large amplitude envelope solitons analytically in epi plasma @xcite . by using the reductive perturbation method , mahmood _ et al . _
@xcite obtained the small amplitude korteweg - de vries ( kdv ) soliton solutions in hot epi plasma and they also considered the influence of temperature ratio of electrons to positrons on solitons .
@xcite studied the manifolds of periodic solutions by poincar section plots , they found that the ion dynamic strongly affect the structure of the phase plot and the ion motion must not be neglected in the high - intensity regime .
some other works are also done , such as the studies of ion - arouse solitary waves @xcite and vortices @xcite in epi plasma .
the propose of this paper is to investigate solitary waves in the cold and unmagnetized epi plasma .
we will focus on three important aspects .
first we have to choose a set of parameters which are relatively practical and physical by the requirement of approximated wave dispersion .
it is noted that this point is always either omitted or less discussed in many publications before .
second we employ the numerical technique to solve the coupling equations of envelop of laser field and potential of plasma electrostatic field , which can provide an exact solitary wave solution .
third we generalize the approximate perturbation analytic method developed by others @xcite for immobile ion case to the case including the ion motion .
motivated by these factors , we study both cases of mobile and immobile ions .
the numerical results are compared to the analytical ones .
we also hope that our results in this paper are helpful to clarify the properties of nonlinear solitary waves in epi plasma . before giving the exact numerical solutions , we have made a detailed stability analysis for fixed point of the system and the parameter region for the existence of solitary waves , like in ref .
@xcite , which can give relatively a deeper understanding of nonlinear coherent structure in multi - component plasma .
the paper is organized as follows .
section ii reviews the theoretical model of laser - plasma interaction for completeness of paper , and presents the parameter region that solitary waves may exist analytically in terms of nonlinear dynamics method .
section iii exhibits the exact numerical solutions in both cases of mobile and immobile ions . and
the numerical results are compared to the approximate analytical solutions .
a conclusion is given in sec .
finally the derivation of analytical method by perturbation series expansion technique generalized to the mobile ion case is given in appendix .
the basic formalism of studied problem is certainly from the maxwell equations and hydrodynamic equations . in the following the involving physical quantities are normalized as time @xmath0 , length @xmath1 , potential @xmath2 , electromagnetic fields @xmath3 and @xmath4 , density @xmath5 , velocity @xmath6 and momentum @xmath7 , respectively , where @xmath8 is the electron plasma frequency
, @xmath9 is the electron rest mass and @xmath10 is the unperturbed electron density .
charges @xmath11 are normalized to @xmath12 , where @xmath13 indicates the particle species ( @xmath14 ) , so for electron @xmath15 , for positron and ion @xmath16 , @xmath17 is defined as @xmath18 , so for electron and positron , @xmath19 , and for ion ( for simplicity we assumed it as proton ) @xmath20 .
we shall consider that the system is initially an equilibrium state which is characterized by @xmath21 to ensure the charge neutrality .
the positron fraction can be defined as @xmath22 , which is the ratio of positron to electron initially . by using the normalized quantities mentioned above
, the maxwell equations and hydrodynamic equations are written in the form as @xmath23 where @xmath24 is the canonical momentum , @xmath25 is the relativistic factor and @xmath26 is particle fluid velocity .
note that we have chosen the coulomb gauge @xmath27 . as in most studies
before , for convenience , we just consider the problem that all the quantities vary only along the direction of propagation @xmath28 .
we look for the solution of vector potential as the form of @xmath29 , where the new variables @xmath30 , @xmath31 are introduced and @xmath32 is the laser group velocity . by the way the wave traveling phase velocity is just @xmath33 . in the quasistatic approximation and the initial plasma conditions of @xmath34 at infinity
, then we get the reduced coupling equations as follows @xmath35 where @xmath36 ( obviously @xmath37 ) and two auxiliary quantities are introduced as @xmath38,\label{eq7}\\ \mathcal{f}_2=\sqrt{1-\varepsilon^2}\bigg[\frac{\psi_e}{r_e}-\frac{\chi\psi_p}{r_p } -\frac{(1-\chi)\psi_i}{r_i } \bigg ] .
\label{eq8}\end{aligned}\ ] ] with @xmath39 and @xmath40 .
now eqs.([eq5 ] ) and ( [ eq6 ] ) constitute a set of coupled ordinary differential equations ( odes ) for the vector potential of the laser pulse and wake potential of plasma , which are our starting point for further investigations in the following .
we shall first discuss the nonlinear dynamics that is hidden in these equations . in order to predict the existence of the solitary waves , eqs.([eq5 ] ) and ( [ eq6 ] )
can be reduced to a set of one - order odes as @xmath41 , @xmath42 , @xmath43 , @xmath44 , where the prime denotes as @xmath45 .
the fixed points can be calculated from @xmath46 , @xmath47 , @xmath48 and @xmath49 .
for example , in the case of mobile ion the fixed point @xmath50 is found .
it is easy to find that the characteristic roots of jacobi determinant are @xmath51/\varepsilon^2-\omega^2 $ ] and @xmath52/(1-\varepsilon^2 ) $ ] .
let us denote @xmath53 .
obviously if @xmath54 , the characteristic roots are @xmath55/\varepsilon^2-\omega^2}$ ] and @xmath56/(1-\varepsilon^2 ) } $ ] . accordingly
the stability analysis mentioned above shows that the fixed point @xmath57 is a saddle - center when @xmath58 .
this means that in this region there exist solitary waves . because ions are much heavier than the electrons , so the ion motion could be neglected for simplified treatment in some situations .
thus for immobile ion case , similar to the mobile ion case , the coupling equations are in the same form except @xmath59 , and with @xmath60,\label{eq9}\\ & & \mathcal{f}_4=\sqrt{1-\varepsilon^2}\bigg[\frac{\psi_e}{r_e}-\frac{\chi\psi_p}{r_p } -\frac{(1-\chi)}{\sqrt{1-\varepsilon^2 } } \bigg ] , \label{eq10}\end{aligned}\ ] ] replacing the former @xmath61 and @xmath62 .
similarly we find that the fixed point is also @xmath63 and the stability analysis shows that it is also the saddle - center .
the solitary waves can be available in the region of @xmath64 .
( color online ) solitary wave solutions of vector potential @xmath65 in ( a ) and electrostatic potential @xmath66 in ( b ) for mobile ion case .
the positron fraction is @xmath67 and the other parameters are seen in text.,width=566 ]
before giving numerical and analytical solutions in the following , it is necessary to see how to choose the system parameters which are important for practical physical system .
it is well known that in non - relativistic frame we have the normalized laser group velocity @xmath68 for conventional ei plasma ( note that the normalized laser frequency is the ratio of real laser frequency @xmath69 to equilibrium plasma frequency @xmath70 , i.e. @xmath71 ) .
this relation comes from the dispersion of transverse electromagnetic waves propagating in plasma .
a further relation is that @xmath72 which will lead to @xmath73 .
however in the weak relativistic frame due to the plasma mass is modified by the relativistic factor so that @xmath74 .
similarly for eip plasma we have @xmath75 for both of mobile and immobile ion cases .
these discussions lead to two facts .
one is that indeed the practical laser and plasma parameters can lie in the region where the solitary waves can exist analyzed in the last section .
the other , maybe more important , is that the production quantity @xmath53 should be near to the value of bifurcation point @xmath76 because the studied problem is the weak relativistic in which the relativistical factor @xmath77 is a little larger than @xmath78 .
in fact our studies below support and confirm these analysis . for convenience
we should remind that the larger parameter @xmath79 ( @xmath80 ) corresponds to the higher plasma density but the smaller laser group velocity .
the 4th order rung - kutta method is employed in our numerical techniques .
we can find a series of numerical solitary wave solutions for eqs.([eq5 ] ) and ( [ eq6 ] ) with different parameters of @xmath81 and @xmath82 if @xmath83 is fixed . for illustration
some typical numerical results are shown in fig.[fig1 ] in mobile ion case .
specifically , fig.[fig1](a ) is the envelop of vector potential @xmath65 and fig.[fig1](b ) is the electrostatic potential @xmath66 .
we have chosen @xmath67 and five sets of @xmath84 and @xmath85 as @xmath86 = @xmath87 ( black line ) , @xmath88 ( red line ) , @xmath89 ( green line ) , @xmath90 ( blue line ) and @xmath91 ( cyan line ) . from the discussion mentioned above these parameters are chosen to near the bifurcation point , i.e. , @xmath92 but @xmath93 .
one can clearly see that with the increase of plasma density the amplitudes of solitary waves are increased accordingly in the mobile ion case . for immobile ion case
, we can get the corresponding solitary wave solutions in the same five sets of @xmath84 and @xmath85 , and the amplitudes of solitary waves increase also with the plasma density .
( color online ) maximum values of vector potential @xmath94 and scalar potential @xmath95 in mobile ( black squares ) and immobile ( red circles ) ion cases.,width=566 ] for comparisons , the maximum values of vector potential and scalar potential in both of mobile ( black squares ) and immobile ( red circles ) ion cases are plotted in fig.[fig2 ] .
it is found that the solitary wave amplitudes of vector potential @xmath65 and electrostatic potential @xmath66 in the mobile ion case are always larger than that in the immobile ion case under the same conditions of @xmath84 , @xmath85 , and @xmath83 .
it is not surprising because only stronger laser can excite the larger plasma motion , which intrigues the larger plasma potential .
this self - consistent physical picture is a typical nonlinear characteristic in laser - plasma interaction which is associated strongly to the ponderomotive of laser field as well as the electrostatic force due to plasma charge density . in order to compare the numerical solutions to analytical results
the derivation of analytical solutions is given in appendix .
we get the concrete analytical expression for solitary wave solutions as eq.([eqa17 ] ) by using a similar perturbation series expansion method which has been developed by many people @xcite but the scheme and results are generalized to include the ion dynamics .
( color online ) maximum values of vector potential @xmath94 and scalar potential @xmath95 for mobile ion case ( upper row ) and immobile ion case ( lower row ) by the numerical results ( black squares ) and analytical results ( red circles).,width=566 ] figure [ fig3 ] demonstrates the maximum values of vector potential @xmath94 and scalar potential @xmath95 in both of mobile ( upper row ) and immobile ( lower row ) ion cases with the numerical results ( black square ) and analytical results ( red circles ) .
the choice of system parameters is the same as in fig.[fig1 ] .
one can find that as the increase of the plasma density , @xmath79 , the amplitudes of solitary waves @xmath65 and @xmath66 are increasing monotonously too .
meanwhile the analytical results agree with the numerical results very well for small values of @xmath79 , i.e. relatively tenuous plasma .
however , as the plasma becomes dense enough , the underestimation of analytical solutions becomes more and more severe .
( color online ) the numerical and analytical solutions of vector potential @xmath65 for @xmath67 , @xmath96 , and @xmath97 in both of mobile ion case ( upper row ) and immobile ion case ( lower row ) .
, width=566 ] for the case of smaller plasma density , @xmath98 , and when parameters @xmath99 and @xmath100 are given , the numerical and analytical solutions of vector potential @xmath65 in both mobile ion case ( upper row ) and immobile ion case ( lower row ) are shown in fig.[fig4 ] . the numerical solutions and the analytical ones
are plotted by black solid line and red solid line , respectively .
indeed the numerical results and analytical ones are almost the same because the analytical expression is valid well for the choice of parameters . on the other hand for the higher plasma density , @xmath101 , and given the other parameters as @xmath99 and @xmath102 , the corresponding numerical solutions and analytical solutions in both of mobile ion case ( upper row ) and immobile ion case ( lower row ) are shown in fig.[fig5 ] .
obviously there are remarkable discrepancy between the numerical and analytical results .
we think it attributes to a fact that the series expansion terms are not enough . if we take higher orders of terms in eq.([eqa4 ] ) , for example @xmath103 or @xmath104 , the analytical solutions will become more accurate which will be in accordance with the numerical ones .
these observations indicate that the limiting condition eq.([eqa19 ] ) in analytical method plays an important role .
the series expansion would lose its validity rapidly if the convergence condition is violated , e.g. , for the higher plasma density .
therefore the numerical solving scheme , presented in this paper , is necessary due to its exactness and simplicity compared to the analytical method .
( color online ) the numerical and analytical solutions of vector potential @xmath65 and electrostatic potential @xmath66 for @xmath105 and @xmath106 in both of mobile ion case ( upper row ) and immobile ion case ( lower row).,width=566 ] to see the effect of positron fraction on solitary waves we plot in fig.[fig6 ] the maximum values of vector potential @xmath94 and scalar potential @xmath95 for mobile ion case ( upper row ) and immobile ion case ( lower row ) with the numerical results ( black square ) and analytical results ( red circles ) .
one can find that as increase of @xmath83 both values of @xmath94 and @xmath95 are increased monotonously , i.e. , the ratio of positron to electron density can enhance the amplitudes of solitary waves .
we also find the analytical results can fit the numerical results very well with small values of @xmath83 but deviate more and more as @xmath83 becomes large .
especially when @xmath83 lies in the region of @xmath107 ( see appendix ) , the analytical method will be invalid severely .
furthermore , by comparing the values of @xmath94 and @xmath95 in fig.[fig6 ] , we can find that the maximum values of @xmath65 and @xmath66 in mobile ion case is larger than that in immobile ion case for the same parameters .
it is just the effects of ion dynamics on the solitary wave solutions .
( color online ) the numerical ( black squares ) and analytical ( red circles ) results of vector potential maximum values @xmath94 and scalar potential maximum values @xmath95 changing with @xmath83 for both of the cases of mobile ion ( upper row ) and immobile ion case ( lower row ) with parameter @xmath108.,width=566 ]
in a summary , in this paper , the solitary waves have been found numerically and analytically in electron - positron - ion plasma with the condition of charge neutrality initially and we have considered both cases of mobile and immobile ions . the existence of solitary wave solutions have been found analytically in a certain parameter region .
we also find that when ion is mobile , compared to immobile ion , both the parameter region and the amplitudes of solitary waves become larger .
in addition , the positron concentration has a significant impact on the solitary waves in the considered plasma system .
it is found that as increase of positron fraction , the parameter region of solitary waves becomes larger and the amplitudes of the waves too .
we also compare the numerical results with analytical solutions , and they are almost the same except for the cases of high plasma density or / and high positron fraction .
it is worthy to note that our research may be helpful to understand charged particle acceleration in plasma , for example , the electrons can be accelerated from the bottom to reach the top in wake plasma potential .
the coupling between laser field @xmath65 and wake field @xmath66 leads to nonlinear localized structure like solitary waves presented in this paper would not only make the effective charged particle acceleration possible but also open a possible new way to increase the acceleration efficiency through adjusting the system parameters , e.g. , the plasma density and the positron fraction , however , which beyond the scope of present paper and need to study in future in detail .
this work was supported by the national natural science foundation of china ( nnsfc ) under the grant no.11175023 and no .
11335013 and partially by the fundamental research funds for the central universities ( frfcu ) .
the computation was carried out at the hscc of the beijing normal university .
by employing the series expansion method which was presented in ref.@xcite , we calculate the analytical solutions of eq.([eq5 ] ) and eq.([eq6 ] ) . in this conservative system , it is not hard to find the integral of motion as @xmath109 where @xmath110 and @xmath111 is the integration constant which can be calculated from the boundary condition . by using the energy integral and eliminating the independent variable @xmath112 in eq.([eq5 ] ) and eq.([eq6 ] ) , * we * can get the following equation : @xmath113 \frac{d\phi}{d\mathcal{a}}-h(\phi,\mathcal{a})=0 \label{eqa3}\end{aligned}\ ] ] we assume that the electrostatic field @xmath114 is the function of field amplitude @xmath115 and use the series expansions as @xmath116 with @xmath117 . for @xmath118 , @xmath119 , @xmath120 and @xmath121 can be expressed by @xmath122 as @xmath123 we also assume the following expansions are valid @xmath124^{\pm\frac{1}{2}}= v^{\pm 1}(1\pm\frac{1}{2 v^2}\sum_n a_n \mathcal{a}^n),\nonumber\ ] ] @xmath125^{\pm\frac{1}{2}}= v^{\pm 1}(1\pm\frac{1}{2 v^2}\sum_n b_n \mathcal{a}^n ) , \label{eqa6}\ ] ] @xmath126^{\pm\frac{1}{2}}= v^{\pm 1}(1\pm\frac{1}{2 v^2}\sum_n d_n \mathcal{a}^n).\nonumber\ ] ] by inserting eq.([eqa4 ] ) and eq.([eqa6 ] ) into eq.([eqa3 ] ) , we have @xmath127\mathcal{a}^2 \bigg(\sum_{n=1}nc_n\mathcal{a}^{n-1}\bigg)^3\nonumber\\ & & + \frac{4}{v^2\mathcal{r}^2}\sum_{n=1}[b_n+\chi a_n+\rho_i(1-\chi)d_n]\mathcal{a}^{n+2}\bigg(\sum_{n=1}nc_n\mathcal{a}^{n-1}\bigg)^3\nonumber\\ & & + \frac{4}{\mathcal{r}^2}\bigg\{-\frac{1}{2v^2}\sum_{n=1}[b_n-\chi a_n-(1-\chi)d_n]\mathcal{a}^{n+1 } + [ 1+\chi+\rho_i(1-\chi)]\sum_{n=1}c_n\mathcal{a}^{n+1}\nonumber\\ & & -\frac{1}{2v^2}\sum_{n=1}c_n\mathcal{a}^{n+1}\sum_{n=1}[b_n+\chi a_n+\rho_i(1-\chi)]\mathcal{a}^n\bigg\}\bigg(\sum_{n=1}nc_n\mathcal{a}^{n-1}\bigg)^2\nonumber\\ & & + 2\bigg\{\bigg[\frac{1+\chi+\rho_i(1-\chi)}{\mathcal{r}}-2\omega\bigg]\mathcal{a}-\frac{1}{\mathcal{r}^2 } \sum_{n=1}\bigg(b_n+\chi a_n+\frac{1-\chi}{\rho_i}d_n\bigg)\mathcal{a}^n\nonumber\\ & & -\frac{1}{2v^2\mathcal{r}}\sum_{n=1}[b_n+\chi a_n+\rho_i(1-\chi)d_n]\mathcal{a}^{n+1}\bigg\}\sum_{n=1}nc_n\mathcal{a}^{n-1}\nonumber\\ & & -\frac{1}{\mathcal{r}}\bigg\{[1+\chi+\rho_i(1-\chi)]\sum_{n=1}c_n\mathcal{a}^n-\frac{1}{2v^2 } \bigg[\sum_{n=1}(b_n-\chi a_n-(1-\chi)d_n)\mathcal{a}^n\nonumber\\ & & + \sum_{n=1}c_n\mathcal{a}^n\sum_{n=1}(b_n+\chi a_n+\rho_i(1-\chi))\mathcal{a}^n\bigg]\bigg\}=0 , \label{eqa7}\end{aligned}\ ] ] and by using eq.([eqa5 ] ) and eq.([eqa7 ] ) , we can get the first two expansion coefficients of @xmath66 , i.e. , @xmath128 and @xmath129 : @xmath130 -8\mathcal{r}(1-\mathcal{r})\omega},\label{eqa8}\\ & & c_2=c_1\kappa_1/\kappa_2 , \label{eqa9}\end{aligned}\ ] ] where @xmath131(1-\chi)-\mathcal{r}(1+\chi ) } { 2\mathcal{r}(1-\mathcal{r})}\nonumber\\ & & + \frac{2c_1\{2(1+\chi)c_1+(1-\chi)[2\rho_ic_1-\mathcal{r}(1-\rho_i^2)]\}}{\mathcal{r}^2(1-\mathcal{r})}\nonumber\\ & & + \frac{6[1+\chi+(1-\chi)\rho_i]-8\mathcal{r}\omega}{\mathcal{r}^2}c_1 ^ 2,\label{eqa10}\\ \kappa_2=&&\frac{[1+\chi+(1-\chi)\rho_i](16 - 15\mathcal{r})}{\mathcal{r}(1-\mathcal{r})}-16\omega . \label{eqa11}\end{aligned}\ ] ] expanding @xmath135 with eq.([eqa4 ] ) and eq.([eqa6 ] ) , then eq.([eqa13 ] ) can be reduced to @xmath136\mathcal{a}^2 -\frac{2}{\mathcal{r}}\sum_{n=1}\frac{b_n+\chi a_n+\rho_i(1-\chi)d_n}{n+1}\mathcal{a}^{n+2}.\label{eqa14}\ ] ] with a limit of @xmath137 , eq.([eqa14 ] ) becomes @xmath138 where @xmath139,\nonumber\\ & & \alpha_2=1+\chi+\rho_i^3(1-\chi)-\frac{2(1-\rho_i^2)(1-\chi)}{\mathcal{r}}c_1,\label{eqa16}\\ & & \alpha_3=-\frac{2}{3\mathcal{r}}\{[1+\chi+\rho_i^3(1-\chi)]c_1 ^ 2 + 2(1-\rho_i^2)(1-\chi)c_2\}.\nonumber\end{aligned}\ ] ] for @xmath140 , the solution of eq.([eqa15 ] ) is got @xmath141 where @xmath142 as the argument in ref.@xcite , for @xmath143 satisfying @xmath144 the series eq.([eqa4 ] ) converged so rapidly that the solution eq.([eqa17 ] ) is almost the same to the exact one , see fig.[fig4 ] .
furthermore , from @xmath140 , we can get @xmath145 which is the same region as we have got with nonlinear dynamics method .
so to keep the series expansion method is valid , the range of values for @xmath146 must satisfy eq.([eqa19 ] ) and eq.([eqa20 ] ) .
another thing should be noted is that when @xmath147 this analytical method is invalid .
|
effects of ion mobility and positron fraction on solitary waves of envelop of laser field and potential of electrostatic field in weak relativistic electron - positron - ion plasma are investigated .
the parameter region for the existence of solitary waves is obtained analytically , and the reasonable choice of parameters is clarified . both cases of mobile and immobile ions
are considered .
it is found that the amplitudes of solitary waves in the former case are larger compared to the latter case . for small plasma density ,
the localized solitary wave solutions in terms of approximate perturbation analytical method are consistent well with that by exact numerical calculations .
however as the plasma density increases the analytical method loses its validity more and more .
the influence of the positron fraction on the amplitudes of solitary waves shows a monotonous increasing relation .
implication of our results to the particle acceleration is also discussed briefly .
|
since the end of october 2008 , there has been a tremendous surge of interest and excitement , and then a consequent flurry of activity among the astroparticle and high energy physicists .
the two works that shook the world of astroparticle and cosmic ray physics in the last one fortnight are the startling revelations made by adriani et al in two consecutive reports@xcite : ( i ) the pamela satellite experiment by adriani et al@xcite observed and convincingly demonstrated a sharply rising ratio of the positron flux measurements upto the studied 10 - 100gev range of ( secondary ) energy ; ( ii ) conversely , the findings on the antiproton - to - proton flux showed no such similar behaviour ; rather the ratio - values showed strong tendencies of flattening in the range of 80 - 100gev secondary energy@xcite .
this striking contrast in the nature of antiparticle - to - particle ratios poses a serious puzzle to the theorists and occupies the centre - stage of astroparticle physics domain today , stimulating more than two dozens of works within a very very short span @xcite-@xcite .
in fact , some previous studies@xcite had already given very careful but strong hints to the possibilities of such discoveries finally reported only very recently , for which the paradox , once formally reported , instantly caught such a large number of physicists to a feverish pitch compelling them to act .
in the domain of astroparticle physics , particularly with regard to the studies on dark matter ( dm ) and the weakly interacting massive particles ( wimps ) , searches for antimatter cosmic rays comprising positrons and antiprotons constitute a very important corner .
generically , the spectra of both positrons and antiprotons are expected to fall with increasing energy , relative to the corresponding matter particles which are here obviously electrons and protons respectively .
any deviation from such standard expectations might be a strong indication of any new primary cosmic rays@xcite .
in fact , such discoveries might unravel new windows to the physics of dark matter ( dm ) and/or provide valuable clues to the sources of ultrahigh energy cosmic rays , both of which are still thoroughly unknown .
these factors explain the reasons for being drawn to such intense activities by the physicists on the issue in question which really remains an enigma to the adherents of the standard models ( sms ) in high energy physics , astroparticle physics and cosmology , with all their ramifications and interconnectedness . in the present work here
we will concentrate on understanding the nature of the positron flux ratio alone in the light of some non - standard ideas , hypotheses and ansatzs about the structure of particles , the mechanism for particle interactions and finally the mode of multiple production of hadrons .
very interestingly , the other ratio of @xmath0 turns , in the light of the model(s ) applied here , to a non - issue . because , there is an exclusivity on the production of positrons arising out of the putatively novel concepts about the structure of hadrons .
this helps us to obtain with some non - standard additional and asymmetric sources for @xmath1-production whereas , for the antiprotons no such asymmetric source exists . before digressing , at the very outset
, we would like to make a few comments .
most surprisingly , the results are no wonder to us ; because we had appreciated and emphasised@xcite-@xcite the importance and impact of such similar findings through a published work roughly twenty years ago@xcite in a chain of related works ; based entirely on the same new paradigms which are outlined in the next section in some detail .
the present work is just the resurrection of some of our past works@xcite-@xcite with the latest and newly obtained data from pamela collaboration .
so neither we build up any new model nor we refurbish the old model .
we simply apply an old model built up by one of the authers ( sb ) to explain the characteristics of the new data produced by pamela collaboration on excess production of positrons . by way
, pamela is the acronym for payload for antimatter - matter exploration and light - nuclei astrophysics"@xcite .
\(i ) in the realm of particle physics we would introduce the concept integer - charged partonic constituents for hadrons ( both mesons and baryons ) with an old five - parton model . and the partons , from the viewpoint of these radical works@xcite-@xcite are identified to be the muonic leptons like positively and negatively charged muons and the muonic neutrinos .
so , borrowing the phrase from baek and ko@xcite , the hadrostructures here are really ` leptophilic ' or , more precisely , ` muonophilic ' .
the details are to be obtained from the works by one of us ( sb)@xcite and the references therein .
\(ii ) in the fields of astroparticle and cosmic ray phenomena we introduced the concept of nucleon - breaking mechanism , while the ultra high energy cosmic rays pass through and collide with intergalactic medium consisting of various light nuclei in highly energised states .
this concept is entirely different from the ideas of proton- or neutron - decay .
in fact , our previous studies on the nature of positron spectrum and estimation of positron flux fractions were based on these two sets of new ideas which mark a radical departure from the standard model(sm)-based points of view .
\(iii ) our approach to the studies on this particular problem of excess production of positron fractions pertains in no way to the existence or annihilation of cosmological dark matter(dm)@xcite - cold or hot , with spin zero , unity or else . nor the little higgs scalars are of any concern to us .
\(iv ) besides , none of the standard model - related ideas like weakly interacting massive particles ( wimps ) , supersymmetric particles ( sps ) or supersymmetry , for that matter , and the kaluza - klein particles in extra dimensions would be our concern here .
\(v ) regarding our choice on the propagation model for the ultra high energy cosmic rays we will avoid the complicated ones and accept the simplest and most commonly used model , called simple leaky box model ( slbm ) .
based on the above - stated assumptions , ansatzs and arguments , we deduced roughly two decades ago the formula for excess production of positrons in one of our previous works@xcite-@xcite .
so , in order to prevent the repetitive presentations of the same calculational steps , we will simply extract@xcite for our present purpose the final working expression therefrom given below .
the standard sources for electrons and positrons are given by @xmath2 @xmath3 @xmath4 @xmath3 @xmath5 @xmath6 @xmath3 @xmath7 @xmath3 @xmath1 where the pions shown here are the secondaries abundantly produced in nature by multiple production of hadrons known as multiparticle phenomena .
similarly , secondary protons and antiprotons are also produced by multiparticle mechanism .
in fact , this is the only method for production of the proton - antiprotons . as stated above
, the standard source of cosmic electron production is the normal route of negatively charged pion - decays , where pions of all varieties are produced by multiple production of hadron mechanism .
there is no other extraneous source of electron production .
but for production of cosmic ray positrons the additional , exotic and non - standard source is the proposed nucleon - breaking mechanism with positive muons residing inside the structure of protons as its constituents . through the proposed proton - breaking mechanism supposed to be operational only at superhigh energies
positive muons are set free , which then could produce positrons through their normal decay channel .
these twin ( standard and non - standard ) sources contribute to the positrons .
the model- based final expression for positron fraction , defined as the ratio of the total flux over total electron plus positron flux , is given by@xcite @xmath8 where @xmath9 is a parameter related to the physics of proton - breaking mechanism , @xmath10 is the cut - off angle of emission or detection of the positrons with the vertical . in the experiments by adriani et al
the cut - off angle is not precisely mentioned for which we have calculated for both small - angle @xmath11(fig.1 ) and large - angle ( fig.2 ) emissions of the positrons .
@xmath12 is a parameter which is to be chosen by the methods of fitting the data .
we have maintained the same value @xmath12 for both small - angle and large - angle scattering .
the parameter values are given in table-1 and table-2 .
.chosen numerical values of the fit parameters in the expression for the positron fraction[with small - angle emission ] . [ cols="^,^,^,^",options="header " , ]
the excellent agreement between the measurements and our model - based results are quite evident in all the cases of assumed both small - angle or large - angle emissions of the positrons in the pamela experiment .
so the anomaly is resolved by virtue of the calculations done on the basis of the conjectures made here which mark a radical departure from the conventional set - patterns of ideas and the fixed standard beliefs . by the same token of argumentative points
entailed in the new paradigm we also predict here the detection of the excess of cosmic muon neutrinos and antineutrinos at very very high energies due to the breaking of the nucleons by the cosmic spallation process in the integergalactic space .
the muon charge ratios at very high cosmic ray energies depict normally a very slowly rising nature .
had the muons been not very decay - prone , it would also have met the similar nature arising out of the same nucleon - breaking mechanism . in any case , both the @xmath13 and @xmath14 ratios rise with energy at very high cosmic ray energies .
this illustrates and manifests one important aspect of the muon - electron universality property : our success in interpreting the pamela - data reinforces our dependence on models different form the standard model(s ) in the particle physics and the related fields , as the entire edifice of the present work is based on the rejection of the standard model which is artificial and thoroughly questionable from the very start . however , some other comments are in order here for the sake of completeness and totality .
( i ) the electron - positrons are charged particles .
we have , however , neglected here the complications of path - deflections suffered by the charged positrons or electrons arising out of the earth s magnetic field .
it is to be stressed upon that the mechanism suggested here plugs automatically the parallel expectations for the cases of antiprotons , as the nucleon-(or , proton-)breaking mechanism can and does in no way give rise to any excess production of antiprotons .
( ii ) as it is a ratio , the systematic uncertainties in the data are cancelled ; that also allows the cancellation of the solar activity . in reality
solar modulations could arise from the phase of the solar cycle and also from the opposite charges of electrons and positrons .
but the fact is we do not include these probable disturbing effects in order to escape initially the complications contributing to some very minor effects with no significant numerical values .
however , the model we apply here takes care of the experimental data quite well and successfully solves the anomaly evinced by the adjoining figures , without assuming , however , the role of pulsars , the dark matter annihilations or taking into account some other phenomena evolved from the standard model(sm)-based points of view . and
the conclusion derived from this work is in perfect accord with what is maintained by morselli and moskalenko@xcite that the excesses imply a source , conventional or exotic , of additional leptonic component , especially of the positive variety .
and this obviously causes a distantiation from the viewpoints expressed by chen et al@xcite . after the completion of our present work on pamela - paradox , our attention was drawn by a scientific colleague to a very important report@xcite and to a few papers@xcite with some concrete findings .
the report by butt@xcite is an exceptionally brilliant one for its unbiasedness or open - mindedness and near - perfect objectivity .
the bump observed by the atic collaboration is not to be electron - specific ; side by side with the signature - electrons , the positrons are also to be produced in roughly equal measure , as was expressly maintained by butt@xcite , if kaluza - klein(kk ) wimps are their progenitor .
but the atic measurements@xcite were singularly aimed at studying the electrons alone .
so , unless the positrons are measured under the same or similar circumstances by these groups@xcite , no definitive comment on the status of kk wimps is possible . in this context , another comment is yet in order .
the measurements by aharonian et al@xcite for detection of cosmic ray electrons at energies beyond 600gev do not report very clearly , such excesses , as is indicated by chang et al @xcite .
this , therefore , tacitly and indirectly puts a question mark to the measurements of chang et al until further scrutiny .
besides , there are a few sharp differences between the two sets of studies[pamela , heat etc . on the one side , and atic , hess etc .
, on the other ] , for which a comparison of the two sets of findings might not be quite significant .
firstly , the studies by pamela collaboration concentrate uniquely on the studies of cosmic antiparticles like positron(s ) and antiproton(s ) , whereas both the atic and the hess collaborations measure only electron spectrum at much higher energies ; they did not report on the measure of corresponding antiparticle production scenario .
secondly , the energy - ranges of the two sets of experiments are grossly different .
so the studies on the excess production of cosmic electrons alone at much higher energies do not disturb the pivot of our present work on pamela- paradox .
unless atic and hess collaborations study and report pointedly on the nature of positron and antiproton productions at the same energy - range and under the stringently same experimental conditions we see no tangible reason to redefine our attitude and reconsider our outlook . however , if asymmetric excess electron production alone is repeatedly reported and confirmed , only then we will have to investigate in to some exotic sources or nor - yet - proposed - or - known mechanism for cosmic electron production .
so for the present , we are still not in favour of attaching too much importance to the hypothesis of dark matter or the postulates kaluza - klein(kk ) wimps .
* o.adriani et al : astro - ph/0810.4995 v1 28 october 2008 .
o.adriani et al : astro - ph/0810.4994 v1 28 october 2008 .
n. arkani - hamed , d.p.finkbeiner , t.slatyer and n.weiner , arxiv : hep - ph/0810.0713 .
n. arkani - hamed and n.weiner , arxiv : hep - ph/0810.0714 .
d.p.finkbeiner , t.slatyer and n.weiner , arxiv : hep - ph/0810.0722 . m.pospelov and a.ritz , arxiv : hep - ph/0810.1502 .
d.hooper , p.blasi and p.d.serpico , arxiv : astro - ph/0810.1527 .
j.hisano , m.kawasaki , k.kohri and k.nakayama , arxiv : hep - ph/0810.1892 .
h.yuksel , m.d.kistler and t.stanev , arxiv : astro - ph/0810.2784 . m.kamionkowski and s.profumo , arxiv : astro - ph/0810.3233 .
m.boezio et al . , arxiv : astro - ph/0810.3508 .
m.masip and i.mastromatteo , arxiv : hep - ph/0810.4468 .
s.khalil and h.okada , arxiv : hep - ph/0810.4573 .
p.d.serpico , arxiv : hep - ph/0810.4846 .
a.e.nelson and c.spitzer , arxiv : hep - ph/0810.5167 .
f.donato , d.maurin , p.brun , t.delahaye and p.salati , arxiv : astro - ph/0810.5292 .
i.cholis , d.p.finkbeiner , l.goodenough and n.weiner , arxiv : astro - ph/0810.5344 . y.nomura and j.thaler , arxiv : hep - ph/0810.5397 . r.harnik and g.d.kribs , arxiv : hep - ph/0810.5557 .
t.hambye , arxiv : hep - ph/0811.0172 .
p.f.yin , q.yuan , j.liu , j.zhang , x.j.bi and s.h.zhu , arxiv : hep - ph/0811.0176 .
k.ishiwata , s.matsumoto and t.moroi , arxiv : hep - ph/0811.0250 .
y.bei and z.han , arxiv : hep - ph/0811.0387 .
p.j.fox and e.poppitz , arxiv : hep - ph/0811.0399 .
c.r.chen , f.takahashi and t.t.yanagida , arxiv : hep - ph/0811.0477 .
k.hamaguchi , e.nakamura , s.shirai and t.t.yanagida , arxiv : hep - ph/0811.0737 . e.ponton and l.randall , arxiv : hep - ph/0811.1029 . a.ibarra and d.tran , arxiv : hep - ph/0811.1555 .
l.bergstrom , t.bringmann and j.edsjo , arxiv : astro - ph/0808.3725 .
m.cirelli and a.strumia , arxiv : astro - ph/0808.3867 .
v.barger , w.y.keung , d.marfatia and g.shaughnessy , arxiv : hep - ph/0809.0162 .
i.cholis , l.goodgnough , d.hooper , m.simet and n.weiner , arxiv : hep - ph/0809.1683 .
j.h.huh , j.e.kim and b.kyae , arxiv : hep - ph/0809.2601 .
p.brun et al : phy . rev . *
d 76 * ( 2007 ) 083506 .
p.picozza et al : astro .
part . phys .
* 27 * ( 2007 ) 296 .
s.bhattacharrya : il nuovo cimento * c 10 * ( 1987 ) 669 .
s.bhattacharrya : il nuovo cimento * c 10 * ( 1987 ) 209 . s.bhattacharrya and p.pal : il nuovo cimento * c 10 * ( 1987 ) 227 .
s.bhattacharyya : j. phys . *
g 14 * ( 1988 ) 9 .
p.bandyopadhyay , r.k.roychowdhury and s.bhattacharyya : phys . rev .
* d 21 * ( 1980 ) 1861 .
p.bandyopadhyay and s.bhattacharyya : il nuovo cimento * a 43 * ( 1978 ) 305 .
p.bandyopadhyay and s.bhattacharyya : il nuovo cimento * a 43 * ( 1978 ) 323 .
p.bandyopadhyay , s.bhattacharyya , r.k.roychowdhury and d.p.bhattacharyya : il nuovo cimento * a 50 * ( 1979 ) 133 .
p.bandyopadhyay and s.bhattacharyya : lett .
nuovo cimento * 17 * ( 1976 ) 393 .
p.bandyopadhyay and s.bhattacharyya : lett .
nuovo cimento * 17 * ( 1976 ) 399 .
p.bandyopadhyay and s.bhattacharyya : hadronic journal * 7 * ( 1984 ) 313 .
s.baek and p.ko , arxiv : hep - ph/0811.1646 .
f.d.steffen , arxiv : hep - ph/0811.3347 . a.morselli and i.v.moskalenko , arxiv : astro - ph/0811.3526 .
c.r.chen , f.takahashi and t.t.yanagida , arxiv : astro - ph/0811.3357 .
y.m.butt : nature * 456 * ( 2008 ) 329 .
j.chang et al ( atic collaboration ) : nature * 456 * ( 2008 ) 362 .
f.aharonian et al ( hess collaboration ) , arxiv : astro - ph/0811.3894 .
|
very recently , pamela collaboration has formally reported two sets of data on positron and antiproton flux measurements done at very high energies and with unprecedented accuracy .
the reports reveal a puzzle of great topical interest and importance : it is decisively found that there is a sharp smooth rise of the positron fraction , whereas for antiproton production no such rises occur ; rather the fraction either flattens , or shows signs of falling off gradually with increasing energies .
the present work is just an attempt to decipher the riddle with the help of a host of radically new ideas about the particle - structure , the multiparticle - production mechanisms and the concept of nucleon break - downs into the constituent partons .
the application of these ideas found remarkable successes in the past ; exactly similar or more striking are the findings by the present study .
keywords : cosmic - ray interactions with the earth , positron emission , cosmic - rays high energy interactions , dark matter .
+ pacs nos . : 94.94 . , 79.20.mb .
, 13.85.tp .
, 95.35.+d
. = currsize = currsize
|
in recent years a significant amount of research has been dedicated to strongly - correlated materials with reduced dimensionality since they exhibit a large variety of fascinating dimension - related properties .
an example is the layered quantum spin system tiocl , where bilayers of ti - o are separated by cl@xmath1 ions .
this system was originally thought to be a possible candidate for a rvb superconductor upon doping@xcite because of its frustrated triangular lattice geometry .
later on , various experimental measurements@xcite revealed that tiocl shows in fact an anomalous spin - peierls behavior with two consecutive phase transitions .
magnetic susceptibility was initially described in terms of a one - dimensional spin-1/2 heisenberg model with a large intra - chain coupling constant @xmath2k.@xcite .
it is though well known that susceptibility is not very sensitive to different models and recent _ ab initio _
dft studies@xcite showed that the underlying interactions for this system can be understood in terms of a spin-1/2 heisenberg model with strong intrachain antiferromagnetic interactions @xmath3 and weaker interchain ferromagnetic interactions @xmath4 , @xmath5 .
this model reproduces the magnetic susceptibility measurements and sets a framework for understanding the puzzling spin - peierls phase transitions in this compound .
only recently , research has also focused on high - pressure studies@xcite as a possible way to drive the system metallic . at room temperature and ambient pressure ,
the system is a mott insulator with a charge gap of about 2ev.@xcite the electronic structure in this high - temperature phase has been examined by angle - resolved photoemission spectroscopy ( arpes).@xcite in agreement with previous experimental evidence , the results show a strong anisotropy of the correlated band structure , with significant dispersion of the ti @xmath6 bands along the chains ( crystallographic @xmath7-direction ) , and almost flat bands perpendicular to the chains . on the theoretical side ,
the electronic properties of tiocl have been studied by means of _ ab initio _ dft calculations within the local density approximation ( lda ) , the lda+u @xcite , b3lyp @xcite and also in combination with the dynamical mean - field theory ( dmft),@xcite which is a modern method for dealing with strong correlations
. it was shown that a proper treatment of non - local correlations is crucial for a reasonable description of the single - particle gap.@xcite however , the momentum dependence of the spectral function @xmath8 seen in arpes is still puzzling .
it has been shown that an _ ab initio _ calculation without proper treatment of correlations is insufficient.@xcite on the other hand , describing the compound by a simplified one - dimensional strongly correlated model was not successful either.@xcite furthermore , lda+dmft could so far only produce the momentum integrated local density of states ( dos ) without any information on the momentum dependence of the spectra.@xcite this situation , having no calculation for the momentum - resolved spectral function @xmath9 at hand , is partly due to the fact that there are only few methods that can deal with all the requirements of such calculations .
this work is intended to fill this gap and investigates the influence of the true two - dimensional band - structure on the momentum - resolved @xmath9 in the presence of strong correlations .
a successful technique for this purpose is the variational cluster approach ( vca)@xcite . in what follows
we apply a two - step procedure to study the spectral function , as has been proposed by chioncel _
et al._.@xcite first , dft calculations within the lda are carried out , and localized wannier functions are constructed by the n - th order muffin - tin - orbital ( nmto)@xcite downfolding technique . using the lda hamiltonian expressed in these wannier functions as the non - interacting part , and adding coulomb and hund interaction terms , we arrive at the correlated low - energy model . by applying vca to this model hamiltonian , we show that the inclusion of the inter - chain processes leads to a significant redistribution of spectral weight from higher to lower binding energies . since these processes enhance the asymmetry of the strongly - correlated band structure , they are crucial for the reproduction of the asymmetric bands seen in arpes measurements .
our calculations show that the hubbard - model description is appropriate for tiocl if effects beyond the one - dimensional description are included
. moreover these results should be valid for a large variety of correlated low - dimensional coupled - chain systems .
the paper is organized as follows : in sect .
[ sec : theory ] we discuss the construction of the low - energy hamiltonian , as well as the vca , which is subsequently used for the calculation of the correlated spectral function .
[ sec : results ] contains our results of the multi - band as well as of the single - band hubbard model and in sect .
[ sec : discuss ] we present our discussions and conclusions .
in many transition metal oxides electronic correlation effects are very important for a proper description of the physical properties .
however , it is a known fact that first - principle calculations suffer from an insufficient treatment of these effects . in order to take the strong correlations into account in our calculation for tiocl
, we apply a two - step procedure ( _ lda+vca _ ) that has first been introduced by l. chioncel _ et al._.@xcite it consists of the construction of the correlated low - energy hamiltonian based on density - functional theory on the one hand , and the solution of the resulting low - energy hamiltonian using the vca on the other hand . in ref . , the authors study the non - quasiparticle states in the half - metallic compound cro@xmath10 and find good agreement with experiments .
moreover , a comparison with lda+dmft calculations showed the applicability of the lda+vca approach .
recently , it has also been used to explain the pseudogap in tin , where also the momentum - resolved spectral function has been calculated.@xcite for a complete description of the electronic structure of a given material it is in principle necessary to consider all electronic degrees of freedom of the underlying constituents .
calculations within dft can to some extent fulfill this requirement
. however , it is clear that only certain states and orbitals contribute to the low - energy physics .
for this reason one is interested in finding an effective model that describes the low - energy physics on the one hand sufficiently accurate and has , on the other hand , not too many degrees of freedom . in the present case of tiocl , dft calculations within the lda approximation have shown that the relevant orbitals at low energies are the ti 3d orbitals , which are split into @xmath11 and @xmath12 manifolds due to the octahedral crystal field provided by the ligands .
since the ti@xmath13 ion is in a 3d@xmath14 configuration , the relevant states closest to the fermi energy are of predominantly @xmath11 character . for the construction of the low - energy hamiltonian , we performed dft calculations within the lda using the linearized muffin - tin - orbitals ( lmto ) basis set .
the localized orbitals , which are the basis of the interacting model , are constructed using the nmto method . by using the downfolding technique@xcite , the hybridization of the ti-@xmath11 orbitals with the ligand orbitals ( o-@xmath15 and cl-@xmath15 )
are taken into account , yielding an effective set of @xmath11 orbitals .
these orbitals represent the lda band structure with great accuracy , and are used as the non - interacting part of the many - body hamiltonian .
the matrix elements of the nmto hamiltonian @xmath16 in the basis set of localized nmto wannier functions give the transfer integrals @xmath17 , and the non - interacting hamiltonian can be written as @xmath18 the indices label the lattice sites by @xmath19 , @xmath20 , as well as the @xmath11 orbitals by @xmath21 , @xmath22 , and @xmath23 denotes the spin . to include correlation effects into the low - energy description ,
we add interaction terms to the hamiltonian , @xmath24 for convenience , we introduced the chemical potential @xmath25 in the hamiltonian .
we will refer to this hamiltonian as @xmath11 model , and give all energies throughout the paper in units of electron volt ( ev ) .
the full low - energy model eq .
( [ eq : ham ] ) consists of the single - particle terms ( @xmath26 and @xmath25 ) , the diagonal ( density - density ) interactions ( @xmath27 , @xmath28 , and @xmath29 ) , and the non - diagonal ( spin - flip ) term ( @xmath30 , third line ) . in this study
we consider only the case @xmath31 .
@xmath32 is the orbital occupation operator , and @xmath33 , @xmath34 , @xmath35 are the components of the spin-@xmath36 operator on site @xmath19 in orbital @xmath21 .
the interaction parameters @xmath27 , @xmath28 , and @xmath30 are not independent , but fulfill the relation @xmath37 . at this point
it is important to note that we can include a full @xmath38 symmetric exchange term .
since the method we consider is not affected by any sign problem , it has no restriction on the type of couplings that can be included.@xcite for the interaction parameters @xmath27 and @xmath30 one can find several values in the literature , ranging from @xmath39ev to @xmath40ev and @xmath41ev to @xmath42ev.@xcite since we want to study also the influence of these parameters on the single - particle properties , we have performed calculations with different values , and indicate the actual value at the corresponding location in the paper . in this work , we also address the question whether the orbital degrees of freedom are important for the low - energy physics or not . since tiocl does not crystallize in a perfect cubic symmetry , the threefold degeneracy of the @xmath11 manifold is lifted .
lda+u@xcite and also lda+dmft@xcite calculations have shown that the ground state shows predominantly d@xmath43 character ( the local reference frame is @xmath44 , and @xmath45 and @xmath46 axes rotated by 45@xmath47 with respect to @xmath7 and @xmath48 ) , with only very small admixture of the other orbital degrees of freedom , a picture that we will also find in our following calculations .
this is in contrast to ipt - dmft calculations,@xcite where a sizable admixture of the other orbital degrees of freedom is found . in order to investigate the effective one - band model that consists of the d@xmath43 orbital
only , we performed a nmto downfolding procedure integrating out all other degrees of freedom , and keeping only the d@xmath43 channel . in this one - band model ,
the only interaction terms are the ones proportional to the hubbard onsite @xmath27 , and the low - energy one - band hamiltonian finally reads @xmath49 where the hubbard interaction @xmath27 is the same as for the @xmath11 -model .
the effective hopping parameters @xmath50 are again the matrix elements of @xmath16 in the wannier basis set .
triangular lattice structure and two possible clusters tiling the lattice .
i ) 4-site cluster including inter - chain self - energies .
ii ) 1d clusters neglecting inter - chain self - energies . full circles and solid lines
mark sites and bonds inside a clusters .
next - next - nearest neighbor hopping @xmath51 is only drawn once for clarity . in the case of the @xmath11 manifold each lattice site consists of three orbitals . ] after having constructed the low - energy hamiltonian using _ ab initio _ techniques , we use the vca@xcite in order to calculate the spectral function of this model .
since we deal with an effective low - energy hamiltonian that involves no other uncorrelated ligand states ( o-@xmath15 , cl-@xmath15 ) , but only the correlated ti-@xmath11 orbitals , the application of vca is straightforward , and is from a technical point of view exactly equivalent to standard multi - orbital calculations for hubbard - model hamiltonians .
the only difference is that the non - interacting part is determined by the procedure discussed in the previous subsection .
furthermore , since there are no explicit ligand states in the hamiltonian , there is no need for a double - counting correction .
it gives just a constant shift in energy which can be absorbed in the chemical potential . as mentioned in the introduction ,
the vca is a quantum cluster method capable of treating strong short - ranged correlations .
the main idea is to approximate the self - energy of the original model , which is defined on an infinite lattice , by the self - energy of a finite cluster , the reference system .
the variational principle states that the optimal solution is given by the stationary points of the grand potential @xmath52 $ ] as a function of the self - energy @xmath53 . parametrizing the self - energy by the single - particle parameters @xmath54 of the reference system ,
we can write the grand potential as @xmath55 where @xmath56 and @xmath57 are respectively the grand canonical potential and the green s function of the reference system and @xmath58 is the non - interacting green s function of the physical ( lattice ) system .
the stationary condition reads @xmath59 it is important to note that the interaction parameters are not variational parameters , since , by construction of the vca , the interaction terms of the reference system and the original lattice model must not differ . in this study , we restrict ourselves to local interactions only since the vca in its strict sense can not be used for models with non - local interactions without further approximations . in its spirit , the vca is closely related to the dynamical mean - field theory ( dmft ) , where in the latter case the self - energy is obtained from an impurity problem .
the actual vca calculation is done in the following steps .
first , we determine the ground state of the reference system , i.e. a cluster of finite size as depicted in fig .
[ fig : cluster ] . the interacting green s function is calculated , and since the non - interacting green s function of the reference system is known , the self - energy can readily be obtained using dyson s equation . using the grand potential @xmath56 of the reference system , the green s function @xmath57 , and the self - energy @xmath60 , eq .
[ eq : ocalc ] is evaluated using the technique of @xmath61-matrices.@xcite note that the green s functions @xmath58 , @xmath57 and the self - energy @xmath62 in eq .
( [ eq : ocalc ] ) are matrices not only in site and spin indices , but also carry an orbital index .
in fact , this is the only difference of the application of vca in the present case compared to the numerous previous applications to the single - band hubbard model . as a reference system solver
, we use the band - lanczos exact diagonalization technique at zero temperature , which means that for the full @xmath11 manifold , we can easily consider clusters with at most 4 sites , yielding a 12-orbital green s function @xmath57 . for the single - band model ,
we consider clusters up to 12 sites .
we exploit particle number and spin conservation , therefore the sizes of the largest hilbert spaces that we have to consider are @xmath63 states in the 4-site multi - orbital case , and @xmath64 states in the 12-site single - orbital case , respectively . since we are considering an exact diagonalization method for solving the cluster problem , all interactions in the hamiltonians eq .
( [ eq : ham ] ) and eq .
( [ eq : ham1b ] ) are treated exactly and on the same footing .
this is a clear advantage compared to , e.g. , using the hirsch - fye quantum monte - carlo method as impurity solver , since in the latter case approximations to the interaction terms of the hamiltonian have to be done.@xcite the vca approach has been tested thoroughly and used successfully for many investigations in recent years .
several studies on the cuprate - based high - temperature superconductors have shown that this approach can reproduce salient features of these materials , such as the ground - state phase diagram,@xcite or the opening of the pseudogap at low hole doping , accompanied with the occurrence of fermi arcs,@xcite in very good agreement with experiments and results obtained by the cellular dynamical mean - field theory ( cdmft ) ( see , e.g , refs . ) .
recently , the vca could also reproduce the pairing symmetry of the iron - based superconductors.@xcite the vca has also been used for multi - orbital systems , which is relevant for the combination with _ ab - initio _ methods . on the pure methodological level ,
the metal - insulator transition in infinite dimensions was studied@xcite , and very good agreement with dynamical mean - field calculations was found .
an application to real materials was done in refs . , where the compounds nio , coo , and mno have been studied and very good agreement with experimental photo - emission data has been found .
details on the practical implementation of the vca , including tests and benchmarking , can be found in refs . . in general ,
all the single - particle parameters @xmath54 are variational parameters of the vca . in practice
, one chooses a physically motivated subset in order to keep the numerical calculations feasible . here
we make the following choice . for a thermodynamically consistent description of the densities ,
it is crucial to consider the onsite energies , i.e. the local terms of the single - particle hamiltonian @xmath65 , as variational parameters.@xcite we define the average @xmath66 and the crystal - field splitting @xmath67 , which are then used as the variational parameters of the vca .
note that in the single - band case , eq . ( [ eq : ham1b ] ) , one has to deal with @xmath68 only .
one has to be aware that the variational parameter @xmath69 _ does not _ impose an artificial orbital polarization of the system , since it is a parameter in the variational procedure and no physical external field .
hence , using @xmath68 and @xmath69 , the orbital occupancies are determined in a fully self - consistent way .
the main property investigated in this work is the single - particle spectral function which we define as @xmath70 since we broke the translational invariance of the system by introducing the cluster tiling , a proper periodization of the lattice quantities is needed in order to restore translational symmetry , an issue also important in cluster dmft calculations.@xcite here , we choose to use the periodization of the green s function since the periodization of the self - energy gives unphysical results in the insulating phase.@xcite in other words , starting from the green s function that depends on two momenta , @xmath71 , one restores the fully translationally invariant green s function @xmath72 , by neglecting the off - diagonal elements , and taking @xmath73 only .
it has been shown that this is a well justified approximation to calculating the momentum - dependent spectral function.@xcite the green s function @xmath74 is in general a matrix in orbital indices and @xmath75 is given by the trace over the orbital degrees of freedom .
comparison of the single - particle spectral function of the @xmath11 model ( top ) and the single - band model ( bottom ) , both calculated with a @xmath76 reference system , see fig .
[ fig : cluster ] i ) . parameters are @xmath77ev , @xmath41ev . for the hopping parameters
the chemical potential has been chosen such that i ) the system is insulating with @xmath78 , and ii ) the position of the occupied states coincide in both calculations .
lorentzian broadening of @xmath79ev was used . ]
( color online ) comparison of the spectral function of the @xmath11 model for two different sets of interaction parameters .
top : @xmath40ev , @xmath80ev .
bottom : @xmath40ev , @xmath41ev .
the vertical dashed lines mark the edges of the single - particle gap .
lorentzian broadening of @xmath79ev was used . ]
before we come to a detailed analysis of the single - band model , we first want to check the validity of the restriction to the lowest @xmath81-orbital . for this reason , we performed _
ab initio _ calculations to determine the full single - particle hamiltonian of the system .
since ti@xmath13 is in a @xmath82 configuration , the @xmath12 orbitals are unoccupied and can be projected out , and the full kinetic hamiltonian is downfolded to the threefold degenerate @xmath11 manifold .
results show , that the system exhibits a strong anisotropy , with the largest hopping integrals in the crystallographic @xmath7-direction , see fig .
[ fig : cluster ] , almost one order of magnitude larger than the other transfer integrals .
moreover , the threefold degeneracy is lifted and the manifold split into the lower @xmath83 and the higher @xmath84 and @xmath85 orbitals .
note that for the orbital designation we consider the same local reference frame as in ref . .
the crystal - field splitting between the ground state and the first excited state obtained from lda is about 0.42ev .
this theoretical value is in reasonable agreement with experimental results.@xcite despite this splitting , the orbital sector is not fully polarized in the lda calculations , and the occupation of @xmath83 is about 0.49 , with 0.51 electrons in the other two orbitals . in order to perform our lda+vca calculations , we take the downfolded hamiltonian of the _ ab initio _ calculations , and add the interaction and exchange terms according to eq .
( [ eq : ham ] ) . in the upper panel of fig .
[ fig : comp_mo_so ] we show the results for the spectral function , eq .
( [ eq : akw ] ) , calculated for this three - band @xmath11 model using typical parameters @xmath77ev and @xmath41ev . the bands which are located just above the fermi level , between
roughly 0.5ev and 2.0ev , have @xmath84 and @xmath85 character , and remain almost unchanged by the strong interactions .
the behavior of the @xmath83 orbital is strikingly different .
it splits into two bands that can be identified with the lower and upper hubbard band , located roughly around @xmath86ev and 2.5ev , respectively . by inspecting the terms of the hamiltonian related to the crystal - field splitting ,
i.e. @xmath87 , we can calculate the orbital polarization @xmath88 .
we find a value of @xmath89 , meaning that the system is almost perfectly polarized into the @xmath83 orbital , which is in agreement with recent lda+cdmft calculations.@xcite interestingly , this polarization is found without any sizable increase of the crystal - field splitting in the variational procedure , i.e. , @xmath90 , but it is only due to the inclusion of strong local interactions .
this result gives rise to the question , to which extent a single - band hubbard model can describe the occupied states relevant for comparison with arpes .
we took parameters from a full downfolding to the ti @xmath83 orbital only , cf .
first column of table i in ref . .
using the same value of @xmath77ev we calculate the spectral function for hamiltonian eq .
( [ eq : ham1b ] ) .
the results are shown in the lower panel of fig .
[ fig : comp_mo_so ] . in order to avoid effects coming from different cluster sizes
, we used also a @xmath76 cluster for this comparison . note that below the fermi level the agreement between the single - band model and the @xmath83 part of the @xmath11 model is excellent .
for this reason we conclude that for a comparison of spectra with experimental arpes measurements the hamiltonian eq .
( [ eq : ham1b ] ) is a reasonable starting point . before turning to a more detailed analysis of the spectra obtained from eq .
( [ eq : ham1b ] ) , let us briefly discuss the single - particle gap @xmath91 , defined as the energy difference between the lowest unoccupied and the highest occupied state . for a comparison of this quantity with experiments , it is clear that the single - band model is not sufficient , since it does not describe the excited states in the @xmath11 manifold .
however , we extracted the gap @xmath91 from the spectral function of the @xmath11 model , for different values of @xmath27 and @xmath30 .
we find that the main quantity that determines the gap is the inter - orbital coulomb interaction @xmath37 , and we get @xmath92ev for @xmath77ev , @xmath41ev . in fig .
[ fig : comp_u_j ] we plot the spectral function for different sets of interaction parameters , and find @xmath93ev for @xmath94ev , @xmath80ev , and @xmath95ev for @xmath94ev , @xmath41ev .
all these values for the gap are a bit smaller than the experimental charge gap of about 2ev,@xcite but nevertheless in reasonable agreement .
( color online ) imaginary part of the self - energies on a reference system consisting of two coupled 4-site chains . left panel : real space @xmath96 .
solid line : local / onsite self - energy @xmath97 . dashed line : inter - chain self - energy @xmath98 .
dash - dotted line : intra - chain self - energy @xmath99 .
right panel : self - energy for selected momenta of the cluster .
momenta are indicated in the plot , and a vertical shift between momenta has been introduced for improved presentation . the self - energy shows causality ( negative definite ) , almost no dependence in @xmath100-direction , and strong dependence in @xmath7-direction .
lorentzian broadening of @xmath79ev was used in both plots . ]
we have shown in sec .
[ sec : t2 g ] that the occupied states of the @xmath11 manifold are well reproduced by a single - band hubbard model . in this section
we want to investigate the spectral function of eq .
( [ eq : ham1b ] ) in more detail .
as mentioned above , we focus on the effect of the additional two - dimensional hopping processes on the quasi - one - dimensional behavior of tiocl .
first , we want to determine the strength of the correlations along the qualitatively different bonds of the lattice , fig . [
fig : cluster ] . this can be done best by inspecting the self - energy @xmath96 , which is , in vca , a quantity defined on the reference system and , thus , can be readily obtained .
the results for three selected matrix elements are shown in fig .
[ fig : selfenergy ] , calculated on an @xmath101 cluster .
this cluster consists of two coupled 4-site chains in @xmath7-direction . in other words , a cluster similar to the one depicted in fig .
[ fig : cluster ] i ) but with doubled extension in @xmath7-direction .
it is obvious that local ( @xmath97 ) and intra - chain correlations ( @xmath99 ) are strong , but the correlations between adjacent chains ( @xmath98 ) are orders of magnitude weaker .
in addition , we show in the right panel of fig . [ fig : selfenergy ] the fourier transformation of the self - energy , for momenta accessible at this small cluster . again , the influence of momenta perpendicular to the chains is hardly visible in the self - energy , whereas it shows significant momentum dependence in the chain direction .
this leads to the conclusion that the spectra of tiocl should be governed by 1d correlations , modified by single - particle effects due to the coupling of the chains .
spectral function of the single - band model with a 12-site chain as reference system .
top : no coupling between chains .
bottom : inclusion of interchain coupling parameters obtained from lda - downfolding .
dark areas mark large spectral weight .
lorentzian broadening of @xmath79ev was used .
the horizontal striped structures occur due to the breaking of translational invariance in the cluster approximation . ] motivated by this result , we use from now on a @xmath102 cluster as reference system .
since the vca approximates the interacting green s function as @xmath103 with @xmath58 the non - interacting green s function of the model hamiltonian and @xmath104 the cluster self - energy , it is easy to see that the inter - cluster coupling is treated in a single - particle ( i.e. , non - interacting ) manner , since it enters just via @xmath58 . on the other hand ,
this procedure gives the best possible description of the correlation effects along the chains in @xmath7-direction . in fig .
[ fig : comp_singleband ] we show a density plot of the spectral function of the hamiltonian eq .
( [ eq : ham1b ] ) for @xmath77ev . in the upper part we included only the intra - chain hopping @xmath105ev in the calculation , leading to flat bands in @xmath100-direction , i.e. , from @xmath106 to @xmath107 , since in this case
the chains are decoupled . by including additional inter - chain parameters @xmath108ev and @xmath109ev as given in ref .
, we notice a slight dispersion in @xmath100-direction . in @xmath7-direction ,
however , the band positions remain almost unchanged ; we find only changes in the spectral weights .
since this can not be seen clearly in the density plots , we show in fig . [ fig : comp_ak0 ] the evolution of the spectral function at the @xmath0 point @xmath110 when longer - ranged hopping processes are included .
spectral function @xmath111 at the @xmath0 point .
top left : without inter - chain coupling .
top right : including nearest - neighbor inter - chain coupling @xmath112 .
bottom right : including next - nearest - neighbor inter - chain coupling @xmath51 .
for comparison , bottom left : next - nearest - neighbor hopping along the chain , but no inter - chain hopping .
lorentzian broadening of @xmath79ev was used . ]
the upper left panel ( a ) is the spectral function for decoupled chains with the spin - charge separation clearly visible . at the @xmath0
point the holon band is located around @xmath113ev , and the spinon band around @xmath114ev .
including the nearest neighbor inter - chain hopping @xmath112 leads to a significant redistribution of spectral weight from the holon to the spinon band , i.e. , from higher to lower binding energies , see the upper right panel ( b ) .
this effect is even enhanced when the next - nearest inter - chain hopping @xmath51 is included , as shown in the lower right panel ( d ) . here
the spectral weight of the low binding energy ( spinon ) excitation is comparable to the weight of the high binding energy ( holon ) excitation for decoupled chains in panel ( a ) , and vice verse . at this point
we would like to mention , that in a strict sense the terminology spinon and holon is not applicable any more , since these are properties of purely one - dimensional systems . anyway , since the spectra resemble to some extent 1d systems , we still use these terms to distinguish the different excitations . from fig .
[ fig : comp_ak0 ] it is clear that the inclusion of inter - chain processes enhances the asymmetry of the @xmath115-resolved spectra .
the low - lying excitation near the @xmath0 point is strongly enhanced , whereas there is no spectral weight transfer to lower binding energies visible around @xmath116 .
note that we define the spectrum to be symmetric if the main excitations at @xmath117-vector @xmath107 and @xmath116 are located at the same binding energy .
one may ask if it is possible to get a similar spectral weight distribution by using only the purely one - dimensional hubbard model , but including longer - ranged intra - chain hopping processes as given by the _ ab initio _ calculations .
in fact , the next - nearest - neighbor hopping term along the chain , @xmath118 , is of similar size of the inter - chain hoppings.@xcite the result for the spectral function at the @xmath0 point in this pure 1d case is shown in the lower left panel ( c ) of fig .
[ fig : comp_ak0 ] . from this result
it is obvious that one can not get an excitation at binding energies of roughly @xmath119ev , as seen in experiments . on the contrary ,
the spectral weight of the excitation at the higher binding energy of about @xmath120ev is even enhanced in the one - dimensional treatment when longer - ranged hopping processes are included .
( color online ) comparison of the theoretical spectral function ( left ) with the experimental arpes spectra ( right ) . only the @xmath7-direction is shown .
lorentzian broadening of @xmath79ev was used . the vertical dashed line marks momentum @xmath121 , and the band widths @xmath122 and @xmath123 are indicated .
dark areas mark large spectral weight , the color scale is normalized in each plot separately . ]
our results support the description of tiocl as a layered two - dimensional compound with strong anisotropy , also on the level of correlations .
there is finite dispersion also perpendicular to the chains , but a backfolding of the bands can only be seen along the chains , where correlations are dominant .
let us now comment on the relation of our results to arpes data .
experiments show@xcite that the dispersion in tiocl shows a strong asymmetric behavior along the crystallographic @xmath7 direction , see also the right plot of fig .
[ fig : comp_exp ] .
the binding energy of the lowest - lying band around @xmath124 is about @xmath125ev , whereas around @xmath126 it is about @xmath127ev .
first attempts to describe the dispersions within _ ab initio _ calculations were not successful .
standard lda calculations do not produce the backfolding of the bands induced by short - range spin fluctuations , and spin - polarized lsda+u calculations can not account for the asymmetry of the spectra .
also the spectra of the one - dimensional hubbard model calculated within the dynamical density - matrix renormalization group ( ddmrg ) do not reproduce the experimental spectral weight distribution.@xcite however , as can be seen in fig .
[ fig : comp_exp ] , our new results show that the hubbard model _ can indeed _ give a good description of the asymmetry , since the inter - chain processes give rise to a spectral weight transfer from the holon to the spinon band around the @xmath0 point , and therefore make the band structure more asymmetric . in order to compare our results more quantitatively , we extract the following numbers related to the band widths of the spectra . the first one , @xmath122 , is the difference of the binding energies at @xmath121 and @xmath116 and is a measure for the overall band width . the second one , @xmath123 ,
is defined to be the difference in binding energies between @xmath121 and @xmath107 .
the larger the difference between these two quantities , the larger is the asymmetry of the spectra .
from experiments@xcite we extract @xmath128ev and @xmath129ev , and for the calculated spectra we get @xmath130ev and @xmath131ev . comparing theory and experiment , we see that the overall band width @xmath122 is well reproduced by the calculation , but the asymmetry is even a bit more pronounced in the theoretical spectra , resulting in a smaller value of @xmath123 .
this result may be improved by including more longer - ranged hopping processes .
although they decrease rapidly with distance , they can change the band widths within a few percent . by inspecting fig .
[ fig : comp_exp ] it is obvious that the width of the spectra is much larger in the arpes data than in the calculated spectra . the most important reason for that is that our calculations are done at @xmath132 , using exact diagonalization techniques .
therefore there are no life - time effects due to finite temperatures included in our calculations . moreover
, additional coupling to lattice degrees of freedom could also lead to a smaller life time , and hence broader excitations . in summary ,
our results imply that it is important to include the inter - chain couplings at least on a single - particle non - interacting level into the effective model , in order to improve the description of the experimental spectra , although the strong correlations are constricted mainly to the chains . with our work we could fill the gap left by previous theoretical studies regarding the momentum - resolved single - particle spectral function of tiocl .
there are , though , still some open questions .
for example , we do see clear signatures of spin - charge separation in our calculated spectra , which have not been found experimentally . also the so - called shadow band , dispersing at around @xmath133ev
has not been seen in the arpes spectra .
a reason for this can be a very small relative spectral weight of the high - energy band that can not be resolved in experiment . in our calculation we also did not include the lattice degrees of freedom , which are supposed to be very important in tiocl@xcite , driving the spin - peierls phase transition . these phonons can also lead to a smearing of the peak structure of @xmath9 . finally , we want to comment also on the differences between the arpes spectra of tiocl and tiobr .
experiments , supplemented with band structure calculations , have shown@xcite that in the latter compound the intra - chain couplings are weaker and the inter - chain couplings stronger compared to tiocl,@xcite e.g. , @xmath134 decreases from @xmath135ev to @xmath136ev , whereas @xmath51 increases from @xmath137ev to @xmath138ev . by inspecting the self - energies in a similar manner as we did in fig .
[ fig : selfenergy ] , we found that also for these parameter values the correlations are predominantly one - dimensional .
there are only changes in the overall bandwidths , but no qualitative changes .
for instance , the band width in @xmath100-direction is enhanced , but there are no signatures of strong inter - chain correlations resulting in a backfolding of the bands . at this point
we want to remark that , in particular for tiobr , one should be very careful with the use of the spinon / holon terminology .
although there are no qualitative changes due to the enhanced couplings , they do change the band widths .
hence , quantitative analysis have to be done including these inter - chain couplings .
by combining _
ab initio _ calculations ( lda ) and the variational cluster approximation , we could study for the first time the momentum - resolved spectral function including a realistic band structure _ and _ strong - correlation effects . in agreement with previous theoretical studies and experimental results , our calculations showed an almost complete polarization in the orbital sector , with 99% of the electrons occupying the ti @xmath83 orbital .
since the orbital degree of freedom is quenched , we could use an effective single - band model for the investigation of the spectral properties .
the most striking result of our study is that the inclusion of inter - chain hopping processes leads to a significant spectral weight redistribution around the @xmath0 point from higher to lower binding energies .
this effect , which makes the spectrum more asymmetric with respect to the points @xmath107 and @xmath116 , can not be reproduced using only the hopping processes along the chains .
this result suggests that the frustrated inter - chain coupling@xcite is one of the main reasons for the strong asymmetry that has been found in experimental arpes measurements .
moreover the calculated spectral features may be extended to a more general class of correlated coupled - chains systems . an open question in tiocl is still the role of phonons .
because of the vicinity of the system to a spin - peierls instability , the phonons are supposed to be important in the system . including these degrees of freedom ,
although theoretically very demanding , could further improve the results with respect to lineshapes and lifetimes of the excitations .
this work has been supported by the deutsche forschungsgemeinschaft , research unit 538 , grant cl 124/6 - 1 , the program sfb / trr49 , and the austrian science fund , grant j2760-n16 .
thanks the max - planck - institue stuttgart , germany , through the partnergroup program . m.a . gratefully acknowledges useful discussions with e. arrigoni , m. potthoff , and a. georges .
|
we present results for the momentum - resolved single - particle spectral function of the low - dimensional system tiocl in the insulating state , obtained by a combination of _ ab initio _ density functional theory ( dft ) and variational cluster ( vca ) calculations .
this approach allows to combine a realistic band structure and a thorough treatment of the strong correlations .
we show that it is important to include a realistic _ two - dimensional _ band structure of tiocl into the effective strongly - correlated models in order to explain the spectral weight behavior seen in angle - resolved photoemission ( arpes ) experiments .
in particular , we observe that the effect of the interchain couplings is a considerable redistribution of the spectral weight around the @xmath0 point from higher to lower binding energies as compared to a purely one - dimensional model treatment .
hence , our results support a description of tiocl as a two - dimensional compound with strong anisotropy and also set a benchmark on the spectral features of correlated coupled - chain systems .
|
quasar continuum in the uv - opt band is powered by the thermal energy of the accretion disk and typically written as a power - law relation @xmath14 , where @xmath0 is the continuum slope and a larger @xmath0 means bluer / flatter continuum .
the standard thin disk ( std ) model predicts a constant value @xmath15 of the slope in the uv - opt band @xcite .
however , in observation , the average @xmath0 of this band is about @xmath16 @xcite , which is far more redder than the theoretical prediction @xcite .
although more sophisticated disk models do predict redder / softer slopes when they include more details such as the comptonization in the disk atmosphere , the deviation from local thermodynamic equilibrium and so on , they are still not as red as the observations ( e.g. , * ? ? ?
* ) . to alleviate the discrepancy between model prediction and observations , several other explanations , e.g. ,
dust extinction , host galaxy contamination have been proposed @xcite . the internal quasar dust ( e.g. , dust torus ) , together with the dust in quasar host galaxy , attenuates the uv - opt spectrum , which prevents us from knowing the intrinsic quasar spectrum . what s more , the reddening effect may be luminosity dependent . inferring from the lack of correlation of lyman @xmath0 equivalent width with the quasar uv slope ,
@xcite concluded that the intrinsic shape of the uv spectrum is independent of luminosity over a range of three orders of magnitude . combining the findings in @xcite
, @xcite further concluded that the dust reddening of the uv continuum is luminosity - dependent in the way that the low luminosity quasars are more reddened .
such a simple dust extinction model , instead of sophisticated accretion disk model , can nicely explain the observational result that quasar uv continuum is redder toward lower luminosity @xcite . in order to derive and then correct the dust extinction of quasars , a `` pair method ''
is often used by comparing a reddened quasar with an unreddened one .
there are many choices of the unreddened quasar spectrum , e.g. , the composite quasar spectrum @xcite , the spectrum of certain blue quasars ( * ? ? ?
* hereafter gb07 ) . however , the dust properties derived from different methods / samples are controversial @xcite . to have a better understanding of the dust properties in this extinction model , a systematical study of the uv continuum slopes of a large and well - defined quasar sample is required .
strictly speaking , all types of quasars , even the type i quasars , suffer some extent of _ local _ dust extinction , either from the accretion disk vicinity or the host galaxy .
@xcite noticed that even the bluest 10% of the low redshift agns targeted by the sdss have significant reddening .
therefore , the choices of composite spectra or blue quasars as the reference for deriving the dust extinction in other reddened quasars still take the risks to suffer from the extinction of _ local _ dust . in this case , the extinction / reddening will be underestimated systematically .
alternatively , one may assume that the quasar continuum predicted by the theoretical model is the one with non - dust extinction and then get constraints on the dust properties of the observed quasars @xcite . in this paper
, we test such an idea with the std model and a large sample of quasars from the sloan digital sky survey ( sdss ) to draw more information of quasar dust .
this paper is organized as follows . in section 2 ,
we introduce the quasar sample , define and measure their continuum slopes .
then , we describe the slope distributions and their luminosity dependence . in section 3 , we take the std model to calculate the intrinsic quasar continuum slope and then make a quantitative comparison with the observational data to evaluate the dust extinction scenario .
we make further discussions on the possible explanations of the luminosity - extinction dependence in section 4 .
finally , we draw a brief conclusion in section 5 . throughout the paper
we assume a concordance @xmath17-cosmology with @xmath18 , @xmath19 and @xmath20 .
our sample is defined from the sdss data release 7 ( dr7 ) quasar catalog ( 105783 objects ) , which are selected to be brighter than @xmath21 and have at least one emission line with fwhm ( full width at half maximum ) greater than 1000 km s@xmath22 or have interesting / complex absorption features . in the following measurements of the continuum slopes , we take the residual sky emission lines subtracted spectra from @xcite and improved redshift measurements from @xcite .
all these spectra are then corrected for the galactic reddening by using the sfd map of @xcite and the extinction curve of @xcite .
we also take bolometric luminosity ( @xmath9 ) , black hole mass ( @xmath23 ) , eddington accretion ratio ( @xmath24 ) values and broad absorption line ( bal ) flags from @xcite . to avoid the influence of emission and absorption lines in measuring the slope , specific continuum windows
are often selected @xcite .
following @xcite , we take three windows , 1350 - 1365@xmath25 , 2210 - 2230@xmath25 and 4200 - 4230@xmath25 .
since the sdss spectra cover the observed wavelength range from @xmath26 to @xmath27 @xmath25 , therefore , when an individual spectrum is transformed to its rest frame , it covers at most two adjacent windows . for quasars with @xmath28 , we can measure their slopes from two longer wavelength windows and label them as @xmath5 .
similarly , for quasars with @xmath29 , we can measure their slopes @xmath6 from two shorter wavelength windows . in practice , for the fuv sample
, we further constrain the redshift range to @xmath30 in order to keep consistence with the same definition of @xmath9 estimation in sec .
[ sec : evector ] .
we exclude spectra of : ( 1 ) bal objects ; ( 2 ) with no calculated @xmath9 or @xmath23 ; ( 3 ) with too many negative flux values in the covered continuum windows .
after that , we have two samples : 19258 quasars in the nuv sample and 19671 quasars in the fuv sample . for simplicity
, we use a log - linear fit for the fluxes in the two covered windows to measure the continuum slope and its uncertainty .
in fact , the spectrum may not follow an explicit power - law .
nevertheless , this simple method provides a good approximation of @xmath0 , and still can fairly reflect the difference between the reddened and unreddened ( model ) spectra that we concern . for the nuv sample ,
the mean and dispersion values of @xmath5 are @xmath3 and @xmath31 , and for the fuv sample the respective values are @xmath4 , and @xmath32 . generally , these mean values agree well with most of the previous works @xcite with the scatter identical to that of @xcite ( @xmath33 ) . on average , the apparent slope of quasars
are much smaller ( redder ) than the theoretically expected value , while the large dispersions imply the complicate mechanism of shaping the spectrum . in fig .
[ fig1 ] , we plot the continuum slopes of each quasar against their bolometric luminosity @xmath34 in gray symbols . to check the luminosity dependence of @xmath0 , we further group the quasars into different luminosity bins with a bin size of 0.2 dex and calculate the median values , dispersions ( the outer error bars shown as 15.9%th to 84.1%th percentiles of distribution that represent the @xmath35 ranges ) . the median measurement errors of @xmath0 for each bins are shown as the inner error bars .
clearly , for both nuv and fuv samples , the median values of the continuum slope increase with luminosity , especially at the low luminosity side , which means less luminous quasars have redder ( softer ) spectra .
meanwhile , the dispersions of both @xmath5 and @xmath6 are decreasing with luminosity , which means the spread of the @xmath0 is narrower toward higher luminosity .
these two luminosity dependence confirm to the phenomena that shown in the fig.4 of @xcite , though their work only contains a much smaller sample of agns .
we notice that the errors of @xmath0 are also related to the luminosity , with less luminous quasars having larger measurement uncertainties . as shown in fig .
[ fig1 ] , for all luminosity bins , the apparent dispersions of @xmath0 are significantly larger than the median errors , so they are definitely _ not _ dominated by errors .
however , to estimate the effect of error , we use the monte - carlo process which includes 100 sets of artificial samples that mimic our observational samples with similar @xmath36 distributions .
these simulated targets are further randomly modified with their @xmath0 values according to the possible errors that depend on @xmath37 and @xmath0 .
then , we can compare the simulated samples with and without errors of @xmath0 , and find that the changes of the median values of @xmath0 for all luminosity bins are neglectable , while the changes of dispersions are all very small with the biggest variation being @xmath38 at the low luminosity side ( see also the ratios of intrinsic to observational dispersions @xmath39 for the simulated sample that plotted in the lower panels of fig.[fig1 ] ) .
thus , we claim that the luminosity dependence of continuum slopes and their spreads are subsistent , which reflects some real physical conditions that drive the spectrum shape of agn
. there is another important thing to note that the continuum slopes of almost all sampled quasars are less than @xmath40 ( shown as the dotted horizonal lines in fig.[fig1 ] ) . to illustrate this , we plot the dashed lines for the 97.7%th percentile ( standing for the @xmath41 level ) .
they also curve the upper / blue envelope of the slopes .
this upper limitation is roughly luminosity independent except for a little increasing at the low luminosity side of @xmath6 .
it could be partly attributed to their larger observational errors of @xmath0 , thus more quasars will probably have the @xmath0 measurements over the blue limitation , though they are intrinsically below it .
there are previous works that claim the luminosity independent cutoff of @xmath0 at the blue side , i.e. the fig.4 of @xcite .
what we want to emphasize here is , based on the large quasar sample from the sdss , this upper limitation is close to the std model prediction , more or less .
on the other hand , we understand that @xmath42 is only a rough approximation of the model prediction of the intrinsic slope .
actually , it should be slightly modified by the central black hole mass and the accretion rate ( see sec.[sec : modelslope ] for details ) .
anyway , the distribution of both @xmath5 and @xmath6 give us a strong hint that the blue envelope of @xmath0 is highly possible to exist , while those bluest ones are bare quasars with little ( or no ) dust extinction so that their continuum slope are in good consistence with the std model .
in this section , we assume a simple scenario that the intrinsic continuum radiation of quasars obeys the std model , then be dimmed and reddened by _
local _ dust , either from the optical thin dust in the internal region or from the interstellar dust in the host galaxy .
then , we take the discrepancy between observational slopes and model predictions , either for @xmath5 or for @xmath6 , to quantify the dust extinction , as well as to transform the apparent @xmath36 trend to the essential @xmath37-extinction relation .
we take the radiatively efficient , geometrically thin and optically thick accretion disc or the so - called standard thin disk ( std ) to model the quasar continuum emission , where the radiation at each radius in the accretion disk is a blackbody @xcite . following @xcite
, we adopt the temperature distribution as function of radius according to @xcite , @xmath43 where @xmath44 is the viscosity parameter , @xmath45 is the black hole mass , @xmath46 is the accretion rate in unit of eddington accretion rate with @xmath47 , @xmath48 ( @xmath49 ) and @xmath50 .
the shape of the spectrum is then calculated with @xmath51 where @xmath52 , @xmath53 is the outer radius of disk and is set to be the place where @xmath54 deceases to 1000k , @xmath55 is the plank constant , @xmath56 is the boltzmann constant and @xmath57 is the speed of light . in fig .
[ fig2 ] , we plot the model slope @xmath6 , as a function of different black hole masses and eddington accretion rates .
it shows that @xmath6 decreases with @xmath58 and increases with @xmath59 .
the grid of @xmath5 is similar to the @xmath6 , with slightly larger slope values and narrower range .
generally , both of them are close to the value of 1/3 , which confirms the expectation from the std model . from 8.5 to 10.5
evenly spaced .
dotted lines from bottom to top have @xmath60 from -1.5 to 0.0 evenly spaced .
the black diamond is an observed data point .
the magenta dashed line is the trail of the luminosity _ vs _ model continuum slope , while the blue and green arrows stand for the luminosity _ vs _ observed continuum slope trails , and their intersections represent the intrinsic luminosities and continuum slopes .
values in the right corner show the extinction needed for the smc extinction curve and the flat curve ( gb07 ) respectively . ] with the _ local _ dust extinction , the quasar luminosity is dimmed and its continuum spectrum is reshaped ( reddened ) .
if we simplify the intrinsic spectrum as @xmath61 , then the reddened spectrum becomes @xmath62 , where the optical depth @xmath63 equals to @xmath64 with @xmath65 refers to the extinction curve .
the reddening of the continuum slope can be derived by @xmath66 \cdot e(b - v)\ ] ] where subscript 1 and 2 stand for the selected continuum windows . according to @xcite , @xmath9
is calculated from a monochromatic continuum luminosity and its bolometric correction ( @xmath67 ) on the basis of redshift , @xmath68 so , the dimming of @xmath9 can be expressed as the change of corresponding continuum luminosity , @xmath69 where @xmath70 for nuv or @xmath71 for fuv samples .
we can regard @xmath72 as @xmath73 components of the `` extinction vector '' ( @xmath74 ) . obviously , for given continuum windows , the direction of @xmath74 only depends on @xmath75 , and its length is in proportional to @xmath76 . here , we test two typical dust extinction curves , the smc curve @xcite and the flat agn curve ( gb07 ) . for the smc curve ,
the unit vector of @xmath74 ( with @xmath77 ) is @xmath78 for nuv sample and @xmath79 for fuv sample .
correspondingly , in the case of the flat curve , these two vectors are @xmath80 and @xmath81 .
we notice that the smc curve is much steeper in the uv band , so it needs much smaller @xmath76 for a given @xmath82 . in fig .
[ fig2 ] , as an example , we illustrate the procedure of dust correction for a typical " quasar , with observed value of @xmath83 , @xmath84 and @xmath85 ( the second point from left in the upper right panel of fig .
[ fig1 ] ) . for a given extinction curve ,
the corrections of @xmath9 and @xmath0 start from the observed value ( the black diamond ) and follow the opposite direction of the extinction vector " , until it reaches its intrinsic @xmath9 and @xmath0 . on the other hand ,
the estimation of virial mass of bh is based on both broad emission line width and continuum luminosity . since the width of emission line is not affected by reddening , according to the eq . 2 of @xcite
, we can easily find that the dust extinction correction on the black hole mass estimation is @xmath86 where @xmath87 is the corresponding continuum luminosity of the broad line and @xmath88 is the empirical coefficient and equals to 0.62 and 0.53 for nuv and fuv samples respectively @xcite . for each observed quasar , when @xmath9 and @xmath89 are derived , the accretion rate @xmath90 is also fixed .
therefore , after a dust correction of @xmath76 , we can also get an estimation of the intrinsic continuum slope from the std model presented in section [ sec : modelslope ] .
such a correction is displayed by the trace of the magenta dashed line in fig .
when the dust extinction in the quasar is properly accounted for , the corrections from the extinction curve and theoretical model will lead to the consistent result of @xmath0 .
in other words , the intersection point of the two above - mentioned traces represents the intrinsic properties of the `` example '' quasar .
thus , we can use this point to calculate the @xmath76 that we concern .
it should be noticed that , for our current sample , the observed @xmath23 , @xmath9 and @xmath0 of individual quasars have large uncertainties .
thus , it seems to be too complicate and uncertain to extract the dust extinction for individual quasars .
alternatively , we use the median values of the 20 binned sub - samples in fig .
[ fig1 ] to be the
median quasars " that represent their typical features .
so in the following analysis , we only focus on these 20 median quasars " and quote their median values of @xmath5 , @xmath6 , @xmath91 and @xmath58 respectively . by applying the algorithm of sec.[sec : ebv ] , we calculate @xmath76 and intrinsic @xmath9 for all 20 `` median quasars '' , for both smc and flat extinction curve situations , and plot the results in fig . [ fig3 ] .
intuitively , one can conclude that : ( 1 ) there is a significant @xmath92 anti - correlation with lower luminosity quasars having more dust extinction , no matter which sample ( nuv or fuv ) or which extinction curve we use ; ( 2 ) for the smc curve , the required @xmath76 value is only about @xmath93 , while for the flat curve it increases to @xmath94 .
the systematical higher @xmath76 values for the flat extinction curve is simply because that the smc extinction curve is much steeper in the uv band , which requires smaller @xmath76 to produce the same reddening of @xmath0 . for comparison
, we also plot the results from gb07 in fig .
[ fig3 ] , which used the `` pair method '' and the flat extinction curve .
these isolated points disperse largely but have the same @xmath92 trend as ours . compared with the same extinction curve ,
these points are mostly below our results .
this result is comprehensible .
the quasar sample of gb07 comes from @xcite which is chosen to be bright in the far - uv , thus it might be biased toward low - reddening objects .
more importantly , the bluest quasars ( pg0953 + 414 , pg1100 + 772 , 3c273 ) they choose as the unreddened spectra may still contain a certain extent of dust extinction .
we have measured the @xmath5 and @xmath6 of these three quasars and find that they are obviously redder than the theoretical values of the bare agns ( see the hollow circles in fig . [ fig1 ] ) .
comparing the quasars with the same intrinsic @xmath9 but in different samples ( nuv and fuv ) , we see that the @xmath76 values derived from the smc extinction curve are more consistent
. however , this apparent consistence does not motivate us to favor the smc type extinction curve in quasars . because the @xmath76 values we derived are strongly dependent on the intrinsic slopes of the quasar continuum , which is far from well - defined in the simple std model we adopted .
if the real intrinsic fuv slope of the quasars are softer than our model calculation , then our derived @xmath76 would be certainly biased to higher values for the flat extinction curve while change less for the smc curve . besides , the nuv and fuv samples are actually in different redshifts and surely have the observational bias with higher redshift sample ( fuv ) containing more luminous quasars .
therefore the direct comparison of their @xmath76 values maybe unfair .
moreover , given the fuv sample at higher redshift , it is expected that these quasars are more attenuated considering the cosmic dust along the line of sight @xcite ( see more discussions in sec . [
sec : cosmicdust ] ) . relations .
red triangles are for the @xmath5 sample and blue diamonds are for @xmath6 , with solid symbols for the smc type dust and hollow ones for the flat dust curves .
dotted lines show results after the cosmic dust correction ( sec .
[ sec : cosmicdust ] ) .
solid and hollow circles ( the three bluest ones ) are the results of individual quasars for gb07 . ]
in this study , we introduce a luminosity dependent dust extinction model and successfully apply it to explain the phenomena that the more luminous quasars have bluer continuum . in this section ,
we further discuss its physical implications and other possibilities .
firstly , the relative bluer continuum slope of more luminous quasar is unlikely to be solely caused by the intrinsic properties of the accretion disk .
the more sophisticated model are also hard to produce such a red spectrum @xcite .
besides , a softer continuum at lower luminosity can hardly explain their larger broad emission line equivalent widths , i.e. the well - known baldwin effect @xcite .
on the other hand , it is quite straightforward to ascribe the observed @xmath36 relationship to the quasar _ local _ dust extinction .
local _ dust extinction may physically be originated from the internal dust around the accretion disk , or from the dust in the host galaxy . for the quasar host galaxies , observations
show that their star formations are co - evolved with the growth of the black holes @xcite .
therefore , we would expect that there are more star formation , so that more dust generated in the host galaxies of high luminosity quasars .
such a scenario is obviously not consistent with the trend we required to explain the observed @xmath36 relation .
therefore , the _ local _ dust in our model is more likely to be dominated by the internal one .
if so , it is worthwhile to further discuss whether the phenomenological internal dust in our model is compatible with the physical dust torus model in the revised unified agn paradigm @xcite . in this revised model , the dust tori are packed with gas / dust into clumps , and the intrinsic attenuation of an agn is determined by the number of the clumps along the line - of - sight , their optical depth and distribution ( e.g. * ? ? ?
* ; * ? ? ?
since our sample of quasars are type i sources , the photons from the accretion disk have only crossed few optical - thin clouds .
these clouds are probably originated from the transition area between the dust torus ( optical thick ) and broad - line - region ( blr , dust - free ) or the low density interclump medium @xcite and provide the attenuation of the type i agns . from the ir reverberation mapping observation , we know that the dust sublimation radius @xmath95 is proportional to the quasar luminosity by @xmath96 @xcite .
thus the distribution of the internal dust will depend on the luminosity of the central engine in a statistical way that the dust tori of higher @xmath9 sources have smaller effective covering factors @xcite .
such a scenario naturally explains the higher number ratio of type i / type ii sources and lower @xmath97 ratios in higher @xmath9 quasars @xcite .
if we assume the sizes of these optical thin clouds are smaller ( so that with smaller optical depths ) for higher @xmath9 agns due to the stronger dust sublimation effect , we would expect that even for type i configuration , the higher @xmath9 agns are averagely less attenuated than their lower luminosity counterparts .
apart from the distribution of dust clumps , the dust properties or in terms of the extinction curve may be another factor which leads to the different reddening in high and low luminosity quasars . from the theoretical consideration
, the larger dust grains are preferred to stay in the severe radiation conditions near the nuclei as the small grains are more easy to be destroyed or blown away by strong radiation @xcite . therefore ,
in the case of low luminosity quasars , the dust component is biased to small dust grains and more uv photons will be absorbed , leading to smaller observed @xmath0 , which is consistent with our finding . in a recent work
, we have found that , within the same luminosity , quasar continuum becomes redder toward higher redshift , which is successfully attributed to the accumulation effect of the cosmic dust extinction @xcite .
therefore , the reddening from the cosmic dust accumulated in the path should be corrected .
we take the cosmic dust extinction model from @xcite ( with @xmath98 and the flat extinction curve , as shown in their fig.5 ) and make corrections on the observed @xmath9 , @xmath23 , @xmath99 and @xmath0 for each quasar . after the correction of cosmic dust extinction
, we repeat the analysis of the _ local _ dust extinction in sec.[sec : ebv ] .
the new results are plotted as dotted lines in fig .
[ fig3 ] for illustration .
as we can see , the involvement of the cosmic dust extinction only slightly reduces the _ local _ dust extinction and does not change the trend of the @xmath92 relation .
that is to say , comparing with the extinction from the _ local _ dust , the accumulated dust extinction along the line of sight is a small quantity .
moreover , we had supposed that the cosmic dust extinction is one of the reasons for the inconsistence of the @xmath92 relation of nuv and fuv samples in the case of the flat extinction curve ( sec.[sec : l - e ] ) .
now we can learn from dotted lines in fig.[fig3 ] to recognize that this factor only slightly alleviates the inconsistence .
in this paper , we investigate the uv continuum slope of a large sample of quasars selected from sdss dr7 .
it is found that for both @xmath5 and @xmath6 , the observed continuum slope are averagely and significantly redder than theoretical expected values . the observed @xmath5 and @xmath6 show dependence on the quasar luminosity , with higher luminosity quasars having bluer continuum .
such findings have been reported in many previous studies . besides
, we notice that the quasar continuum slopes show an upper limit ( @xmath42 ) either for @xmath5 or @xmath6 at any luminosity range .
this strongly implies intrinsic quasars ( or perhaps the observed bluest quasars ) obey the standard thin disk model .
we further test the assumption that all observed quasar spectra intrinsically follow the standard model prediction , but get dimming and reddening by the dust surrounding the accretion disk .
we calculate the @xmath76 needed to achieve the intrinsic value of @xmath0 by simultaneously considering the correction of the bolometric luminosity and the continuum slope due to dust extinction and reddening .
our estimation of @xmath76 under the flat agn extinction curve case is consistent with the previous finding @xmath100 @xcite . for the smc type extinction curve , because of its much steeper shape in uv wavelength , the @xmath76 values required is even smaller @xmath101 .
our results also show that there is a dependence of @xmath76 on intrinsic quasar luminosity @xmath9 .
the higher @xmath9 quasars are averagely less reddened than the lower luminosity ones .
we have discussed that such a luminosity dependence is compatible with the revised unified agn model .
we argue that the type i quasars are slightly attenuated by the optical thin clouds in the transition area where the dust particles are sublimated by the central engine or the interclump medium .
we have also checked the effect of the cosmic dust extinction and conclude that compared with the _ local _ dust extinction , the accumulated dust extinction along the line of sight is a small quantity .
we gratefully acknowledge the referee dr .
martin gaskell for very helpful comments and detailed suggestions .
the authors thank prof .
minfeng gu and dr .
shuangliang li in shao for helpful discussions .
this work was supported by the strategic priority research program `` the emergence of cosmological structures '' of the chinese academy of sciences ( cas ; grant xdb09030200 ) , the national natural science foundation of china ( nsfc ) with the project numbers 11390373 , 11433003 , and the `` 973 program '' 2014 cb845705 .
baldwin , j. a. 1977 , , 214 , 679 carballo , r. , gonzlez - serrano , j. i. , benn , c. r. , snchez , s. f. , & vigotti , m. 1999 , , 306 , 137 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 cheng , f. h. , gaskell , c. m. , & koratkar , a. p. 1991
, , 370 , 487 czerny , b. , li , j. , loska , z. , & szczerba , r. 2004 , , 348 , l54 czerny , b. 2007 , the central engine of active galactic nuclei , 373 , 586 davis , s. w. , woo , j .- h . , & blaes , o. m. 2007 , , 668 , 682 ferrarese , l. , & merritt , d. 2000 , , 539 , l9 gaskell , c. m. , goosmann , r. w. , antonucci , r. r. j. , & whysong , d. h. 2004 , , 616 , 147 gaskell , c. m. , & benker , a. j. 2007 , arxiv:0711.1013 ( gb07 ) gaskell , c. m. 2015 , arxiv:1512.09291 gebhardt , k. , bender , r. , bower , g. , et al .
2000 , , 539 , l13 gordon , k. d. , clayton , g. c. , misselt , k. a. , landolt , a. u. , & wolff , m. j. 2003 , , 594 , 279 gu , m. 2013 , , 773 , 176 hewett , p. c. , & wild , v. 2010 , , 405 , 2302 hnig , s. f. 2013 , arxiv:1301.1349 hubeny , i. , agol , e. , blaes , o. , & krolik , j. h. 2000 , , 533 , 710 kato , s. , fukue , j. , & mineshige , s. 1998 , black - hole accretion disks .
edited by shoji kato , jun fukue , and sin mineshige .
publisher : kyoto , japan : kyoto university press , 1998 .
isbn : 4876980535 , koratkar , a. , & blaes , o. 1999 , , 111 , 1 kormendy , j. , & ho , l. c. 2013 , , 51 , 511 li , a. 2007 , the central engine of active galactic nuclei , 373 , 561 li , s .- l . , & cao , x. 2008 , , 387 , l41 li , z. , zhou , h. , hao , l. , et al . 2015 , , 812 , 99 lynden - bell , d. 1969 , , 223 , 690 nenkova , m. , sirocky , m. m. , nikutta , r. , ivezi , . , & elitzur , m. 2008 , , 685 , 160 netzer , h. 2015 , , 53 , 365 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 schneider , d. p. , richards , g. t. , hall , p. b. , et al .
2010 , , 139 , 2360 shakura , n. i. , & sunyaev , r. a. 1973 , , 24 , 337 shankar , f. , calderone , g . , knigge , c. , et al .
2016 , , 818 , l1 shang , z. , brotherton , m. s. , green , r. f. , et al .
2005 , , 619 , 41 shen , s. , shao , z. , & gu , m. 2010 , , 725 , l210 shen , y. , richards , g. t. , strauss , m. a. , et al .
2011 , , 194 , 45 simpson , c. 2005 , , 360 , 565 suganuma , m. , yoshii , y. , kobayashi , y. , et al .
2006 , , 639 , 46 telfer , r. c. , zheng , w. , kriss , g. a. , & davidsen , a. f. 2002 , , 565 , 773 toba , y. , oyabu , s.,matsuhara , h. , et al .
2014 , , 788 , 45 vanden berk , d. e. , richards , g. t. , bauer , a. , et al . 2001 , , 122 , 549 wild , v. , & hewett , p. c. 2010 , arxiv:1010.2500 xie , x. , shen , s. , shao , z. , & yin , j. 2015 , , 802 , l16
|
we investigate the uv continuum slope @xmath0 of a large quasar sample from sdss dr7 . by using specific continuum windows , we build two samples at lower ( @xmath1 ) and higher ( @xmath2 ) redshifts , which correspond to the continuum slopes at longer ( nuv ) and shorter ( fuv ) rest wavelength ranges respectively . overall , the average continuum slopes are @xmath3 and @xmath4 for @xmath5 and @xmath6 with similar dispersions @xmath7 . for both samples ,
we confirm the luminosity dependence of the continuum slope , i.e. , fainter quasars have redder spectra .
we further find that both @xmath5 and @xmath6 have a common upper limit ( @xmath8 ) which is almost independent of the quasar luminosity @xmath9 .
this finding implies that the intrinsic quasar continuum ( or the bluest quasar ) , at any luminosity , obey the standard thin disk model .
we propose that the other quasars with redder @xmath0 are caused by the reddening from the dust _
locally_. with this assumption , we employ the dust extinction scenario to model the observed @xmath10 relation .
we find that , a typical value of @xmath11 to @xmath12 mag ( depending on the types of extinction curve ) of the quasar _ local _ dust is enough to explain the discrepancy of @xmath0 between the observation ( @xmath13 ) and the standard accretion disk model prediction ( @xmath8 ) .
|
the properties of magnetic impurities in metals , and of small dots attached to metallic leads share many characteristics , due to the similar role played by electron electron interactions .
the electrostatic repulsion leads to the formation of local moments@xcite , which are quenched at sufficiently low temperatures .
quantum dots are more complicated structures@xcite and an effective kondo hamiltonian can only be defined in the limit of large spacing between the electronic levels within the dot .
then , the main physical processes at low temperatures are due to the changes of the occupancy of the level closest to the fermi energy of the leads .
even if we assume this restriction , terms beyond the kondo hamiltonian can arise due to the finite extension and inhomogeneities of the electron states inside a mesoscopic quantum dot .
the simplest of these additional interactions is an assisted hopping term , extensively studied in relation to bulk correlated systems@xcite .
it can be shown that , within a systematic expansion in the intradot conductance , the leading correction to the intradot capacitance , or averaged coulomb interaction , has such a form@xcite . the inhomogeneous screening and effective `` orthogonality catastrophe " which give rise to this term
also lead to interesting non equilibrium effects in the metallic limit , when the level spacing is much less than the temperature@xcite .
the present work analyzes the model of a correlated impurity , or quantum dot , in the limit where the low temperature physical properties are determined by a single electronic state within the dot , using the flow equations first proposed in refs . and .
the method transforms the initial hamiltonian into a family of related hamiltonians by means of a sequence of unitary transformations .
if adequately chosen , these transformations lead to simplified models , in a similar way to a renormalization group transformation .
the method has proven useful for a variety of problems@xcite , and it has also been used to obtain the kondo hamiltonian from the anderson model@xcite . in the following , we will apply the flow equation technique to the anderson hamiltonian with assisted hopping .
this problem has been analyzed in@xcite using mean field techniques , but there are no more accurate studies in the literature .
thus , besides the intrinsic interest of the model , this work provides a test case where the flow equation method is used to obtain information not available by other techniques .
the next section presents the model , and describes how the flow equation method is applied .
the results are discussed in section iii .
finally , section iv analyzes the main properties of the model , as derived by this method .
we will study the hamiltonian : @xmath2 where @xmath3 is the anderson hamiltonian , and the assisted hopping terms are included in @xmath4 .
we have also defined @xmath5 and @xmath6 .
we now perform a sequence of infinitesimal unitary transformations : @xmath7 \label{transformation}\end{aligned}\ ] ] these transformations are characterized by the anti - hermitian generator @xmath8 .
the parameters of the initial hamiltonians become functions of the flow parameter @xmath9 and , if there are no upper indices on the parameters , it will be implied that they depend on @xmath9 .
our main goal is to reduce the initial hamiltonian , eq .
( [ hamil ] ) to another whose physical properties are well understood .
we assume that this is achieved if the assisted hopping term is made to flow to zero .
hence , we impose the requirement @xmath10 this condition is satisfied if @xmath11 $ ] .
more generally , we will employ the following generator : @xmath12 this expression differs from the generator used to cancel the hybridization with the band , @xmath13 in eq .
( [ hamil ] ) , in that it contains an additional operator related to the localized orbital , @xmath14 .
it is interesting to note that for the analytic treatment of the flow equations the explicit expression of the parameters @xmath15 is not needed .
we only require that eq .
( [ limit ] ) holds .
the physical properties of the impurity are determined by the flow of the on - site energy @xmath16 and the correlation energy @xmath17 as @xmath18 .
notice that the one particle hybridization in the hamiltonian , eq .
( [ hamil ] ) , is not changed during the flow , i.e. , @xmath19 this is a consequence of our choice of the generator given in eq .
( [ generator ] ) .
the commutator @xmath20 $ ] generates new interactions which have to be included in the hamiltonian @xmath21 , which , in turn , will lead to other interactions .
this infinite hierarchy needs to be decoupled in order for the method to be useful .
this point is carefully discussed in@xcite , and we will follow the same procedure . we do not include the new operators in the flow , although they will appear in the fixed point hamiltonian . the role of these `` marginal '' operators will be discussed in section iiic .
the resulting flow equations thus take on the simple form @xmath22 where we have introduced the occupation number @xmath23 , which arises from normal ordering the neglected operators@xcite .
the procedure used here explicitly breaks particle - hole symmetry .
note that there is no renormalization of the energies of the conduction electrons , @xmath24 , which is a typical feature of impurity problems@xcite . from eq .
( [ floww ] ) , we obtain the general relation between the flow of the assisted hopping amplitudes and the generator of the infinitesimal transformations , @xmath25 inserting this expression in eqs .
( [ flowd ] ) and ( [ flowu ] ) we obtain @xmath26 note that the above equations imply that @xmath27 where @xmath28 and @xmath29 , and @xmath30 and @xmath31 denote universal functions .
these functions only depend on the coupling constants @xmath13 and @xmath32 .
the above flow equations can also be treated semi - analytically by substituting the @xmath9-dependent parameters @xmath33 and @xmath34 in the denominator by their fixed point values @xmath0 and @xmath1 .
this strategy has proven to be successful in several related models.@xcite we can then integrate the equations ( [ flowe ] ) and ( [ flowu ] ) from @xmath35 to @xmath36 , using the constraint that @xmath37 .
we define @xmath38 where we assume that the hybridization @xmath13 and the assisted hopping amplitude @xmath32 differ in sign@xcite .
the spectral function is positive for the physically relevant parameter regime , where the hybridization amplitude is larger or comparable to the assisted hopping amplitude . with the above definition
we obtain the final result @xmath39 where we have defined the fermi function @xmath40.@xcite
in order to make the analysis more quantitative , we need to specify the density of states in the surrounding medium , described by the function @xmath41 , eq .
( [ spectral ] ) .
the width of the conduction band in the leads is @xmath42 , and the fermi energy is @xmath43 .
we also define @xmath44 we further assume that both coupling functions , @xmath13 and @xmath32 , are semi - elliptic .
this choice resembles a realistic environment and facilitates the computation . to obtain the universal functions @xmath30 and @xmath31 from eqs .
( [ unif ] ) and ( [ unig ] ) , we numerically integrate the flow equations ( [ flowd ] - [ floww ] ) with @xmath45 this choice guarantees that @xmath46 vanishes exponentially for @xmath47 and algebraically for @xmath48.@xcite by this , we introduce an energy - scale dependent decoupling scheme with respect to the renormalized energy of the double occupancy of the dot , @xmath49 .
we now turn to the semi - analytical solution . the spectral function of eq .
( [ spectral ] ) reads @xmath50 with the resonance @xmath51 .
evaluating eqs .
( [ intepsilon ] ) and ( [ intu ] ) we have to distinguish two cases : for @xmath52 , we obtain the following self - consistent equations : @xmath53\quad,\\ \label{urenormalizedlld } u^*-u^0&=-\frac{2\gamma}{d}(\epsilon_d^*+u^*)\end{aligned}\ ] ] for @xmath54 , we obtain : @xmath55\quad,\\ u^*-u^0&=-\frac{2\gamma}{d}\left[(\epsilon_d^*+u^ * ) -\text{sgn}(\epsilon_d^*+u^*)\sqrt{(\epsilon_d^*+u^*)^2-d^2}\right]\quad . \label{urenormalizedggd}\end{aligned}\ ] ] in fig .
[ fixedpoint ] , the numerically exact universal functions @xmath30 and @xmath31 of eqs .
( [ unif ] ) and ( [ unig ] ) are shown as functions of the scaling variable @xmath56 ( solid line ) .
the coupling strengths were chosen to be @xmath57 and @xmath58.@xcite the semi - analytical solution ( dashed line ) given in eqs .
( [ erenormalizedlld ] - [ urenormalizedggd ] ) qualitatively agrees with the exact solution .
the main difference is that the kink in @xmath59 near @xmath60 which is present in the numerical solution is not adequately reproduced by our analytical approach , eq .
( [ urenormalizedlld ] ) .
this is due to the cruder decoupling scheme which does not incorportate energy - scale separation.it is interesting to note that at this point , the numerical solution yields @xmath61 .
it will lead to significant differences for the phase diagram which will be discussed in the next subsection .
( [ urenormalizedlld ] ) yields @xmath62 hence , for @xmath63 , the renormalized interaction can become negative .
we thus define the critical initial correlation energy @xmath64 such that the renormalized correlation energy @xmath1 becomes zero , i.e. , @xmath65 @xmath64 depends on the inital on - site energy @xmath16 and we further restrict @xmath64 to be positive .
the critical correlation energy @xmath64 is enhanced for positive on - site energies @xmath66 .
this property can most clearly be seen from the analysis of eqs .
( [ intepsilon ] ) and ( [ intu ] ) for @xmath67.@xcite then , the fermi function is constant , i.e. @xmath68 , and the integrals yield simple expressions . for the semi - elliptical coupling function ,
we obtain for @xmath69 @xmath70 on the left hand side of fig .
[ phasedia ] , the critical correlation energy @xmath64 is shown as a function of the initial on - site energy @xmath16 as it follows from the numerical ( solid line ) and semi - analytical ( dashed line ) solution .
the coupling strengths were again chosen to be @xmath57 and @xmath58 .
the main difference between the two curves stems from the spike of @xmath71 around @xmath72 , seen on the right hand side of fig .
[ fixedpoint ] .
it is now interesting to consider the possibility of local pairing correlations .
pair formation does not rely on particle - hole symmetry , in fact , asymmetries can actually enhance this phenomenon.@xcite but local pairing can only occur if - in addition to an effective negative correlation energy @xmath73 - the renormalized on - site energy @xmath0 lies sufficiently close to the chemical potential @xmath74.@xcite the local pairing regime is thus determined by the tradeoff between a high critical correlation energy @xmath64 and an effective on - site energy @xmath0 sufficiently close to the chemical potential . to discuss the phase diagram of this local pairing regime we will only consider the numerical solution .
this approach is more reliable since it incorporates a systematic decoupling of the assisted hopping term , starting with large and ending with small energy differences with respect to @xmath49 .
the results are presented on the right hand side of fig .
[ phasedia ] , where the critical correlation energy @xmath64 is shown as function of @xmath16 in the regime where @xmath75 .
furthermore , we distinguish between positive ( @xmath76 ) and negative ( @xmath77 ) renormalized on - site energy @xmath0 .
our analysis predicts local pairing for values of @xmath16 which favor double occupancy , and @xmath78 . the mean field analysis in ref .
gives local pairing for @xmath79 and a similar range of values of @xmath16 ( note that we are using here comparable hybridization with the band ) .
it is interesting to note that the strong renormalization of @xmath1 for @xmath80 only occurs at large values of @xmath81 .
therefore , it is not present in the semi - analytical solution which does not incorporate energy - scale separation . furthermore , there are fixed points for which @xmath82 , i.e. , a degeneracy between the zero - charge and two - charge state of the impurity .
since the marginal interactions also include pseudo spin - pseudo spin couplings , this might give rise to the kondo effect in the charge channel , i.e. , in the mixed valence regime .
we now analyze the marginal terms , which have been neglected so far .
one term will be interpreted as a renormalization of the hybridization term @xmath13 .
this identification is crucial since the assisted hopping term @xmath32 contains an effective hybridization term which must be present in the fixed point hamiltonian .
another term will explicitly favor local pairing .
as in the anderson model , the flow generates spin - spin and pseudo spin - pseudo spin couplings : @xmath83\quad,\end{aligned}\ ] ] in addition to this interaction , the following term is also generated : @xmath84 the existence of these terms is independent of the assisted hopping interaction and has been discussed in detail in relation with the anderson model@xcite .
there , @xmath85 gives rise to the well - known kondo behavior in the spin and charge sector , respectively .
@xmath86 is usually neglected in the fixed point hamiltonian since the first term of eq .
( [ neglectterms ] ) represents potential scattering of the conduction electrons and will be irrelevant for our discussion .
the last term renormalizes the on - site energy by @xmath87 . for the semi - elliptical coupling function ( [ semicircular ] ) ,
we obtain @xmath88 according to eq .
( [ vkk ] ) within the semi - analytical treatment .
our approximation - not to include newly generated coupling terms into the flow of the hamiltonian - is thus consistent for @xmath89 . as in ref .
, we will ignore all terms of eq .
( [ neglectterms ] ) in the fixed point hamiltonian and assume that they can be taking into account by redefining the initial parameters .
finally , the commutator @xmath90 $ ] also gives rise to the following pairing term : @xmath91\end{aligned}\ ] ] which has the same shape as some of the contributions in @xmath85 , eq . ( [ hsw ] ) .
the additional contributions of eq .
( [ hsw ] - [ pair ] ) need to be included in the flow equation scheme . as these terms contain an even number of operators associated to the localized level , but the generator , eq .
( [ generator ] ) , contains an odd number , these terms can not directly change the flow equations of the parameters @xmath34 and @xmath33 , eqs.([flowd],[flowu ] ) .
they can , however , change the flow of the assisted hopping term , eq .
( [ floww ] ) , which will eventually modify eqs.([flowd ] ) and ( [ flowu ] ) .
we will neglect this second order modification of the flow equations for @xmath34 and @xmath33 .
this approximation was initially introduced in@xcite and is based on the fact that @xmath92 and @xmath93 for @xmath94 .
note that , provided that the flow equations are unchanged , the presence of these `` marginal '' terms in the fixed point hamiltonian does not alter the main conclusions of the paper .
we now approximate the magnitude of these terms at the fixed point , neglecting their influence on the flow equations .
we obtain : @xmath95 as mentioned earlier , @xmath85 arises in the analysis of the anderson hamiltonian , due to the effects of the hybridization between the state at the impurity and the conduction band , and can be derived by means of a schrieffer - wolff transformation@xcite .
hence , we can interpret these fixed point interactions as arising from an effective hopping term in the initial hamiltonian , which , indeed , can be obtained from a mean field decoupling of the assisted hopping term .
@xmath96 will enhance the tendency towards pair formation .
the kondo interactions generated in the anderson hamiltonian when the hybridization flows from @xmath97 to @xmath98 is@xcite : @xmath99 comparing this result with eq .
( [ vkk ] ) , we obtain the following relation : @xmath100 where we introduced the parameters @xmath101 .
the values of @xmath102 and @xmath103 define the initial and final values of the effective hybridization needed to take into account the terms in @xmath85 , eq .
( [ hsw ] ) .
( [ gamma ] ) only makes sense for @xmath104 . in order to make a more systematic analysis , the flow of the hybridization amplitude @xmath13 needs to be included from the beginning .
but for the relevant regime of local pairing , we have from the exact solution @xmath105 .
we thus find the following renormalization : @xmath106 this is consistent with the mean field renormalization of the assisted hopping term with @xmath107 .
@xmath107 also follows from @xmath108 .
this shows that we can indeed identify eq .
( [ hsw ] ) as a renormalization of the hybridization .
we have studied a model of an impurity , or a quantum dot , with a single electronic level , an on - site repulsion term , hybridization , and assisted hopping .
the flow equation method allows us to map the model onto an effective anderson model without assisted hopping , and with additional kondo couplings and pair hopping terms .
the most interesting outcome of our calculation is that the value of the on - site electron - electron interaction can become attractive , because of the effect on the flow of the initial assisted hopping term .
although the anderson model with attractive interactions has not been as extensively studied@xcite as the repulsive anderson model , we think that our mapping suffices to give a qualitative description of the physical properties of the model for the interesting case of a relatively strong assisted hopping term .
note that , in addition to a renormalized on - site interaction term , terms which describe pair hopping between the impurity and the conduction band are generated .
we expect that significant local pairing correlations will develop in this regime .
this result is in agreement with general arguments which suggest that assisted hopping interactions tend to favor superconducting ground states@xcite , and with mean field studies of the same impurity model@xcite .
a related tendency towards phase rigidity can be found in the limit where the level spacing within the impurity is negligible with respect to the other parameters@xcite .
the model does not have electron - hole symmetry , and local pairing is mainly favored when the number of electrons at the impurity exceeds one ( half filling ) , and when the direct hopping and the assisted hoping terms have opposite signs , also in agreement with general properties of bulk systems .
t.s . is supported by the daad - postdoctoral program .
we thank j. hirsch for a critical reading of the manuscript .
financial support from mcyt ( spain ) , through grant no .
mat2002 - 04095-c02 - 01 is gratefully acknowledged .
99 j. kondo , prog .
theor . phys . *
32 * , 37 ( 1964 ) .
p. w. anderson , phys .
rev . * 124 * , 41 ( 1961 ) .
y. alhassid , rev .
phys . * 72 * , 895 ( 2000 ) .
i. l. aleiner , p. w. brouwer , and l. i. glazman , phys . rep . * 358 * , 309 ( 2002 ) .
j. hirsch , phys .
b * 48 * , 3327 ( 1993 )
. j. hirsch , and f. marsiglio , phys .
b * 43 * , 424 ( 1991 ) .
f. guinea , phys .
b * 67 * , 195104 ( 2003 ) . m. ueda and f. guinea , zeits . fr phys .
b * 85 * , 413 ( 1991 ) .
e. bascones , c. p. herrero , f. guinea , and g. schn , phys .
b * 61 * 16778 ( 2000 ) .
f. wegner , ann .
( leipzig ) * 3 * , 77 ( 1994 ) ; f. wegner , physics reports * 348 * , 77 ( 2001 ) .
s.d . gazek and k.g .
wilson , phys .
d * 48 * , 5863 ( 1993 ) ; ibid .
* 49 * , 4214 ( 194 ) .
s.k . kehrein and a. mielke , j. phys .
a : math . gen . * 27 * , 4259 ( 1994 ) ; ann .
( ny ) * 252 * , 1 ( 1996 ) .
p. lenz and f. wegner , nucl .
b * 482 * [ fs ] , 693 ( 1996 ) ; a. mielke , ann .
( leipzig ) * 6 * , 215 ( 1997 ) .
we set the chemical potential equal to zero - even for finite temperatures .
kehrein and a. mielke , ann .
( leipzig ) * 6 * , 90 ( 1997 ) .
m. crisan , c.p .
moca , and i. tifrea , j. supercond . *
10 * , 251 ( 1997 ) ; c.p .
moca , m. crisan , and i. tifrea , _ ibid . _ * 11 * , 719 ( 1998 ) .
the coupling strengths correspond to @xmath109 and @xmath110 for constant coupling functions as used in ref . .
a. taraphder and p. coleman , phys .
lett . * 66 * , 2814 ( 1991 ) .
d. van der marel and j.e .
mooij , phys .
b * 45 * , 9940 ( 1992 ) .
the limit @xmath111 is mainly academical since the contribution of assisted hopping decreases with finite temperature , i.e. @xmath112 for @xmath113 , due to the broadening of the discrete spectrum of the quantum dot , see ref . .
schrieffer and p.a .
wolff , phys . rev . * 149 * , 491 ( 1966 ) .
|
we investigate the effect of assisted hopping in the anderson impurity model .
we use the flow equation method , which , by means of unitary transformations , generates a sequence of hamiltonians in order to eliminate the assisted hopping terms .
this approach yields a renormalized on - site energy @xmath0 , a renormalized correlation energy @xmath1 , and other terms , which include pair hopping . for some parameter values , the initial hamiltonian flows towards an attractive anderson model .
we argue that this result implies a tendency towards local pairing fluctuations .
|
circinus x1 ( cir x1 ) is a puzzling x ray binary .
because of its rapid variability a black hole was supposed to be the compact object in this system ( toor , 1977 ) , until type i x ray bursts were detected ( tennant et _ al .
_ , 1986 a , b ) , suggesting that the compact object is a neutron star .
a modulation with a period of 16.6 days was observed at several wavelengths in the emission of the source .
the high stability of this period was a strong indication that it is the orbital period of cir x1 .
the orbit is highly eccentric ( @xmath7 ) ( murdin et al .
1980 ; tauris et al . 1999 ) .
periodic radio flares near the zero phase ( i.e. at the periastron , for recent ephemeris see stewart et _ al .
_ , 1991 ) are also accompanied by drastic changes in the x ray lightcurve ( x ray flares and dips ) .
the position of the probable radio counterpart is close to the supernova remnant g321.90.3 ( clark , parkinson & caswell , 1975 ) , suggesting that cir x1 could be a runaway binary .
case & bhattacharya ( 1998 ) have revised the distance to the supernova remnant g321.9 - 0.3 , and hence to the associated cir x1 , to 5.5 kpc . therefore in the following
we adopt 5.5 kpc as the distance to the source .
quasi periodic oscillations ( qpos ) at 100200 hz were observed in this source ( tennant , 1987 ; shirey et al . , 1996 ) , but no khz qpos were detected .
recently iaria et al .
( 2001 ) detected the presence of highly ionized matter around cir x1 , with hydrogen column density of @xmath8 @xmath9 , at the orbital phase 0.110.16 .
the continuum emission was fitted using a comptonization model plus a power
law tail at energies higher than 10 kev .
the luminosity of the power law is @xmath10 of the total .
brandt & schulz ( 2000 ) , analysing chandra observations taken at the periastron , detected , for the first time in the x - ray band , the presence of p cygni lines in x ray band .
they explain these features as due to the presence of high velocity outflows from this system ( probably a moderate temperature , @xmath11 kev , wind from the x - ray heated accretion disk ) .
johnston et al .
( 1999 ) , using several infrared ( ir ) and optical data from 1976 to 1997 , observed along the whole orbit an asymmetric h@xmath12 emission line .
they argued that the broad component of the emission line arises in a high velocity ( @xmath13 km s@xmath1 ) , optically thick , flow near the neutron star .
brandt et al . (
1996 ) , using asca data taken on 1994 august 4 - 5 , found a sudden variation of the flux at the periastron phase , the count rate increasing from @xmath14 counts s@xmath1 to @xmath15 counts s@xmath1 . in the low - count rate state they obtained a good fit using a partial covering component , in addition to a two blackbody model , with a corresponding hydrogen column of @xmath16 @xmath9 ; in the high - count rate state the partial covering of the x - ray spectrum was not needed . because of the similarity with the spectra of seyfert 2 galaxies with compton - thin tori , brandt et al .
( 1996 ) suggested that matter at the outer edge of the accretion disk , together with an edge - on disc orientation , could explain the partial covering of the spectrum in cir x1 .
here we report the results of the spectral analysis in the energy range 0.610 kev performed on cir x1 data from the asca satellite taken in 1998 .
these data cover a whole orbital period allowing to study the corresponding changes in the properties of the matter around the system .
we performed a systematic spectral study of cir x1 along its orbital period using asca archival data taken in 1998 march between 3 and 18 .
the instruments on board asca ( tanaka , inoue , & holt , 1994 ) consist of two solid state imaging spectrometers ( sis ) and two gas imaging spectrometers ( gis , ohashi et al . , 1996 ; makishima et al . , 1996 ) .
in our analysis we used only data from gis instruments .
gis have an energy resolution of @xmath17 ( fwhm ) at 6 kev and their detection efficiency above @xmath18 kev is higher than with sis ( it is twice as large as the sis detection efficiency at energies of 6.57 kev ) .
we used data at high telemetry from the asca archive , screened using current standard filter criteria .
the gis photon extraction region was limited to the standard circular region of @xmath19 radius centered on the source .
the background was taken from a @xmath19 radius circle free of emission from cir x1 in the external region of the gis field of view and after an accurate investigation of the uniform background distribution . because of the high luminosity of the source it was necessary a deadtime correction after the extraction of the lightcurves and the spectra . to correct for this effect we used the ftools _ ldeadtime _ and _ deadtime _ for lightcurves and spectra ,
respectively , using the corresponding mkf files ( which contain the time - histories of various parameters from which good data can be identified and screened ) for each observation .
a systematic error of @xmath20 was added to the spectra .
the observations used for the analysis cover an entire orbital period from 1998 march 3 to march 19 . in table 1
we report for each of the seven analysed observations ( labelled a , b , ... , g ) the start and stop times , the exposure time and the relative orbital phase assuming the orbital ephemeris as reported by stewart et al .
( 1991 ) .
in figure 1 we plotted the cir x1 lightcurve in the energy band 0.610 kev ( upper panel ) , and the corresponding hardness ratio , i.e.the ratio of the count rate in the energy band 37 kev to that in the band 13 kev ( lower panel ) . in the lightcurve large variabilities
are observed , especially at orbital phase 00.2 , where sudden flux variations are visible , with the count rate ranging from @xmath21 counts s@xmath1 to @xmath22 counts s@xmath1 . in the same phase interval
we observe a rapid decrease of the hardness ratio from @xmath23 to @xmath24 . in the phase interval
0.30.6 the lightcurve does not show any significant variability and the count rate is @xmath15 counts s@xmath1 , while the hardness ratio increases from 0.7 to 0.9 . in the phase interval 0.750.88 the lightcurve shows a count rate slightly higher than 200 counts s@xmath1 and the hardness ratio is constant at @xmath25 .
the longest observation on march 3 shows a large variability in the hardness ratio .
therefore we divided this observation into four intervals ( labelled a1 , a2 , a3 and a4 ) , the first corresponding to the orbital phase 0.95 , the second to 0.981 , the third to 0.010.04 and the fourth to 0.050.09 . in table 2
we report the intervals in which the lightcurve was divided and the corresponding orbital phases .
we also show for each of the intervals the flux of the source extrapolated in the 2.525 kev energy range , in order to compare these values with previous observations .
the flux in the band 2.525 kev is very similar to that measured by shirey et al .
( 1999 ) with the rxte satellite . in the same table
we also report the flux of the source extrapolated in the 0.1100 kev energy range ( using the best fit spectral model discussed in the next section ) , and the corresponding luminosity of the source .
for each of the ten intervals described above , corresponding to different orbital phases , we extracted energy spectra , in the band 0.610 kev , in order to study the spectral variations along the orbit .
the continuum emission was fitted using a comptonization model ( _ comptt _ , titarchuk 1994 ) , modified , at low energies , by photoelectric absorption by cold matter . in a previous paper ( iaria et al . , 2001 ) the broad band ( 0.1100 kev ) spectrum from the bepposax satellite did not allow to distinguish between a simple comptonization and the so - called eastern model , consisting of a multicolor disk blackbody plus a blackbody component ( mitsuda et al . , 1984
) . however , because of the large value of the neutron star radius obtained from the fit ( @xmath26 km ) , we preferred the comptonization model to fit the spectrum
. therefore we use the same model here .
the average photoelectric absorption , obtained from the fit , is @xmath27 @xmath9 .
the seed photon temperature varies from 0.51 kev near the periastron to 0.37 kev at orbital phase 0.89 .
the electron temperature varies from 1 kev ( at orbital phase @xmath28 ) up to 1.7 kev ( at orbital phase 0.89 ) .
the compton optical depth varies between 16 and 20 along the orbital period . in spectra a1 ,
f , and g we improved the fit adding a partial covering to the model .
the probability of chance improvement for the addition of this component , estimated using an f - test , was 0.08 , @xmath29 and @xmath30 , respectively .
note that this component is essentially needed to fit an iron absorption edge .
in fact we obtain equivalently good fits if we use an absorption edge , at energies between 7.6 and 7.9 kev , instead of the partial covering for these intervals . in this case
, the edge optical depth ranges from 0.008 for interval a1 to 0.12 for interval f. however , in agreement with the results of brandt et al .
( 1996 ) , in the following we adopt and discuss the partial covering interpretation . in some spectra ( a4 to e ) we detected the presence of an absorption edge from highly ionized iron . in the spectra where the partial covering is present there is not evidence of a such feature . in a2 ,
a3 and d neither the partial covering nor the absorption edge are present ; in these cases we could only find an upper limit on the optical depth of the absorption edge .
an emission line from ionized iron is also needed to fit some of the spectra ( a2 to a4 and c to e ) . in table 2
we show the spectral features that were included in the best fit model for each of the ten spectra , and in tables 3 and 4 we report the parameters of the best fit model , together with the probability of chance improvement of the fit when the indicated component is included in the spectral model .
we analyzed data of cir x1 from seven asca observations ( spanning an entire orbital period of the x - ray source ) in the energy range 0.610 kev . from these observations we extracted ten spectra corresponding to different orbital phases .
the lightcurve shows large variability at the orbital phase interval 00.2 . at phases 0.30.6
the lightcurve is constant and at phases 0.780.89 the count rate decreases .
the equivalent hydrogen absorption column , @xmath31 , derived from the best fit model to these spectra is @xmath32 @xmath9 , in agreement with previous observations ( brandt et _ al .
_ , 1996 ; iaria et al . , 2001 ) .
for a distance to the source of 5.5 kpc the visual extinction in the direction of cir x1 is @xmath33 mag ( hakkila et al .
1997 ) . using the observed correlation between visual extinction and absorption column ( predehl & schmitt 1995 ) we find that the expected hydrogen column for cir x1 is @xmath34 @xmath9 .
it is slightly smaller than the value obtained from the fit , probably because of the presence of obscuring matter close to the x ray source .
the small fluctuations of the measured galactic absorption along the orbital phase , which seem to be statistically significant for the intervals a3 and a4 ( see tab .
3 ) , can be caused by fluctuations in the density ( and/or in the ionization level ) of this excess of matter around the system .
we fitted the continuum using a comptonization model .
the values of the seed photon and electron temperatures with respect to the orbital phase are plotted in figure 2 ( upper and middle panel , respectively ) .
the spectra a1 , f and g require in addition a partial covering component to fit an iron edge at 7.67.9 kev .
an absorption edge from ionized iron is present in spectra a4 , b , c and e. in spectrum d the addition of this component is not statistically significant and we find only an upper limit for its optical depth . the unabsorbed flux , extrapolated in the band 0.1100 kev , is reported in figure 3 . in this figure we marked the corresponding eddington flux for a neutron star of @xmath35 at a distance of 5.5 kpc .
the flux seems to be super eddington at the orbital phases 00.2 and 0.951 ( near the periastron phase ) . for each spectrum we calculated the radius , @xmath36 , of the seed - photon wien spectrum using the parameters reported in tables 3 and 4 . assuming a spherical geometry , this radius can be expressed as @xmath37 km ( in t zand et _ al .
_ 1999 ) , where @xmath38 is the distance to the source in kpc , @xmath39 is the unabsorbed flux of the comptonization spectrum in erg @xmath9 s@xmath1 , @xmath40 is the seed photon temperature in kev , and @xmath41 is the relative energy gain due to the comptonization .
the obtained radii , which are reported in tables 3 and 4 , are in the range between 60 and 120 km . for the sake of clarity @xmath36
is also plotted versus the orbital phase in figure 2 ( lower panel ) .
following iaria et al .
( 2001 ) we interpret this radius as the inner radius of the disk .
tennant ( 1987 ) , using exosat data , and shirey et al .
( 1996 ) , using rxte data , observed the presence of a qpo at frequencies between 100 and 200 hz in cir x1 .
this qpo has been tentatively identified with the lower khz qpo based on the observed correlations among the qpo frequencies in ns lmxbs and black hole binary systems ( psaltis , belloni , & van der klis 1999 ) , although the interpretation of this feature needs to be confirmed .
however , we observe that assuming that the frequencies of this qpo are related to the keplerian velocity of the plasma in the accretion disk , we deduce that the corresponding radii are between 60 and 120 km , similar to the obtained radius of the seed photon emitting region .
note also that the other luminous low mass x
ray binaries ( lmxbs ) of the z class show a radius of the seed
photon emitting region much closer to the neutron star radius as compared to cir x1 ( see di salvo et al .
, 2000 ; di salvo et al . , 2001 ) , and , accordingly , these z sources present qpos at khz frequencies in their power spectra . note , however , that interpreting the seed - photon radius as the inner radius of the disk would imply a quite high accretion rate through the disk of @xmath42 g / s , i.e. @xmath18 times the eddington accretion rate , for @xmath43 km and @xmath44 kev . a possibility to explain the large seed photon radius in cir x1 is the presence of a relatively large magnetic field : we obtain a strength of the magnetic field in cir x1 of @xmath45 g , assuming that the measured seed - photon radius is the magnetospheric radius .
this strength is one order of magnitude higher than the supposed magnetic field for lmxbs of the z class .
moreover , the interpretation of @xmath36 as the magnetospheric radius would imply lower values of @xmath36 at higher flux levels , that is not observed ( see fig .
3 ) , although the error bars are large especially at phases between 0.3 and 0.9 .
another possibility to explain the large value of the seed
photon radius is that the inner region of the disk is not directly visible , for the presence of optically thick material in the central part of the system , such as the optically thick corona that probably produces the comptonization spectrum , or it is absent , because , for instance , its matter is expelled in a jet or wind .
we observe that the seed
photon radius seems to increase during the flaring activity at the orbital phase between 0 and 0.2 .
we could suppose that because of the high ( probably super eddington ) accretion rate , a shock is produced in the inner region of the system , pushing the accreting matter towards the external regions .
a similar scenario was proposed by haynes et al .
( 1980 ) , who explained the observed radio flares at the periastron as coming from a shock front formed by the super eddington accretion . also , as cir x1 is thought to be observed at a high inclination ( brandt et al .
1996 ) , it is possible that the optically thick material at the shock intercepts and reprocesses the radiation from the inner regions , which can therefore not be directly observed
. the comptonized spectrum could originate in an accretion disk corona ( adc ) that could be formed by evaporation of the outer layers of the disk illuminated by the emission of the central object ( white & holt , 1982 ) .
the radius of the corona can be written as @xmath46 ( white & holt , 1982 ) , where @xmath47 is the mass of the compact object , @xmath48 and @xmath49 are mass and radius of the sun , and @xmath50 is the adc temperature in units of @xmath51 k. under this hypothesis , using the values reported in tables 3 and 4 , we find that the radius of the adc , @xmath52 , varies along the orbit in the interval @xmath53 km
. we can infer the density of the adc using the relation @xmath54 , where @xmath55 is the optical depth obtained from the fit , @xmath56 is the thomson cross - section , @xmath57 is the number of particles per unit volume and @xmath52 is the adc radius calculated above .
note that we are considering @xmath57 constant along the radius of the corona , that is a rough approximation . under
this hypothesis @xmath57 varies between @xmath58 @xmath59 .
these are typical values of density for an adc ( vrtilek et al . , 1993 ) .
we observe that the electron temperature increases in the orbital phase interval 0.20.9 while it decreases in 0.91 and 00.1 .
this behavior can be due to the fact that at orbital phase 0.9 the the accretion rate increases and consequently the inner temperature of the accretion disk increases ( as is seen in fig .
2 , upper panel ) . the injection of more soft photons in the corona enhances the compton cooling and the electron temperature decreases . in the orbital interval 00.2 the radiation pressure moves the shock to @xmath60 km , the seed photon temperature decreases , and so does the electron temperature .
after , at orbital interval 0.20.9 , the injection of soft photons decreases , the cooling of the corona becames less efficient and the electron temperature increases . an iron emission line at energies 6.66.8 kev is present in some spectra ( a2 , a3 , a4 , c , d and e ) , corresponding to emission from fe xxiii xxv
. the equivalent width of the line is @xmath61 ev .
we have investigated the hypothesis that this emission line could be produced by reprocessing of x
rays , coming from the corona or the inner part of the system , by the disk surface ( e.g. brandt & matt , 1994 ) . if cir x1 is observed at high inclination ( as suggested by brandt et al .
1996 ) an iron emission line at @xmath62 kev could be produced by the mechanism described above , but the corresponding equivalent width should be @xmath63 ev and the width of the line , @xmath64 , between 0.60.8 kev ( see fig . 2 , 3 , and 4 in brandt & matt , 1994 ) .
the predicted equivalent width seems too low to explain the observed line in cir x1 .
therefore , in agreement with chandra results ( brandt & schulz 2000 ) we will consider in the following that the iron emission line is produced in a outflow of ionized matter from the accretion disk ( see below ) . in this section
we summarize our interpretation of the spectral evolution of cir x1 along the orbit based on the results from the analysis described above , as well as the bepposax results ( iaria et al . , 2001 ) , the chandra results ( brandt & schulz , 2000 ) and the detection in ir and optical bands of emission lines along the orbital period of the source ( johnston et al . , 1999 ) . the broader component of the h@xmath12 emission line , observed along the whole orbit of the source , can be interpreted as produced in an outflow of matter with a velocity of @xmath13 km s@xmath1 ( johnston et al . , 1999 ) .
this detection gives us the information that an outflow of matter , from the compact system , is present in cir x1 .
the lack of the red wing of the line is associated to an optically thick outflow .
chandra satellite observed , near the periastron ( phase @xmath65 ) , the presence of discrete features with p
cygni profiles in the x ray band ( brandt & schulz , 2000 ) .
the inferred velocity of the outflow is again @xmath13 km s@xmath1 , in agreement with the results of the ir observations .
moreover , the fact that p
cygni profiles are observed for fe xxiv and other highly ionized elements , suggests that the matter in the outflow at the periastron phase is highly ionized .
iaria et al .
( 2001 ) observed near the periastron ( phase @xmath66 ) the presence of an absorption edge at energy @xmath5 kev corresponding to highly ionized iron .
brandt & schulz ( 2000 ) argue that the absorption edge could be associated to the outflow , probably consisting of a thermally driven wind formed by the x rays from the inner region which heat the disk surface .
it is not straightforward to infer the geometry ( radial or vertical ) of the outflow from these data .
however , we can see that the model of thermally driven wind proposed by begelman et al .
( 1983 ) could explain the observed outflow .
in this model the outflow is near the disk surface and mainly in the direction of the observer ( begelman et al .
1983 ; begelman & mckee 1983 ) .
this seems reasonable in the case of cir x1 given that brandt et al .
( 1996 ) argued that cir x1 is almost edge on , and the p - cygni profiles ( to date the only ones observed in a lmxb ) indicate that we are observing both the approaching and the receding part of the outflow .
this geometry could also explain the large optical depth of the iron absorption edge ( @xmath67 ; iaria et al . , 2001 ) .
however , the model of thermally driven wind does not take into account the radiation pressure , and is valid for luminosities of a few percent of the eddington luminosity or less ( begelman et al .
1983 ) , while the radiation pressure could be significant in cir x1 because of the high luminosity of the source , especially near the periastron .
the study of cir x1 along the entire orbit allows to describe the evolution of the iron absorption edge and emission line present in the spectra with a scenario where an outflow from the system is almost always present , but its physical conditions change as function of the accretion rate along the orbital period . we distinguish three states of the source along the orbit : 1 . a super eddington state at phases 0.991 and 00.20 , when the outflow of highly ionized matter ( and probably a shock too ) is formed ; 2 . a quiet state between 0.30.7 ,
when the source has a sub eddington luminosity and the outflow is weaker ; 3 . a partial covering state between 0.770.95 , when the atmosphere of the companion star and/or the accreting matter from the companion , which probably forms a bulge in the outer disk , partially cover the emission from the compact object .
the spectra a2 , a3 , a4 and b belong to the same state of the source , corresponding to the phase interval around the periastron , 0.991 and 00.2 . in these spectra
there is not evidence of partial covering , suggesting the absence of neutral matter around the system ( contrarily to what is observed at previous orbital phases , between 0.780.96 , see below ) .
the flux is super
eddington for each spectrum , for the assumed distance and in the hypothesis that the emission from the source is mainly isotropic . in this state probably
a shock forms in the inner region of the system , and radio and x ray flares might be associated to the shock front ( haynes et al .
cygni profiles observed by brandt & schulz ( 2000 ) could be produced by the high luminosity ( and then the high radiation pressure ) that sweeps away and ionizes the accreting matter around the system . in spectra a2 and a3
we do not detect the absorption edge , but an iron emission line is present at energy @xmath68 kev ( produced by fe xxiii , turner et al .
1992 ) and @xmath62 kev ( fe xxiv fe xxv ) , respectively , implying a gradual increase of the ionization level in the outflow .
the equivalent width of the iron energy line at @xmath69 kev is @xmath14 ev , that is smaller than the equivalent width of the iron emission line in the other spectra , probably because the fluorescence yield associated to fe xxiii is smaller than that associated to fe xxiv and fe xxv ( turner et al , 1992 ) . in the hypothesis that the iron emission line is associated to the outflow , we deduce that also during intervals a2 and a3 the matter in the outflow is highly ionized .
we argue that the outflow has a low density at these orbital phases because the absorption edge is not observed ( the corresponding equivalent hydrogen column is lower than @xmath70 @xmath9 ) . the presence of the h@xmath12 line , detected at these phases by johnston et al .
( 1999 ) , suggests that the matter of the outflow is either photoionized or heated by a non thermal emission .
in fact , if the matter in the outflow would be heated by a thermal emission , we should not observe an h@xmath12 emission , because recombination to neutral hydrogen should not be possible . in spectra a4 and b ( orbital phase 0.050.19 )
an absorption edge is present .
the energy of the edge is 8.46 kev and 8.71 kev , respectively .
these energies are compatible with iron ionization levels of fe xxiii xxv ( turner et _ al .
the best fit value for the optical depth @xmath71 , considering the photoionization cross section for the k shell of fe xxiii ( krolik & kallman , 1987 ) and assuming cosmic abundance of iron , corresponds to a hydrogen column density of @xmath72 @xmath9 and @xmath73 @xmath9 for a4 and b respectively , in agreement with the results reported by iaria et al .
the iron ionization level is similar to that in spectra a2 and a3 and the presence of the edge could be explained by an increase of the density and/or optical depth of the outflow .
the iron emission line is still observed in spectrum a4 ( together with the absorption edge ) , but it is not detected is spectrum b. the spectra c , d and e , corresponding to the orbital phase 0.30.6 , show a sub eddington flux : the lightcurve is quite constant at @xmath74 counts s@xmath1 and the hardness ratio increases from @xmath75 up to 0.9 .
photon radius is constant at @xmath76 km , the seed photon temperature is constant at @xmath77 kev , while the electron temperature increases from 1.2 kev up to 1.5 kev . in these spectra
an absorption edge is present at @xmath78 kev , with optical depth decreasing from 0.12 to 0.06 .
the energy of @xmath6 kev indicates that the matter is less ionized and the low values of the optical depth indicate that the absorption column decreases .
the equivalent hydrogen column goes from @xmath79 @xmath9 down to @xmath80 @xmath9 .
an iron emission line is also present , the energy of which is stable at @xmath62 kev .
this indicates that the outflow is still present , although it is now probably much weaker .
cir x1 has a highly eccentric orbit ( @xmath81 ; murdin et al . , 1980 ; tauris et al , 1999 ) .
it is possible that , for the geometry of the system ( see e.g. fig . 4 in johnston et al . , 1999 )
, near the periastron , the neutron star is occulted by the outer layers of the companion - star atmosphere .
also , the high eccentricity of the orbit might produce large tidal interactions at these phases , which expand and/or deform the stream of the accreting matter , maybe forming a bulge in the outer accretion disk .
the excess of neutral matter modifies the observed emission of the x - ray source , producing a partial covering of the energy spectrum of the source ( see also inoue 1989 ) .
the partial covering component is detected in spectra a1 , f and g , corresponding to the phase interval between 0.78 and 0.97 . in this state the ( ionized ) iron absorption edge and emission line are not detected . at orbital phase
@xmath82 the hydrogen column , inferred by the partial covering , is @xmath83 @xmath9 , and it increases up to @xmath84 @xmath9 at the orbital phase @xmath85 .
this is probably due to the approach of the x - ray source to the companion star . at 0.95
the hydrogen column decreases to @xmath86 @xmath9 , and the partial covering component disappears at following orbital phases .
the decrease of the hydrogen column can be caused by the increase of the accretion rate to super
eddington values , which photoionizes the surrounding neutral matter .
in fact the x - ray flux increases from spectrum f to spectrum g ( see fig .
3 ) , indicating that the accretion rate increases at these phases .
recently several z sources were analysed using a broad band ( 0.1200 kev ) energy range ( e.g. gx 17 + 2 , di salvo et al . , 2000 ;
gx 349 + 2 , di salvo et al . , 2001 ) .
their spectra show an absorption edge associated to highly ionized matter , as it is observed in cir x1 .
the hydrogen column associated to the absorption edge in these z sources is @xmath87 @xmath9 and the equivalent widths of the iron emission line present in these spectra are @xmath88 ev .
these equivalent widths , according to vrtilek et al .
( 1993 ) who suppose that emission lines are produced in a photoionized adc , could be expected in sources with a low inclination angle .
the phenomenon of the outflow from the accretion disk connected to eddington
sources is partially confirmed by the detection , in ir band , of lines with p
cygni profiles in sco x1 and gx 13 + 1 ( bandyopadhyay et al . , 1999 ) . because the outflow of ionized matter should be mostly along the disk surface
, the absorption edge should be more evident in sources with high inclination .
therefore the small values of the optical depth of the absorption edge in gx 17 + 2 and gx 349 + 2 could be associated to low inclinations . in this case
the absorption edge might be considered an indicator of the inclination angle of the source .
we have analyzed observations of cir x1 from asca archive , performed in 1998 march and corresponding to a whole orbital period .
cir x1 is most probably a high inclination source , and this might explain why we do not observe the blackbody emission of the neutron star or inner accretion disk .
we observe a comptonized component probably coming from a adc .
we find that the optical depth of the comptonizing region is @xmath63 , the seed photon temperature ranges from 0.4 to 0.5 kev , the electron temperature increases from 1 kev at phase 0.05 up to 1.8 kev at phase 0.9 .
the radius of the seed photons is around 90 km .
this could be explained either by a magnetic field of @xmath89 g or , more probably , by the presence of optically thick material in the inner part of the system which hides the innermost regions . during the observations , which span an entire orbital period of the x - ray source
, it is possible to distinguish three states of the source : a high - accretion rate ( probably super eddington ) state , where x ray flares are present , near the periastron ; a quiet state , where the system accretes at a sub eddington rate , at orbital phases from 0.3 to @xmath90 , and a partial covering state , at orbital phases 0.780.97 .
the first state is characterized by the presence of a prominent iron absorption edge , probably produced in an outflow of highly ionized matter .
the matter is swept away and photoionized by the high luminosity of the x - ray source in this state .
the second state shows an absorption edge of less ionized iron with an equivalent hydrogen column smaller than @xmath70 @xmath9 ; the outflow is probably still present in this state , although much weaker .
the third state is characterized by the presence of partial covering , explained as due to absorption by the atmosphere of the companion star and/or by the accreting matter from the companion , which , at these orbital phases , can fall in the line of sight between the x - ray source and the observer .
this work was supported by the italian space agency ( asi ) , by the ministero della ricerca scientifica e tecnologica ( murst ) .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center .
asai , k. , dotani , t. , nagase , f. , mitsuda , k. , 2000 , apjs , 131 , 571 bandyopadhyay , r. m. , shahbaz , t. , charles , p. a. , naylor , t. , 1999 , mnras , 306 , 417 begelman , m. c. , mckee , c. f. , shields , g. a. , 1983 , apj , 271 , 70 begelman , m. c. , mckee , c. f. 1983 , apj , 271 , 89 brandt , w. m. ; matt , g. , 1994 , mnras , 268 , 1051 brandt , w. n. , fabian , a. c. , dotani , t. , nagase , f. , inoue , h. , kotani , t. , segawa , y. , 1996 , mnras , 283 , 1071 brandt , w. n. ; schulz , n. s. , 2000 , apjl , 544 , l123 case , g. l. , bhattacharya , d. , 1998 , apj , 504 , 761 clark , d. h. , parkinson , j. h. , caswell , j. l. , 1975 , nature , 254 , 674 di salvo t. et al . , 2000 , apj ,
544l , 119 di salvo t. et al . , 2001 ,
apj in press fender , r. p. , et al . , 1998 , apj , 506 , l121 haynes , r. f. , lerche , i. , murdin , p. , 1980 , a&a , 87 , 299 hakkila j. , myers j. m. , stidham b. j. , hartmann d. h. , 1997 , aj , 114 , 2043 johnston , h. m. , fender , r. , wu , k. , 1999 , mnras , 308 , 415 iaria , r. ; burderi , l. , di salvo , t. , la barbera , a. , robba , n. r. , 2001 , apj , 547 , 412 inoue , h. , the 23rd eslab symposium on two topics in x ray astronomy .
volume 1 : x ray binaries , 109 in t zand , j. j. m. et al . , 1999 ,
a&a , 345 , 100 krolik , j. h. , kallman t. r. , 1987 , apjl , 320 , l5 makishima , k. , et al .
, 1996 , pasj , 48 , 171 mitsuda , k. , et al . , 1984 ,
pasj , 36 , 741 murdin , p. , jauncey , d. l. , haynes , r. f. , lerche i. , nicolson , g. d. , holt , s. s. , kaluzienski , l. j. , 1980 , a&a , 87 , 292 ohashi , t. , et al . , 1996 , pasj , 48 , 157 predehl p. & schmitt j. h. m. m. , 1995 , a&a , 293 , 889 psaltis , d. , belloni , t. , van der klis , m. 1999 , apj , 520 , 262 shirey r. e. , bradt , h. v. , levine , a. m. , morgan , e. h. , 1996 , apjl , 469 , l21 shirey r. e. , bradt , h. v. , levine , a. m. , 1999 , apj , 517 , 472 stewart , r. t. , nelson , g. j. , penninx , w. , kitamoto , s. , miyamoto , s. , nicolson , g. d. , 1991 , mnras , 1991 , 253 stewart , r. t. , caswell , j. l. , haynes , r. f. , nelson , g. j. , 1993 , mnras , 261 , 593 tanaka , y. , inoue , h. , holt , s. s. , 1994 , pasj , 46 , l37 tauris , t. m. , fender , r. p. , van den heuvel , e. p. j. , johnston , h. m. , wu , k. , 1999 , mnras , 310 , 1165 tennant , a. f. , fabian , a. c. , shafer , r. a. , 1986a , mnras , 219 , 871 tennant , a. f. , fabian , a. c. , shafer , r. a. , 1986b , mnras , 221 , 27 tennant , a. f. , 1987 , mnras , 226 , 971 titarchuk l. , 1994 , apj , 434 , 570 toor , a. , 1977 , apjl , 174 , l57 turner , t. j. , done , c. , mushotzky , r. , madejski g. , 1992 , apj , 391 , 102 vrtilek , s. d. , soker , noam , raymond , j. c. , 1993 , apj , 404 , 696 white , n. e. , holt , s. s. , 1982 , apj , 257 , 318
0.5 cm c|c|c|c|c observation & start time & stop time & exposure time ( ks ) & orbital phase + a & 1998 - 03 - 03 12:12:49 utc & 1998 - 03 - 05 18:50:33 utc & 38.2 & 00.09 , 0.951 + b & 1998 - 03 - 06 23:33:13 utc & 1998 - 03 - 07 09:00:43 utc & 7.9 & 0.170.19 + c & 1998 - 03 - 09 04:19:38 utc & 1998 - 03 - 09 15:10:39 utc & 10.2&0.300.33 + d & 1998 - 03 - 11 02:21:16 utc & 1998 - 03 - 11 13:30:52 utc & 6.9 & 0.410.44 + e & 1998 - 03 - 13 02:12:29 utc & 1998 - 03 - 13 13:20:47 utc & 8.4 & 0.540.56 + f & 1998 - 03 - 17 00:44:30 utc & 1998 - 03 - 17 11:20:24 utc & 4.9 & 0.780.80 + g & 1998 - 03 - 18 16:24:23 utc & 1998 - 03 - 19 00:21:19 utc & 3.9 & 0.870.89 + 0.5 cm c|c|c|c|c|c|c|c interval & absorption & partial & emission & flux & flux & luminosity & orbital + & edge & covering & line & 2.525 kev & 0.1100 kev & 0.1100 kev & phase + a1 & no & yes & no & 3.8 & 5.8 & 2.1 & 0.95 + a2 & no & no & yes & 3.8 & 5.4 & 1.9 & 0.99 + a3 & no & no & yes & 4.1 & 7.0 & 2.5 & 0.010.04 + a4 & yes & no & yes & 2.5 & 5.9 & 2.1 & 0.050.09 + b & yes & no & no & 2.8 & 6.4 & 2.3 & 0.170.19 + c & yes & no & yes & 2.4 & 4.3 & 1.5 & 0.300.33 + d & yes & no & yes & 2.7 & 4.6 & 1.6 & 0.410.44 + e & yes & no & yes & 2.7 & 4.3 & 1.6 & 0.540.56 + f & no & yes & no & 3.1 & 4.6 & 1.7 & 0.780.80 + g & no & yes & no & 3.5 & 5.0 & 1.8 & 0.870.89 + 0.5 cm l|c|c|c|c|c spectra & a1 & a2 & a3 & a4 & b + orbital phase & 0.95 & 0.99 & 0.010.04 & 0.050.09 & 0.170.19 + & & & & & + @xmath91 @xmath92 & 1.630 ( fixed ) & @xmath93 & @xmath94 & @xmath95 & @xmath96 + @xmath97 @xmath98 & @xmath99 & & & &
+ @xmath100 & @xmath101 & && & + e@xmath102 ( kev ) & & 8.3 ( fixed ) & 8.3 ( fixed ) & @xmath103 & @xmath104 + @xmath105 & & @xmath106 & @xmath107 & @xmath108 & @xmath109 + @xmath110 ( kev ) & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 + @xmath116 ( kev ) & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 + @xmath55 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 + n@xmath127 & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath132 + f@xmath133(@xmath134 ) & 5.8 & 5.4 & 7.0 & 5.9 & 6.4 + r@xmath135 ( km ) & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 + e@xmath141 ( kev ) & & @xmath142 & @xmath143 & @xmath144 & + @xmath145 ( kev ) & & @xmath146 & @xmath147 & @xmath148 &
+ i@xmath149 @xmath150 & & @xmath151 & @xmath152 & @xmath153 & + @xmath154 ( ev ) & & @xmath155 & @xmath156 & @xmath157 & + @xmath158/d.o.f . &
309/391 & 252/389 & 244/389 & 259/387 & 263/390 + f test ( line ) & & @xmath159 & @xmath160 & @xmath160 & + f test ( edge ) & & @xmath161 & @xmath161 & @xmath160 & @xmath162 + f test ( pc ) & 0.08 & & & &
+ 0.5 cm l|c|c|c|c|c spectra & c & d & e & f & g + orbital phase & 0.300.33 & 0.410.44 & 0.540.56 & 0.780.80 & 0.870.89 + & & & & & + @xmath91 @xmath92 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 + @xmath97 @xmath98 & && & @xmath168 & @xmath169 + @xmath100 & && & @xmath170 & @xmath171 + e@xmath102 ( kev ) & @xmath172 & 8.3 ( fixed ) & @xmath173 & & + @xmath105 & @xmath174 & @xmath175 & @xmath176 & &
+ @xmath110 ( kev ) & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 + @xmath116 ( kev ) & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 + @xmath55 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 + n@xmath127 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath196 + f@xmath133(@xmath134 ) & 4.3 & 4.6 & 4.3 & 4.6 & 5.0 + r@xmath135 ( km ) & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 + e@xmath141 ( kev ) & @xmath202 & @xmath203 & @xmath204 & & + @xmath145 ( kev ) & @xmath205 & @xmath206 & @xmath207 & &
+ i@xmath149 @xmath150 & @xmath208 & @xmath209 & @xmath210 & & + @xmath154 ( ev ) & @xmath211 & @xmath212 & @xmath213 & & + @xmath158/d.o.f . &
256/387 & 251/387 & 250/387 & 285/390 & 309/390 + f test ( line ) & @xmath160 & @xmath214 & @xmath215 & & + f test ( edge ) & @xmath216 & @xmath161 & @xmath217 & & + f test ( pc ) & & & & @xmath29 & @xmath30 +
* figure 1 * : upper panel : cir x1 lightcurve in the energy band 0.610 kev . the intervals in which we divided these observations
are also indicated with the corresponding labels .
lower panel : the ratio of the count rate in the energy band 37 kev to that in band 13 kev .
the bin time is 5595 s. + * figure 2 * : the best fit values of the seed photon temperature , electron temperature and radius of the seed photons are plotted versus the orbital phase of cir x1 in the upper , middle and lower panel , respectively . + * figure 3 * : unabsorbed flux of cir x1 extrapolated in the band 0.1100 kev .
we marked with a solid line the flux of @xmath218 erg s@xmath1 @xmath9 , which is the eddington flux for a neutron star with a mass of 1.4@xmath48 at a distance of 5.5 kpc .
|
we report on the spectral analysis of circinus x1 observed by the asca satellite in march 1998 along one orbital period .
the luminosity of the source ( in the 0.1100 kev band ) ranges from @xmath0 erg s@xmath1 at the periastron ( orbital phase @xmath2 ) to @xmath3 erg s@xmath1 at orbital phase @xmath4 . from the spectral analysis and the lightcurve
we argue that cir x1 shows three states along the orbital evolution .
the first state is at the orbital phase interval 0.970.3 : the luminosity becames super eddington and a strong flaring activity is present . in this state
a shock could form in the inner region of the system due to the super eddington accretion rate , producing an outflow of ionized matter whose observational signature could be the prominent absorption edge at @xmath5 kev observed in the energy spectrum at these phases . in the second state ,
corresponding to the orbital phase interval between 0.3 and 0.7 , the accretion rate is sub eddington and we observe a weaker outflow , with smaller hydrogen column : the absorption edge is now at @xmath6 kev with an optical depth a factor of 2.5 to 6 smaller .
the third state corresponds to the orbital phase interval 0.780.97 . in this state the best fit model to the spectrum requires the presence of a partial covering component , indicating that the emission from the compact object is partially absorbed by neutral matter , probably the atmosphere of the companion star and/or the accreting matter from the companion .
|
linear colliders , lc , at the center of mass energy from 100 gev up to several tev will be one of the central instruments in experimental high energy physics in the next 2 to 3 decades .
four projects are being developed : nlc @xcite , tesla @xcite , jlc @xcite , and clic @xcite .
tesla team has published in march 2001 the technical design of the superconductiing linear collider on the energy 90800 gev @xcite .
the unique feature of the @xmath0 linear colliders is the possibility to construct on its basis a photon collider using the process of the compton backscattering of laser light off the high energy electrons @xcite .
this option is considered now for all linear colliders projects : nlc @xcite ; tesla @xcite ; jlc @xcite ; clic @xcite . in the following
we discuss the physics programme , possible laser and optical schemes , the expected and luminosities and some other interaction region aspects .
examples are given mostly for the photon collider at tesla , other projects are also discussed where they have principle differences .
the basic scheme of the photon collider is shown in fig .
[ ggcol ] .
discussion of the photon collider scheme and basic principles can be found elsewhere @xcite .
[ ggcol ] let us remind main features of photon colliders .
the maximum energy of the scattered photons is @xcite @xmath1\left[\frac{\omega_0}{{\mbox{\,ev}}}\right]= 19\left[\frac{e_0}{{\mbox{\,tev}}}\right ] \left[\frac{{\mbox{\,$\mu$m}}}{\lambda}\right ] , \label{eq1}\ ] ] where @xmath2 is the electron beam energy and @xmath3 the energy of the laser photon .
for example , for @xmath4 , @xmath5 ( @xmath6 ) ( powerful solid state lasers ) we obtain @xmath7 and @xmath8 ( it will be somewhat lower due to nonlinear effects in compton scattering . the value @xmath9 is a good choice for photon colliders , because for @xmath10 the produced high energy photons create qed pairs in collision with the laser photons , and as result the luminosity is reduced @xcite .
hence , the maximum center of mass system energy in collisions is about 80% , and in collisions 90% of that in collisions .
if for some study lower photon energies are needed , one can use the same laser and decrease the electron beam energy . the same laser with @xmath11
can be used for all tesla energies . at @xmath12 the parameter @xmath13 , which is larger than 4.8 .
but nonlinear effects at the conversion region effectively increase the threshold for production , so that production is significantly reduced @xcite .
the luminosity distribution in collisions has a high energy peak and a low energy part .
the peak has a width at half maximum of about 15% .
the photons in the peak can have a high degree of circular polarization .
this peak region is the most useful for experimentation .
when comparing event rates in and collisions usually we use the value of the luminosity in this peak region @xmath14 where @xmath15 .
the energy spectrum of high energy photons becomes most peaked if the initial electrons are longitudinally polarized and the laser photons are circularly polarized .
this gives almost a factor of 34 increase of the luminosity in the high energy peak .
the average degree of the circular polarization of the photons within the high - energy peak amounts to 9095% .
the sign of the polarization can easily be changed by changing the signs of electron and laser polarizations .
the luminosity in the high energy luminosity peak for tesla is just proportional to the geometric luminosity @xmath16 of the electron beams : @xmath17 .
the latter can be made larger for collisions than the luminosity because beamstrahlung and beam repulsion are absent for photon beams .
it is achieved using beams with smallest possible emittances and stronger beam focusing in the horizontal plane ( in collisions beams should be flat due to beamstrahlung ) . for current tesla parameters @xcite ( see details in section 3 ) @xmath18 approximately similar luminosities are expected in collisions .
typical cross sections in collisions are about one order of magnitude higher than those in collisions therefore the number of interesting events at photon colliders will be similar of even higher than in collisions .
the physics potential of the photon collider is very rich and complements in an essential way the physics program of the mode . a few examples : * in collisions , resonances with @xmath19 are produced as single resonances .
the effective cross section of the @xmath20 production at photon colliders is even higher that in collisions @xcite .
one of the most important examples is the higgs boson .
the precise knowledge of its two photon width is of particular importance .
it is sensitive to heavy virtual charged particles .
supersymmetry predicts three neutral higgs bosons .
photon colliders can produce the heavy neutral higgs bosons with masses about 1.5 times higher than in collisions at the same collider ( because heavy higgs bosons @xmath21 and @xmath22 have almost equal masses and in collisions are produced in pairs ) . *
a collider can produce pairs of any charged particles ( charged higgs , supersymmetric particles etc . ) with a cross section about one order of magnitude higher than those in collisions ( see graphs elsewhere @xcite .
moreover , the cross sections depend in a different form on various physical parameters .
the polarization of the photon beams and the large cross sections allow to obtain valuable information on these particles and their interactions .
* at a collider charged particles can be produced with masses higher than in pair production of collisions ( like a new @xmath23 boson and a neutrino or a supersymmetric scalar electron plus a neutralino ) .
* photon colliders offer unique possibilities for measuring the fusion of hadrons for probing the hadronic structure of the photon .
a short list of physics processes for the photon collider is presented in table [ processes ] @xcite .
more detail consideration of the physics program at photon colliders can be found elsewhere @xcite .
[ processes ]
the basic scheme for photon colliders is shown in fig .
[ ggcol ] . the distance between the conversion point ( cp ) and the ip , @xmath24 ,
is chosen from the relation @xmath25 , so that the size of the photon beam at the ip has equal contributions from the electron beam size and the angular spread from compton scattering . at tesla @xmath26
gives @xmath27 at @xmath28 .
larger @xmath24 values lead to a decrease of the luminosity . for smaller @xmath24 values the high energy luminosity increases only a little while the low energy photons give a larger contribution to the luminosity ( which is not useful for the experiment but causes additional backgrounds ) .
the removal of the disrupted beams can best be done using the crab - crossing scheme fig .
[ ggcol ] , which is foreseen in the nlc and jlc projects for collisions . due to the collision angle the outgoing disrupted beams
travel outside the final quads .
the value of the crab crossing angle is determined by the disruption angles and the final quad design ( diameter of the quad and its distance from the ip ) .
simulation shows that for tesla the maximum disruption angle is about 1012 mrad . above this angle
the total energy of particles in disrupted beams is smaller than from unremovable pairs backgrounds . in the present tesla design @xmath29 . at first sight
, one may think that there are no collision effects in and collisions because at least one of the beams is neutral .
this is not correct because during the beam collision electrons and photons are influenced by the field of the opposite electron beam , which leads to the following effects @xcite : @xmath30 in : conversion of photons into pairs ( coherent pair creation ) ; @xmath30 in : coherent pair creation ; beamstrahlung ; beam displacement .
beam collision effects in and , collisions are different . in particular , in collisions there are no beamstrahlung or beam instabilities which limit the horizontal beam size in collisions on the level 550 ( 350 ) nm for tesla ( nlc / jlc ) . therefore , it was of interest to study limitations of the luminosity at photon colliders due to beam collision effects .
the simulation , which include all collision effects , was done for the tesla beams and the horizontal size of the electron beams was varied .
the results are the following ( see graphs in @xcite ) : all curves for the luminosity follow their natural behavior ( as without collision effects ) : @xmath31 at least down to @xmath32 ( smaller values were not considered because too small horizontal sizes may introduce problems with the crab crossing scheme ) .
the luminosity ( in the high energy part ) can reach the value @xmath33 .
note that while in collisions @xmath34 , in collisions the attainable @xmath35 with the planned injector ( damping ring ) is about @xmath36 . in collisions
the luminosity at small @xmath35 is lower than follows from the geometric scaling due to beamstrahlung and displacement of the electron beam during the beam collision .
its maximum value ( high energy part ) is about ( 23)@xmath37 .
corresponding numbers and curves for nlc / jlc beam parameters can be found in @xcite .
so , we can conclude that for collisions at c.m.s .
energies below about one tev one can use beams with @xmath35 much smaller than in collisions . as a result , the luminosity in the high energy peak can be , in principle , several times higher than the luminosity .
contrary to collisions , where the luminosity is determined by unremovable beam - beam collision effects , at photon colliders the achievable luminosity is limited only by technical problems in obtaining beams with small transverse sizes .
attainable luminosity depends strongly on the emittances of the electron beams .
there are two methods of production low emittance electron beams : damping rings and low emittance rf photo guns ( without damping rings ) .
the second option is promising , but at the moment there are no such photo guns producing polarized electron beams .
polarization of electron beams is very desirable for photon colliders .
so , there is only one choice now damping rings . especially for a photon collider
the possibility of decreasing the beam emittances at the tesla damping ring has been studied @xcite and it was found that the horizontal emittance can be reduced down to @xmath40 m. the luminosity also depends on the @xmath41functions at the interaction points : @xmath42 .
the vertical @xmath43 is usually chosen close to the bunch length @xmath44 ( this is the design for collisions and can also be realized for collisions ) .
some questions remain about the minimum horizontal @xmath41function . for collisions ,
@xmath45 which is larger than the bunch length @xmath46 , because beams in collisions must be flat to reduce beamstrahlung . in collisions
, @xmath47 could be about @xmath48 ( or even somewhat smaller ) .
there are two fundamental limitations : the beam length and the oide effects ( radiation in final quads ) .
the latter is not important for the beam parameters considered .
there is also a certain problem with the angular spread of the synchrotron radiation emitted in the final quads .
but , for the photon collider the crab crossing scheme will be used and in this case there is sufficient clearance for the removal of the disrupted beams and synchrotron radiation .
studies of the existing scheme for the tesla final focus have shown @xcite that chromo geometric aberrations dominate at @xmath49 .
fortunately , a new scheme for the final focus system proposed at slac @xcite ( see also graphs in @xcite ) allows to obtain @xmath50 with small aberrations and further optimization is possible .
the parameters of the photon collider at tesla for @xmath51 200 , 500 and 800 gev are presented in table [ sumtable ] .
for comparison the luminosity at tesla is also included .
it is assumed that the electron beams have 85% longitudinal polarization and that the laser photons have 100% circular polarization .
the thickness of the laser target is one compton scattering length for @xmath51 500 and 800 gev and 1.35 scattering length for @xmath51 200 gev , so that @xmath52 0.4 and 0.55 , respectively ( @xmath53 is the @xmath54 conversion coefficient ) .
the laser wave length is 1.06 for all energies .
the distance between conversion and interaction points is @xmath55 for @xmath56 and 800 gev and @xmath57 for @xmath58 gev .
simulation results presented below include nonlinear effects in the compton scattering @xcite .
corresponding parameters @xmath59 for @xmath60 gev , respectively .
[ sumtable ] from table [ sumtable ] one can see that for the same energy @xmath61 [ ldist250e ] simultaneously with collisions there are also collisions with somewhat lower luminosity , so one can study both types of collisions simultaneously .
residual electron - electron luminosity is very small due to the beam repulsion .
the normalized luminosity spectra for @xmath62 gev and 800 , 200 gev are shown in fig .
[ ldist250 ] .
the luminosity spectra are decomposed in two parts : with the total helicity 0 and 2 .
[ ldist250 ] shows also luminosity spectra with additional cuts on the longitudinal momentum of the produced system , which suppress the low energy luminosity to a low level . in the case of two jets
one can restrict the longitudinal momentum using the acollinearity angle between jets ( @xmath63 , etc . ) . for the higgs
the production rate is proportional to @xmath64 at @xmath65 . for the case
considered , @xmath66 , and @xmath67 , @xmath68 note that the corresponding peak luminosity for nlc at the time of snowmass2001 was 7 times smaller@xcite and more recent revised number is 2.8 times smaller @xcite . in turn
, tesla also has large resources for further improvement of the luminosity
. it will be more clear after detailed optimization of the final focus system .
the normalized luminosity spectra for @xmath56 and parameters from table [ sumtable ] are shown in fig .
[ ldist250e]-left . for dedicated @xmath69 experiments one can convert only one electron beam ,
increase the distance between the conversion and the interaction points and obtain the luminosity spectrum with suppressed low energy part , fig .
[ ldist250e]-right .
the luminosity distributions , including their polarization characteristics , can be measured using processes @xmath70 ( @xmath71 , @xmath72 ) @xcite . for measurement of one can use the processes @xmath39e@xmath73e @xcite , @xmath74 @xcite and @xmath75 .
a key element of photon colliders is a powerful laser system which is used for the @xmath76 conversion .
the required parameters and possible schemes of lasers are discussed below .
there are two main parameters characterizing compton scattering : @xmath77 defined by eq.[eq1 ] and @xmath78 characterizing nonlinear effects in compton scattering . at @xmath79
the electron scatters on one laser photon , at @xmath79 on several . with grows of @xmath80
the spectrum becomes wider and is shifted to lower energies @xcite , therefore it is preferable to have @xmath81 .
however , decrease of @xmath80 in the conversion leads inevitably to the increase of the required laser flash energy .
for tesla project we assume @xmath82 for @xmath83 gev when change of the luminosity spectra is still acceptable .
another ( now `` good '' ) consequence of the nonlinear effects is the shift of the threshold for pair creation in collision of laser and high energy photons . instead of @xmath84 @xcite for the linear compton
it is @xmath85 . for the maximum tesla energy @xmath86 gev and laser wavelength 1.06 @xmath87 . for this energy
we assume @xmath88 and @xmath89 . as result ,
pair creation is practically absent and we can use the same laser in the whole tesla energy region . for optimization of conversion efficiency we have performed simulation asuming that the final focusing mirrors are situated outside the electron beams ( without holes in mirrors for electron beams ) .
the tilte of beams due to the crab crossing angle was taken into account .
the result of the simulation @xcite of the conversion probability for the electron bunch length @xmath90 , @xmath6 , @xmath91 as a function of the rayleigh length @xmath92 for various flash energies and values of the parameter @xmath80 can be found elsewhere @xcite . in order to obtain a conversion probability of @xmath93 at all tesla
energies a laser with the following parameters is required : [ cols= " < , < " , ] for nlc / jlc colliders the electron bunch is shorter by a factor of two .
calculations show that for similar assumptions ( laser optics is outside the electron beams ) the required flash energy is 3.3 j and it is about 1.75 j for the case head - on collisions ( mirrors with the holes ) which is considered for these projects . main difference for lasers between tesla and nlc projects is connected with structures of electron bunch trains , see fig.[trains ] . at superconducting tesla
the distance between bunches is 100 times larger which leads to different possible approaches for laser schemes .
all parameters for lasers are reasonable for exception of the repetition rate ( average power ) . to overcome the `` repetition rate '' problem it is quite natural to consider a laser system where one laser bunch is used for the @xmath94 conversion many times .
two ways of multiple use of one laser pulse are considered for the photon collider at tesla : an optical storage ring and an external optical cavity . in the first scheme @xcite , each laser bunch is used for the @xmath95 conversion about 12 times .
the laser pulse is sent to the interaction region where it is trapped in an optical storage ring .
this can be done using pockels cells ( p ) , thin film polarizers ( tfp ) and 1/4-wavelength plates ( @xmath96 ) .
the maximum number of cycles is determined by reflection coefficients of mirrors and attenuation in the pockels cell . in the second scheme @xcite
, an `` external '' optical cavity is used . using a train of low energy laser pulses
one can create in the `` external '' cavity ( with one mirror having small transmission ) an optical pulse with an energy higher than in a laser pulse by a factor q ( cavity quality factor ) and this pulse collides with electron bunches many times . as result
the required laser power can be lower than in the one - pass case by factor of 50100 .
recently , i.will @xcite has suggested to place the final focusing mirrors outside the detector and added an `` active '' mirror to the optical cavity which should compensate attenuation of the laser pulse .
initial 5 j pulse in the cavity is created using a train of specially prepared pulses from the external laser ( about 200 pulses , 0.05 j , 3 ps long , with frequencies matched to the cavity ) .
then this pulse is circulating in the optical ring cavity and losses of its intensity are compensated by an active mirror .
schematically this approach is illustrated in fig.[cavity ] .
the optical scheme includes also ( not shown ) movable mirrors , adaptive optics , fast feedback .
there are many questions to this scheme which need further study . for nlc / jlc
the train is short and the distance between electron bunches is only 2.8 ns therefore exploiting of the optical cavity is not effective .
small distance between electron bunches makes also a problem for combining pulses from several lasers to one train using pockels cells .
current solution is the following .
the laser system is based on about mercury diode pumped lasers ( @xmath11 ( yb : s - fap ) ) developed for the fusion program .
the laser produces 100 - 200 j , several ns long pulses with 10 hz rep .
each pulse is split using semitransparent mirrors into 96 pulses with 2.8 ns delays and which then are compressed down to 2 ps using grating pairs . in this way
one mercury pulse is sufficient for production of one @xmath97 j pulse train for nlc photon collider .
twelve such lasers working sequentially are needed to provide 120 hz rep .
the estimated cost of such laser `` plant '' according to the report at snowmass2001 is of the order of 200 m$ @xcite .
this approach does not look optimum because only one mercury laser of twelve is used in a given moment .
this means , for example , that one need 12 independent diode - pumping systems which are very costly .
it seems more reasonable to use about 5 special @xmath9830 j lasers which work in parallel with 120 hz rep .
each 30 j pulse is split in the same way as in @xcite to twenty 1.5 j pulses , ten with vertical and ten with horizontal direction of linear polarization .
subtrains with equal polarization from all lasers are combined using pockels cells and two long trains with perpendicular polarizations can be joined to one train using a thin film polarizer .
note that in this case the required switching time for pockels cells is 33 ns which looks possible .
the number of lasers and corresponding switching times can be optimized .
the optical system inside the detector for one nlc approach is shown in fig.[fgron ] @xcite .
it is one pass laser scheme .
the final focusing mirrors of 38 cm diameter are situated at the distance 4 m from the ip .
the mirror has a hole with 15 cm diameter for incoming beams and outgoing disrupted beams .
development of laser technologies is being driven by several large programs , such as inertial fusion .
this is a fortunate situation for photon colliders as we may benefit from the laser technology developments of the last @xmath99@xmath100 years which cost hundreds m$ per year . in the last decade
the technique of short pulse powerful solid state lasers made an impressive step and has reached petawatt ( @xmath101 ) power levels and few femtosecond durations .
obtaining few joule pulses of picosecond duration is not a problem using modern laser techniques .
the wave length of the most powerful lasers about 1 which is just optimum for the tesla photon collider .
photon colliders have one very specific background problem .
high energy photon beams are produced just inside the detector . after the @xmath94 conversion the electron beams have a very broad spectrum , @xmath102 , and an angular spread of about 10 mrad which is due to deflection of the low energy particles by the opposing electron beam . in the case of head - on beam collisions large fraction of the beam
would hit the final quadrupoles and produce enormous backgrounds in the detector .
this problem is solved by using crab - crossing scheme of beam collisions @xcite . the maximum disruption angle for tesla
photon collider is about 12 mrad @xcite . in the present design the crab - crossing angle in the photon collider ip is 34 mrad .
these values put restriction on possible quadrupole design .
other backgrounds are pairs and hadrons produced in beams collisions . consideration of these and some other backgrounds can be found elsewhere @xcite .
qed background at photon colliders are very similar to collsions ( or even smaller due to the large crab crossing angle and exit hole for disrupted beams .
strong longitudinal detector magnetic field confines low energy pairs near the collider axis and considerably reduce backgrounds in the vertex detector .
some problem presents a hadronic background .
the effective average rate of the reaction @xmath103 is about one event per bunch collisions .
it can lead to some worsening of the resolution for jets at low angles @xcite .
as the ecfa panel in europe and the snowmass study and the hepap panel in us have recommended the linear collider on the energy about 500 gev as the next large hep project ( asian physicists have also similar plans ) , it is very likely that in about one decade physicists will get a new very powerful instrument for study of matter : , , , collider .
the photon collider with , collisions is very natural and technically feasible option for linear colliders .
the physics potential of the photon collider is very rich and give information complementary to that in collisions .
the next step in development of the photon collider , designing the second interaction region and detector needs active participation of experimentalists , detector physicists .
2001 report on the nlc submitted to snowmass 2001 , fermilab - conf 01 - 075 , lbll-47935 , slac - r-571 ; linear collider physics resource book for snowmass 2001 , by american linear collider working group ( t. abe et al . ) , slac - r-570 , may 2001 .
|
photon colliders ( , ) are based on backward compton scattering of laser light off the high energy electrons in linear colliders .
all projects of linear colliders include this option . in this paper physics motivation , possible parameters and some interaction region aspects of photon colliders
are discussed .
pacs : 29.17.+w , 41.75.ht , 41.75.lx , 13.60.fz photon collider ; linear collider ; photon photon ; gamma gamma ; electron photon ; photon electron ; compton scattering ; backscattering ;
|
in the last few years the study of galaxy formation and evolution has seen tremendous progress .
instruments at many different wavelengths have begun to penetrate to the relevant redshifts .
one important prediction of models of galaxy formation and evolution is the nature of the radiation field produced by star formation .
one way to probe the resulting extragalactic background light ( ebl ) is to measure the attenuation through pair production of gamma rays from distant sources .
however , without a large sample of sources distributed across a wide redshift range , it is difficult to distinguish between extragalactic absorption and characteristics peculiar to individual sources . the large area telescope ( lat ) instrument on the gamma - ray large area space telescope ( glast ) will observe gamma rays with energies from 20 mev to @xmath1 gev .
the glast lat will be the first instrument able to probe the intergalactic radiation field by observing the absorption of gamma rays from a large number of extragalactic point sources as a function of redshift over a wide range .
ground - based telescopes can measure the attenuation of tev emission by intergalactic ir radiation ( stecker , dejager , & salamon 1992 ; macminn & primack 1996 ; madau & phinney 1996 ) .
however , these telescopes will measure the spectra of relatively small number of sources , making it more difficult to resolve the question of whether differences between sources are due to intergalactic attenuation or intrinsic peculiarities .
furthermore , the high pair production opacity of the ir radiation limits tev probes of the ebl to a narrow , low - redshift range .
glast , on the other hand , will observe thousands of sources , and will measure less drastic attenuation of gev photons by optical and uv radiation .
the energy range and capabilities of glast are thus ideal for probing the ebl to cosmological distances .
this paper reports our first modeling of the ability of glast to measure the extragalactic background light absorption . in order to do this , we need 1 ) models of the intergalactic radiation field , 2 ) the luminosity function of extragalactic gamma - ray sources , and 3 ) parameters of the instrument . in section 2.1
we briefly review models for the intergalactic radiation field and the resulting gamma - ray opacity as a function of redshift . in section 2.2
, we describe the two gamma - ray blazar luminosity functions used . in section 2.3
we describe the parameters used to simulate glast . in section 3 we discuss the simulation procedure , including the two different models of blazar input spectra and the two models for the intergalactic radiation field . in section 4
we present our results and conclusions .
gamma rays with @xmath2 gev traveling through intergalactic space will interact through pair production with the extragalactic background starlight ( ebl ) emitted by galaxies .
the total center - of - mass energy must be high enough to produce the electron - positron pair , and , for a wide range of ebl models , the attenuation becomes significant only above @xmath3 gev .
the cross section is maximized when the ebl photon energy @xmath4 ev , with @xmath5 in gev ( stecker , dejager & salamon 1992 ) . for 10 gev to tev gamma rays this corresponds to @xmath6 in the optical - uv range .
salamon & stecker ( 1998 ) calculated the opacity of high energy gamma - rays to redshift @xmath7 . to estimate the stellar emissivity and spectral energy distributions vs. redshift they adapted the analysis of fall , charlot , & pei ( 1996 ) , consistent with the canada - france redshift survey , and included corrections for metallicity evolution .
they found that the stellar emissivity peaks between @xmath8 and 2 before falling off , leading to a significant redshift - dependent absorption below @xmath7 .
other models , _
e.g. _ by primack et al .
( 1999 ) , provide for significant attenuation at even larger redshifts .
more recently , bernstein , freedman , & madore ( 2002a , b ) have made the first direct measurement of the optical - uv ebl integrated over redshift .
as shown in section 3 , our technique is a powerful discriminator among models , giving information about the era of galaxy formation and evolution .
the energetic gamma ray experiment telescope ( egret ) detected more than 60 blazar - type quasars ( mukherjee et al .
1997 ) emitting gamma rays with @xmath9 mev .
these sources are flat - spectrum radio - loud quasars ( fsrqs ) and bl lac objects , often exhibiting non - thermal radio continuum spectra , violent optical variability and/or high optical polarization .
they are also highly variable and powerful gamma - ray sources .
the egret blazars whose optical redshifts have been measured lie between @xmath10 and 2.28 .
the redshift distribution is consistent with the observed distribution of fsrqs , which extends up to @xmath11 .
however , since the luminosity function determines the statistical power of our technique versus redshift , and since this function is still relatively unconstrained , we use two different models for the blazar luminosity function .
the first model , by salamon & stecker ( 1996 ) , makes the assumption that blazars seen in gamma - rays above 100 mev are also seen in the radio as fsrqs .
this model assumes that the gamma and radio ray luminosity functions are linearly related as@xmath12 where @xmath13 is a parameter of the model and @xmath14 with @xmath15 .
the units of the comoving density @xmath16 are mpc@xmath17unit interval of @xmath18 and the units of @xmath19 are w hz@xmath20 sr@xmath20 . using cosmological parameters @xmath21 and @xmath22 km / s / mpc ,
the model is constrained to predict the number of blazars observed by egret .
the number of sources with redshift in the interval @xmath23 seen at earth with a flux for @xmath24 mev in the interval @xmath25 is given by ( salamon & stecker 1996)@xmath26 with @xmath27 , where @xmath28 is the hubble expansion rate . combining the choice of parameters given by salamon & stecker ( 1996 ) with a glast flux sensitivity of @xmath29 photons @xmath30s@xmath20
, a number of @xmath31 blazars is expected to be observed with redshifts up to @xmath32 .
the second model , by chiang & mukherjee ( 1998 ) , does not assume a correlation between luminosities at gamma ray energies and at other wavelengths .
this model parametrizes the luminosity function as @xmath33 with de - evolved luminosity @xmath34 and a maximum cutoff redshift of @xmath35 .
the energy range of this integrated luminosity is @xmath24 mev .
the best fit found for this broken power law is parametrized by @xmath36 , @xmath37 erg / s and @xmath38 , with cosmological parameters @xmath21 and @xmath39 km / s / mpc .
each model was separately fit in a self - consistent fashion to the egret data to produce the luminosity functions .
more recent cosmological data suggest a non - zero value for @xmath40 .
the impact on the luminosity function , however , is small ; we therefore retain the original model , along with the fit to the data , for our calculations .
of course , one of the important goals of the glast mission will be to constrain the luminosity function .
the spectra of the blazars observed by egret are well characterized in the @xmath24 mev range by power laws with an average photon spectral index of @xmath41 ( mukherjee et al .
the spectra of some individual blazars have a measured index significantly different from the mean value , suggesting true scatter in the distribution of blazar spectra , which our simulation takes into account as described below .
more importantly , most of the egret blazars have not been detected by tev telescopes ; for many of these sources , this implies a spectral break or rolloff at some energy between the egret and tev energy ranges .
intergalactic attenuation , the very effect explored in this paper , would account for the lack of detection of high - redshift objects , but there are relatively low - redshift blazars that are bright in the egret range and undetected in the tev range . more tellingly ,
most of the tev blazars belong to the same subset of blazars , the x - ray selected bl lac objects ( xbls ) .
since only a small fraction of the egret blazars are xbls , this implies that the non - xbl blazars may have spectra with intrinsic rolloffs independent of any intergalactic attenuation effects .
finally , blazars that have been detected in both the gev and tev ranges have tev fluxes that are lower than simple extrapolations of the egret power laws would suggest .
of course , such an extrapolation over such a wide range of energies is unreasonable .
most of the models for blazar spectra attribute both the gev and tev emission to the same inverse compton component of the emission .
however , with little observational data in the 30 - 300 gev range , no firm conclusions can be drawn about the precise shape of the spectra .
indeed , this is one of the motivations for the next generation of experiments .
our technique , as described in section [ sec-3 ] , is to form the ratio of the observed fluxes for @xmath42gev and @xmath43gev , @xmath44 this ratio is simple , robust , and insensitive to rolloffs above @xmath45 gev for most ebl models as shown in section 3.1 .
we attempt to bracket the range of possible spectra by first analyzing a sample of blazars whose power law spectral indices are normally distributed around a mean of -2.15 with standard deviation 0.04 , representing a situation where there is a range of spectral indices but no intrinsic rolloff in this energy range . to model intrinsic rolloffs
, we then repeat the analysis with a sample of blazars whose unredshifted spectra have a broken power law with mean index -2.15 below 50 gev and -3.15 above , again with a standard deviation of 0.04 in each case .
the gamma - ray large area space telescope ( glast ) is under development with a planned launch in 2006 ( michelson 2001 ) .
the large area telescope ( lat ) of glast will observe gamma rays with energies from 20 mev to @xmath1 gev .
glast will have a much larger effective area than egret , especially at higher energies ( peak effective area @xmath46 @xmath47 at @xmath48 gev ) , a larger field of view , and sub - arcminute scale source localization .
glast should be able to reach a @xmath49 point source flux sensitivity of less than @xmath29 photons @xmath30 s@xmath20 for @xmath9 mev within five years . as noted above
, using the distribution of blazars observed by egret and extrapolating to lower fluxes , it is estimated that glast will detect thousands of blazars .
improved angular resolution should allow a high percentage of optical identifications and redshift measurements , depending on the available ground - based resources .
improved high - energy performance should yield accurate flux determinations above 10 gev for many of these sources . note that our modeling is based on the generic parameters outlined in the glast science requirements document ( michelson 2001 ) ; the performance of the flight instrument may be substantially better .
to simulate the gamma - ray sources observable by glast , we need a reasonable extrapolation of the egret source distribution to the glast flux limit .
we used the two luminosity functions described in section [ sec-2.2.1 ] for this purpose , but our main conclusions do not depend significantly on this choice .
we note that any predictions made now will be supplanted by the data glast itself provides . before any observational selection , according to the luminosity function by salamon & stecker ( 1996 ) , @xmath50 blazars in principle will have fluxes in the range detectable by glast .
each one was assigned a random luminosity and redshift according to this model . with the luminosity function by chiang & mukherjee ( 1998 ) we generated 10,000 blazars , between redshifts 0 and 5 according to figure 6 of their paper . for both samples , the flux of each blazar
was then calculated according to @xmath51 where @xmath52 is the photon spectral index and @xmath53 is the cosmological luminosity distance @xmath54 $ ] . only blazars with observed flux greater than @xmath29 photons @xmath55 for @xmath56 are allowed in the sample .
the @xmath2 gev flux of each blazar was calculated by adding two effects .
first , each blazar was given a random , normally distributed spectral index , @xmath41 .
an index of -2.15 yields a flux ratio @xmath57 of @xmath58 .
also included was the redshift - dependent absorption above 10 gev .
the form of the dependence was parameterized from figure 6 of salamon & stecker ( 1998 ) , with metallicity corrections . in this ebl model ,
@xmath59 , @xmath60 , and the value of @xmath61 scales out .
we set the absorption for @xmath62 for this model equal to the absorption at @xmath7 , both because it is a conservative assumption and because it is physically plausible ( little stellar emissivity , and smaller scale and path lengths , for @xmath62 ) . to produce observed fluxes from these intrinsic fluxes ,
each blazar was assigned a random position on the sky and , assuming an exposure equivalent to two full years , galactic and extragalactic backgrounds were added .
the galactic backgrounds were derived from the diffuse model used in egret analysis ( hunter et al .
1997 ) . to take into account the extragalactic background
, we added a second , fixed , background component , with intensity @xmath63 photons @xmath30 s@xmath20 sr@xmath20 for @xmath9 mev and a power law index of -2.15 , under the assumption that glast may resolve a significant fraction of the egret isotropic background ( stecker & salamon 1999 ) .
this component represents the sum of the flux from unresolved blazars and any truly diffuse background contribution .
any blazar within @xmath64 of the galactic plane and any blazar whose observed flux was less than @xmath49 above the background flux at @xmath65 gev was removed from the sample , leaving @xmath66 blazars ( stecker & salamon 1996 ) or @xmath67 blazars ( chiang & mukherjee 1998 ) .
figure 1 shows a histogram of the number of blazars in each 0.5 redshift bin .
the model by chiang & mukherjee ( 1998 ) predicts a population of blazars that are intrinsically brighter when compared to the model by stecker & salamon ( 1996 ) . in that case
, glast would detect more blazars at higher redshift as can be observed from the graph .
we note that , with no ebl attenuation , for @xmath62 and requiring more than 5 detected photons @xmath68 , glast would see @xmath69 blazars using the stecker & salamon luminosity function , or @xmath70 blazars using the chiang & mukherjee luminosity function .
the integrated fluxes of each blazar for @xmath65 gev and @xmath2 gev were used to generate observed fluxes using poisson distributions equivalent to two full years of exposure . for each blazar , we calculated the ratio between these fluxes .
the error in each flux ratio was set to @xmath71 , where @xmath72 is the statistical error of the flux measurement in each energy range .
the crosses in figure [ fig2 ] show the weighted mean ratio in each redshift bin . to avoid
the bias of small number poisson statistics toward lower values , the flux ratio of each source was weighted by the poisson error of the @xmath65 gev flux , rather than the formal , propagated error of the flux ratio .
the diamonds show the same ratio when the intergalactic absorption is removed from the observed blazar fluxes . in all cases
the error bars are statistical , obtained by computing the rms scatter within each redshift bin and dividing by @xmath73 .
the analytically derived flux ratio using the opacity model of salamon & stecker is plotted as a solid curve . for comparison ,
the dashed lines in figure [ fig2 ] show the same results with no intergalactic absorption .
we repeated the entire analysis with the blazar spectra changed from single power laws with mean index -2.15 to broken power laws with mean index -2.15 below 50 gev ( at the source ) and -3.15 above .
the results are plotted as crosses in figure [ fig3 ] .
although fewer blazars have detected flux above 10 gev , the effects of absorption are still apparent .
note that sources with no detectable flux above 10 gev ( zero photons ) still provide important information ; indeed , neglecting them introduces a bias .
the modified @xmath74 statistic used here ( mighell 1999 ) accounts for these sources .
the ratio obtained without ebl absorption is presented as diamonds , along with the analytically derived flux ratio ( dashed line ) . as can be easily seen
, this flux ratio is not constant as a function of redshift .
this is a consequence of defining the break in the index for a given energy at the source .
primack and collaborators combined theoretical modeling with observational data to develop semi - analytic models of galaxy formation and evolution ( primack et al .
their models permit a physical treatment of the processes of galaxy formation and evolution in a cosmological framework , including gravitational collapse , mergers , etc . , rather than relying on pure luminosity evolution of the galaxies existing today .
we use their calculations of opacities to gamma rays at redshifts up to @xmath75 .
the cosmological parameters used are @xmath76 , @xmath77 , and @xmath78 km / s / mpc .
the luminosity functions use a different value for @xmath61 , but for our purposes this difference does not significantly affect the results ; as shown by blanch and martinez ( 2003 ) , the gamma - ray horizon has a relatively weak dependence on @xmath61 . note that these opacities , with their different cosmological parameter sets , should not be thought of as predictions , but rather as another set of reasonable values to illustrate the discriminating power of our technique .
the results are shown as triangles in figure [ fig2 ] and figure [ fig3 ] , along with the lines representing the analytical prediction .
the fact that the flux seems to be more highly attenuated is not important .
what is more interesting is that the decrease in flux ratio from @xmath79 to @xmath75 is observable .
this indicates , assuming the availability of gamma - ray sources and sufficient ebl density , that ebl absorption can effectively probe galaxy formation at those redshifts , a regime of intense theoretical interest .
more recently , oh ( 2001 ) performed an independent calculation of the opacity of gamma - ray blazar emission to pair production by uv photons as a function of redshift . while not addressing the detectability of high - redshift blazars by glast in detail , he obtains attenuation factors that vary strongly with redshift in a manner roughly consistent with the calculations we have used .
extragalactic attenuation of gamma - rays by low - energy background photons produces a distortion in the spectra of gamma - ray blazars that increases with increasing redshift . because we can not distinguish the difference between extragalactic attenuation and intrinsic effects in individual blazar spectra , statistical analysis of a large sample of blazars such as those presented in this paper is a powerful tool to study ebl absorption .
although agile , the next gev mission ( tavani et al . 2001 ) , will produce a significant increase in the total number of blazars and therefore refine the blazar luminosity function and evolution , glast will be the first mission to observe a large sample of high redshift blazars with sufficient statistics to separate intrinsic differences between blazars from redshift dependence of ebl absorption .
our results indicate that the redshift dependence of the attenuation should be easily detectable by glast even when the diffuse background is taken into account and possible high - energy intrinsic rolloffs are considered .
selection effects , both from glast itself and from optical coverage of redshift determinations , will primarily affect sources with low flux .
these sources will have poorly measured flux ratios , and will suffer from optical selection effects due to their more poorly determined positions .
other biases include the locations of optical telescopes , source clustering , and other effects
. it will be important to catalog these effects explicitly ; in particular , insuring adequate optical coverage may require active preparation and participation .
glast will be able to measure the differences in blazar attenuation in the cosmologically interesting range in redshift from @xmath8 up to @xmath75 .
this is in contrast to ground - based observations of tev attenuation by ir radiation , which will only be able to measure differences well below @xmath8 , where the ir becomes opaque .
as the energy threshold of the ground based experiments drops over time , their redshift range will increase , but will remain limited to low redshifts except for exceptional , statistically insignificant special cases , especially given their generally small fields of view . more than establishing that ebl attenuation occurs , glast will be able to distinguish between different ebl models .
this would validate ebl attenuation as a direct cosmological probe .
we emphasize that this analysis will require redshift determinations of a large fraction of glast blazars .
this is another example of the importance of cross - wavelength studies : by using optical measurements of blazar redshifts , gamma - ray measurements can uniquely probe the optical - uv ebl .
a redshift measurement for thousands of high - redshift sources is not a trivial undertaking , but the effort will be well rewarded .
even after observation of a redshift - dependent effect , the possibility would remain that the spectral evolution of gamma - ray blazars might coincidentally mimic redshift - dependent ebl absorption .
for example , if blazars that formed in the early universe suffered more internal attenuation than blazars that formed later , the same effect could be produced .
note that blazars are variable , and there are some indications that their spectra can become harder when they flare ( sreekumar et al .
evolution in flaring probability could produce the same effect as actual spectral evolution from a statistical standpoint ( for example , a higher percentage of high - redshift blazars might be observed in a quiescent phase ) , although one would expect the glast flux limit to produce a selection effect in the opposite direction . in any case ,
observation of a redshift - dependent spectral softening will provide an important constraint .
theorists will have to decide the likelihood of an evolutionary conspiracy .
we acknowledge useful conversations with bill atwood , who first suggested using the large statistics of glast agns to look for systematic effects of extragalactic background light attenuation with redshift .
we would also like to thank david thompson , seth digel , floyd stecker , and mike salamon for their useful comments , and james bullock and joel primack for providing calculations from their models .
we would like to thank the referees for many useful comments , in particular drawing our attention to the blanch & martinez article on cosmological effects in emissivity evolution .
bernstein , r.a . ,
freedman , w.l .
, & madore , b.f .
2002a , apj , 571 , 56 bernstein , r.a . , freedman , w.l . , &
madore , b.f .
2002b , apj , 571 , 107 blanch , o. , & martinez , m. 2001 , astro - ph/0107582 chiang , j. , & mukherjee , r. 1998 , apj , 496 , 752 chiang , j. , et al .
1997 , apj , 452 , 156 hunter , s.d .
1997 , apj , 481 , 205 macminn , d. & primack , j.r .
1996 , space sci .
75 , 413 madau , p. & phinney , e.s .
1996 , apj , 456 , 124 michelson , p. 2001 , in aip conf .
587 , @xmath80ray astrophysics , eds .
s. ritz , n. gehrels & c.r .
schrader , ( melville , n.y . :
aip ) , 713 , http://glast.gsfc.nasa.gov/srd mighell , k.j . 1999 ,
apj , 518 , 380 mukherjee , r. , et al .
1997 , apj , 490 , 116 oh , s.p .
2001 , apj , 553 , 25 primack , j.r . , bullock , j.s . ,
somerville , r.s . , macminn , d. 1999 , aph , 11 , 93 salamon , m.h .
, & stecker , f.w .
1998 , apj , 493 , 547 sreekumar , p. et al .
1996 , apj , 464 , 628 stecker , f.w .
, & salamon , m.h . 1999 ,
space sci .
261 , 327 stecker , f.w .
, & salamon , m.h .
1996 , apj , 464 , 600 stecker , f.w . ,
dejager , o.c . & salamon , m.h .
1992 , apj , 390 , l49 tavani , m. , et al .
2001 , in aip conf .
587 , @xmath80ray astrophysics , eds .
s. ritz , n. gehrels & c.r .
schrader , ( melville , n.y .
: aip ) , 729
|
gamma rays with energy above 10 gev interact with optical - uv photons resulting in pair production .
therefore , a large sample of high redshift sources of these gamma rays can be used to probe the extragalactic background starlight ( ebl ) by examining the redshift dependence of the attenuation of the flux above 10 gev .
glast , the next generation high - energy gamma - ray telescope , will have the unique capability to detect thousands of gamma - ray blazars to redshifts of at least @xmath0 , with sufficient angular resolution to allow identification of a large fraction of their optical counterparts . by combining established models of the gamma - ray blazar luminosity function , two different calculations of the high energy gamma - ray opacity due to ebl absorption , and the expected glast instrument performance to produce simulated fluxes and redshifts for the blazars that glast would detect ,
we demonstrate that these gamma - ray blazars have the potential to be a highly effective probe of the optical - uv ebl .
|
the statistical analysis of point processes is well understood , provided that the data is modeled by a homogeneous or inhomogeneous poisson point process .
if there are interactions among the points , the models become more complicated and so does the inference . in the case of many sets of data , using tests to see if probabilistically defined models fit is very difficult , both in terms of finding models to fit and tests which are sensible . in the absence of either , there is little that can be done using existing theory . taking a more experimental approach based on statistical learning has the potential to circumvent these difficulties . for the most part
, the non - poisson point processes will be assumed to be regular point processes , in which no two points can be any closer than some distance @xmath0 .
the methods developed here could be applied to any point process that defies modeling by a straightforward probabilistic model .
the methods discussed in this paper could be applied to four different types of problem in point process inference . 1
. in a pathology study , images of tissue are available in which the nuclei of cells have been stained .
the centres of the stained nuclei can be modeled by a regular point pattern , but the pattern is different in healthy tissue and in cancerous tissue .
can a classification rule be found which can identify single samples of tissue as being healthy or cancerous ? 2 .
a formal model for a non - poisson process is proposed .
it is mathematically intractable , but algorithms are proposed which allow realizations to be simulated after some simplifying assumptions are made .
there are two such algorithms , built using slightly different simplifying assumptions .
is there any evidence that the two algorithms are sampling from different point processes ?
two plants are constructed in different locations by different constructors to make a cement - fibre composite with fibres aligned in parallel .
cross sections are taken through specimens of the composite from both plants .
is there any evidence that the two processes are producing composite material with the same cross - sectional pattern to the fibres ?
if there is a difference , will it affect the specified properties of the composite ?
a model is proposed for a point pattern in the form of a simulation algorithm .
there is no probabilistic definition of the model , although the algorithm is well - defined .
an example of this would be a program which re - arranges a poisson realization into a realization where each point is at the centre of a packed sphere [ @xcite , @xcite ] .
does the model fit the data ?
the first of these examples is a standard problem in classification , but the other three would conventionally be seen as testing problems .
unfortunately , there are no classical tests which could be used in these examples .
if a pattern is modeled by a poisson process , then many tests are available to determine whether or not the process is poisson [ @xcite ] . none of these are directly of any use in the last three problems .
these tests are often based on a statistic or plot which can show signs of how the pattern is not poisson in one particular way .
some of these statistics may be of use in summarizing aspects of non - poisson point patterns .
formally expressed likelihood tests are not available for non - poisson processes , on account of the lack of a formal and tractable expression for the joint distribution of the point locations .
a test could be based upon the distribution of a descriptive statistic when the null hypothesis in the test was true .
this requires a statistic whose null hypothesis distribution can be found . in general , there are no results beyond finding what the mean of the statistic should be .
these results are few in number , and tend to be found in regular point processes with weak dependence among point locations within the realization .
the pair correlation of the matern hard - core point process can be explicitly found [ @xcite , @xcite ] .
this function could be used as a basis for a test based on estimates of the pair correlation function , provided that the variation of the estimates around the pair correlation function could be modeled .
simulation - based tests are of more use , yet still present difficulties
. if a density is available for configurations of points ( e.g. for strauss processes ) , then likelihood ratio tests can be undertaken , based on mcmc methods [ @xcite ] .
if no density is available , then a simulation test can be based on a statistic which is expected to detect important differences between the model and the data .
this statistic is calculated from many simulations of the null hypothesis model , and then compared with the value of the statistic found from the data [ @xcite ] .
if the model and the data truly do differ , then finding that difference depends on the choice of the right statistic . in the comparison of a poisson process with a baddeley - silverman cell process [ @xcite ]
, a comparison of @xmath1functions would reveal no difference unless a type i error occurred .
tests of this kind can be further complicated if the null hypothesis model requires that some parameters be estimated from the data .
assessing the fit of point process models to data using tests fails due to intractability or to the non - existence of useful probabilistic models .
this approach also encounters difficulties with finding a single statistic that can identify the differences that do exist .
often , there is no guidance at all beforehand as to which statistics will best identify differences . in all three of the testing problems , one is trying to determine if there is any evidence of difference between two point processes . in the absence of a formal model for one or both of the processes , any information about differences has to come from looking at joint distributions of every statistic that could be calculated from their realizations .
if the distributions of all possible finite collections of statistics are the same in both models , then one must conclude that the processes are indistinguishable .
this can be called the _ strong inference _
approach to model fitting , and is what physicists aspire to when they give up trying to show that a physical model ( such as the quantum mechanical model for the hydrogen atom ) does nt fit .
if differences are found , then it may be the case that these differences do not affect what the researcher is interested in .
for example , if electrical conductance only depends on few other descriptive statistics of a sphere packing , then a model could be used if the joint distribution of electrical conductance and those statistics were the same for both processes .
this approach would allow simple models to be used on very complicated materials ( such as concrete ) .
it represents the engineering approach to finding the simplest usable model that fits the data faithfully in a particular case , rather than seeking a comprehensive theory .
this approach can be called the _ weak inference _
approach to model fitting for a specific property ( electrical conductance , in this case ) , and it allows a lack of model identifiability to work in the scientist s favour . in almost no cases
is it possible to formally derive the distributions of the summarizing statistics , so it is necessary to estimate them from the data .
this requires that any inference be based on large independent and identically distributed samples of realizations from both processes being compared .
single realization inference is not possible in the absence of a model
. given two large samples from both processes , statistical learning can be used as an exploratory tool to find evidence of differences between the two processes .
the goal is to seek out statistics which best show the differences between the two sets of realizations , and then use these to build classification rules which distinguish single realizations from each process .
if a classification rule [ @xcite , @xcite ] can be constructed from a training set , this constitutes evidence that there could be a difference between the processes .
if the classification rule continues to perform well on test sets of new realizations , then a low misclassification rate would be further evidence in favour of a difference and against the possibility that the classification rule was fit to noise .
if the statistics required to identify a difference do not affect the distribution of a statistic of interest , then this could be taken to be evidence of a weak fit with respect to the statistic of interest . if no such classifier could be found after much effort , then this would suggest ( in a manner aspiring towards strong inference ) that the same process generates both sets of realizations . since there is no end to the number of statistics that could be tried
, establishing model fit by strong inference is impossible .
the main use of the idea is to show how difficult it is to establish that a model fits comprehensively .
all of these methods are based on the analysis of data , and so there is always the risk that the samples of realizations will be unrepresentative , and that the differences found will be due to chance variation .
the use of test sets and classifiers based on large samples of realizations reduces the chances of this happening , but can not eliminate it . insisting that the differences be so large as to give almost no misclassifications on large test sets
further reduces the chance of this happening , although this would overlook differences in cases where the statistics found could only identify a minor but real difference in the distribution of statistics . using summarization and classification to build classification rules to distinguish between processes known to be different
is also a good way to develop new and more powerful summarizing statistics .
if there is a model as elegant and useful as the gibbs or poisson models for describing sphere packings , it may be based on statistics which have not yet been found .
experimental investigation may prove to be a more useful way of finding such models , rather than attempting to formally advance existing theory . in classification problems , this aggressive approach to finding differences using new statistics
may be more effective than trying to build classification rules based on the classical statistics of point processes . to make this approach to fitting work ,
more is needed than just access to effective classification rules .
there must also be a very large library of descriptive statistics which can be applied to point patterns , many more than are currently in use now .
distinguishing realizations of non - poisson processes requires new statistics , which will need to be invented or borrowed from mathematical models used in other disciplines .
most descriptive statistics for point patterns were developed in the context of inference for poisson processes .
the intensity , estimated for a homogeneous poisson process as the number of points per unit area , describes the first moment properties of the underlying random measure .
the second moment properties can not be expressed by a single statistic , but can be described by the @xmath1 function , the pair correlation function , and the nearest neighbour function .
the last is related to the palm distribution for the process , as is the empty space function .
third moment - based statistics can also be defined [ @xcite ] .
all standard estimators of these statistics are based upon the assumption that the point process is stationary .
when the realizations come from a point process which is not stationary , then these statistics can still be used descriptively , as long as they are applied to each realization in a consistent fashion .
a realization of a point process can be transformed into a realization of a random set [ @xcite ] known as a germ grain model by attaching spheres at each point . if the process is regular with all points separated by distance @xmath2 , then using discs of radius @xmath2 would be useful , although other radii could be used .
the most basic random set statistics are estimates of the @xmath3 moments of the realization , which are defined by @xmath4\ ] ] where @xmath5 is the union of all of the spheres and @xmath6 .
when the point pattern is stationary , the reduced moments can be defined as @xmath7 \label{m1 } \\
m_2(r ) & = \text{pr}[0 , r \in \phi ] \label{m2 } \\
m_3(r , s ) & = \text{pr}[0 , r , s \in \phi ] , \label{m3}\end{aligned}\ ] ] where @xmath8 is the origin in @xmath9 .
given a window of observation @xmath10 onto a realization of a stationary point process , the first moment can be estimated by finding the proportion of the area of a covered by spheres .
higher moments such as the third can be estimated by superimposing translates of the realization by @xmath11 and @xmath12 onto the original and finding the proportion of the window intersection occupied by points where three spheres intersect .
the third moment can be shown to be a smoothed function of the three point analogue of the pair correlation function [ @xcite ] .
since estimates of the third moment involve information used in estimating the second moment , some form of moment decomposition such as the third cumulant @xmath13 should be estimated in order to emphasize those aspects of the third moment that are dependent on the use of all three defining points . given a realization of a point process
, a voronoi triangulation and its dual tessellation [ @xcite ] can be constructed .
statistics based on tessellations and triangulations are often used in the study of point processes arising from sphere packings [ @xcite , @xcite , @xcite ] . in the case of regular point processes
, the tessellation will provide a skeleton for the void structure around the points .
in general point patterns , the near neighbours of a given point can be defined as all points linked to the given point by triangulation edges .
the triangulation is a slightly more useful source of statistics , since each cell has a fixed shape .
given a triangulation or a tessellation , various statistics can be calculated directly .
these include means , standard deviations , maxima , and minima of cell volume , cell surface area , cell perimeter , longest side , and largest angle . for tessellations , the mean number of sides is also useful .
other aspects of tessellations can also be used as statistics .
given any listing of the @xmath14 points in a realization , the adjacency matrix @xmath10 is a sparse @xmath15 matrix whose non - zero - entries @xmath16 iff points @xmath17 and @xmath18 are connected by an edge of the triangulation .
the three largest eigenvalues of this matrix , as well as the coefficients of its characteristic polynomial can be used as descriptive statistics .
the coefficients of the characteristic polynomial include the number of triangles in the triangulation [ @xcite ] , which can be used as a statistic when the counts from all realizations are adjusted for any scale or window size differences .
germ - grain models based on regular point processes can be used as models for composite materials [ @xcite ] , for the liquid state [ @xcite ] , and for granular materials [ @xcite ] .
physicists generally want to evaluate some bulk property on each individual realization , such as viscosity ( in liquids ) , bulk electrical conductance ( in composites ) , or yield stress ( in granular materials ) .
the mathematical models and algorithms for finding these properties from realizations can be used out of their physical context in order to provide statistics [ @xcite ] .
simplifications of the structure of a realization using tessellations and triangulations are often needed to make these statistics calculable .
summarizing statistics can be divided into two classes , based on how they extract information from a realization of point process . in both cases
, some form of measurement is taken at many different locations in the observation .
a [ _ smoothly defined _ ] statistic reduces the list of measurements to a single number through use of some function that is invariant to the ordering of the measurements .
one example is the volume fraction of a realization , which can be estimated as the average of a set of indicator function values on a grid of points within the realization .
all moment estimates , all histograms of measurements of this kind , and generally all statistics based on the order statistic of the measurements are smoothly defined .
a _ structurally defined _
statistic is one whose treatment of the measurements is determined by large scale random structures within the realization .
if a triangulation is assumed to be electrically conducting with edge resistances proportional to their length , then the bulk resistance of that circuit is proportional to a weighted average of lengths of loop - free paths between the electrodes in the circuit .
bulk resistance can be used to define a measure of anisotropy for sphere packings [ @xcite ] .
the weights are determined by a physical model for electrical conduction , but the paths are determined by the overall structure of the entire realization . smoothly defined statistics make up most of the statistics currently used to summarize point patterns and random sets .
they are of great use for describing poisson process realizations , but they are difficult to interpret when applied to regular point processes . structurally averaging statistics are expected to be of great use in the study of regular point processes , since their form is determined by the very disorder whose detail is smoothed away by the smoothly defined statistics . to build a classification rule , a large collection of labeled realizations from both processes is needed for use as a training set . if the right summarizing statistics have been found and the realizations are sufficiently large , then the vectors of @xmath19 summarizing statistics for each element in the training set should form two distinct clouds of points in @xmath20 .
the goal of classification is to find a hyperplane or hypersurface which separates the two clouds of points . ideally , the hypersurface will completely separate the two clouds of points .
it is impossible to clearly visualize the two clouds of points that the classification rule separates if the number of statistics exceeds 3 .
theory can provide some idea of the shape of the clouds , provided that the statistics involved are smoothly defined averages . then , even in the case of a point pattern arising from a sphere packing , there are results which suggest that the between - realization distribution of those statistics will be normal [ @xcite ] .
this suggests that in those circumstances , the clouds would roughly ellipsoidal and that linear or quadratic discrimination should be effective .
if some of the statistics are not smoothly defined averages and are instead based on extrema , then the limit laws may be quite different and the shape of the point clouds would be unclear . in these circumstances , non - linear methods such as @xmath21nearest - neighbours , cart , or boosted or bagged cart
would be required to separate the two clouds .
classification rules were developed to separate realizations of point processes in which the processes were known to be different .
three pairs of processes were considered . in each pair ,
one process was a hard - core strauss process [ @xcite , @xcite , @xcite ] chosen to have an intensity close to that of the other process in the pair .
the processes compared to the strauss processes were the dead leaves ( dl ) process [ @xcite ] , the simple sequential inhibition ( ssi ) process [ @xcite ] , and the diggle gratton ( dg ) process [ @xcite ] with parameters @xmath22 , @xmath23 , and @xmath24 .
all processes were simulated on a 44 by 44 unit window with the central 39 by 39 unit portion used as the realization .
no points were allowed to be closer than 2 units . in the diggle - gratton process
, there was an additional weak exclusion effect in a region beyond the unit hard - core exclusion zone around each point . when discs of unit radius are attached to each point , the mean estimated coverage fractions are 0.24 for the dead leaves model , 0.39 for the ssi model , and 0.51 for the diggle - gratton model .
the nearest neighbour functions @xmath25 were estimated for each realization in spatstat [ @xcite ] and the distances from all training set realizations were combined .
the model @xmath26 where @xmath27 and @xmath28 was found to fit these distances well , and so the model was fit to each realization .
the fit was quadratic without intercept , and the quadratic term used was the second - order legendre polynomial .
the estimates of @xmath29 , @xmath30 , and @xmath31 for each realization were used as descriptive statistics .
the remaining summarizing statistics were found from the delaunay triangulation of the points , as determined by spatstat .
once the triangulation had been cleaned of artificial vertices and edges , the number of triangles was determined .
the longest side , largest angle , and area were found for each triangle , and the maxima of each were determined for each realization .
the three largest eigenvalues of the adjacency matrix derived from the triangulation were also used as statistics . in all cases ,
linear discriminant analysis produced the lowest misclassification rates .
logistic discrimination performed about as well , but cart and 5-nearest - neighbours produced rules which often did no better than guessing .
this was attributed to the relatively small size of the training set .
the example shows that summarization and classification is useful way to build classifiers .
it does depend very heavily on having powerful summarizing statistics , which were available only in the diggle - gratton case .
in the other cases , further experimentation is needed to see if an effective rule can be found based on larger realizations , or if more powerful statistics are needed .
the classification rule in the diggle - gratton case was not based on many more nearest - neighbour distances than was the rule for the ssi case .
extra classification power may have been derived from the diggle - gratton case data having an intensity closer to that of sphere packings , which can be shown to be less disordered than other regular point processes [ @xcite ] . as with all conclusions that are based solely on data
, their strength can only be determined through analysis of more realizations . if the investigator is clever enough to find the right summarizing statistics and has enough data , then very powerful classifiers can be built .
while the statistics are different than those used by the eye and mind of a trained pathologist ( in the case of pathological specimens ) , the fact that a classification rule can be built which can correctly classify 200 realizations of the kinds seen in fig . [ discern ] suggests that a rule built by summarization and classification could do at least as well . c. geyer .
likelihood inference for spatial point processes . in o. barndorff - nielsen , w. kendall , and m.n.m .
van lieshout , editors , _ stochastic geometry : likelihood and computation_. chapman and hall , 1999 .
|
fitting models for non - poisson point processes is complicated by the lack of tractable models for much of the data . by using large samples of independent and identically distributed realizations and statistical learning ,
it is possible to identify absence of fit through finding a classification rule that can efficiently identify single realizations of each type .
the method requires a much wider range of descriptive statistics than are currently in use , and a new concept of model fitting which is derive from how physical laws are judged to fit data .
|
there are two ways in which bosons can condense : via bose - einstein condensation or via spontaneous symmetry breaking .
bose - einstein condensation occurs due to a conserved current associated with the bosons . since the total number of particles ( or other conserved quantity )
can then be fixed , at low temperatures and high densities energy levels become overcrowded . " as a result the charge of the system must be stored the zero - momentum ground state . in theories with spontaneous symmetry breaking , a conserved current is not responsible for the condensation . instead , the interactions of the bosons make it energetically favorable for them to condense . in this paper
we consider the first mechanism , bose - einstein condensation , in the case that there is both a global @xmath0 symmetry corresponding to a conserved scalar number density as well as a local gauge symmetry .
in particular , we consider scalar quantum electrodynamics in which the conserved scalar current allows us to introduce a nonzero chemical potential for the scalars . to ensure the neutrality of the system we consider scalars that reside in a constant background charge density of the opposite charge .
in this work we treat this density as a classical , external background . to study the phase transitions of the system , we calculate the grand canonical partition function at the one - loop level . we find the critical temperature @xmath1 below which the ground state of the system is macroscopically occupied .
below this temperature some fraction of the scalar charge must reside in a condensate .
we find also an implicit expression for the critical temperature @xmath2 above which all of the scalar charge is stored in thermally excited states . in between these two temperatures both condensed and uncondensed solutions exist .
for the usual neutral bose - einstein condensate these two temperatures coincide . at the critical temperature @xmath3 the expectation value of the scalars @xmath4 which is the order parameter of the phase transition goes continuously from zero to nonzero as the temperature of the system
is lowered .
this is characteristic of a second order phase transition . for the charged condensate we find markedly different behavior depending on the relative values of the number density @xmath5 and the mass @xmath6 . for low densities , when @xmath7 , we find evidence of a first order phase transition : in between the critical temperatures @xmath1 and @xmath2 the value of @xmath4 jumps discontinuously from zero .
the lower critical temperature @xmath1 coincides with the critical temperature @xmath3 for the neutral condensate , while the upper critical temperature @xmath2 can be several times larger than @xmath3 in the low density regime .
this implies that for the charged condensate , condensation can occur at higher temperatures than for the neutral condensate . for high densities , when @xmath8 , then @xmath2 approaches @xmath1 and the transition appears to be quite nearly second order . the latter behavior can be understood through simple arguments .
when the average kinetic energy of a boson is much greater than its coloumb energy , the gas of bosons can be treated as nearly ideal . for this to be the case near the transition temperature
, it is necessary that @xmath8 .
thus the physics of the charged condensate in the high density regime should be well approximated by the ideal , neutral condensate .
the main interest of this paper is the low density regime , when @xmath7 . in this regime
the mass of the gauge boson in the condensate @xmath9 becomes comparable with or greater than the transition temperature @xmath3 and a first order phase transition is possible .
such a system of charged scalars has been considered recently in the context of helium - core white dwarf stars @xcite . in these works
helium-4 nuclei play the role of the scalars while a degenerate electron gas is the background charge density .
the number density of the nuclei @xmath10 is such that the average interparticle separation falls between the nuclear and atomic scales .
thus the nuclei are unable to form neutral atoms , while nuclear effects are negligible .
it is also the case that the number density satisfies @xmath11 , as @xmath12 .
while standard theory holds that white dwarfs with carbon or oxygen cores will crystalize as they cool , it was argued in refs .
@xcite that for helium - core white dwarfs quantum effects will become significant before the crystallization point is reached . thus instead of crystallizing , the helium nuclei can condense . in these works the condensation of helium-4 nuclei was described using a low energy effective field theory .
the work of this paper is relevant to the helium - core white dwarfs
if we take scalar qed to be a relativistic effective field theory .
the relativistic theory is not the most appropriate description of the helium-4 condensate .
it is overly restrictive in that it enforces lorentz invariance - a symmetry we do not expect the low energy system of helium-4 nuclei and electrons to preserve .
in addition it contains a heavy mode which we would expect to be beyond the scope of a low energy theory .
however , it captures the significant features of the condensate including equivalent dispersion relations for the relevant degrees of freedom .
it was recently argued in ref .
@xcite that , in a neutral system of deuterons and electrons , at sufficiently high densities and low temperatures the spin-1 deuterons will condense .
such a system may be relevant to certain low - mass brown dwarf stars in which it is expected that such matter can exist in localized regions .
in addition , because deuterons can condense at lower densities than helium , it may be possible to produce such high - density deuteron matter in laboratories in future shock wave compression experiments .
the methods of this paper should be applicable to studying the phase transitions of these systems as well .
let us make a few comments on the literature .
gauge theories at finite temperature and symmetry breaking in gauge theories at finite temperature have been studied extensively , early works being @xcite ( see also @xcite and references therein ) . in ref .
@xcite spontaneous symmetry breaking in the abelian higgs model at finite temperature was studied .
it was found that a first order phase transition could take place if @xmath13 where @xmath14 is the coupling of the @xmath15 interaction and @xmath16 is the charge of the scalar . for @xmath17
the phase transition is second order .
the effect of a nonzero background fermionic current density at zero temperature was also considered in this work .
it was found that an increase of the background charge density could lead to an increase in symmetry breaking . in ref .
@xcite the abelian higgs model with a background fermionic current density at finite temperature was studied in the limit that @xmath18 .
the same model was also considered in ref .
@xcite using a somewhat different method : instead of introducing a background fermionic current , a chemical potential associated with the conserved scalar current was introduced .
the @xmath18 limit was taken here as well .
taking this limit allows one to ignore terms that mix the scalar and gauge bosonic fields . in our work
we assume that the @xmath15 interaction is negligible compared to the electrostatic interactions .
thus this mixing term plays a significant role , changing the spectrum of the theory in the condensed phase .
moreover , the emphasis of these two works is the effect of a conserved current on spontaneous symmetry breaking , rather than the bose - einstein condensation of the scalars due to the conserved current .
the latter is the focus of this work .
the effects of nonzero chemical potentials associated with conserved charges on symmetry breaking at finite temperature were also considered in ref .
@xcite . in ref .
@xcite an exact expression for the one - loop effective potential for @xmath15 theory with a global @xmath0 charge at finite temperature was obtained .
@xcite consider the screening of electric charge and photon polarization in scalar qed at finite temperature and density .
phase transitions in scalar qed with a nonzero chemical potential at zero temperature due to external magnetic fields were considered in @xcite .
the organization of this work is as follows . in the following section we briefly review the condensation of charged scalars at zero temperature .
a more thorough discussion can be found in the recent review @xcite . in section 3
we compute the grand canonical partition function at the one - loop level .
we also treat the uv divergences of the theory . in section 4
we find the critical temperature @xmath1 below which the scalars have condensed and we find an implicit expression for @xmath2 above which all of the scalars are in thermally excited states .
we plot the expectation value of the scalars @xmath4 as a function of temperature for various values of the mass @xmath6 and number density @xmath5 of the scalars .
we discuss the evidence for a first order phase transition when @xmath19 . in the concluding section
we make some comments on the validity of the perturbative expansion and the relevance of our findings to helium - core white dwarf stars .
we start by considering the lagrangian of a charged scalar field @xmath20 of mass @xmath6 and a gauge field @xmath21 at zero temperature . the scalars carry charge @xmath16 which we take to be some multiple of the electric charge @xmath22 .
we include a background external charge density @xmath23 which has a charge opposite that of the scalars : [ lagr0 ] = - f_^2 + d _ ^2 - m^2 ^ -g a^ j _ .
the covariant derivative for the scalars is defined as @xmath24 .
this lagrangian can arise from a gauge invariant theory in which fermions with charge @xmath25 play the role of the background charge density : = - f_^2 + d _ ^2 - m^2 ^ + | ( i ^d_-m_f ) . in this work
we consider such a theory in the limit that the fermions are non - dynamical ( i.e. , frozen in by hand " or by some other dynamics ) . at distance scales that are greater than the average separation between the fermions ,
their spatial distribution can be assumed to be uniform .
then , the background charge density can be approximated as @xmath26 . due to the global @xmath0 symmetry
, there is a conserved scalar current : j_^s -i [ ( d_)^ * -^ * ( d _ ) ] .
we can associate a chemical potential @xmath27 with the conserved scalar current . to incorporate the presence of a conserved current into the hamiltonian density we make the usual shift @xmath28 .
for the lagrangian density the equivalent shift is [ lagr ]
= - f_^2 + ( d _ - i _ 0 ) ^2 - m^2 ^ -g a^ j _ . for convenience
we switch notation and write the scalar field @xmath20 in terms of a modulus and a phase @xmath29 . in terms of these variables the lagrangian density becomes [ lagr1 ] =- f_^2 + ( _ ) ^2 + ( ga_-_+_0)^2 ^2- m^2 ^2 - g a^ j _ . written in this form , it is evident that the chemical potential acts as a tachyonic mass for the scalars . in these variables the scalar number density is given by [ j ] j_0^s = ( ga_0-_0 + ) ^2 .
varying the lagrangian with respect to @xmath21 gives the following equation of motion : [ eom1 ] -^f _ = g(ga_-_+_0 ) ^2-gj _ .
the system is electrically neutral when the scalar charge density is equal in magnitude to the background charge density : g j_0^s = g j_0 .
varying the lagrangian with respect to @xmath30 gives : [ eom2 ] = [ ( ga_-_+_0)^2-m^2 ] , while varying with respect to @xmath31 gives the conservation equation for the scalar current : [ j0scalar ] ^j_^s = ^= 0 .
let us take @xmath32 so that the number density of scalars is constant in time : @xmath33 .
subject to this constraint , the equation of motion for the scalars ( [ eom2 ] ) becomes [ v ] = -m^2 .
this system has a constant , static solution = .
the nonzero expectation value for @xmath30 indicates that the scalars are in the condensed phase .
it follows from ( [ j ] ) that in the condensed phase @xmath34 . for an electrically neutral system in which the gauge - independent quantity @xmath35 is equal to zero
, condensation occurs when @xmath36 .
thus at zero temperature we have a neutral system in which the charged scalars are condensed into a zero - momentum , macroscopic state . for further discussion of the zero temperature condensate
see @xcite .
in order to study the properties of the charged condensate at finite temperature we start by computing the grand canonical partition function @xmath37 .
we use the functional integral representation of the partition function as it is most suited to the field - theoretic approach adopted above : [ z ] z = n ( ) ( f ) . here
@xmath38 is an irrelevant normalization constant , @xmath39 is the inverse temperature and @xmath40 is the imaginary time .
we have introduced a gauge fixing condition @xmath41 in order to evaluate the functional integral over the gauge fields .
the action is defined by s - _
0^dd^3x l , where the lagrangian density @xmath42 is given by ( [ lagr ] ) and contains the scalar chemical potential as well as the external current @xmath43 .
we also make the replacement @xmath44 .
let us decompose @xmath20 into real and imaginary parts : = ( _ 1+i_2 ) .
the functional integral over the fields is constrained to be periodic so that @xmath45 . given this constraint , we can fourier expand the fields as follows : [ p1 ] _
1(x ) = v+ ^_n=- _ ^i ( _ n + ) _ 1;n ( ) , + [ p2 ] _
2(x ) = ^_n=- _ ^i ( _ n + ) _ 2;n ( ) , + [ a ] a_(x ) = ^_n=- _ ^i ( _ n + ) a_;n ( ) , where @xmath46 and @xmath47 is the volume of the system . for the field @xmath48 we have separated out a constant part @xmath4 which is independent of @xmath49 and @xmath50 so that @xmath51 .
thus @xmath4 represents the thermal average of the field : @xmath52 .
we do this in anticipation of the condensation of the scalars into the @xmath53 , @xmath54 state .
the partition function will be at a minimum with respect to the free parameter @xmath4 : = 0 . without loss of generality we can set @xmath55 . one could also separate out an expectation value for the scalar potential : @xmath56 .
the partition function will also be at a minimum with respect to @xmath57 : [ za ] = 0 .
it can be seen from the lagrangian ( [ lagr1 ] ) that a nonzero expectation value for the scalar potential acts as a shift in the chemical potential , except in its coupling to the external current @xmath58 .
thus equation ( [ za ] ) is equivalent to [ zmua ] g -g j_0=0 .
the first term on the right hand side is the scalar charge density : [ zmu ] = j_0^s .
thus minimization of @xmath59 with respect to @xmath57 gives the condition of charge neutrality : @xmath60 .
equation ( [ zmua ] ) depends only on the total effective chemical potential , i.e. , on the sum @xmath61 and not on @xmath57 alone .
thus minimization does not fix the value of @xmath62 independently of @xmath27 . instead ,
equation ( [ zmu ] ) can be used to determine @xmath63 , given a fixed charge density .
one could consider a configuration with a nonzero @xmath57 acting as an external chemical potential for the scalars , corresponding to some uncompensated charge on a surface at infinity .
such a setup was studied in earlier works @xcite in the zero - temperature limit . for simplicity
, we will assume a configuration with no surface charge and take @xmath64 in what follows . in the following expressions .
the results are otherwise unaffected . ]
terms in the lagrangian that are linear in the excitations of the fields above their background values will contribute to the action an amount proportional to @xmath65 or @xmath66 for the scalar and gauge fields respectively .
thus given the above definitions , these linear terms in the lagrangian will not contribute to the partition function .
terms that are linear in @xmath67 do not appear in the lagrangian as long as the external current @xmath68 is zero . in order to compute the functional integral
we must first choose a gauge @xmath41 .
the most natural choice is the unitary gauge in which the phase of the scalar field is set to zero : @xmath69 . in this gauge
the physical content of the theory is apparent and no unphysical degrees of freedom need to be introduced .
however , the unitary gauge is known to give incorrect results at the one - loop level .
this is attributed to the fact that lagrangian in the unitary gauge does not correspond to a renormalizable theory @xcite .
we choose instead a renormalizable @xmath70 gauge .
a family of covariant gauges is given by f = _
-f(x)=0 , for some arbitrary function @xmath71 . with
this choice the partition function ( [ z ] ) becomes [ z2 ] z = n ( + g^2v^2 ) ( _ a^+gv_2-f(x ) ) ^-s .
the delta function in the partition function can be incorporated into the lagrangian via [ lagr2 ] l -(_a^+gv_2)^2 . in what follows
we will keep the gauge parameter @xmath72 general in order to check that our results are independent of the gauge fixing condition . to compute the partition function
we proceed via the mean field approximation .
we take the field expansions ( [ p1 ] ) , ( [ p2 ] ) and ( [ a ] ) and substitute them into the lagrangian ( [ lagr2 ] ) .
we expand the lagrangian to second order in the fields , which we assume to be small fluctuations above the mean field values .
we neglect the terms that are linear in the fields , as per the discussion above . for convenience
we split the gauge boson into transverse and longitudinal components , the transverse components being given by a_j^tr = a_j- ( _ k a_k ) . upon integrating , the action ( to second order )
can be written as a sum of the tree - level component , the contribution coming from the transverse photons and a contribution coming from the scalars and remaining photon degrees of freedom : s_2 = s_0+s_tr+s_s .
the tree - level action is given by s_0= v ( m^2-^2 ) v^2 .
the contribution to the action coming from the transverse components of the gauge boson is s_tr = ^2 _
i=1,2 a_i ;-
n^tr(- ) ( _ n^2+^2 + g^2 v^2 ) a_i;n^tr ( ) , + where the sum over @xmath73 is the sum over both transverse degrees of freedom .
these components acquire a mass when the scalar is condensed , i.e. , when @xmath74 .
the remaining scalar and photon degrees of freedom mix with each other .
their contribution to the action is given by s_s= ^2 _ n _ ( _ 1;-n(-),_2;-n(-),|a_0;-n(-),a_l ;- n(- ) ) d ( c_1;n ( ) + _ 2;n ( ) + a_l ; n ( ) ) , + where @xmath75 [ d ] ( cccc _ n^2+^2+m^2-^2 & -2 _ n & -2igv & 0 + + 2 _ n & _ n^2+^ 2 + m^2-^2 + g^2 v^2 & + igv_n ( 1- ) & + igv|| ( 1- ) + -2igv & -igv_n ( 1- ) & _ n^2+^2 + g^2 v^2 & -_n || ( 1- ) + 0 &
-igv|| ( 1- ) & -_n || ( 1- ) & _ n^2 + ^2 + g^2 v^2 ) . +
let us make a few comments on this matrix . in the decoupling limit @xmath76 the scalar degrees of freedom decouple from the photon degrees of freedom as expected . in this limit
we recover the theory for the neutral condensate together with decoupled , massless photons . when @xmath77 but the scalar field is not condensed , i.e. , when @xmath78 , the same decoupling takes place . in the uncondensed phase
the partition function for the charged scalars is the same as that for the neutral scalars and decoupled photons , at the one - loop level .
an important element of this matrix is the term that mixes the scalar field @xmath48 with the @xmath79-component of the gauge field : @xmath80 . in the absence of a chemical potential for the scalars ( i.e. , when @xmath81 ) , an appropriate choice of gauge will decouple the gauge degrees of freedom from the scalars , even when @xmath4 is nonzero . when @xmath82 this decoupling no longer occurs , because of this term . due to this mixing ,
the spectrum of the theory in the condensed phase is significantly different from that of the neutral condensate . carrying out the functional integrations ,
the logarithm of the partition function ( [ z2 ] ) can be expressed as a sum of the various components = _ 0 + _
s+ _ gh . there is also a constant term coming from the overall normalization which is temperature independent and which we can neglect .
the first term comes from the tree - level action _ 0 = - v ( m^2-^2 )
v^2 , the second term from the two transverse photon polarizations _ tr= -(2 ) ^2 _ n _ ( _ n^2+^2 + g^2 v^2 ) , and the third term from the scalar and remaining photon degrees of freedom [ zs ] _
s=- ^2_n _ ( d ) .
the last term arises from the ghost determinant in the partition function ( [ z2 ] ) .
_ gh=^2 _ n _ ( _ n^2+^2 + g^2 v^2 ) .
formally , the ghost term exactly cancels the contribution coming from the transverse degrees of freedom .
this allows us to simplify our calculations .
the determinant of @xmath83 can be written in the following form [ det ] d = ( _ n^2 + _
+ ^2)(_n^2 + _ -^2)(_n^2 + _ 1 ^ 2)(_n^2 + _ 2 ^ 2 ) , where the dispersion relations @xmath84 , @xmath85 , @xmath86 and @xmath87 are functions of @xmath88 and @xmath72 .
we will denote them by @xmath89 . due to their length
we will not write the full expressions here .
below we will give their exact expressions when evaluated on the solutions to the tree - level equations of motion for the scalar .
we will also find expressions for when the tree - level scalar equations of motion are not satisfied .
this form for @xmath90 allows us to carry out the sum over @xmath5 in ( [ zs ] ) ( see @xcite for more details ) .
we replace the sum over @xmath91 by the appropriate integral .
the thermodynamic potential @xmath92 is then : [ om ] = ( m^2-^2 ) v^2 + _ _ + _ ( 1-^- _ ) .
the first term is the tree - level potential .
the second term is due to the zero point energies of the fields .
the third term comes from the thermal excitations of the fields .
the thermodynamic potential is subject to two constraints .
first , it should be at a minimum with respect to the free parameter @xmath4 : [ con1 ] = 0 .
for the tree - level potential @xmath93 we see that the first condition is satisfied in two ways : in the condensed phase when @xmath74 and @xmath36 or in the uncondensed phase when @xmath78 and @xmath94 .
we will refer to these solutions as the on - shell " solutions since they satisfy the tree - level equations of motion for the scalar .
second , the thermodynamic potential should give the fixed number density of scalars when differentiated with respect to @xmath27 : [ con2 ] n j_0^s = - .
when applied to the tree - level potential , this constraint gives n= - = v^2 . in the condensed phase this is satisfied
when @xmath95 , as we found in the previous section for the condensate at zero temperature . in the uncondensed phase ( @xmath78 )
this expression can not be satisfied .
this is unsurprising as we have neglected contributions to the number density coming from thermal fluctuations . in the absence of these fluctuations ,
i.e. at zero temperature , the charge of the system must be stored in the condensate .
the dispersion relations @xmath89 simplify greatly when evaluated on the on - shell solutions . for the uncondensed phase we set @xmath78 in the determinant of @xmath83 ( equation ( [ d ] ) ) .
then @xmath96 can be factored according to equation ( [ det ] ) .
we find : [ ov0 ] _
= , _
the first two dispersion relations are for the scalar particle and antiparticle and are the same as those for a neutral system at finite chemical potential .
the second two are for the photon degrees of freedom . in the uncondensed phase ,
the spectrum of the theory and thus the one - loop thermodynamic potential is the same as for neutral scalars and decoupled , massless photons as mentioned above . to find the dispersion relations in the condensed phase we set @xmath36 in the determinant of @xmath83
this gives [ opm ] _
^2 = ^2 + g^2 v^2 + 2m^2 , [ o12 ]
_ 1,2 ^ 2 = ^2+g^2 v^2 .
these dispersion relations agree with those found for the charged condensate at zero temperature in @xcite , with the photon mass @xmath97 replaced by the more general @xmath98 . in the condensed phase the first two dispersion relations ( [ opm ] ) no longer have the simple interpretation of corresponding to the scalar particle and antiparticle . in the condensate
the gauge symmetry is broken .
the gauge boson becomes massive by eating one of the scalar degrees of freedom .
thus we can think of these two dispersion relations as corresponding to the longitudinal component of the photon and the remaining scalar degree of freedom , though in reality they correspond to a linear combination of the scalar and gauge fields .
their masses are found by taking @xmath54 : @xmath99 .
both modes are massive .
thus unlike the neutral condensate which contains a massless particle in the condensed phase , the charged condensate has a mass gap .
for @xmath100 the relations ( [ opm ] ) simplify to _
+ + m , and _
g^2 v^2 + , ^2 2 g v m , + _ - -m , ^2 2 g v m . the solutions to equation ( [ con1 ] ) calculated from the tree - level potential will be modified by finite temperature effects . in particular , we will find that at non - zero temperature , @xmath36 no longer holds identically in the condensed phase .
moreover , in order to solve the second constraint equation ( [ con2 ] ) at finite temperature , we need to know the dispersion relations @xmath89 as a function of @xmath27 , away from @xmath101 . thus to understand the full thermodynamic potential and its constraints , we must generalize the @xmath89 given above to the case that the tree - level solutions to the scalar equations of motion are not satisfied , when @xmath74 and @xmath94 . to compute the thermodynamic potential we have used the standard background field method . for gauge theories
this method gives a unique result only when the background field is a solution to the tree - level equations of motion .
if this condition is not satisfied , the background field method can give gauge dependent results for physical quantities .
this effect can be seen here by considering the matrix @xmath83 given in ( [ d ] ) . calculating the determinant of @xmath83 ,
one finds terms that depend on the gauge fixing parameter in the form @xmath102 .
these terms vanish on - shell , i.e. , when either @xmath103 or @xmath36 .
thus the dispersion relations found above are independent of @xmath72 as we would expect . however , when @xmath74 and @xmath94 the determinant of @xmath83 and thus the more general dispersion relations become dependent on the gauge fixing condition . in order to determine the general , gauge - independent dispersion relations from @xmath96 we would first have to find the unique off - shell potential . for our purposes , however , it is possible to determine the more general @xmath89 , given our knowledge of their behavior when @xmath103 or @xmath36 , in combination with constraints coming from a ward identity .
we will do this now .
the photon degrees of freedom do not contribute to the number density of the scalars .
thus the generalized dispersion relations @xmath104 must be independent of @xmath27 .
it follows that they are the same when @xmath94 as when @xmath36 : _ 1,2 ^ 2 = ^2+g^2 v^2 . to find the dispersion relations for the scalar degrees of freedom , we start by parametrizing @xmath105 by [ gen ] _
^2 = a . by requiring that the general dispersion relations reduce to those found above when @xmath78 or when @xmath36 we find a & = & ^2+m^2+^2 + g^2 v^2 + a(m , gv , ) g^2 v^2 , + b&= & 4 ^ 2 ^2 + 4m^2 ^2 - 2m^2g^2v^2+g^4 v^4 + b(m , gv , ) ( m^2-^2)g^2 v^2 . the functions @xmath106 and
@xmath107 are not fixed by this requirement .
they can in general represent a series expansion in @xmath108 and @xmath109 with undetermined coefficients .
we take @xmath106 and @xmath107 to be independent of @xmath110 , in order to maintain the appropriate high - momentum behavior of the theory . for the neutral condensate , @xmath36 identically in the condensed phase , even at finite temperature . for the charged condensate
this is not necessarily the case .
this is due to the fact that for the charged condensate the dispersion relations depend on @xmath4 .
however , any deviation of @xmath27 away from @xmath6 should be suppressed by @xmath111 and vanish in the limit that @xmath76 . thus for the low density systems that are the primary interest of this paper , the @xmath106 and @xmath107 terms
are subdominant as they are @xmath111-suppressed compared to the leading order terms in @xmath112 and @xmath113 . for high density systems they can be relevant ,
when @xmath114 .
a ward identity can be used to fix @xmath106 and @xmath107 .
the sum over the zero point energies in the thermodynamic potential ( [ om ] ) is uv divergent and thus must be renormalized . because of the ward identity associated with the conserved scalar current
, one can show that the conserved current is not subject to infinite renormalization , even in cases of spontaneous symmetry breaking ( see , e.g. , [ 19 ] ) .
it follows that the divergences in the potential should be independent of @xmath27 . expanding the dispersion relations ( [ gen ] ) for large @xmath115 we find : ( _ + + _ - ) || , where @xmath116 and @xmath117 depend on @xmath106 and @xmath107 .
the three terms on the r.h.s . represent the quartic , quadratic and logarithmic divergences of the scalar sector .
these divergences can be made independent of @xmath27 for appropriate choices of @xmath106 and @xmath107 .
this uniquely fixes @xmath118 and @xmath119 .
the generalized dispersion relations are thus : [ oalpha1 ] _ ^2 & = & ^2+m^2+^2 + g^2 v^2 .
+ let us now treat the remaining , @xmath27-independent divergences in the thermodynamic potential . to regularize the divergent terms we introduce an ultraviolet cutoff @xmath120 and take @xmath121 .
after integration , and in the limit of large @xmath120 , the divergent contributions to the thermodynamic potential coming from both the scalar and gauge degrees of freedom are _ _ & = & + & & + g^2 v^2 + & & -g^4 v^4 + finite terms . + the first term renormalizes the vacuum energy density .
this term is independent of temperature and can be trivially subtracted .
the second term renormalizes the mass of the scalar and the third term renormalizes the coupling of the @xmath15 interaction , had we included one at the start .
we can add to the lagrangian counterterms of the form _
= -m^2 ^ -(^ ) ^2 , to absorb these divergences . to renormalize the zero - temperature thermodynamic potential to leading order in @xmath111
, we can evaluate the contributions from the zero - point energy on the tree - level solution @xmath36 .
we also , for the moment , ignore the divergences that multiply @xmath122 as they are also higher order in @xmath111 .
we choose our renormalization conditions so that at zero temperature the thermodynamic potential is finite and is at a minimum when with respect to @xmath4 when @xmath123 .
this fixes @xmath124 and @xmath125 so that _ tree+_z.p.+_c.t . =
( m^2-^2 ) v^2 + o(_em^2 ) .
the presence of a @xmath122 term in the potential will shift the zero - temperature values of @xmath4 and @xmath27 .
moreover , the dependence of the divergent term that multiplies @xmath122 on the logarithm of @xmath4 means that the coupling of the @xmath122 interaction will run mildly with temperature .
however this term is suppressed by @xmath111 compared to both the other zero temperature terms and the finite temperature terms .
we will neglect its contribution to the potential when studying finite temperature effects in what follows .
in section 3 we found that the minimization of the tree - level potential with respect to @xmath4 has two solutions : one when @xmath78 and one for arbitrary @xmath4 when @xmath123 .
we also observed that at zero temperature the uncondensed solution @xmath78 can not satisfy the constraint equation ( [ con2 ] ) . at zero temperature
all of the scalars reside in the condensate .
more generally , there exists a critical temperature @xmath1 below which some fraction of the conserved scalar charge must be in the condensed phase .
likewise , there exists a critical temperature @xmath2 above which @xmath78 is the only solution to both equations ( [ con1 ] ) and ( [ con2 ] ) . between these two temperatures
the scalar field will undergo a phase transition into the condensed state . in this section
we will determine @xmath1 and @xmath2 by applying the constraints ( [ con1 ] ) and ( [ con2 ] ) to the full thermodynamic potential .
we will also discuss features of the phase transition .
let us consider the nonrelativistic limit @xmath126 . in this case @xmath127 .
since @xmath128 in this limit , the @xmath84 term in the thermodynamic potential is exponentially suppressed compared the @xmath85 and @xmath104 terms .
thus we can neglect its contribution : [ om2 ] ( m^2-^2 ) v^2+t ( 1-^- _ - ) + 2 t ( 1-^- ) .
the thermodynamic potential should be at a minimum with respect to @xmath4 : @xmath129 .
since @xmath4 appears in the dispersion relations only in the form @xmath130 we can write this condition as [ omdv ] ( m^2-^2 ) v+2g^2 v ( + ) = 0 .
+ this equation has two solutions .
the first solution is the same as for the tree - level potential , when @xmath78 .
the other solution is given by [ sol2 ] ^2=m^2 + 2g^2 ( + ) .
at finite temperature , @xmath27 is shifted away from its tree - level value @xmath6 by an amount proportional to @xmath111 .
it is straightforward to see that @xmath85 as given by ( [ oalpha1 ] ) increases monotonically as a function of @xmath109 , i.e. , @xmath131 for arbitrary @xmath4 .
thus at finite temperature , the second solution to @xmath129 requires @xmath132 . for the neutral condensate ,
the integral in the thermodynamic potential corresponding to the @xmath85 mode is convergent only when @xmath133 .
for the charged condensate however , when @xmath74 , solutions with @xmath132 are possible because of the contribution @xmath134 makes to the mass of this mode . still , when @xmath78 we must have @xmath133 , as in the neutral case .
it follows that @xmath78 is not a solution to the above equation ( [ sol2 ] ) .
the absence of a second solution at @xmath78 implies that @xmath4 must change discontinuously when going from the uncondensed phase to the condensed phase .
this is indicative of a first order phase transition .
we will see that this discontinuity is more pronounced in the low density regime .
let us consider the @xmath78 solution .
the general dispersion relations given by ( [ oalpha1 ] ) reduce to those found above in ( [ ov0 ] ) .
in particular _ - = - .
differentiating the potential with respect to @xmath27 gives the usual result for number density : n=- = .
all of the particles are in thermally excited states .
fixing @xmath5 gives an implicit expression for the chemical potential as a function of temperature .
the critical temperature @xmath1 is defined to be the minimum temperature at which all the particles are still in excited states .
it is found by taking @xmath36 and solving for @xmath135 .
the result is the same as for the critical temperature of the neutral condensate : [ tc1 ] t_c1 = ( ) ^2/3 for n^1/3 m .
the temperature @xmath1 puts a lower bound on the condensation temperature .
below this temperature no @xmath78 solution exists .
however , unlike in the neutral case , it may be possible for the charged scalars to condense before @xmath1 is reached , as we will now discuss .
when @xmath74 the number density is given by [ n ] n = v^2- , where @xmath85 is given by ( [ oalpha1 ] ) . the first term on the right hand side represents the fraction of particles in the condensate while the second term gives the particles that remain in thermally excited states .
this equation combined with ( [ sol2 ] ) gives an implicit expression for the value of the condensate @xmath4 as a function of temperature for fixed @xmath5 .
as noted above , the finite temperature correction to @xmath136 is suppressed by @xmath111 . as a lowest order approximation
we can evaluate the above expression ( [ n ] ) on the tree - level solution @xmath36 to find @xmath137 . .15 in + in figure 1 we plot @xmath4 as a function of temperature for various values of @xmath138 using expression ( [ n ] ) and taking @xmath36 . in the relativistic regime
we also include the contribution to number density coming from the @xmath84 mode .
we take the charge of the scalars to be @xmath139 with a system of helium nuclei in mind .
the values of @xmath4 at @xmath140 and of the critical temperature @xmath1 are of course different for different values of @xmath6 and @xmath5 , but we rescale the results for easy comparison . in figure 1a
we plot @xmath137 in the nonrelativistic regime when @xmath141 .
the orange line corresponds to @xmath142 , the red line to @xmath143 , and the pink line to @xmath144 . for comparison
we have also plotted @xmath137 for the neutral condensate in the nonrelativistic limit .
this is the black dashed line . in figure 1b
we plot @xmath137 in the relativistic regime when @xmath145 .
the purple line corresponds to @xmath146 , the dark blue line to @xmath147 and the light blue line to @xmath148 .
the black dotted line is @xmath137 for the neutral condensate in the relativistic limit .
all of the curves share the solution @xmath78 for @xmath149 .
the shape of the contours for @xmath150 is characteristic of a first order phase transition . at sufficiently high temperatures the only solution is @xmath78 .
but as the temperature drops , at some temperature @xmath151 , the @xmath137 curve ceases to be single - valued .
both the uncondensed solution and the condensed solution exist simultaneously .
moreover , in between @xmath2 and @xmath1 the value of @xmath4 must change discontinuously from zero as the system cools . at temperatures below @xmath1
only the condensate solution exists .
the temperature @xmath2 is the maximum temperature at which the condensate solution exists .
it corresponds to the point on the above plots where .
|_t = t_c2 = .
we see from figure 1 that lowering the value of @xmath138 appears to increase @xmath2 relative to @xmath1 .
lower densities also appear to correspond to larger discontinuities in @xmath4 . for the charged condensate ,
as @xmath152 increases relative to @xmath153 the function @xmath137 appears to asymptotically approach a single curve that is similar to the curve for the neutral condensate .
the phase transition appears to be nearly very second order with @xmath4 increasing more - or - less continuously as @xmath135 decreases .
even though the phase transition may technically be first order the jump in @xmath4 becomes negligibly small .
the above contours were drawn using the approximation @xmath36 in the condensed phase .
the actual condensate solutions are slightly to the left of these contours , when @xmath154 .
the existence of a first order phase transition at low densities can be understood in the following way . in the low density regime
the mass of the bosonic excitation @xmath134 becomes comparable to or greater than the transition temperature .
thus the thermal contributions of this mode to the number density become exponentially suppressed by a factor of @xmath155 in the condensed phase . in order to maintain the fixed charge density ,
@xmath4 must jump discontinuously so that charge that had been stored in thermal excitations is stored instead in the condensate . as the density is lowered , @xmath3 continues to decrease relative to @xmath134 and the effect grows more pronounced .
let us now find @xmath2 .
in the nonrelativistic limit we can approximate the number density as n mv^2 + , where @xmath156 . here
we have taken @xmath157 in the condensed phase .
changing to dimensionless variables x , and y , the number density becomes n mv^2 + ( ) ^3/2 _ 0^ .
let us assume that at the critical temperature @xmath2 the expectation value of the condensate as given by the above expression is roughly the same order of magnitude as it is at zero temperature : @xmath158 .
let us also assume that @xmath2 is roughly the same order of magnitude as @xmath1 as given by ( [ tc1 ] ) .
then when @xmath159 we have @xmath160 . in terms of our dimensionless variables
this is @xmath161 . in this regime
we can do a low temperature expansion of the above integral @xcite .
the first two terms of the low temperature expansion give : [ nmum ] n m v^2 + ( ) ^3/2 , + where li is the polylogarithm .
@xmath2 corresponds to the maximum temperature at which this expression can be satisfied for a fixed @xmath5 .
equivalently , it corresponds to the critical point of the function @xmath137 .
so to find @xmath2 we differentiate both sides of the above expression with respect to @xmath4 and we set @xmath162 .
we find that at the critical temperature , @xmath2 and @xmath4 are related by v_c2 - .
substituting this expression for @xmath4 into ( [ nmum ] ) gives an implicit expression for @xmath2 in terms of @xmath5 and @xmath6 .
in figure 2 we plot the ratio of @xmath2 to @xmath1 as a function of @xmath163 , from @xmath164 to @xmath165 .
again we set @xmath166 .
as @xmath5 decreases the ratio @xmath167 grows .
since @xmath1 coincides with the critical temperature for the neutral condensate , this implies that for the charged condensate at low densities , condensation can occur at temperatures several times higher than for the neutral condensate . in high density regimes
where @xmath168 , the above approximations break down .
we expect that in this limit the coulomb energy of the scalars should be small compared to their kinetic energy and thus @xmath167 should approach @xmath169 .
we have calculated the one - loop grand canonical partition function for scalar electrodynamics at finite temperature and chemical potential .
we considered an electrically neutral system in which the charged scalars reside in a background density of the opposite charge . using the background field method we found a result that was gauge condition dependent when the background fields did not satisfy the tree - level equations of motion . to address this
, we required consistency with a ward identity in order to determine the off - shell thermodynamic potential . using this potential we could determine , to lowest order in perturbation theory , the thermal expectation value of the charged scalars @xmath4 as a function of temperature .
plotting @xmath137 for various values of the scalar mass @xmath6 and number density @xmath5 , we found evidence of a first order phase transition in the low density regime , when @xmath7 . in between two critical temperatures @xmath1 and @xmath2 ,
@xmath4 jumps discontinuously from zero .
the lower temperature @xmath1 coincides with the critical temperature for a neutral condensate , while @xmath2 can be several times higher in the low density regime .
thus the phase transition for the charged condensate can occur at higher temperatures than for the neutral condensate .
the strength of the first order phase transition appears to increase with decreasing density . the critical temperatures @xmath1 and @xmath2 give only lower and upper bounds on the transition temperature .
finding the exact transition temperature as well as other physically interesting quantities , such as the latent heat involved in the phase transition or the timescale for the transition to occur , requires a more thorough understanding of the energetics of these solutions and their stability or metastability .
investigation of these questions would entail using the constraint equation @xmath170 to integrate out " @xmath27 so that the free energy density @xmath171 .
this is analogous to our procedure for the zero - temperature potential in equation ( [ v ] ) .
the above conclusions were based on a one - loop calculation .
it is reasonable to consider the effects of higher order corrections in perturbation theory . in thermal field theory at high temperatures , the perturbative expansion can break down , even for weakly coupled theories .
this can occur when the temperature of the system is greater than the masses of the particles running in the loops . for theories with massless particles ,
this breakdown is often signaled by the appearance of infrared divergences at higher loops . for the system described here , because the gauge symmetry is spontaneously broken in the condensed phase , the gauge bosons are massive .
thus one can compute to any order in perturbation theory without encountering infrared divergences .
moreover , in the condensate , as long as the temperature is sufficiently below the gauge boson mass @xmath134 and the mass of the remaining scalar degree of freedom @xmath172 , the loop expansion parameter should remain small . in the uncondensed phase ,
the massless photons interact with each other only through the massive scalars , the lightest of these scalars having mass @xmath173 .
thus as long as @xmath174 our considerations in the uncondensed phase should be valid . for a second order phase transition or for a weakly first order phase transition
these criteria will be violated near the phase transition point , where @xmath175 and @xmath176 .
thus perturbation theory can not be considered reliable in the vicinity the phase transition . however , for a strongly first order phase transition , which is precisely the interest of this work , this region is avoided . in the low density regime
the condensate solution does nt pass near @xmath177 ; the temperature of the system is always small compared to the masses of the excitations in the condensate .
thus higher order corrections to the results found above should be small .
to illustrate this we plot @xmath137 in figure 3 for the values of @xmath6 and @xmath5 that are relevant for helium - core white dwarf stars ( @xmath178 ) .
the orange region corresponds to where @xmath179 and thus perturbation theory can be unreliable . the condensate solution which is the upper solid blue curve is well outside this region .
we also mark the point where , when @xmath78 , @xmath180 . to the left of this point ,
i.e. , when @xmath181 , calculations of , say , the free energy density in the _ uncondensed _ phase using the expressions obtained above should not to be trusted .
this may place some limitations on calculating the exact transition temperature , if the phase transition occurs in this region .
however , it is possible that methods such as resummation can be applied in this region to obtain more reliable results .
let us end with some comments on physical applications
. in helium - core white dwarf stars the number density of the helium nuclei is such that @xmath11 .
thus based on the arguments given above , we might expect the condensation of these nuclei to be a first order phase transition .
we may also expect the condensation temperature for the helium nuclei to be somewhat greater than that for neutral bosons .
the latent heat associated with the phase transition could potentially delay the cooling of the star . for a carbon - core white dwarf star of mass @xmath182 which crystallizes
, the latent heat associated with crystallization can increase the classical cooling time by a factor of @xmath183 @xcite . however , one of the main features of a condensed - core white dwarf is that it cools much more rapidly than white dwarfs with uncondensed or crystallized cores @xcite .
it is possible that any contribution to the cooling history coming from the latent heat will be overwhelmed by the rapid cooling in the condensed phase . in this work we treated the background charge density as static and classical . for the system of nuclei and electrons in the cores of white dwarfs ,
this is not a good approximation .
the electrons , i.e. the background charge density , form a degenerate fermi gas with gapless excitations . due to these excitations ,
the electrons could play a significant role in the phase transition .
i am thankful to marcus berg , t. h. hansson and fawad hassan for stimulating conversations and especially to gregory gabadadze for conversations and comments on the draft .
this work was supported by the swedish research council ( vr ) through the oskar klein centre .
g. gabadadze and r. a. rosen , phys lett .
b * 658 * , 266 ( 2008 ) .
g. gabadadze and r. a. rosen , jcap * 0810 * , 030 ( 2008 ) .
g. gabadadze and r. a. rosen , jcap * 0902 * , 016 ( 2009 ) .
g. gabadadze and r. a. rosen , jhep * 0907 * , 093 ( 2009 ) . g. gabadadze and d. pirtskhalava , jcap * 0905 * , 017 ( 2009 ) .
l. berezhiani , g. gabadadze and d. pirtskhalava , jhep * 1004 * , 122 ( 2010 ) . c. w. bernard , phys .
d * 9 * , 3312 ( 1974 ) .
l. dolan and r. jackiw , phys .
d * 9 * , 3320 ( 1974 ) .
s. weinberg , phys .
d * 9 * , 3357 ( 1974 ) .
j. i. kapusta and c. gale , `` finite - temperature field theory : principles and applications , '' _ cambridge ,
uk : univ .
pr . _ ( 2006 ) .
d. a. kirzhnits and a. d. linde , annals phys . *
101 * , 195 ( 1976 ) .
a. d. linde , phys .
d * 14 * , 3345 ( 1976 ) .
j. i. kapusta , phys .
d * 24 * 426 ( 1981 ) .
h. e. haber and h. a. weldon , phys .
d * 25 * , 502 ( 1982 ) .
k. m. benson , j. bernstein and s. dodelson , phys .
d * 44 * , 2480 ( 1991 ) .
a. d. dolgov , a. lepidi and g. piccinelli , jcap * 0902 * , 027 ( 2009 ) .
a. d. dolgov , a. lepidi and g. piccinelli , phys .
d * 80 * , 125009 ( 2009 ) .
g. gabadadze and r. a. rosen , jcap * 1004 * , 028 ( 2010 ) .
g. gabadadze and r. a. rosen , phys .
b * 666 * , 277 ( 2008 ) .
cheng and l. f. li , `` gauge theory of elementary particle physics , '' _ oxford , uk : clarendon _ ( 1984 ) .
|
we calculate the grand canonical partition function at the one - loop level for scalar quantum electrodynamics at finite temperature and chemical potential .
a classical background charge density with a charge opposite that of the scalars ensures the neutrality of the system . for low density systems we find evidence of a first order phase transition .
we find upper and lower bounds on the transition temperature below which the charged scalars form a condensate .
a first order phase transition may have consequences for helium - core white dwarf stars in which it has been argued that such a condensate of charged helium-4 nuclei could exist .
0.9 cm * phase transitions of charged scalars * * at finite temperature and chemical potential * 0.7 cm rachel a. rosen _ oskar klein centre for cosmoparticle physics , department of physics _ _ stockholm university , albanova se-10691 , stockholm , sweden _ 0.5 cm
|
the evolution of the density of a test particle moving in a configuration of obstacles is described at mesoscopic level by linear kinetic equations .
they are obtained from the microscopic hamiltonian dynamics under a kinetic scaling of space and time , namely @xmath0 , @xmath1 and a suitable rescaling of the density of the obstacles and the intensity of the interaction . accordingly to the resulting frequency of collisions , the mean free path of the particle can have or not macroscopic length and different kinetic equations arise .
typical examples are the linear boltzmann equation and the linear landau equation
. the first rigorous result appeared in 1969 in the paper of gallavotti @xcite , who derived a linear boltzmann equation starting from a random distribution of fixed hard scatterers in the boltzmann - grad limit ( low density ) , namely when the number of collisions is small , thus the mean free path of the particle is macroscopic .
the result was improved by spohn @xcite . in the weak - coupling regime ,
when there are very many but weak collisions , a linear landau equation appears @xmath2 where @xmath3 is the laplace - beltrami operator on the @xmath4-dimensional sphere of radius @xmath5 .
it describes a momentum diffusion , i.e. the velocity process is a brownian motion on the ( kinetic ) energy sphere .
this intuitively follows from the facts that there are many elastic collisions with obstacles isotropically distributed .
the diffusion coefficient @xmath6 is proportional to the variance of the transferred momentum in a single collision and depends on the shape of the interaction potential .
the first result in this direction was obtained by kesten and papanicolau in 1978 for a particle in @xmath7 and by drr , goldstein and lebowitz in 1987 for a particle in @xmath8 for sufficiently smooth interaction potentials .
the linear landau equation yields also in an intermediate scale between low density and weak - coupling regime , namely when the ( smooth ) interaction potential @xmath9 rescales according to @xmath10 , @xmath11 and the density of the obstacles is of order @xmath12 ( @xcite , @xcite ) .
the limiting cases @xmath13 and @xmath14 correspond respectively to the low density limit and the weak - coupling limit . in the present paper
we want to investigate the limit @xmath15 in the intermediate case , namely when @xmath16 but sufficiently small , for an interaction potential no more smooth given by a circular potential barrier , in dimension two .
the physical interest of this problem is connected to the geometric optics since the trajectory of the test particle is that of a light ray traveling in a medium ( say water ) in presence of circular drops of a different substance with smaller refractive index ( say air ) .
the opposite situation , namely drops of water in a medium of air , can be described as well by the circular well potential .
our analysis applies also to this case with minor modifications , but we consider only the case of potential barrier for sake of concreteness .
the novelty of this choice is that in this case the diffusion coefficient @xmath6 diverges logarithmically . roughly speaking ,
the asymptotic equation for the density of the lorentz particle reads @xmath17 which suggests to look at a longer time scale @xmath18 .
as expected , a diffusion in space arises .
the proof follows the original constructive idea , due to gallavotti @xcite , for the low - density limit of a hard - sphere system .
this approach is based on a suitable change of variables which leads to a markovian approximation described by a linear boltzmann equation .
this presents some technical difficulties since some of the random configurations lead to trajectories that `` remember '' too much preventing the markov property of the limit . in the two - dimensional case
the probability of those bad behaviors producing memory effects ( correlation between the past and the present ) is nontrivial .
thus we need to control the unphysical trajectories : we estimate explicitly the set of bad configurations of the scatterers ( such as the set of configurations yielding recollisions or interferences ) showing that it is negligible in the limit ( see @xcite ) .
the control of memory effects still holds for a longer time scale @xmath19 which allows to get the heat equation from the rescaled linear boltzmann equation . + we remark that the diffusive limit analyzed in the present paper is suggested by the divergence of the diffusion coefficient for the particular choice of the potential we are considering .
however the same techniques could work in presence of a smooth , radial , short - range potential @xmath9 .
also in this case we obtain a diffusive equation as longer time scale limit of a linear boltzmann equation ( section 5 ) .
this is in the same spirit of @xcite and @xcite .
consider a point particle of mass one in @xmath8 , moving in a random distribution of fixed scatterers whose center are denoted by @xmath20 .
the equation of motion are @xmath21 where @xmath22 denote position and velocity of the test particle , @xmath23 the time and , as usual , @xmath24 indicates the time derivative for any time dependent variable @xmath25 . +
finally @xmath26 is a given spherically symmetric potential .
+ to outline a kinetic behavior of the particle , we usually introduce a scale parameter @xmath27 , indicating the ratio between the macroscopic and the microscopic variables , and rescale according to @xmath28 with @xmath29 $ ] .
then eq.ns become @xmath30 we assume the scatterers @xmath31 distributed according to a poisson distribution of intensity @xmath32 , where @xmath33 .
this means that the probability density of finding @xmath34 obstacles in a bounded measurable set @xmath35 is given by @xmath36 where @xmath37 . +
now let @xmath38 be the hamiltonian flow solution of eq.n with initial datum @xmath22 in a given sample @xmath31 of obstacles ( skipping the @xmath39 dependence for notational simplicity ) and , for a given initial probability distribution @xmath40 , consider the quantity @xmath41,\ ] ] where @xmath42 is the expectation with respect to the measure @xmath43 given by .
+ in the limit @xmath44 we expect that the probability distribution solves a linear kinetic equation depending on the value of @xmath45 . more precisely if @xmath13 ( low - density or boltzmann - grad limit ) then @xmath46 converges to @xmath47 , the solution of the following linear boltzmann equation @xmath48 where @xmath49 and where @xmath50 here we are assuming @xmath9 of range one i.e. @xmath51 if @xmath52 , and @xmath53 is the unit vector obtained by solving the scattering problem associated to @xmath9 .
this result was proven and discusses in @xcite,@xcite,@xcite,@xcite .
+ on the other hand , if @xmath14 , the corresponding limit , called weak - coupling limit , yields the linear landau equation ( see @xcite and @xcite ) @xmath54 where @xmath55 and @xmath56 note that @xmath57 is real and spherically symmetric . in the present paper
we want to investigate the limit @xmath15 , in case @xmath16 sufficiently small , when the diffusion coefficient @xmath6 given by is diverging .
actually we consider the specific example @xmath58 namely a circular potential barrier .
+ for a potential of the form
a simple computation shows that @xmath6 defined in diverges logarithmically .
therefore we are interested in characterizing the asymptotic behavior of @xmath59 , given by , under the scaling illustrated above .
the main result of the present paper can be summarized in the following theorem .
[ th : main th ] suppose @xmath60 a continuous , compactly supported initial probability density .
suppose also that @xmath61 , where @xmath62 is any partial derivative with respect to @xmath63 and @xmath64 .
finally assume @xmath65 .
the following statements hold 1 . if @xmath66 , for all @xmath67 $ ] , @xmath68 , @xmath69 the convergence is in @xmath70 .
2 . if @xmath71 , for all @xmath67 $ ] , @xmath68 , @xmath72 where @xmath47 solves the landau equation with a renormalized diffusion coefficient @xmath73 the convergence is in @xmath70 .
+ 3 . if @xmath66 , defining @xmath74 , for all @xmath75 , @xmath68 , @xmath76 where @xmath77 solves the following heat equation @xmath78 with @xmath79 given by the green - kubo formula @xmath80\,dt , \ ] ] where @xmath81 is the stochastic process dictated by the generator of the landau equation starting from @xmath82 and @xmath83 $ ] denotes the expectation with respect to the invariant measure on @xmath84 .
the convergence is in @xmath70
. some comments to theorem [ th : main th ] are in order .
as we shall prove in section 4 , the asymptotic behavior of the mechanical system we are considering is the same as the markov process ruled by the linear landau equation with a diverging factor in front of @xmath85 .
this is equivalent to consider the limit in the euler scaling of the linear landau equation , which is trivial .
the system quickly thermalizes to the local equilibrium just given by @xmath86 .
this is point 1 ) . to detect something non - trivial we have to exploit longer times in which the local equilibrium starts to evolve ( according to the diffusion equation ) , see point 3 ) .
note however that , rescaling differently the density of the poisson process , we can recover the kinetic picture given by landau equation ( with a renormalized diffusion coefficient @xmath6 ) as in @xcite , see point 2 ) .
we finally remark that this picture is made possible because the recollisions set ( see below for the precise definition ) is negligible , as established in section 4 .
we believe that the present result could be recovered also in high - density regimes @xmath87 $ ] , namely also when the recollisions are not negligible anymore .
however in this case different ideas and techniques are indeed necessary .
the plan of the paper is the following . in the next section
we illustrate our strategy and establish some preliminary results . in section 3
we prove theorem 1.1 .
finally in section 4 we prove a basic lemma showing that our non - markovian system can indeed be approximated by a markovian one , easier to handle with .
we follow the explicit approach in @xcite , @xcite and @xcite .
+ by we have , for @xmath88 , @xmath89 , @xmath90 where @xmath38 is the hamiltonian flow generated by the hamiltonian @xmath91 where @xmath9 is given by , and initial datum @xmath22 .
finally @xmath92 , where here and in the following , @xmath93 denotes the disk of center @xmath63 and radius @xmath94 .
+ the explicit solution to the equation of motion is obtained by solving the single scattering problem by using the energy and angular momentum conservation ( see figure below ) . here
we represent the scattering of a particle entering in the ball @xmath95 toward a potential barrier of intensity @xmath96 .
we have an explicit expression for the refractive index @xmath97 where @xmath82 is the initial velocity , @xmath98 the velocity inside the barrier , @xmath45 the angle of incidence and @xmath99 the angle of refraction .
the scattering angle is @xmath100 and the impact parameter is @xmath101 .
( see appendix [ scatt ] for a detailed analysis of the scattering problem . )
formula makes sense if @xmath102 or @xmath103 .
+ when one of such two inequalities is violated , the outgoing velocity is the one given by the elastic reflection .
after the scaling @xmath104 the scattering process takes place in a disk of radius @xmath39 , but the velocities ( and hence the angles ) are invariant . a picture of a typical trajectory is given as in figure 3 .
here we are not considering possible overlappings of obstacles .
the scattering process can be solved in this case as well . however , as we shall see , this event is negligible because of the moderate densities we are considering .
+ coming back to eq.n , we distinguish the obstacles of the configuration @xmath105 which , up to the time @xmath23 , influence the motion , called internal obstacles , and the external ones .
more precisely @xmath106 is internal if @xmath107 while @xmath106 is external if @xmath108 here @xmath109 .
note that the integration over the external obstacles can be done so that @xmath110 here and in the sequel @xmath111 is the characteristic function of the event @xmath112 .
+ moreover @xmath113 is the tube : @xmath114 note that @xmath115 instead of considering @xmath46 we introduce @xmath116 where @xmath117 obviously @xmath118 following @xcite,@xcite,@xcite we would like to perform the following change of variables @xmath119 @xmath120 where , after ordering the obstacles @xmath121 according to the scattering sequence , @xmath122 and @xmath123 are the impact parameter and the entrance time of the light particle in the protection disk around @xmath124 .
more precisely , fixed an impact parameter @xmath77 and an entrance time @xmath23 we construct @xmath125 , the center of the obstacle .
then we perform the backward scattering and iterate the procedure to construct a trajectory @xmath126
. however @xmath127 ( therefore the mapping is one - to - one ) only outside the following pathological situations .
+ i ) * overlapping*. + if @xmath124 and @xmath128 are both internal then @xmath129 .
+ ii ) * recollisions*. + there exists @xmath124 such that for @xmath130 , @xmath131 , @xmath132
. + iii ) * interferences*. + there exists @xmath124 such that @xmath132 for @xmath130 , @xmath133 .
+ we simply skip such events by setting @xmath134 @xmath135 @xmath136 and defining @xmath137 note that @xmath138 .
note also that in we have used the change of variables for which , outside the pathological sets i ) , ii ) , iii ) , @xmath139 .
+ next we remove @xmath140 by setting @xmath141 we can prove : [ th : prop ] @xmath142 where @xmath143 as @xmath44 for all @xmath144.$ ] [ th : remarkprop ] proposition [ th : prop ] still holds for longer times , namely : @xmath145,\quad t>0.\ ] ] we postpone the proof of the above proposition in the last section . + next
we consider the limiting trajectory @xmath146 obtained by considering the collision as instantaneous . + more precisely , for the sequence @xmath147 @xmath148 consider the sequence @xmath149 of incoming velocities before the q collisions .
then @xmath150 we define @xmath151 due to the lipschitz continuity of @xmath152 we can assert that @xmath153 where @xmath154}|{\varphi}_2(x , v , t)|\leq c{\varepsilon}^{1 - 2\alpha}t.\ ] ] for more details see @xcite , section 3 . as matter of facts , since we realize that @xmath155 is the solution of the following boltzmann equation @xmath156 where @xmath157 we have reduced the problem , thanks to proposition 1 , to the analysis of a markov process which is an easier task .
let be @xmath158 .
we rewrite the linear boltzmann equation in the following way @xmath159 where @xmath160 , namely @xmath161,\qquad f\in l^1({\mathbb{r}}^2).\ ] ] we will show that for @xmath162 we get a trivial result ( theorem [ th : main th ] , item 1 ) ) , then we should look at the solution for times @xmath163 , namely in the diffusive scaling . denoting by @xmath164 , where @xmath165 solves , @xmath166 solves @xmath167 it is convenient to introduce the cauchy problem associated to the following rescaled landau equation : @xmath168 where @xmath169 .
we observe preliminarily that eq .
propagates the regularity of the derivatives with respect to the @xmath63 variable and , due to the presence of @xmath85 , gains regularity with respect to the transverse component of the velocity .
indeed , for any fixed @xmath5 , denoting by @xmath84 the circle of radius @xmath5 , under the hypothesis of theorem [ th : main th ] on @xmath152 , the solution @xmath170 satisfies the bounds @xmath171 $ ] , where @xmath172 and @xmath173 is the derivative with respect to the transverse component of the velocity .
in particular the solutions of we are considering are classical . + before analyzing the asymptotic behavior of the solution of we first need a preliminary lemma .
[ th : preliminary ] let @xmath174 be the average of @xmath175 with respect to the invariant measure @xmath176 , namely @xmath177 under the hypothesis of theorem [ th : main th ] * @xmath178;l^2({\mathbb{r}}^2\times s_{|v|})).$ ] moreover , setting @xmath179 for @xmath180 then * @xmath181 where @xmath182 let @xmath183 .
we have @xmath184 where @xmath185 we can estimate the last quantity by : @xmath186 therefore by we have @xmath187 where @xmath188 is the first positive eigenvalue of @xmath85 . here
we used that @xmath189 in @xmath190 .
hence @xmath191 so that ( 1 ) is proven .
+ to prove ( 2 ) observe that , thanks to the fact @xmath85 is negative , we have @xmath192 therefore @xmath193 which vanishes as @xmath194 . finally , recalling that @xmath195
, we have @xmath196 } \|g_{\eta}-\langle g_\eta\rangle\|_{l^2}+\|\langle g_\eta(t_\eta)\rangle-\langle f_0\rangle\|_{l^2}\\ & \leq \sup_{t\in ( 0,t]}\|g_{\eta}-\langle g_\eta\rangle\|_{l^2}+c\|g_{\eta}(t_\eta)-f_0\|_{l^2}. \end{split}\ ] ] by and ( 1 ) we conclude the proof . [
th : diff ] let @xmath175 be the solution of
. under the hypothesis of theorem [ th : main th ] for the initial datum @xmath152 , for @xmath162 @xmath175 converges to the solution of the diffusion equation @xmath78 where @xmath197 and @xmath198 convergence is in @xmath199;l^2({\mathbb{r}}^2 \times s_{|v|}))$ ] .
the proof of the above lemma is rather straightforward ( see e.g. @xcite ) .
+ suppose for the moment that the initial datum depends only on the position variables , namely the initial datum has the form of a local equilibrium .
we assume that @xmath175 has the following form @xmath200 where @xmath201 , @xmath202 are the first three coefficient of a hilbert expansion in @xmath203 , and @xmath204 is the reminder . comparing terms of the same order in @xmath203 we obtain the following equations : @xmath205 with @xmath206 . since @xmath207 is an odd function of @xmath82 , the integral with respect to @xmath82 of the left hand side of ( i ) vanishes .
then we can invert the operator @xmath208 and set @xmath209 , where @xmath210 is an odd function of the velocity .
now we integrate the second equation with respect to the velocity . by observing that @xmath211 ,
since @xmath212 is proportional the invariant measure , we obtain @xmath213 we define the @xmath214 matrix @xmath79 as @xmath215 and we observe that @xmath216 for @xmath217 and @xmath218 , where @xmath219 therefore @xmath220 where @xmath221 satisfies the initial condition @xmath222 . moreover , the @xmath190-norm of @xmath210 is bounded .
if we show that also the @xmath190-norm of @xmath223 and @xmath224 are bounded , we deduce that @xmath175 converges to @xmath221 for @xmath194 . + from equation @xmath225 and the diffusion equation for @xmath221 we derive that the integral with respect to @xmath82 of the left hand side of @xmath225 vanishes . therefore we can invert the operator @xmath208 and obtain @xmath226 . \end{split}\ ] ] therefore the @xmath227-norm of @xmath223 is bounded .
we derive from equation @xmath228 @xmath229 where @xmath230 denotes the scalar product in @xmath190 . using positivity of @xmath231 and cauchy - schwartz we deduce @xmath232
recall the explicit expression for @xmath233 , namely @xmath234 . by direct computation @xmath235 \delta_{|_{s_{|v|}}}^{-1}v_k , \end{split}\ ] ]
from which we deduce that the @xmath190-norm of @xmath236 is bounded .
similarly , one can easily show that the @xmath190-norm of @xmath237 is bounded , and then @xmath238 is uniformly bounded in @xmath239 $ ] and @xmath240 . to complete the proof we consider more general initial data @xmath152 depending also on the velocity variable
let @xmath241 .
we compare @xmath175 with @xmath242 , the solution with initial datum @xmath86 . by the same argument as in lemma [ th : preliminary ] , item ( 2 ) , we have that @xmath243 @xmath244 where @xmath245 depends on the @xmath190-norm of @xmath86 and @xmath246 .
since @xmath247 and @xmath248 we derive @xmath249 thus , by lemma [ th : preliminary ] , item ( 2 ) , we obtain that @xmath250 and @xmath251 have the same asymptotics and this concludes the proof of lemma .
[ th : prop 2 ] let @xmath152 be an initial datum for @xmath166 solution of . under the hypothesis of theorem [ th : main th ] @xmath166 converges to @xmath252 as @xmath44 , where @xmath253\to{\mathbb{r}}_+$ ] is the solution of the diffusion equation @xmath254 with @xmath197 .
the diffusion coefficient is @xmath79 given by the green - kubo formula .
convergence is in @xmath255 uniformly in @xmath256 $ ] .
let @xmath257 be solution of with @xmath258 and initial condition @xmath152 .
we look at the evolution of @xmath259 , namely @xmath260 where @xmath261
. then we obtain @xmath262 \big)\\ & + \eta_{\varepsilon}^2 \big({\tilde h_{\varepsilon}}-g_{\eta_{\varepsilon } } , \
, \big[\tilde{\text{l}}_{\varepsilon}-{\mathcal{l}}\big ] g_{\eta_{\varepsilon } } \big ) , \end{split}\ ] ] from which , using positivity of @xmath263 and cauchy - schwartz , @xmath264 recalling that @xmath265,\ ] ] we set @xmath266 with @xmath267 . integrating with respect to @xmath82 and using symmetry arguments we obtain @xmath268 observe that @xmath269 , then by direct computation ( see appendix ) @xmath270 and @xmath271 therefore @xmath272 , which vanishes for @xmath44 . in order to complete the proof of the item 3 ) of theorem [ th : main th ] , we need to show that @xmath273 converges to @xmath274 in @xmath275 , for every @xmath276 $ ] . by proposition [ th : prop ] and remark [ th : remarkprop ]
we have that @xmath277 defined in converges to @xmath278 , , in @xmath279 , for every @xmath276 $ ] .
moreover , using and the fact that the initial datum has compact support , we have that @xmath280 converges to @xmath274 in @xmath279 , for every @xmath276 $ ] . under hypothesis of theorem [
th : main th ] , convergence in @xmath281 norm implies convergence in @xmath190 .
since @xmath282 and using the fact that at @xmath283 the equality holds and the linear boltzmann equation [ be1 ] preserves the total mass , then also @xmath284 converges to @xmath274 in @xmath285 , for every @xmath276 $ ] .
now we go back to equation . using the same strategy of the proof of proposition [ th : prop 2 ] we can replace @xmath286 with @xmath287 , and we denote @xmath288 the solution of @xmath289 with initial datum @xmath152 . by the same arguments as in lemma [ th : preliminary ] , item ( i )
, one can prove that for @xmath194 @xmath290 and @xmath291 .
we observe that @xmath292 therefore @xmath293 converges to @xmath294 as @xmath194 , which concludes the proof of item 1 ) .
proof of item 2 ) is included in the proof of proposition [ th : prop 2 ] .
in this section we prove proposition [ th : prop ] .
+ clearly @xmath295 and we estimate separately all the events in the right hand side of .
+ we denote by @xmath296 the backward markov process defined , for @xmath297 , in section 2 and we set @xmath298 for any measurable function @xmath299 of the process @xmath300 .
we have @xmath301 for @xmath302 , @xmath303 and @xmath39 sufficiently small . + here and in the sequel @xmath23
is allowed to behave as @xmath304 .
+ estimate is obvious . indeed if @xmath305 the first or the last collision must satisfy either @xmath306 or @xmath307 .
hence follows easily .
+ a similar argument can be used to estimate @xmath308 . indeed if @xmath309 it must be @xmath310 for some @xmath311 . therefore proceeding as before @xmath312 for some @xmath302 , @xmath313 and @xmath39 sufficiently small .
+ next we pass to the control of the recollision event .
we proceed similarly as in @xcite and in @xcite .
let @xmath123 the first time the light particle hits the i - th scattering , @xmath314 the incoming velocity , @xmath315 the outgoing velocity and @xmath316 the exit time .
moreover we fix the axis in such a way that @xmath315 is parallel to the @xmath63 axis ( see figure [ fig:7 ] ) .
we have @xmath317 where @xmath318 if and only if @xmath124 ( constructed via the sequence @xmath319 ) is recollided in the time interval @xmath320 .
+ note that , since @xmath321 , where @xmath322 is the i - th scattering angle , in order to have a recollision it must be an intermediate velocity @xmath323 , @xmath324 such that @xmath325 is almost orthogonal to @xmath326 ( see the figure ) . then @xmath327 where @xmath328 if and only if @xmath318 and is fulfilled .
+ fix now all the parameters @xmath329 , @xmath147 but @xmath330 and perform such a time integration .
the two branches of the trajectory @xmath331 are rigid so that , if the recollision happen the time integration with respect to @xmath330 is restricted to a time interval proportional to @xmath332 .
more precisely it is bounded by @xmath333 performing all the other integrations and summing over @xmath334 we obtain @xmath335 for some @xmath302 , @xmath336 and @xmath39 sufficiently small .
+ we finally estimate the event @xmath337 . to do this
we fix a sequence of parameters @xmath329 , @xmath147 . for instance consider the case in figure [ fig:8 ] in which we exhibit an unphysical trajectory .
consider the integral @xmath338 here @xmath339 for those values of @xmath340 for which an interference takes place and @xmath341 by the liouville theorem we can integrate over the variables @xmath342 as independent variables @xmath343 where @xmath344 note that @xmath345 since a backward interference is a forward recollision .
clearly @xmath346 where @xmath347 is the ball of radius @xmath348 in @xmath349 .
+ thus @xmath350 therefore , by using estimate and @xmath351 this concludes the proof of proposition [ th : prop ] .
the diffusive limit analyzed in the present paper is suggested by the divergence of @xmath6 for the particular choice of the potential we are considering .
however the same techniques could work in presence of a smooth , radial , short - range potential @xmath9 .
[ th : main th2 ] under the same hypothesis of theorem [ th : main th ] , assume @xmath352)$ ] .
scale the variables , the density and the potential according to @xmath353 then , for @xmath89 and @xmath44 , there exists @xmath354 s.t . for @xmath355 @xmath356 solution of the heat equation @xmath78 with @xmath79 given by the green - kubo formula
@xmath357 the convergence is in @xmath358 .
the significance and the proof of the above theorem is clear .
the kinetic regime describes the system for kinetic times @xmath359 .
one can go further to diffusive times provided that @xmath188 is not too large .
indeed the distribution function @xmath46 `` almost '' solves @xmath360 for which the arguments of section 3 do apply . in other words
there is a scale of time for which the system diffuses .
however such times are not so large to prevent the markov property . obviously the diffusion coefficient is computed in terms of the limiting markov process .
we can give an estimate , certainly not optimal , of the coefficient @xmath188 appearing in .
estimating recollisions and interferences as in section 4 , setting @xmath361 , the condition on @xmath188 is ( see ) @xmath362 although the scaling we are considering in theorem is quite particular , the aim is the same as in @xcite where the same problem has been approached for the weak - coupling limit ( @xmath363 ) of a quantum system .
+ recently we were aware of a result concerning the diffusion limit of a test particle of a hard - core system at thermal equilibrium @xcite . also in this case the quantitative control of the pathological trajectories allows to reach larger times in which a diffusive regime is outlined . * acknowledgments .
* + we are indebted to s. simonella and h. spohn for illuminating discussions .
the potential energy for a finite potential barrier is given by @xmath364 the light particle , of unitary mass , moves in a straight line with energy @xmath365 .
let @xmath77 be the impact parameter .
for small impact parameters the particle will pass through the barrier , for large ones the particle will be reflected .
inside the barrier the velocity is a constant @xmath366 ( @xmath367 ) .
the complete trajectory of the light particle which passes through the barrier consists of three straight lines and is symmetrical about a radial line perpendicular to the interior path .
let @xmath45 be the angle of incidence ( the inside angle between the trajectory and a radial line to the point of contact with tha barrier at @xmath368 ) and @xmath99 the angle of refraction ( the corresponding external angle ) .
we assume that the radius of the circle is @xmath368 .
according to the geometry of the problem @xmath45 and @xmath99 are such that @xmath369 where @xmath370 .
the angle of deflection is @xmath371 thanks to the energy and angular momentum conservation the expression for the refractive index becomes @xmath372 and so we have a scattering angle defined in the following way : @xmath373 in the first case the particle passes through the barrier ( for @xmath374 ) , and in the second one the particle is reflected ( for @xmath375 ) .
the maximum scattering angle @xmath376 is the angle at which the particle scatters tangentially to the barrier .
the differential scattering cross section @xmath377 is then : @xmath378[1-n\cos(\theta/2)]}{(1+n^2 - 2n\cos(\theta/2))^{3/2 } } & \text{if } \theta\leq 2\arccos(n ) \\ [ 1-\cos^2(\theta/2)]^{1/2 } & \text{otherwise } \end{cases}\ ] ] scaling now the potential as @xmath379 , the previous formulas still hold . thus , according to this scaling , the refractive index becomes , @xmath380 to replace into , and .
in this section we show that the diffusion coefficient is divergent for the circular potential barrier . at this level
we assume that @xmath381 to simplify the following expressions .
our aim is to perform a taylor expansion of the first branch of @xmath385 for @xmath386 , + @xmath387 , with @xmath388 .
we have @xmath389 where @xmath390 then , looking at the first integral in the r.h.s of , we have to split it as @xmath391 thus @xmath392 ^ 2\,d\rho}\\ & \leq4{\varepsilon}^{-2\alpha}\left[\int_{0}^{n_{{\varepsilon}}(1-\delta)}{\frac{(1-n_{{\varepsilon}})^2}{n_{{\varepsilon}}^2}\frac{\rho^2}{(1-\rho^2)}\,d\rho}+\int_{0}^{n_{{\varepsilon}}(1-\delta)}{r_1\left(\rho / n_{{\varepsilon}}\right)^2\,d\rho}\right]+\\&4{\varepsilon}^{-2\alpha}\left[\left(\int_{0}^{n_{{\varepsilon}}(1-\delta)}{r_1\left(\rho / n_{{\varepsilon}}\right)^2}\right)^{\frac{1}{2}}\left(\int_{0}^{n_{{\varepsilon}}(1-\delta)}{\frac{(1-n_{{\varepsilon}})^2}{n_{{\varepsilon}}^2}\frac{\rho^2}{(1-\rho^2)}\,d\rho}\right)^{\frac{1}{2}}\right ] . \end{split}\ ] ] it is sufficient to compute the first two integrals .
let @xmath393 and @xmath394 be the first and the second integrals respectively .
we have @xmath395 .
\end{split}\ ] ] using that @xmath396 , from it is clear that @xmath397 a straightforward computation shows that the right hand side of the previous expression is @xmath398 choosing @xmath399 with @xmath400 , it follows @xmath401 .
in order to compute @xmath394 , we need the following estimate for the remainder term @xmath402 then @xmath403 also in this case , the only significant contribution is given by @xmath404 again for @xmath399 with @xmath400 .
this shows that @xmath405 now we compute @xmath6 in , namely @xmath406 ^ 2\,d\rho}\\&={\varepsilon}^{-2\alpha}\int_{n_{{\varepsilon}}(1-\delta)}^{n_{{\varepsilon}}}\left(\int_{\rho}^{\frac{\rho}{n_{{\varepsilon}}}}\,dx \frac{1}{\sqrt{1-x^2}}\right)^2\,d\rho .
\end{split}\ ] ] since @xmath407 in we have @xmath408\simeq { \varepsilon}^{-2\alpha}{\varepsilon}^{2\alpha}\delta . \end{split}\ ] ] again , with the previous choice for @xmath409 , this term vanishes in the limit for @xmath44 .
|
we consider a point particle moving in a random distribution of obstacles described by a potential barrier .
we show that , in a weak - coupling regime , under a diffusion limit suggested by the potential itself , the probability distribution of the particle converges to the solution of the heat equation .
the diffusion coefficient is given by the green - kubo formula associated to the generator of the diffusion process dictated by the linear landau equation .
|
in july 2012 , the atlas @xcite and cms @xcite collaborations announced the discovery of a new boson , with a mass in the @xmath1 gev ballpark .
although detailed studies of its properties need to be done in order to confirm its identity , one is tempted to interpret this new particle in terms of the long - awaited higgs boson . in any case , this constitutes a decisive breakthrough in high energy physics that , once correctly understood , will surely shed light on the dynamics of the electroweak scale .
the discovery of the higgs boson , and the measurement of its mass , reminds us a long - standing theoretical problem in particle physics : the famous hierarchy problem @xcite .
this can be expressed as the high sensitivity that fundamental scalars have to physics at high energies .
unless one accepts a very precise fine - tuning of the parameters of the theory , the higgs mass is naturally pushed to those high energies , distabilizing the electroweak scale .
supersymmetry ( susy ) is one of the most popular solutions to the hierarchy problem .
if the susy breaking scale is low ( not far from the tev scale ) , the electroweak scale is kept under control by virtue of the cancellations between bosonic and fermionic contributions to the higgs mass .
+ when constructing a supersymmetric model one finds new gauge and susy invariant renormalizable interactions , not present in the standard model , that lead to lepton ( l ) and baryon ( b ) number violation . with the particle content of the minimal supersymmetric standard model ( mssm )
, these are @xmath2_f,\quad [ l l e]_f,\quad [ l q d]_f,\quad [ u d d]_f,\label{rpvt}\ ] ] where @xmath3 stands for f - terms .
if they were simultaneously present in the lagrangian , the proton would have a fast decay rate unless very small coefficients are introduced .
for that reason , one usually introduces a discrete symmetry called r - parity , @xmath4 ( where @xmath5 is the spin of the particle ) , that forbids the l and b violating terms shown above @xcite .
the conservation of r - parity has very relevant phenomenological implications .
this discrete symmetry stabilizes the lightest supersymmetric particle ( lsp ) . as a consequence of that
, supersymmetric events at colliders contain large amounts of missing energy in the final state .
furthermore , if neutral , the lsp would be the perfect example of a weakly interacting massive particle ( wimp ) and a good dark matter candidate .
r - parity has , however , some drawbacks .
first of all , r - parity is introduced by hand in the mssm , without a theoretical explanation for its origin . and second
, there are additional non - renormalizable interactions which , even though they break lepton and/or baryon numbers , are perfectly allowed by r - parity . even if these operators are generated at the planck scale , they would lead to unnacceptably fast proton decay unless their coefficients are tiny @xcite . +
many theoretical ideas have been proposed in order to explain the origin of r - parity .
most of them consider r - parity as a remnant after the breaking of a larger symmetry group , see for example @xcite .
however , in most cases there is no explanation for the suppression of the higher - dimensional operators . on the other hand , the origin of fermion masses hierarchies and mixings , the so - called flavor problem , is another long - standing mystery in particle physics . one interesting possibility to address
the flavor problem is the introduction of an horizontal symmetry between the three generations of fermions .
the flavor symmetry , which can be either continuous or discrete , imposes some structures on the yukawa couplings . in this paper
we propose that r - parity may be a consequence of the flavor symmetry . in this case , quark and lepton mixings , as well as the stability of proton and dark matter , can be explained in a common framework .
we introduce a flavor model where all renormalizable as well as dimension five l or b violating operators are forbidden by the non - abelian flavor symmetry @xmath0 . charged femion masses and mixings can be fitted to their observed values . since the weinberg operator is also forbidden by the flavor symmetry , one is forced to go beyond minimal models if neutrino masses and mixings are to be explained .
this turns out be a non - trivial task due to ( 1 ) the restrictions imposed by @xmath0 , and ( 2 ) our main motivation of not inducing r - parity violating terms .
we find that this can be achieved by extending the lepton sector without inducing dangerous dimension five operators .
neutrino mass arises from a variation of the inverse seesaw mechanism @xcite , called inverse type - iii seesaw mechanism @xcite , that provides a correct description of neutrino mixings .
the rest of the paper is organized as follows : we present the model and main idea in section [ sec : model ] : how the introduction of a @xmath0 flavor symmetry can automatically lead to r - parity conservation and proton stability . afterwards , we discuss the flavor structure of the resulting quark and lepton sectors in sections [ sec : quarksector ] and [ sec : leptonsector ] , respectively . in the latter case
we extend the original model in order to account for neutrino masses and show how the resulting neutrino mixing pattern can accommodate the data coming from oscillation experiments .
finally , we conclude with a short summary and a brief discussion in section [ sec : summary ] .
let us consider the mssm extended by a @xmath0 flavor symmetry , see for instance @xcite .
this discrete group is a subgroup of @xmath6 ( for a classification see @xcite ) that belongs to the series @xmath7 .
it has @xmath8 irreducible representations , namely two triplets @xmath9 , @xmath10 and 9 singlets @xmath11 .
the product rules for the triplet representations are @xmath12 and @xmath13 . from these rules , it is clear that the product @xmath14 is invariant under @xmath0 whereas @xmath15 is not .
we assign @xmath0 representations to the mssm particle content as shown in table [ model ] .
it is straightforward to check that this assignment forbids all the aforementioned r - parity violating couplings in eq .
( [ rpvt ] ) . therefore , r - parity results as an accidental symmetry originated by the underlying flavor symmetry of the model . .@xmath0 charges of the mssm superfields . [ cols="^,^,^,^,^,^,^,^",options="header " , ] when the scalar component of @xmath16 gets a vev , an effective majorana mass for the @xmath17 triplet ( the fermionic component of the @xmath18 superfield ) is generated .
this leads to an inverse seesaw mechanism @xcite induced by @xmath19 triplets ( for other realizations of the _ inverse type - iii seesaw _ see @xcite ) . in the basis @xmath20
, we obtain the @xmath21 mass matrix for the neutral fermions @xmath22 which , assuming @xmath23 , leads to @xmath24 where @xmath25 and @xmath26 .
in fact , it is worth emphasizing some advantages that our model has with respect to the conventional inverse seesaw .
typically , in the context of inverse seesaw models it is quite difficult to forbid the @xmath27 mass term after ones allows for lepton number violation . here
its absence is a direct consequence of the @xmath0 symmetry .
moreover , in order to have new physics at the tev scale , that is @xmath28tev , and @xmath29 yukawa couplings , the parameter @xmath30 must be of the order of @xmath31 ev . in the original inverse seesaw mechanism
the @xmath32 parameter is expected to be naturally small in the t hooft sence since the limit @xmath33 enhances the symmetry of the lagrangian ( lepton number is recovered ) .
although this is not the case here , @xmath32 can be suppressed by means of the vev of the scalar @xmath34 . moreover ,
several mechanisms to obtain dynamically a small @xmath32 parameter are known in the literature @xcite . for these reasons
, we consider a small @xmath32 parameter a natural choice .
as mentioned above , the flavor symmetry forbids the weinberg operator .
in fact , in this model neutrino masses are generated due to the effective operator @xmath35_f \ , , \ ] ] which is obtained after integrating out the @xmath36 superfields , as depicted in figure([mnu - dim6 ] ) .
therefore , the proposed model can be seen as a concrete renormalizable realization of this operator , which is the minimal implementation of majorana neutrino masses after the weinberg operator is forbidden . _
f$ ] realization . ]
lepton number is explicitly broken by the trilinear superpotential term @xmath37 and thus no majoron , the goldstone boson associated to the spontaneous breaking of @xmath38 , appears in the spectrum .
however , r - parity remains unbroken after the addition of this term , since @xmath34 carries two units of lepton number .
one may also wonder about the appearance of new unwanted interactions after the introduction of the @xmath16 superfield .
it is straightforward to check that no l or b violating renormalizable operators are allowed , apart from the aforementioned @xmath39 term .
regarding higher dimensional operators , we note that the operator @xmath40_f$ ] is the only dimension 5 operator ( with l or b violation ) that is allowed by the symmetries .
however , it can not be generated at tree - level , since @xmath41 does not couple to the quark superfields , and only some specific high - energy completions of our model would generate it at the ( less dangerous ) loop level .
let us now discuss the resulting neutrino mixing pattern . using the @xmath0 contraction rules
, one finds that the mass matrix @xmath42 is proportional to the identity matrix .
the dirac neutrino mass matrix @xmath43 and the matrix @xmath44 have the structures @xmath45 when all the scalar fields take vev in the @xmath46 direction . in this case , the charged lepton mass matrix fermions .
this leads , however , to numerically irrelevant corrections in the leptonic mixing matrix . ]
is diagonalized by the magic matrix @xmath47 , see eq . .
one can now perform a @xmath47 rotation in order to go to the basis where the charged lepton mass matrix is diagonal . in this basis , the neutrino mass matrix can be written as @xmath48 where the parameters @xmath49 and @xmath50 are functions of @xmath51 . in general @xmath52 . by setting @xmath53
we recover the result of the babu - ma - valle model @xcite . in this limit , the mass matrix in eq .
gives maximal atmospheric angle and degenerate neutrino mass spectrum , while the solar and reactor mixing angles are zero . as observed in @xcite , eq .
is corrected by wave - function renormalizations of @xmath54 , @xmath55 , and @xmath56 , as well as the corresponding vertex renormalizations , lifting the neutrino degeneracy and the solar / reactor mixing angles .
the resulting neutrino mass matrix can be written as @xmath57 where @xmath58 parametrize the radiative corrections . when @xmath53 ( as in @xcite ) and assuming real @xmath58 corrections , the resulting neutrino mass matrix is @xmath59 symmetric giving maximal atmospherc angle and zero @xmath60 reactor angle .
the solar angle is a free parameter and can be fitted .
if the corrections are allowed to be complex , a non - zero value for @xmath60 can be obtained . in this case , cp violation is predicted to be maximal @xcite . regarding the nature of the corrections , these come from flavor mixing in the slepton / sneutrino sector .
a detailed study can be found in @xcite .
note that our model has more freedom since @xmath61 in general .
this can be used to relax some of the restrictions in the parameter space .
nevertheless , large @xmath62 mixing is necessary , typically predicting br(@xmath63 ) close to its experimental limit .
furthermore , in the susy inverse seesaw one expects large z - penguin contributions in lepton flavor violating processes @xcite .
therefore , observables such as br(@xmath64 ) , with @xmath65 , are also expected to set important constraits on the susy parameter space .
finally , violation of lepton flavor universality in observables such as @xmath66 and @xmath67 is also an important test of the model @xcite .
in summary , we have proposed a supersymmetric flavor model based on the @xmath0 symmetry .
all lepton and/or baryon number violating operators of dimension 4 and 5 are forbidden by the flavor symmetry , thus providing a single explanation for the proton and dark matter stabilities .
the extension to account for neutrino masses and mixings turned out to be non - trivial due to the restrictions imposed by the @xmath0 symmetry .
in fact , naive extensions of the leptonic sector typically spoil the nice features of the original model and depart from the main motivations for this work .
we found that one can indeed generate neutrino masses by means of an inverse type - iii seesaw , while preserving r - parity and keeping the proton stable .
the phenomenology of the complete model has been briefly discussed . in particular , the observed neutrino mixing pattern ( with a large reactor angle )
can be well accommodated . before concluding the paper we would like to comment on a difference that our setup has with respect to the canonical mssm : the existence of three higgs doublets .
this feature , the replication of the minimal higgs sector , is shared by all flavor models where the usual higgs doublet is promoted to a multiplet of the flavor group .
this typically implies very complicated scalar potentials , whose minimization leads to viable vacua ( from a phenomenological point of view ) thanks to the soft susy breaking terms .
an example in the context of the @xmath68 symmetry can be found in ref .
although the model under consideration here involves a different flavor symmetry , similar results are expected .
a detailed minimization of the potential and determination of the scalar spectrum is clearly beyond the scope of the paper .
furthermore , if the new scalar / pseudoscalar states can be made relatively light , one may expect sizable higgs - mediated contributions to flavor processes such as @xmath69 , @xmath70 or @xmath71 ( where @xmath72 denotes a neutral pseudoscalar meson ) . ]
@xcite .
the extended higgs sector provides additional freedom to accommodate the observed @xmath1 gev higgs - like resonance found by atlas and cms .
in fact , one can easily find regions in parameter space very similar to the so - called _ decoupling limit _ of the mssm , where the lightest scalar particle behaves as the sm higgs boson .
we note that the non - standard corrections coming from the triplets are typically very small due to the relatively small yukawa couplings ( see for example ref .
@xcite ) .
although most of the phenomenological features of our setup can be found in other flavor models , their combination is quite distinctive .
in particular , the presence of the light @xmath73 @xmath19 fermion triplet leads to a very rich collider phenomenology , since it can be pair produced at the lhc due to its gauge interactions .
this possibility , not present in singlet extensions of the mssm , may lead to very clear signatures with additional leptons and/or lepton flavor violation in the @xmath73 decays @xcite .
it is well - known that the assumption of r - parity in the mssm serves to stabilize the lsp .
here we obtain the same result , _ r - parity _ , without assuming it .
therefore , the usual mssm candidate , a neutralino lsp , can play the role of the dark matter of the universe .
the authors are grateful to werner porod for useful comments on the manuscript .
this work was supported by grants fpa2011 - 22975 , multidark csd2009 - 00064 , prometeo/2009/091 , eu itn unilhc pitn - ga-2009 - 237920 , s.m has been also supported by dfg grant wi 2639/4 - 1 . a.v .
acknowledges support by the anr project cpv - lfv - lhc nt09 - 508531 .
e. gildener , phys .
d * 14 * ( 1976 ) 1667 .
g. r. farrar and p. fayet , phys .
b * 76 * ( 1978 ) 575 .
s. weinberg , phys .
d * 26 * ( 1982 ) 287 ; n. sakai and t. yanagida , nucl .
b * 197 * ( 1982 ) 533 ; l. e. ibanez and g. g. ross , nucl .
b * 368 * ( 1992 ) 3 .
nath and p. fileviez perez , phys . rept .
* 441 * ( 2007 ) 191 [ hep - ph/0601023 ] .
r. n. mohapatra , phys .
d * 34 * ( 1986 ) 3457 ; s. p.
martin , phys .
d * 46 * ( 1992 ) 2769 [ hep - ph/9207218 ] .
r. n. mohapatra and j. w. f. valle , phys .
d * 34 * ( 1986 ) 1642 .
e. ma , mod .
a * 24 * ( 2009 ) 2491 [ arxiv:0905.2972 [ hep - ph ] ] ; d. ibanez , s. morisi and j. w. f. valle , phys .
d * 80 * ( 2009 ) 053015 [ arxiv:0907.3109 [ hep - ph ] ] ; f. -x .
josse - michaux and e. molinaro , arxiv:1210.7202 [ hep - ph ] . c. luhn , s. nasri and p. ramond , j. math .
phys . * 48 * , 073501 ( 2007 ) [ hep - th/0701188 ] ; h. ishimori , t. kobayashi , h. ohki , h. okada , y. shimizu and m. tanimoto , lect .
notes phys .
* 858 * ( 2012 ) 1 .
h. ishimori , t. kobayashi , h. ohki , y. shimizu , h. okada and m. tanimoto , prog .
* 183 * ( 2010 ) 1 [ arxiv:1003.3552 [ hep - th ] ] .
d. piriz and j. wudka , phys .
d * 56 * ( 1997 ) 4170 [ hep - ph/9707314 ] . c. d. carone , l. j. hall and h. murayama , phys .
d * 54 * ( 1996 ) 2328 [ hep - ph/9602364 ] .
h. k. dreiner , c. luhn and m. thormeier , phys .
d * 73 * ( 2006 ) 075007 [ hep - ph/0512163 ] ; h. m. lee , s. raby , m. ratz , g. g. ross , r. schieren , k. schmidt - hoberg and p. k. s. vaudrevange , phys .
b * 694 * ( 2011 ) 491 [ arxiv:1009.0905 [ hep - ph ] ] .
i. de medeiros varzielas , s. f. king and g. g. ross , phys .
b * 648 * ( 2007 ) 201 [ hep - ph/0607045 ] .
k. s. babu , b. dutta and r. n. mohapatra , phys .
d * 60 * ( 1999 ) 095004 [ hep - ph/9812421 ] .
l. lavoura and h. kuhbock , eur .
j. c * 55 * , 303 ( 2008 ) [ arxiv:0711.0670 [ hep - ph ] ] .
g. bhattacharyya , i. de medeiros varzielas and p. leser , arxiv:1210.0545 [ hep - ph ] .
e. ma , phys .
d * 80 * ( 2009 ) 013013 [ arxiv:0904.4450 [ hep - ph ] ] ; p. s. b. dev and r. n. mohapatra , phys . rev .
d * 81 * ( 2010 ) 013001 [ arxiv:0910.3924 [ hep - ph ] ] ; f. bazzocchi , phys . rev .
d * 83 * ( 2011 ) 093009 [ arxiv:1011.6299 [ hep - ph ] ] ; a. g. dias , c. a. de s.pires , p. s. rodrigues da silva and a. sampieri , phys . rev .
d * 86 * ( 2012 ) 035007 [ arxiv:1206.2590 [ hep - ph ] ] .
k. s. babu , e. ma and j. w. f. valle , phys .
b * 552 * ( 2003 ) 207 [ hep - ph/0206292 ] .
m. hirsch , j. c. romao , s. skadhauge , j. w. f. valle and a. villanova del moral , phys .
d * 69 * ( 2004 ) 093006 [ hep - ph/0312265 ] .
m. hirsch , f. staub and a. vicente , phys .
d * 85 * ( 2012 ) 113013 [ arxiv:1202.1825 [ hep - ph ] ] ; m. hirsch , w. porod , l. reichert and f. staub , phys .
d * 86 * ( 2012 ) 093018 [ arxiv:1206.3516 [ hep - ph ] ] ; a. abada , d. das , a. vicente and c. weiland , jhep * 1209 * ( 2012 ) 015 [ arxiv:1206.6497 [ hep - ph ] ] .
a. abada , d. das , a. m. teixeira , a. vicente and c. weiland , arxiv:1211.3052 [ hep - ph ] .
m. hirsch , s. morisi , e. peinado and j. w. f. valle , phys .
d * 82 * ( 2010 ) 116003 [ arxiv:1007.0871 [ hep - ph]];m .
s. boucenna , m. hirsch , s. morisi , e. peinado , m. taoso and j. w. f. valle , jhep * 1105 * ( 2011 ) 037 [ arxiv:1101.2874 [ hep - ph ] ] .
f. bazzocchi , s. morisi , e. peinado , j. w. f. valle and a. vicente , jhep * 1301 * ( 2013 ) 033 [ arxiv:1202.1529 [ hep - ph ] ] .
a. dedes , j. r. ellis and m. raidal , phys .
b * 549 * ( 2002 ) 159 [ hep - ph/0209207 ] ; k. s. babu and c. kolda , phys .
* 89 * ( 2002 ) 241802 [ hep - ph/0206310 ] ; a. abada , d. das and c. weiland , jhep * 1203 * ( 2012 ) 100 [ arxiv:1111.5836 [ hep - ph ] ] .
r. franceschini , t. hambye and a. strumia , phys .
d * 78 * , 033002 ( 2008 ) [ arxiv:0805.1613 [ hep - ph ] ] ; f. del aguila and j. a. aguilar - saavedra , nucl .
b * 813 * ( 2009 ) 22 [ arxiv:0808.2468 [ hep - ph ] ] ; f. del aguila and j. a. aguilar - saavedra , phys . lett .
b * 672 * ( 2009 ) 158 [ arxiv:0809.2096 [ hep - ph ] ] .
a. elsayed , s. khalil and s. moretti , phys .
b * 715 * ( 2012 ) 208 [ arxiv:1106.2130 [ hep - ph ] ] .
|
proton stability is guaranteed in the mssm by assuming a discrete symmetry , r - parity . however , there are additional r - parity conserving higher dimensional operators which violate lepton and baryon numbers and induce fast proton decay . here
we study the possibility that all renormalizable , as well as the most dangerous non - renormalizable , r - parity violating operators are forbidden by a flavor symmetry , providing a common origin for fermion mixing and proton and dark matter stability .
we propose a specific model based on the @xmath0 discrete symmetry .
|
the famous st . petersburg game is easily formulated as a simple coin tossing game .
the player s gain @xmath8 in a single game can be expressed by means of the stopping time @xmath9 of repeated independent tosses @xmath10 of a fair coin until it first lands heads . for a sequence of gains
@xmath11 in independent st .
petersburg games the partial sum @xmath12 denotes the total gain in the first @xmath1 games . to find a fair entrance fee for playing the game
is commonly called the st .
petersburg problem , frequently raised to the status of a paradox .
since the expectation @xmath13=\infty$ ] is infinite , a fair premium can not be constructed by the help of the usual law of large numbers .
we refer to jorland @xcite and dutka @xcite for the history of the st .
petersburg game and for early solutions of the 300 year old problem .
the first step towards a mathematically satisfactory solution has been achieved by feller @xcite who showed that a time - dependent premium can fulfill a certain weak law of large numbers @xmath14 where @xmath15 denotes logarithm to the base @xmath16 .
however , feller s result does not tell if the game is dis- or advantageous for the player , i.e. if @xmath17 is likely to be negative or positive .
this question can only be answered by a weak limit theorem and the first theorem of this kind has been shown by martin - lf @xcite for the subsequence @xmath18 @xmath19 the limit @xmath20 is infinitely divisible with characteristic function @xmath21 , where @xmath22 and the lvy measure @xmath23 is concentrated on @xmath24 with @xmath25 for @xmath26 .
hence @xmath20 is a semistable random variable and the corresponding lvy process @xmath27 with @xmath28 is a ( non - strictly ) semistable lvy process fulfilling the semi - selfsimilarity condition @xmath29 for details on semistable random variables and lvy processes we refer to the monographs @xcite .
the nature of semistability is that there exists in general a continuum of possible limit distributions . for the st .
petersburg game the possible limit distributions have been characterized by csrg and dodunekova @xcite who proved that for any subsequence @xmath30 with @xmath31\ ] ] we have @xmath32 where @xmath33 ; cf . also @xcite .
the object of our study are local fluctuations of the sample paths of the stochastic process @xmath34}$ ] consisting of all the possible weak limits of normalized total gains in repeated st .
petersburg games .
figure [ pb ] shows typical ( approximative ) sample paths of @xmath35}$ ] generated by @xmath36 simulated st .
petersburg games . .
for better visibility the jumps are shown as vertical lines.,title="fig : " ] .
for better visibility the jumps are shown as vertical lines.,title="fig : " ] . for better visibility
the jumps are shown as vertical lines.,title="fig : " ] . for better visibility
the jumps are shown as vertical lines.,title="fig : " ] note that the sample paths do only have upward jumps due to the fact that the lvy measure @xmath23 is concentrated on @xmath24 .
the main goal of our paper is to determine the hausdorff dimension of the range @xmath37)=\{y(t):\,t\in[\frac12,1]\}$ ] and the graph @xmath38)=\{(t , y(t)):\,t\in[\frac12,1]\}$ ] of the stochastic process @xmath39 encoding all the possible distributional limits of st .
petersburg games .
for an arbitrary subset @xmath40 the @xmath41-dimensional hausdorff measure is defined as @xmath42 where @xmath43 denotes the diameter of a set @xmath44 .
it can now be shown that there exists a unique value @xmath45 so that @xmath46 for all @xmath47 and @xmath48 for all @xmath49 .
this critical value is called the hausdorff dimension of @xmath50 .
specifically , we have @xmath51 for details on the hausdorff dimension we refer to @xcite .
an alternative fractal dimension is the so called box - counting dimension ( see , e.g. , @xcite ) .
for this purpose let @xmath52 be the smallest number of closed balls of radius @xmath53 that cover the set @xmath40 .
the lower and the upper box - counting dimensions of an arbitrary set @xmath40 are now defined as @xmath54 and the box - counting dimension of @xmath50 is given by @xmath55 provided that this limit exists .
the different fractal dimensions are related as follows : @xmath56 note that there are plenty of sets @xmath40 where these inequalities are strict . in section 2
we will determine the hausdorff and box - counting dimension of the range @xmath37)$ ] and the graph @xmath38)$ ] for almost all sample paths of the stochastic process @xmath39 .
additionally , in section 3 we will also consider a deterministic sequence introduced by steinhaus @xcite which is called the `` _ steinhaus sequence _ ''
according to @xcite .
the steinhaus sequence @xmath57 is defined by @xmath58 if @xmath59 for some @xmath60 and @xmath61 . alternatively , as in vardi @xcite , one can define @xmath62 to be twice the highest power of @xmath16 dividing @xmath1 .
the steinhaus sequence is explicitly given by @xmath63 and has relative frequencies @xmath64 for @xmath60 .
the sequence @xmath57 has been considered as time - dependent entrance fees for repeated st .
petersburg games in @xcite and has been proven to be a sequence of nearly asymptotically fair premiums in a certain sense .
for details we refer to @xcite .
in contrast to @xcite we will consider the steinhaus sequence as a sequence of possible gains in repeated st .
petersburg games .
again , we will determine the hausdorff and box - counting dimension of the range and the graph of the specific sample path of @xmath39 resulting as a limiting object of the steinhaus sequence .
to do so , we will employ results for iterated function systems as presented in @xcite .
in this section we evaluate the hausdorff dimension of the range of the stochastic process @xmath65}$ ] . we employ common techniques used to calculate hausdorff dimensions of selfsimilar lvy processes ( see @xcite ) and adapt them to our situation .
note that the given process @xmath39 is neither a lvy process nor does it have the selfsimilarity property of a semistable process .
the result is stated in the theorem below .
[ dimranger ] we have @xmath66)=1 $ ] almost surely .
note that theorem [ dimranger ] together with yields @xmath66)=\dim_{\rm b } y([\frac{1}{2},1])=1 $ ] almost surely .
since @xmath39 is a process on @xmath67 it is obvious that @xmath66)\leq1 $ ] almost surely . for the proof of theorem [ dimranger ] it is hence sufficient to prove the following lemma .
[ dimrangerlow ] we have @xmath66)\geq1 $ ] almost surely .
as mentioned above we can write @xmath68 where @xmath69 is a semistable lvy process . to prove the proposition we will apply frostman s theorem @xcite with the probability measure @xmath70}$ ] , where @xmath71 denotes lebesgue measure .
for this purpose let @xmath72 and note that @xmath73 is an admissible measure for frostman s lemma , i.e. @xmath74 by frostman s theorem it is now sufficient to show that @xmath75ds\,dt < \infty.\ ] ] for @xmath76 $ ] let @xmath77 be a lebesgue density of @xmath78 chosen from the class @xmath79 by proposition 2.8.1 in @xcite .
then we have @xmath80}\sup_{x\in{{\mathbb r}}}\left|g_r(x)\right|<\infty$ ] as in lemma 2.2 of @xcite . by symmetry of the integrand we get @xmath81ds\,dt\\ & = 2 \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{t}{{\mathbb e}}\left[\left|s^{-1}x(s)-\log_2s - t^{-1}\left(x(s)+(x(t)-x(s)))+\log_2t\right)\right|^{-\gamma}\right]ds\,dt\\ & = 2 \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{t}\int_{{{\mathbb r}}^2}\left|s^{-1}x-\log_2s - t^{-1}(x+y)+\log_2t\right|^{-\gamma}g_s(x)\,g_{t - s}(y)\,d\lambda^2(x , y)\,ds\,dt\\ & = 2 \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{t}\int_{{{\mathbb r}}^2}\left|\frac{t - s}{st}\,x+\log_2\left(\frac{t}{s}\right)-\frac{y}{t}\right|^{-\gamma}g_s(x)\,g_{t - s}(y)\,d\lambda^2(x , y)\,ds\,dt\\ & = 2 \int_{\frac{1}{2}}^{1}\int_{0}^{t-\frac{1}{2}}\int_{{{\mathbb r}}^2}\left|\frac{w}{t(t - w)}\,x+\log_2\left(\frac{t}{t - w}\right)-\frac{y}{t}\right|^{-\gamma}g_{t - w}(x)\,g_w(y)\,d\lambda^2(x , y)\,dw\,dt,\end{aligned}\ ] ] where in the last equality we substituted @xmath82 .
now we write @xmath83 $ ] as @xmath84 with @xmath85 and @xmath86 $ ] .
this leads us to @xmath87 using the substitutions @xmath88 and @xmath89 we get @xmath90 where @xmath91 denotes the set @xmath92 .
we now estimate the two integrals separately .
first , @xmath93 and secondly , @xmath94 this leads us to @xmath95 taken all together , we obtain @xmath96ds\,dt \leq 2k\int_\frac{1}{2}^1\int_0^{t-\frac{1}{2}}t^\gamma2^{m(w)\gamma}\,dw\,dt\\ & \quad=2k\int_0^\frac{1}{2}\int_{\frac{1}{2}+w}^1 t^\gamma2^{m(w)\gamma}\,dt\,dw = 2k\sum_{m\in{{\mathbb n}}}\int_{2^{-(m+1)}}^{2^{-m}}\int_{\frac{1}{2}+w}^1 t^\gamma2^{m\gamma}\,dt\,dw\\ & \quad\leq 2k\sum_{m\in{{\mathbb n}}}\int_{\frac{1}{2}}^{1}\int_\frac{1}{2}^1 t^\gamma2^{m\gamma}\,dt \;2^{-m}\,dr = k\sum_{m\in{{\mathbb n}}}\left(2^{\gamma-1}\right)^m\int_{\frac{1}{2}}^{1}t^\gamma\,dt < \infty,\end{aligned}\ ] ] since @xmath97 .
this concludes our proof . in this section
we show that the dimension result for the range of the stochastic process @xmath39 also holds for its graph @xmath98)$ ] .
we will split the proof into two parts , first verifying @xmath99 as an upper bound and secondly as a lower bound for the hausdorff dimension of the graph .
+ we first calculate the upper bound for the hausdorff dimension of the graph of the semistable lvy process @xmath20 and later on transfer the result to the process @xmath39 . as @xmath20 is not strictly semistable we ca nt use the dimension results of @xcite , without modifying it according to our situation .
note that for the strictly stable ( symmetric ) cauchy process on @xmath67 it is known that the hausdorff dimension of the range coincides with those of an asymmetric ( non - strictly ) stable cauchy process ; see @xcite .
[ dimxup ] let @xmath100
. then almost surely @xmath101)\leq 1.\end{aligned}\ ] ] let @xmath102 denote the sojourn time of the lvy process @xmath103 up to time @xmath41 in the closed ball @xmath104 with radius @xmath105 centered at the origin . to prove theorem [ dimxup ] we need the following lemma .
[ sojour ] let @xmath103 be the stochastic process from above .
there exists a positive and finite constant @xmath106 such that for all @xmath107 and @xmath108 we have @xmath109\geq ka.\end{aligned}\ ] ] fix @xmath107 and let @xmath110 , to be specified later , so that @xmath111 for some @xmath112 .
we have @xmath113=\int_{0}^{s}\mathbb{p}\left(\|z(t)\|<a\right)dt \geq \int_{0}^{s}\mathbb{p}\left(|x(t)|<\frac{a}{\sqrt{2 } } , t<\frac{a}{\sqrt{2}}\right)dt\\ & = \int_{0}^{\frac{a}{\sqrt{2}}}\mathbb{p}\left(|x(t)|<\frac{a}{\sqrt{2}}\right)dt \geq \int_{0}^{2^{-i_0}}\mathbb{p}\left(|x(t)|<\frac{a}{\sqrt{2}}\right)dt\\ & = \sum_{i = i_0 + 1}^{\infty}\int_{2^{-i}}^{2^{-i+1}}\mathbb{p}\left(|x(t)|<\frac{a}{\sqrt{2}}\right)dt = \sum_{i = i_0 + 1}^{\infty}2^{-i}\int_{1}^{2}\mathbb{p}\left(|x(2^{-i}r)|<\frac{a}{\sqrt{2}}\right)dr\\ & = \sum_{i = i_0 + 1}^{\infty}2^{-i}\int_{1}^{2}\mathbb{p}\left(|2^{-i}(x(r)-ir)|<\frac{a}{\sqrt{2}}\right)dr\\ & = \sum_{i = i_0 + 1}^{\infty}2^{-i}\int_{1}^{2}\mathbb{p}\left(|x(r)-ir|<\frac{2^ia}{\sqrt{2}}\right)dr.\end{aligned}\ ] ] the probability from above can be estimated from below by @xmath114 if we choose @xmath112 large enough so that @xmath115 for all @xmath116 . as @xmath20 is a lvy process , we can assume that it has cdlg paths and thus both @xmath117 and @xmath118 are random variables
. hence we can choose @xmath110 from above small enough ( i.e. , @xmath112 even bigger ) so that we have @xmath119 note that @xmath53 does not depend on @xmath105 .
it follows that @xmath109\geq \sum_{i = i_0 + 1}^{\infty}2^{-i}\int_{1}^{2}\frac12\,dr = \frac12\sum_{i = i_0 + 1}^{\infty}2^{-i } = \frac12 2^{-i_0}\geq \frac12\delta a=:ka,\end{aligned}\ ] ] which concludes the proof .
let @xmath120 be a fixed constant .
a family @xmath121 of cubes of side @xmath122 $ ] in @xmath123 is called @xmath124-nested if no ball of radius @xmath105 in @xmath123 can intersect more than @xmath124 cubes of @xmath121 . for any @xmath125
let @xmath126 be the number of cubes hit by the lvy process @xmath103 at some time @xmath127 $ ] .
then a famous covering lemma of pruitt and taylor ( * ? ? ?
* lemma 6.1 ) states that @xmath128\leq2k_1s\cdot({{\mathbb e}}[t(\tfrac{a}3,s)])^{-1}.\ ] ] lemma [ sojour ] now enables us to construct a covering of @xmath129)$ ] whose expected @xmath41-dimensional hausdorff measure is finite for every @xmath130 .
the arguments are in complete analogy to the proof of part ( i ) of lemma 3.4 in @xcite and thus omitted . in order to transfer the result of theorem [ dimxup ] to the process @xmath39
we can now write all elements @xmath131)$ ] as @xmath132 it can easily be shown that for a fixed constant @xmath133 the function @xmath134\times[-c , c]\rightarrow { \operatorname{im}}(t ) , \quad\left(\begin{array}{c } t \\ x \end{array}\right)\mapsto t(t , x ) = \left(\begin{array}{c } t \\
t^{-1}x-\log_2 t \end{array}\right)\end{aligned}\ ] ] is bi - lipschitz .
since @xmath20 is a lvy process , it can be assumed that all paths are cdlg and hence that for all fixed @xmath135 there exists a constant @xmath136 such that @xmath137 \quad \text { for all } t\in[\tfrac{1}{2},1].\end{aligned}\ ] ] this means that for @xmath138}$ ] and all @xmath135 we have @xmath101)(\omega)=\dim_{\rm h } t(z([\tfrac{1}{2},1]))(\omega)=\dim_{\rm h } g_y([\tfrac{1}{2},1])(\omega)\end{aligned}\ ] ] by lemma 1.8 in @xcite .
since we have shown in theorem [ dimxup ] that @xmath139)\leq1 $ ] almost surely , we have thus proven the following upper bound .
[ dimhgup ] we have @xmath140)\leq1 $ ] almost surely . to prove the lower bound for the hausdorff dimension of the graph we can use the same technique as for the lower bound in case of the range of @xmath39 .
[ dimhglow ] we have @xmath140)\geq1 $ ] almost surely .
let @xmath72 .
by we get @xmath141ds\,dt\\ & \quad= \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{1}{{\mathbb e}}\left[\left((s - t)^2+(y(s)-y(t))^2\right)^{-\frac{\gamma}{2}}\right]ds\,dt\\ & \quad\leq \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{1}{{\mathbb e}}\left[|y(s)-y(t)|^{-\gamma}\right]ds\,dt<\infty.\end{aligned}\ ] ] the rest of the proof runs exactly as the proof of lemma [ dimrangerlow ] . with similar techniques it is also possible to proof the following dimension result for the box - counting dimension of the graph of the st .
petersburg process @xmath39 .
[ dimbg ] we have @xmath142)=1 $ ] almost surely .
the lower bound follows directly from the almost sure inequalities @xmath143)\leq\operatorname{\underline{dim}}_{\rm b}g_y([\tfrac{1}{2},1])\leq\operatorname{\overline{dim}}_{\rm b}g_y([\tfrac{1}{2},1]).\ ] ] for the upper bound it is now sufficient to verify @xmath144)\leq1 $ ] almost surely . with similar arguments as in the proof of theorem [ dimxup ]
we can show that @xmath145)\leq1 $ ] almost surely ; see also the proof of lemma 3.5 in @xcite .
with the bi - lipschitz invariance of the upper box - counting dimension ( see section 3.2 in @xcite ) the proof concludes .
if one prefers to flip an unfair coin this naturally leads to so called generalized st .
petersburg games as treated in @xcite .
let @xmath146 be the probability of the coin falling heads and let @xmath147 .
then a gain of @xmath148 in a single st .
petersburg game results in the limit theorem @xmath149 in distribution , whenever @xmath150,\ ] ] where @xmath27 is a semistable lvy process with the semi - selfsimilarity property @xmath151 we emphasize that with the above techniques our theorems [ dimranger ] , [ dimhgup ] , [ dimhglow ] and [ dimbg ] also hold for the process @xmath152}$ ] in this generalized situation when replacing the interval by @xmath153 $ ] .
recall the definition of the steinhaus sequence @xmath57 given in the introduction .
the asymptotic properties of @xmath57 have been analyzed in full detail by csrg and simons @xcite .
let @xmath154 and @xmath155 $ ] then by theorem 3.3 in @xcite we have for any @xmath156 @xmath157 where the function @xmath158\to[0,2]$ ] is defined by @xmath159 and the sequence @xmath160 is given by the dyadic expansion @xmath161 of @xmath162 $ ] with the convention that @xmath163 for infinitely many @xmath60 . by theorem 3.1 in @xcite the function @xmath164 is cdlg with @xmath165 and has jumps precisely at the dyadic rationals in @xmath166 $ ] .
all these jumps are upward and the largest jump occurs from @xmath167 to @xmath168 .
the graph of @xmath164 seems to inhere fractal properties as can be seen in figure [ gxi ] below , a replication of figure 1 in @xcite .
it follows directly from that the sequence @xmath170 of total gains satisfies the asymptotic property of feller @xmath171 as @xmath172 ; see @xcite .
moreover , for any sequence @xmath173 with @xmath174 $ ] we get from @xmath175 where @xmath176 denotes the set of accumulation points .
hence we may consider the function @xmath164 as the corresponding limiting sample path of @xmath35}$ ] .
note that shows that the steinhaus sequence is an exceptional sequence of gains when considering almost sure limit behavior , since feller s law of large numbers does not hold in an almost sure sense . according to classical results in @xcite
it is known that @xmath177 more precisely , by corollary 1 in @xcite we have @xmath178 $ ] almost surely , but there is a version of the strong law of large numbers by @xcite when neglecting the largest gain @xmath179 a comparison of and shows that the steinhaus sequence belongs to an exceptional nullset concerning almost sure limit behavior of the total gain in repeated st .
petersburg games .
we will now show that the steinhaus sequence is not exceptional concerning the local fluctuations of the limiting sample paths measured by the hausdorff or box - counting dimension .
it follows directly from the above stated properties of @xmath164 given in theorem 3.1 of @xcite that the range @xmath180)$ ] is equal to the interval @xmath181 $ ] and hence @xmath182)=1 $ ] by theorem 1.12 in @xcite .
this shows that @xmath182)$ ] coincides with the hausdorff dimension of the range of a typical sample path of @xmath35}$ ] .
clearly , by we also have @xmath182)=\dim_{\rm b}\xi([\frac12,1])=1 $ ] . a look at figure [ gxi ]
suggests that it is merely the graph and not the range of @xmath164 that should inhere fractal properties . in the sequel
we will argue that also the graph @xmath183)$ ] is typical concerning the almost sure dimension properties of the sample graph of @xmath35}$ ] . to this aim
we will again apply the bi - lipschitz function @xmath184 from section 2 whose inverse is given by @xmath185\times[0,2]\to{\operatorname{im}}t^{-1}$ ] with @xmath186 .
applied to the graph of @xmath164 we get for any @xmath162 $ ] @xmath187 and by bi - lipschitz invariance we have @xmath188)=\dim _ { \rm h}t^{-1}(g_\xi([\tfrac12,1])).\ ] ] the same equality holds for upper and lower box - counting dimensions ; e.g. , see @xcite .
the image @xmath189 is illustrated in figure [ gtxi ] and shows perfect selfsimilarity .
to see this , we may write @xmath190 with @xmath191 . for the dyadic expansion @xmath195 of @xmath193 we necessarily have @xmath196 .
consequently , @xmath197 it follows that @xmath198 this shows @xmath199 and furthermore we get @xmath200 concluding the proof .
let @xmath201\times[0,\frac12]\to[\frac12,1]\times[0,\frac12]$ ] be the affine contractions given by @xmath202 then it follows from lemma [ selfsimss ] that for any @xmath193 @xmath203 these contraction properties are illustrated in figure [ congtxi ] and show that the image @xmath189 can be generated by an iterated function system . by hutchinson @xcite there exists a unique non - empty compact set @xmath204\times[0,\frac12]$ ] , called the attractor , such that @xmath205 which fulfills @xmath206\times[0,\tfrac12]).\ ] ] our construction shows that for @xmath193 with dyadic expansion @xmath195 we have @xmath196 and @xmath207\times[0,\tfrac12 ] ) , t^{-1}(\gamma,\xi(\gamma))\big)\to0\ ] ] as @xmath208 , where @xmath209 for @xmath210 and @xmath211 .
since we required @xmath163 for infinitely many @xmath60 , the only limit points missing are those with @xmath212 for all but finitely many @xmath213 . for these we have @xmath214\times[0,\tfrac12 ] ) , t^{-1}(\gamma,\xi(\gamma-))\big)\to0\ ] ] for a dyadic rational @xmath215 $ ] .
the above arguments show that @xmath50 is the closure of @xmath189 and since the dyadic rationals are countable , by elementary properties of the hausdorff dimension and we get @xmath216).\ ] ] the same equality holds for upper and lower box - counting dimensions ; e.g. , see @xcite . a common way to calculate the fractal dimension of the self - affine invariant set @xmath50 is by means of the singular value function . for on overview of such methods we refer to @xcite . the linear part of both affine mappings @xmath217 and @xmath218 is equal to the linear contraction with associated matrix @xmath219 by induction one easily calculates for @xmath220 @xmath221 and the singular values of @xmath222 are the positive roots of the eigenvalues of @xmath223 which calculate as @xmath224 these determine the singular value function of @xmath222 for @xmath220 given by @xmath225 now the affinity dimension of @xmath50 is defined by @xmath226 and the special form of the singular values in shows that @xmath227 . since the union @xmath205 is disjoint , by proposition 2 in@xcite we get a lower bound for the hausdorff dimension of @xmath50 @xmath228 again , by induction one easily calculates for @xmath220 @xmath229 and the singular values of @xmath230 are @xmath231 which shows that @xmath232 by .
since by @xcite we have @xmath233 altogether the above calculations show : this shows that the graph of @xmath164 , being the limiting object of the steinhaus sequence ( considered as a possible sequence of total gains in repeated st .
petersburg games ) , is not exceptional concerning the hausdorff or box - counting dimension of the sample graph @xmath38)$ ] calculated in section 2 .
csrg , s. ; and dodunekova , r. ( 1991 ) limit theorems for the petersburg game . in : m.g .
hahn et al .
_ sums , trimmed sums and extremes .
_ progress in probability , vol .
23 , birkhuser , boston , pp .
285315 .
falconer , k.j .
( 2013 ) dimension of self - affine sets : a survey . in : j. barral and s. seuret ( eds . ) _ further developments in fractals and related fields .
_ trends in mathematics , vol .
13 , birkhuser , basel , pp . 115134 .
|
let @xmath0 be the total gain in @xmath1 repeated st .
petersburg games .
it is known that @xmath2 converges in distribution to a random element @xmath3 along subsequences of the form @xmath4 with @xmath5 and @xmath6 $ ] .
we determine the hausdorff and box - counting dimension of the range and the graph for almost all sample paths of the stochastic process @xmath7}$ ] .
the results are compared to the fractal dimension of the corresponding limiting objects when gains are given by a deterministic sequence initiated by hugo steinhaus .
|
the discovery and confirmation of terrestrial bodies orbiting other stars ( e.g. @xcite ) has generated enormous public as well as scientific interest .
it has shown that after a mere two decades of exoplanetary research finding potentially habitable worlds around other stars seems to be almost within our grasp .
close - by stars and stellar systems are thereby premium targets , as they tend to offer reasonable signal to noise ratios ( snrs ) for photometry and radial velocity as well as comparatively large astrometric amplitudes @xcite . as more than half of the stars in the solar neighborhood are members of binary or multiple systems @xcite , it is not surprising that more than 70 planets in or around binary - stars have been discovered @xcite despite the current observational focus on single star systems . even - though nasa s _ kepler _ mission has been quite successful in finding circumbinary planets ( e.g. @xcite ) we will focus on binary - star systems with potential _
habitable zones ( hzs ) in this study .
in fact , most of the planets discovered in double stars are in these so - called s - type configurations @xcite , where the planet orbits one star only .
the telluric companion of @xmath0 cen b is such an example @xcite .
an interesting question in this regard is doubtlessly : can s - type binary stars harbor habitable worlds ?
already @xcite and @xcite and more recently @xcite investigated the effects such configurations have on the insolation hypothetical planets would receive .
@xcite ( in the following referred to as eg12 ) were able to derive analytic expressions to find hzs in binary star systems unifying dynamical and radiative balance models for s - type binary star - planet systems .
while the exact manner in which planets form in tight binary star systems is still hotly debated in astrophysical literature - see for instance @xcite and references therein , the discovery of has made the existence of terrestrial planets in s - type binary star systems an observational fact .
opinions still differ on whether it is theoretically possible that planets in @xmath0 centauri s hzs can form on stable orbits . even though classical n - body simulations with best case accretion scenarios seem to be able to produce terrestrial planets near the hzs of the @xmath0 centauri system @xcite
, @xcite concluded that even when gas drag is included the encounter velocities between kilometer sized planetesimals would lead to erosive collisions , thus making constant accretion unlikely .
however , in their model they did not include a self - consistent evolution of the gas disc , nor did they consider planetesimal self gravitation or re - accretion of collisional debris .
@xcite used a self consistent disc model with planetesimals .
they were able produced accretion friendly scenarios when the collision frequency was sufficiently high to prevent orbital dephasing .
other possible solutions to the problem of high encounter velocities range from including planetesimal and embryo migration @xcite over mild inclination of planetesimal discs with respect to the binary s orbit @xcite to more realistic radiative modeling of the system s gaseous disc @xcite .
@xcite show that even if additional earth - like planets in @xmath0 centauri do exist , it is not an easy task to find them given current observational limitations in radial velocity resolution .
the rv signal semi amplitude of @xmath0 cen b b was near the current edge of feasibility with , whereas accuracies lower than @xmath2 cm / s would be necessary to discover telluric planets in @xmath0 cen b s hzs .
astrometry is not much more helpful in this case , as the necessary astrometric amplitudes to detect habitable worlds in @xmath0 centauri will only be available near the end of the gaia mission s lifetime @xcite . in this study
we tackle the question whether there are s - type systems in the solar neighborhood that might make for easier targets . for this purpose
we select 19 s - type binary systems from `` the washington visual double star catalog '' ( wdc ) @xcite with well determined stellar parameters that lie within a distance of 31 pc from the solar system , and calculate hzs for each stellar component using the analytic method presented in @xcite .
we provide estimates on the rv and astrometric ( am ) root mean square ( rms ) signal strengths expected for an earth - like planet orbiting at the borders of a system s hzs .
furthermore , we present likely transit depths for potentially transiting habitable planets in co - planar s - type double star systems . this article is structured as follows : first , we will discuss the selection criteria for the 19 systems investigated ( section [ sec : sel ] ) . after a brief summary of the main factors that determine habitability for terrestrial planets in binary star systems ( section [
sec : hz ] ) , the issue of dynamical stability of planets in such configurations is addressed in section [ sec : stab ] .
our results - tables with hz borders and signal strength estimates - are presented and discussed in sections [ sec : res ] and [ sec : trans ] .
current problems in modeling tidal locking of planets in binary systems are mentioned in section [ sec : tides ] .
a summary ( section [ sec : sum ] ) concludes this study .
we preselected all detached binaries with semimajor axes @xmath3 au using the stellar orbital parameters provided in the wdc , in order to find suitable s - type systems in the solar neighborhood where earth - like planets in hzs could be detectable . hereby , we only considered systems within a distance of @xmath1 pc from our solar system as determined by the hipparchos mission @xcite . together with the prerequisite that the binaries orbital elements had to be available
, the afore mentioned restrictions reduced the number of admissible double star systems to 313 .
furthermore , only double star configurations with known spectal types of both components were used .
peculiar spectra that might have been classified incorrectly by hipparchos like hip 17544 & 73695 were also excluded , narrowing the set of candidate systems from 313 to 35 .
the ultimate selection criterion consisted of calculating the binaries periods using bolometric luminosity derived masses together with the semimajor axes given in wds and comparing them to the observed binary periods . the stellar bolometric luminosities and masses required for this purpose were derived as follows : with the distances available through hipparchos data the systems absolute visual magnitudes could be calculated . in order to assess the bolometric luminosities of the binary sample , we performed bolometric corrections ( bc ) of the absolute visual magnitudes using the polynomial fits by @xcite .
the required effective temperatures were estimated via spectral type and luminosity using the catalog of stellar model atmospheres @xcite .
we then calculated the binaries periods using the masses derived via the mass - luminosity relations given in @xcite . only those systems who s derived periods
did not deviate more than 11% from the observed periods were selected for the final sample .
stellar and orbital parameters for the final set of 19 s - type binary systems are presented in table [ tab1 ] . in the next section
we will briefly discuss the main points on how to determine hzs for these s - type binary systems .
[ cols=">,>,>,>,>,^,^,>,>,^,^,<,<",options="header " , ]
following @xcite & @xcite the rv amplitude a planet causes on its host star is given by @xmath4 where @xmath5 denotes the gravitational constant , @xmath6 and @xmath7 are the host - star s and planet s masses .
the quantities @xmath8 , @xmath9 , @xmath10 and @xmath11 denote the planet s semimajor axis and eccenticity , the system s inclination to the plane of the sky and the planet s argument of pericenter respectively .
the planet s true anomaly is represented by @xmath12 .
we can write the maximum possible radial velocity ( rv ) amplitude caused by a terrestrial planet in a circumstellar orbit around one binary component as follows : @xmath13 the maximum possible eccentricity the planet can acquire due to gravitational interaction with the double star is denoted by @xmath14 @xcite .
expressions for the root mean square ( rms ) values of the rv signal are given as follows @xcite : @xmath15^{1/2}= \frac{\sqrt{\mathcal{g}}\;m_1|\sin{i}|}{\sqrt{2 a ( m_0 + m_1)}}\ ] ] in a similar manner we can use the formalism applied in @xcite to determine maximum astrometric ( am ) signal strengths : @xmath16 where @xmath17 is the observers distance to the observed system , and @xmath18 is the planet - host - star system s mass ratio .
the astrometric rms amplitudes are given by : @xmath19^{1/2}\label{eq : rhorms2}\ ] ] here , @xmath20 is the averaged squared planetary eccentricity .
an analytic expression for @xmath20 can be found in @xcite . for a detailed derivation of all rms and maximum signal amplitudes
see @xcite .
|
binary and multiple systems constitute more than half of the total stellar population in the solar neighborhood @xcite .
their frequent occurrence as well as the fact that more than 70 @xcite planets have already been discovered in such configurations - most noteably the telluric companion of @xmath0 cen b @xcite - make them interesting targets in the search for habitable worlds .
recent studies @xcite have shown , that despite the variations in gravitational and radiative environment , there are indeed circumstellar regions where planets can stay within habitable insolation limits on secular dynamical timescales . in this article
we provide habitable zones for 19 near s - type binary systems from the hipparchos and wds catalogues with semimajor axes between 1 and 100 au .
hereby , we accounted for the combined dynamical and radiative influence of the second star on the earth - like planet . out of the 19 systems presented , 17 offer dynamically stable habitable zones around at least one component .
the 17 potentially habitable systems contain 5 f , 3 g , 7 k and 16 m class stars .
as their proximity to the solar system ( @xmath1 pc ) makes the selected binary stars exquisite targets for observational campaigns , we offer estimates on radial velocity , astrometric and transit signatures produced by habitable earth - like planets in eccentric circumstellar orbits .
[ firstpage ] astrobiology habitable zones binaries
|
according to the hierarchical model , dwarf galaxies are the building blocks of larger structures , consequently their study is crucial to understand galaxy formation and evolution .
although dwarf galaxies are the most numerous objects in the universe , there are some questions that remain apparently unresolved . in fact three of them deal with the same idea about their evolution .
the first question is related to observed mass metallicity relation : do these objects have a metallicity according to what we can expect ? .
in other words : why less massive objects have lower metallicity ?
since @xcite many efforts have been devoted to explain it .
the second question is related to a theoretical idea : are dwarf galaxies able to retain their gas in spite of their shallow potential wells ?
although dwarf galaxies are simpler than bigger ones , they exhibit different structures that avoid a simple gravitation calculus .
we can not find the answer only in the scape velocity .
and the third question is related to their star formation history : are they really young objects ? what have they been doing during the last 13 gyr ?
the main features of the model are : a ) gas outflow is not considered ; b ) there is a continuous infall of primordial gas from the halo to the galaxy c ) the star formation process is continuous , but not constant , so bursts are not taken into account , and d ) a galaxy is modeled only as a single region , so no spatial gradients will be obtained .
the model works with a scenario that begins with a primordial gas sphere .
the gas begins to fall to the central part , controlled by the collapse time ( inverse to the infall rate ) . in this collapse
molecular clouds are formed , and a proportion of them are transformed into stars .
these processes are controlled by different efficiencies .
the star formation process is continuous , but it is not constant , thus the sfr has different values along time . basically the model has two engines : the first one that calculates the chemical evolution , based on the work by @xcite , in its later version of @xcite , and the second one that provides the spectroscopic and photometric evolution by @xcite all the models use the same stellar yields and sn rates . for massive stars we have chosen those of @xcite and for low and intermediate mass stars we have selected the ones by @xcite .
the sn rates have been provided personally by pilar ruiz - lapuente .
the things that change from one model to another one are the efficiencies and the collapse time ( or the infall rate ) . to be consistent ,
we have joined the efficiencies in pairs , in the sense that if there are good conditions to form molecular clouds it can be expected that the conditions would be good for the stars as well .
the results are six different values , from low to high efficiency .
for the collapse time we have selected four values : 8 , 20 , 40 , and 60 gyr .
we have run the grid of models for six different galaxy masses , varying from @xmath0 to @xmath1 , representative for the objects under study .
to be consistent all over this work we will compare our results with a sample of well known galaxies with some prescriptions : a ) they have to be dwarf , b ) it has to be clearly known if they are isolated or belong to a cluster , c ) they have to be gas rich objects ( dsph and de are not considered ) . with all these prescriptions we have selected a small sample of _ good objects _ : some dirr , a few field bcd and some bcd of the virgo cluster .
dirr data are from @xcite . for the field bcd , we have used data of @xcite . for virgo
cluster bcd we have used data of @xcite and @xcite .
[ [ gas - fraction - vs - oxygen - abundance . ] ] gas fraction vs oxygen abundance .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the gas fraction is the ratio between the mass in gas form and the mass in gas and stars , and in a system that evolve as a closed box , obviously the more mass in stars it has the less the gas fraction will be .
but we have to take into account that our model incorporates infall , so the galaxy can increase its stellar mass without an appreciable decrease of the gas . in fig . [ zmu ]
we show the relation between the gas fraction and the metallicity of three models for the most massive galaxy with different star formation efficiencies : extra - low , medium and extra - high .
as can be seen in the figure , dirr lie nearer the models than bcd , except for ngc 1569 , that is considered a _
post - star - burst _ galaxy , in fact some authors consider it as a dirr while others classify it as a bcd .
cluster bcd are spread in a bigger area than isolated galaxies .
only two objects present extreme values : vcc802 and vcc213 .
vcc802 is clearly located in the _ forbidden _ area , the place for objects that have little gas despite their low metallicity .
has this object suffered any interaction during its life ?
this low metallicity can be due to a gas loss ?
the case of vcc213 is different , it has too high metallicity although it has a lot of gas .
this could be an effect of an enriched infall due to an interaction with other companion , or it could be produced by a violent star burst . in both cases
it is difficult to explain these values with models of continuous star formation .
[ [ nitrogen - abundance ] ] nitrogen abundance + + + + + + + + + + + + + + + + + + in the absence of knowledge of the sfh , nitrogen can give some clues about the age of an object , and is generated in low and mainly intermediate mass stars , and owing to their long life , it returns to the ism hundreds of mega - years after oxygen does . in fig .
[ no ] the relation between nitrogen and oxygen is plotted . as in the previous case , we only present the extreme cases , and the rest of the models lay between these two curves .
strong bursting star formation histories would produce high oxygen abundances soon ; in contrast , a low and continuous sfr keeps the oxygen abundance low for a long time ; thus , a large quantity of the primary nitrogen may be ejected.therefore , the upper line corresponds to the case of a smooth star formation history , thus the primary nitrogen can bee seen in the ism , and the slope is very flat ; the opposite situation is represented by the lower line . in this case , the sf process is closer to a burst , and oxygen increases its abundance very soon , so the slope is similar to a secondary behavior .
isolated dirr data are located all over the plot , but taking out ic 1613 most of them are near the model lines .
for example , the magellanic clouds are very close to the extreme sfr line , corresponding to objects with a young population , or with recent star formation .
other dirr galaxies are in the transition area .
once again , ic1613 and vcc802 can not be fitted with the models . both present very high
( n / o ) ratio for their poor oxygen abundance .
if the galaxy loses gas because of a burst , this gas will be oxygen rich , and when the nitrogen goes to the ism , it will find less oxygen than can be expected , and consequently n / o will be high .
vcc213 may have different explanation .
it has high oxygen abundance , so if it presents high ( n / o ) is because it has a lot of nitrogen as well .
this can be reached through two bursts : the first that has enough time to generate the nitrogen , and the second one , and more recent , that increases the oxygen , but has not had time to expel the nitrogen .
[ [ nitrogen - vs - color ] ] nitrogen vs color + + + + + + + + + + + + + + + + + following the idea of @xcite about the relation between ( n / o ) and color , in fig [ col_no ] we present this relation for our models and the data .
solid circles are the final values reached by models , and they would match the data if the age of the objects would be 13.2 gyr , in other cases , the data will be located between the lines , as the majority of dirr do .
this is not the case of isolated bcd galaxies , because they have lower ( n / o ) abundance , so they are located in the lower part of the plot .
probably , models with star formation bursts would be needed to match them .
vs ( n / o ) relation .
dashed line corresponds to the relation obtain by @xcite .
grey dots are the final points of each model.,width=377 ] cluster bcd have a different behavior from that of isolated ones and , as can be seen in the plot , they fit the models better .
nevertheless vcc213 is bluer than the general tendency .
this can be due to a second burst population .
this population has to be very young , since it has not had time to eject its nitrogen , so its color should be blue . [
[ luminosity - metallicity - relation ] ] luminosity - metallicity relation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + finally , in fig [ lboh ] we show the luminosity - metallicity relation , as a way to estimate the @xcite relation . in order to have a clearer plot , we have separated dirr data ( in the upper panel ) from bcd ( in the lower panel ) .
lines show time evolution of the relation , and solid circles correspond to the final values .
the dashed line shows the relation found by @xcite for nearby dwarf galaxies with [ oiii ] lines detected . in the case of dirr ,
the slope shown by the data is flatter than what we obtain with the models , wich in fact , do not reproduce the relation calculated by @xcite .
the models with higher efficiencies present a very flat evolution track .
they reach high oxygen abundances very soon and , as they decrease their sfr along time , the ( o / h ) level remains constant , but this is not a realistic scenario for dwarf irregular galaxies . we find a better fit with models with lower efficiencies , and in these cases , the tracks followed by them are similar to the empirical relation . in bcd we find again different trends in isolated and cluster bcd .
isolated bcd show a slope very similar to dirr , that is flatter than our models , but cluster bcd agree quite well with the final points of the evolution .
we have to consider that only five objects are represented .
more cluster bcd data would be needed to obtain robust conclusions .
from our results of gas , metallicity and photometry , we can conclude that most isolated dirr galaxies can be modeled by continuous star formation processes , if primordial gas infall is considered . in this scenario
most of them seem to be young objects .
this is not the case for bcd , neither isolated nor cluster ones , but it is surprising that more coincidences between models and data have been found for cluster bcd than for isolated ones , mainly for color results .
17 , m. , garca - vargas , m. l. , bressan , a. , & gmez - lvarez , p. 2008 , in astronomical society of the pacific conference series , vol .
390 , pathways through an eclectic universe , ed .
j. h. knapen , t. j. mahoney , & a. vazdekis , 312+
|
we have developed a grid of chemical evolution models applied to dwarf isolated galaxies , using @xcite yields .
the input data enclose different star formation efficiencies , galaxy mass and collapse time values .
the result is a wide collection of solutions that vary from objects with low metallicity and great amount of gas , to those with little gas and high metallicity .
no environmental effects like tidal or galactic winds have been treated , so these objects are expected to be close to field dwarf galaxies , more than cluster ones .
we have studied the time evolution of the abundance of oxygen and nitrogen and the amount of gas , related to their star formation history , as well as the possibility of gas losses by sn winds .
|
a star with sufficient remnant mass ( @xmath0 ) , on completion of its nuclear fuel cycle , must enter the phase of a continuous gravitational collapse . once the nuclear fuel is exhausted gravitational forces become all powerful and hence star s internal pressure can not sustain the equilibrium resulting in a continued collapse @xcite . in the late stages of collapse
the gravitational forces become dominant and the physics of collapse is determined mainly by the theory of general relativity . under quite general and physical situations general relativity predicts that such a collapse must end in a singularity , i.e. , a region of spacetime with extreme curvatures @xcite .
physically one could describe singularity as a region of space with vanishing volume and unbounded gravitational forces .
general relativity , however , does not say anything about the nature or physical properties of such a singularity .
this is partially due to the fact that mathematical structure breaks down preventing analysis at and beyond the singularity .
one could perhaps argue that as collapse progresses and matter is condensed in a region comparable to planck length the quantum physical properties of spacetime would become dominant , thus preventing the formation of singularity . but this picture may not hold also since gravity as a force is very different in its nature in comparison to other forces and has a geometrical interpretation as curvature of spacetime . moreover , despite numerous efforts , a viable quantum theory of gravity is not in sight .
hence for such regions of spacetime , whether relativity theory or quantum physics would determine the physics is still an open question . to fill in the gap in our understanding of spacetime singularities in a mathematical consistent manner , a cosmic censorship conjecture , that all gravitational collapse must end in a black hole was proposed @xcite .
the physical consequence of such a hypothesis is that even before the formation of a singularity a trapped surface develops covering the singularity from the outside world .
hence from a physical point of view singularity is hidden from the outside world .
initial studies in censorship were directed towards formulating the conjecture in a mathematically precise manner which could then possibly be proven @xcite .
this also led to formulation of other conjectures like , hoop conjecture by kip thorne and siefert s conjecture @xcite .
however , extensive studies in collapse with various forms of matter fields have shown that under fairly generic reasonable physical conditions both naked singularity and black holes would form as an end state of collapse , depending on various initial and boundary conditions @xcite .
it is still not very clear how to classify either matter or the initial and boundary conditions in a satisfactory way which would end in either state of singularity ( naked or covered ) .
thus from the studies this far almost all physically reasonable matter fields lead to both naked and covered singularities during collapse ( see , @xcite , and references therein ) .
considerable work has since been done on naked singularities from the point of view of giving counterexamples to cosmic censorship but also on the study of their nature and structure .
having established their existence it is important to study the phenomena of formation of naked singularity from a more astrophysical perspective .
one could look for a possible observable signature of naked singularity distinguishing them from other compact strong gravity objects , like black holes . in the studies carried out this
far the stress has been towards showing that for a naked singularity to be `` observable '' a family of lightlike geodesics must terminate at the singularity @xcite . optical appearance and redshift for such possible radiation
has also been studied @xcite . however , from the point of view of general relativity the first null ray coming out of singularity forms a cauchy horizon ( ch ) , and the spacetime model can not remain valid after its formation . therefore , without any consistent extension of spacetime beyond ch the validity and usefulness of all such geodesic analysis becomes doubtful .
the basic question of the existence of the spacetime structure after the ch is unaddressed ( it is difficult to provide extensions of spacetimes , for example , even for shell - crossing singularities which are gravitationally weak @xcite ) which is of utmost importance if we want to talk about families of geodesics ending at singularity in the past , making it a possible astrophysical source . in this paper
we wish to study the structure of the spacetime from this perspective .
is it possible to connect the two spacetimes before and after with cauchy horizon as the boundary ?
whether the resulting spacetime after the ch has formed can still have the same symmetry ?
does relativity theory allows such continuation of spacetime through ch and whether boundary conditions pose any restrictions ?
furthermore , can these boundary layers carry energy from naked singularity to a distant observer ?
earlier hiscock et al .
has considered a model spacetime in which cauchy horizon ultimately becomes the event horizon of the schwarzschild black hole with non vanishing surface energy density and where it could be visible to observers falling into the blackhole @xcite .
if indeed the formation of a naked singularity is a physical phenomenon then the ch would represent a null surface layer emanating from the naked singularity , and reaching the distant observer separating the two spacetimes .
it has been suggested in various studies that naked singularities may be responsible for various high energy phenomena in our universe ( for example gamma ray bursts etc .
it has also been suggested that in the late stages of collapse , when spacetime shrinks to size of the order of planck length quantum effects would play a dominant role resulting in either a burst of particle creation or preventing the formation of singularity all together @xcite .
our aim in this paper is to examine two examples of naked singularities within the frame work of general relativity and whether this allows such a scenario as emission of a impulsive null wave carrying energy from the naked singularity .
the result of such a study would have manifold implications .
first does there exist a spacetime after the formation of a naked singularity which can be joined satisfactorily together with the original model separated by the null shockwave ( ch ) ?
if such a spacetime exists then whether it allows the existence of outgoing families of geodesics terminating at the singularity in the past .
second , and equally important , question is the structure of the ch itself .
whether this null surface ` boundary layer ' is allowed to carry huge amounts of energy along the null ray to distant observer ? and , if the answer is in affirmative , what is its structure and whether this scenario can be called a valid solution to the einstein equations ?
despite numerous exact solutions of the field equations , very few exact solutions of the field equations exist which can describe a physically reasonable collapsing matter cloud .
in fact , in nearly all the studies of spherically symmetric collapse , the key models are either lematre -tolman - bondi metric ( ltb ) @xcite or the vaidya spacetime @xcite .
both these spacetimes have been well studied , and very well may be the only physically reasonable exact solution available . in all such studies
it has been shown that there are out going null and time like geodesics which terminate at the naked singularity in the past .
the visibility of the singularity in terms of a _ roots equation _ whose roots are tangents to the out - going radial null geodesics with past end points at the singularity
. therefore , problem of relating initial data with end state of collapse is reduced to finding roots of a polynomial equation @xcite .
we would first take up the vaidya spacetime .
existence of naked singularity in this model is well established @xcite . in particular for the case of a imploding shell with a linear mass function @xmath1 for @xmath2 singularity
is known to be a naked singularity .
the metric describing a spherically symmetric vaidya space - time is given by @xmath3 where @xmath4 . for linear mass function case @xmath5 , and the singularity formed at @xmath6 ,
@xmath7 is naked iff : @xmath8 has real and positive root where @xmath9 .
it follows that for @xmath10 the above has two real and positive roots namely @xmath11 , @xmath12 given by once the singularity forms the spacetime below the ch is described by the metric above .
however , if the further analysis of family of geodesics is to be valid than the spacetime beyond ch must also be described by similar metric with ch as the boundary between the two solutions .
if such a collapse scenario is to be called a solution of the field equations , the two vaidya spacetimes separated by the ch ( null hypersurface layer ) must form a smooth solution .
hence , the thin null shell with the stress energy should be matched with two spacetimes before and after .
barrabs and israel ( and poisson ) have analyzed in detail the conditions for immersion of such null surface layers between two general spherically symmetric spacetimes . to implement our model
we follow the prescription of matching across null hypersurface by barrabs and israel @xcite ( see also poisson @xcite ) .
let the two spacetimes separating the first singular light ray ( ch ) be given by @xmath17 before and @xmath18 after ( see fig [ fig : vaidya ] ) .
we can describe the spacetime metric across ch in the following form @xmath19 where @xmath20 and @xmath21 . here
region i and region ii correspond to spacetime before and after formation of cauchy horizon , respectively .
in order to glue these two vaidya spacetimes along the null hypersurface @xmath22 ( ch ) we should have @xmath23,\nonumber \\
\frac{d u_+}{dr } & = & \left.\frac{2}{(1-\lambda_+ x_+)}\right|_\sigma , \end{aligned}\ ] ] in spacetime i , and @xmath24 , \nonumber\\ \frac{d u_-}{dr } & = & \left.\frac{2}{(1-\lambda_- x_-)}\right|_\sigma\end{aligned}\ ] ] in spacetime ii . on the boundary we have from continuity
@xmath25 defining tangent vectors on the cauchy horizon : @xmath26\right|_\sigma , \ ; \mbox{region i } \nonumber \\
k^a\equiv e^a_{(1)}&= & \left.\left [ \frac{2 r}{(r-\lambda_- u_- ) } , 1 , 0 , 0 \right]\right|_\sigma , \ ; \mbox{region ii } \nonumber\end{aligned}\ ] ] and @xmath27 for region i & ii . where @xmath28 are the intrinsic coordinates on @xmath29 and we take @xmath30 to be the parameter of the null generator .
the transverse vectors completing the basis for region i and ii are given by @xmath31\right|_\sigma , \mbox{region i } \nonumber \\ n^a & = & \left.\left[0,-\frac{1}{2}\left(1-\frac{\lambda_- u_-}{r}\right),0,0 \right]\right|_\sigma , \mbox{region ii } \nonumber \end{aligned}\ ] ] satisfying @xmath32 and where @xmath33 .
the transverse curvature @xmath34 and the intrinsic metric of the surface layer ( @xmath35 ) is given by @xmath36 and @xmath37 we find the surface energy density and pressure of the null layer for vaidya case as @xmath38}{4 \pi r^2}= \frac{\lambda_-u_- - \lambda_{+ } u_+}{4 \pi r^2}\ ; , \nonumber \\ p & = & [ h(\lambda , m ) ] \;.\end{aligned}\ ] ] first we note that @xmath39 quantifies jump in pressure across the ch .
since it is transverse component to cauchy surface it does not affect the physics of energy propagating along the ch which is of interest to us here .
if energy is transported along the ch the energy density @xmath40 , of the null layer must be positive definite .
it follows from the continuity of the boundary layer from eqs .
( [ bli ] ) and ( [ blii ] ) that @xmath41}{4 \pi r^2}= \frac{u_+}{\gamma_+}\left[\frac{\sqrt{1 - 8\lambda_+ } - \sqrt{1 - 8\lambda_-}}{8 \pi r^2 } \right]\;.\ ] ] hence , ch can carry energy to either a local or a distant observer . therefore , as a result the rate of collapse slows down ( @xmath42 ) , which results in a net positive energy density on the ch . furthermore
, this surface energy on the ch has a clear physical interpretation . to see this
consider the motion of a freely falling timelike observer ( four velocity @xmath43 ) in vaidya spacetime given in eq .
( [ eq : vaidya ] ) ( analysis of timelike trajectories in vaidya spacetime has been worked out @xcite ) . where @xmath45 is a constant labeling different timelike geodesics and s is the affine parameter .
positive sign solutions terminate at the singularity @xmath46 with a positive definite tangent @xmath15 and hence do not intersect the ch . for all timelike radial observers intersecting the ch we have it follows that in imploding null dust collapse ch can be immersed between two vaidya spacetimes ( with linear mass function ) with the parameter @xmath49 . in case
when @xmath50 the matching across @xmath22 is smooth and no energy is carried along the first ray . in the case otherwise
the rate of collapse slows down and facilitates the positive surface density on the null boundary .
it has been shown that there is a family of out going geodesics which terminate in the past at the singularity with a definite tangent @xmath51 .
the path of such outgoing null geodesics has been calculated earlier ( see @xcite ) and is given by eq .
( [ geo - fam ] ) . the problem can be considerably simplified if we can write the metric in terms of out - going null geodesics . in this representation ch
corresponds to a constant value of one of the coordinates .
let us consider a general spherically symmetric spacetime @xmath52 given by @xmath53 here @xmath54 are functions of @xmath55 and @xmath30 .
null layers given by @xmath56 are out going if @xmath57 and ingoing if @xmath58 . the density and pressure of the null shell surface immersed in the two spacetimes is @xmath59}{4 \pi r^2 } , \nonumber \\ p & = & -\zeta \frac{1}{8\pi } \frac{\partial \psi}{\partial r } .\end{aligned}\ ] ] in order to analyse the case of family of null geodesics let us consider a coordinate transformation for the spacetime given in equation ( 1 ) @xmath60 .
the vaidya metric the for spacetime @xmath61 now becomes @xmath62 + r^2 d \omega^2,\nonumber\ ] ] with metric function @xmath63 @xmath64 where @xmath65 and @xmath66 .
here @xmath67 are outgoing singular geodesics with normal @xmath68 .
hence outgoing singular null layers immersed between the two vaidya spacetimes with different mass functions ( @xmath69 ) have non vanishing surface density @xmath40 and @xmath70 allowing energy to escape .
though for lightlike shells there is no rest frame and therefore @xmath40 and p can not be given an absolute meaning as surface density and pressure nonetheless as rightly pointed out by israel they serve perfectly well to determine the results of measurement by any observer in this regard as shown by israel that for a radially freely falling observer momentum normal to the shell is continuous and the energy density associated with the shell as measured by this radially freely falling observer ( @xmath71 ) $ ] is given by @xmath72 } { 4\pi r^2 } \delta ( \tau)(k^a u_a)\ ] ] and is accompanied by equal energy flux . here
@xmath73 is the equation of @xmath22 .
we briefly consider now another scenario , namely the inhomogeneous dust collapse .
the metric describing a spherically symmetric space - time is given by @xmath74 where @xmath75 , and @xmath76 and @xmath77 are arbitrary functions of @xmath78 and @xmath30 .
the metric in ( [ ltbmetric ] ) has to satisfy field equations which can be put in a form @xmath79 the functions , @xmath80 , @xmath76 and @xmath81 are @xmath82 functions throughout the cloud .
notation ( ) and @xmath83 are used to denote partial differentiation with respect to @xmath30 and @xmath78 .
consider marginally bound case @xmath84 , and @xmath85 .
existence of naked singularity in this case ( @xmath86 ) is well established ( see @xcite and ref . therein ) .
function @xmath87 is interpreted as the mass function and for physical reasons @xmath88 , @xmath89 , and gives mass enclosed in a given shell of co - moving radius @xmath30 .
the ch is a null ray @xmath90 where @xmath91 is lowest of the real and positive root of the algebraic equation the two dust spacetimes @xmath61 separating the first singular light ray ( ch ) be given by respective mass functions , @xmath93 before , and @xmath94 after , where @xmath95 satisfies @xmath96 where @xmath97 is a constant and coefficients can be determined as @xmath98 , @xmath99 and @xmath100 .
an argument similar to one for vaidya model shows that that cauchy horizon can be immersed between these two two different dust solutions with a non vanishing surface energy density given by different values of constants @xmath97 and @xmath101 .
s. w. hawking and g. f. r. ellis , _ the large scale structure of space - time _
( cambridge : cambridge university press , 1973 ) ; c. j. s. clarke , _ the analysis of space - time singularities _ , cambridge lecture notes , ( cambridge : cambridge university press , 1993 ) ; j. m. m. senovilla , gen . relativ . and grav .
, * 30 * , 701 ( 1998 ) .
see for example , r. penrose , _ gravitational radiation and gravitational collapse _
iau symp.(iau symposium no .
64 ) ed c dewitt - morette ( dordrecht : reidel 1974 ) ; r. hegedorn , nuovo cimento a*56 * 1027 ( 1968 ) .
r. m. wald , _ gravitational collapse and cosmic censorship _
[ arxiv : gr - qc/9710068 ] ; r. penrose , j. astrophys .
astr . * 20 * , 233 ( 1999 ) ; t. p. singh , j. astrophys .
* 20 * , 221 ( 1999 ) ; d. christodoulou , class . quantum grav . * 16 * , a23 ( 1999 ) ; s. jhingan and g. magli , [ arxiv : gr - qc/9903103 ] ; p. s. joshi , pramana * 55 * , 539 ( 2000 ) ; t. harada , pramana * 63 * , 741 ( 2004 ) .
p. s. joshi , _ global aspects in gravitation and cosmology _ ( clarendon press , oxford , 1993 ) ; p. s. joshi , _ gravitational collapse and spacetime singularities _ ( cambridge monographs on mathematical physics , cambridge , 2007 ) .
s. w. hawking and j. m. stewart , nucl .
b*400 * , 393 ( 1993 ) ; t. harada , h. iguchi and k. nakao , progress of theoretical physics , * 107 * , 449 ( 2002 ) ; r. goswami , p. s. joshi and p. singh , phys .
lett . * 96 * , 031302 ( 2006 ) .
i. h. dwivedi and p. s. joshi , commun .
phys . * 146 * , 333 ( 1992 ) ; i. h. dwivedi and p. s. joshi , class .
* 14 * , 1223 ( 1997 ) ; i. h. dwivedi and p. s. joshi , class .
* 16 * , 41 ( 1999 ) .
a. papapetrou , in _ a random walk in general relativity _ , ed .
n. dadhich , j. krishna rao , j. v. narlikar and c. v. vishveshwara , ( wiley eastern , new delhi , 1985 ) ; y. kuroda , prog .
* 72 * , 63 ( 1984 ) ; k. lake , phys .
a * 116 * , 17 ( 1986 ) ; i. h. dwivedi and p. s. joshi , class . quantum grav . *
6 * , 1599 ( 1989 ) ; i. h. dwivedi and p. s. joshi , class .
quantum grav .
* 8 * , 1339 ( 1991 ) ; k. lake , phys .
rev . d*43 * , 1416 ( 1991 ) .
|
a distinguishable and observable physical property of naked singular regions of the spacetime formed during a gravitational collapse has important implications for both experimental and theoretical relativity . we examine here whether energy can escape physically from naked singular regions to reach either a local or a distant observer within the framework of general relativity .
we find that in case of imploding null dust collapse scenarios field outgoing singular null geodesics including the cauchy horizon can be immersed between two vaidya spacetimes as null boundary layers with non vanishing positive energy density .
thus energy can transported from the naked singularity to either a local or a distant observer . and
example illustrating that similar considerations can be applied to dust models is given .
|
a prominent paradigm in quantum information processing ( qip ) @xcite is to employ flying qubits ( fqs ) and static qubits ( sqs ) as carriers and registers of quantum information , respectively @xcite .
key to such idea is the ability to write and read - out the information content of a sq by means of a fq . by this , here we mean that efficient quantum state transfer ( qst ) between these two types of qubits must be possible on demand . in this picture
, _ control over memory allocation _ appears a desirable if not indispensable requirement .
for instance , one can envisage the situation where only one or a few sqs are available , e.g. because the remaining ones are encoding some information to save . on the other hand , one may need to carry away only the information saved in certain specific sqs .
alternatively , only a restricted area of the register of sqs may be interfaced with some external processing network where one would like to eventually convey information or from which output data are to be received . in such cases , the ability of _ selecting _ the exact location where the information content of the fq should be uploaded or downloaded is demanded . ideally , according to the schematics in fig .
[ fig1 ] , one would like the fq to reach the specific target sq , then fully transfer its quantum state to this and eventually fly away . evidently , this picture is implicitly based on the assumption that , firstly , the motional degrees of freedom of the fq are in fact fully classical and , secondly , that these can be accurately controlled . despite its simplicity , although interesting research along this line is being carried out mostly through so called surface acoustic waves ( see e.g. ref .
@xcite and references therein ) , such an approach calls for a very high level of control .
if we set within a fully quantum framework , the most natural situation to envisage is the one where the fq , besides bearing an internal spin , moves in a quantum mechanical way and hence propagates as a wavelike object .
such a circumstance substantially complicates the dynamics in that , besides the complex spin - spin interactions , intricate wavelike effects such as multiple reflections between the many sqs occur either .
this appears an adverse environment to accomplish _ selective _ qst : while ideally one would like to focus the fq s wave packet right on the target sq , the former is expected to spread throughout the sqs register .
thereby , not only it is non - trivial what strategy would enable selective qst but even the mere possibility that this could occur can be questioned .
( c ) sqs .
the fq @xmath0 incomes from the left with a given wave vector @xmath1 , undergoes multiple scattering between the sqs and the perfect mirror and eventually moves away from the register with the same @xmath1 . ] in this work , we consider a paradigmatic hamiltonian memory read - out model where the fq propagates along a 1d line comprising a collection of ( fixed ) spatially - separated non - interacting sqs and couples to them via a contact - type spin - spin heisenberg interactions ( see fig .
[ fig2 ] ) .
we start with a single sq and prove that a unitary swap between the itinerant and static spins is unattainable .
the insertion of a perfect mirror along the 1d line , however , makes it possible . at the same time , since the transmission channel is suppressed there is no uncertainty over the final path followed by the fq .
next , we find that even for a pair of sqs this can be achieved with either of the two sqs through an _ ad hoc _ setting of distances and coupling strengths . surprisingly enough
, this means that feynman paths entering multiple reflections can combine so as to effectively decouple one sq while enabling at the same time a unitary swap involving the other one .
even more surprisingly , the working principle behind this phenomenon is such that it is naturally generalized to the case of an arbitrarily large register of sqs , as we rigorously prove .
consider the case where a single memory static qubit @xmath2 lies on the @xmath3-axis close to position @xmath4 . to read - out the quantum information stored in @xmath2 ( or write it there ) a fq @xmath0 is injected along the axis with momentum @xmath1 , say from the left - hand side .
we model the @xmath0-@xmath2 interaction as a contact - type spin - dependent scattering potential having the heisenberg coupling form .
the system hamiltonian can thus be expressed as @xmath5 , where @xmath6 is the momentum operator of @xmath0 ( its mass being set equal to one for simplicity ) and @xmath7 is the coupling potential with associated strength @xmath8-shaped potential is a standard one , and for the present setup it relies on the usually met condition that the fq s wavelength is significantly larger than the characteristic sq size . ] . here
, @xmath3 is the spatial coordinate of @xmath0 while @xmath9 and @xmath10 are the spin operators of qubits @xmath0 and @xmath2 , respectively , i.e. @xmath11 with @xmath12 having eigenvalues @xmath13 ( we set @xmath14 throughout ) .
we ask whether or not , when @xmath0 will emerge from the scattering process , the internal degree of freedom ( i.e. the spin ) of the two qubits have been exchanged according to the mapping @xmath15 where @xmath16 is the ( joint ) input spin state of @xmath0 and @xmath2 , while @xmath17 is the usual swap two - qubit unitary operator exchanging the states of qubits @xmath18 and @xmath19 @xcite .
while there are in fact counterexamples @xcite showing that this is impossible , the scattering process between @xmath0 and @xmath2 can never lead to @xmath20 .
owing to conservation of @xmath21 , this is equivalent to state that the transformation @xmath22 is unattainable . ] , we give next the general proof that such swap operation can not occur .
for this purpose , let us define @xmath23 , where for each qubit , either flying or static , @xmath24 and @xmath25 stand for the eigenstates of @xmath26 with eigenvalues @xmath27 and @xmath28 , respectively ( from now on , we omit particle subscripts whenever unnecessary ) .
state @xmath29 is the well - known singlet , while the triplet subspace is spanned by @xmath30 .
using the identity @xmath31 , where @xmath32 , the interaction hamiltonian can be written as @xmath33 , entailing @xmath34= 0 $ ] @xcite . within the singlet ( triplet )
subspace the effective interaction is thus _ spinless _ and reads @xmath35 [ @xmath36 : the problem is reduced to a scattering from a ( spin - independent ) @xmath37-barrier . for a @xmath37-potential step @xmath38 and a particle incoming with momentum @xmath1 , the reflection and transmission probability amplitudes @xmath39 and @xmath40 , respectively ,
are found through a textbook calculation as @xmath41 where we have introduced the rescaled parameter @xmath42 .
these functions allow to calculate the reflection coefficient for the singlet and triplet sectors as @xmath43 where we have set @xmath44 . evidently , @xmath45 for any @xmath46 .
this is the very reason which forbids one from using the above scattering process for implementing _ any _ unitary gate on the spin degree of freedom of @xmath0 and @xmath2 , hence , in particular , the swap gate ( [ swapgate ] ) enabling perfect writing / read - out of @xmath2 .
observe in fact that , once the orbital degree of freedom of the fq are traced out , the final spin state @xmath47 of the joint system @xmath0-@xmath2 can be related to the initial one @xmath16 ( in general mixed ) through the completely positive , trace - preserving map @xcite @xmath48 where the first contribution refers to the @xmath0-wave component emerging from the right of the 1d line ( transmission channel ) , while the second to the one emerging from the left ( reflection channel ) .
the kraus operators @xcite @xmath49 and @xmath50 describing these two complementary events are provided , respectively , by the transmission and reflection operators of the model , namely @xmath51 where @xmath52 and @xmath53 are the projector operators associated with the singlet and triplet subspaces , respectively , of the @xmath0-@xmath2 system .
notice that in the computational basis @xmath54 ( @xmath55 ) a matrix element @xmath56 yields the probability amplitude that , given the initial joint spin state @xmath57 , @xmath0 is reflected back and the final spin state is @xmath58 @xcite ( an analogous statement holds for @xmath59 ) . via the identities ( [ defrandt1 ] ) and ( [ defrandt2 ] ) one can easily verify that eq .
( [ defrt ] ) immediately entails the proper normalization condition @xmath60 .
furthermore , expressed in this form it is now easy to see why the mapping ( [ mapping1 ] ) is never unitary : in fact for this to happen , @xmath50 and @xmath49 should be mutually proportional , i.e. @xmath61 .
this is impossible since it requires @xmath62 , which can be fulfilled only provided that @xmath63 ( conflicting with @xmath64 proven above ) . a strategy to get around
this hindrance is to insert a _ perfect mirror _ at @xmath65 beyond the sq located at @xmath66 at a distance @xmath67 as sketched in fig .
[ fig2](a ) ( this is inspired by ref .
@xcite , where , however , a somewhat different system was addressed ) .
first of all , such modified geometry suppresses the transmission channel eliminating the uncertainty in the direction along which @xmath0 propagates after interacting with @xmath2 .
specifically , in the presence of the perfect mirror we have @xmath68 and eq .
( [ mapping1 ] ) thus reduces to @xmath69 where now the reflection matrix @xmath70 is always unitary @xmath71 .
more interestingly , eq .
( [ mapping2 ] ) allows for the perfect swap gate ( [ swapgate ] ) to be implemented . to see this ,
observe that since the squared total spin is still a conserved quantity as in the no - mirror case , the problem reduces to a spinless particle scattering from a _ spinless _ barrier @xmath72 and a perfect mirror which , via a simple textbook calculation , gives the reflection amplitude @xmath73/[1+i\gamma(1-e^{2 i k d_1 } ) ] \ ] ] ( recall that @xmath42 ) .
therefore , a reasoning fully analogous to the previous case leads to @xmath74 with @xmath75 observe that @xmath76 is unitary because @xmath77 has unit modulus . to work out the conditions for realizing an @xmath0-@xmath2 swap gate ( [ swapgate ] ) , we use the fact that this unitary can be written as @xmath78 . evidently , @xmath79 can be made coincident with @xmath80 ( up to an irrelevant global phase factor ) if and only if @xmath81 .
this identity is fulfilled provided that @xmath82 and @xmath83 are related to each other according to the function @xmath84 which is plotted in fig .
[ fig3](a ) . in eq .
( [ g(kd0 ) ] ) [ panel ( a ) ] and @xmath85 in eq . ( [ 2qstf1 ] ) [ panel ( b ) ] , which set the conditions for perfect swap between @xmath0 and the static memories .
either function is periodic of period @xmath86 .
note , in particular , that as the optical distance @xmath87 approaches @xmath88 ( @xmath89 ) condition ( [ g(kd0 ) ] ) can be satisfied only in the asymptotic limit of infinite spin - spin coupling .
moreover , there is a threshold @xmath90 [ dashed line in panel ( a ) ] that @xmath82 must exceed to ensure the existence of values of @xmath87 allowing for the implementation of the swap gate . ]
interestingly , @xmath91 means that @xmath82 must exceed the threshold @xmath90 to ensure occurrence of the swap . to summarize , in the presence of a single sq and for a given spin - spin coupling strength , for any @xmath92 [ see fig .
[ fig3](a ) ] there always exists a corresponding coupling constant @xmath93 ensuring the occurrence of the @xmath0-@xmath2 swap .
conversely , as long as @xmath8 is strictly larger than @xmath1 , there are always two distinct values of @xmath87 enabling the perfect swap between @xmath0 and @xmath2 . before concluding this section , we point out that , based on the form of @xmath94 , when the optical distance @xmath87 is an integer multiple of @xmath86 ( i.e. @xmath95 ) the above coefficients reduce to @xmath96 and hence @xmath97 independently of the coupling strength
this situation is indeed equivalent to _ moving the mirror to @xmath2 s location _ : the chance for the fq to be found at such position then vanishes and its spin is thus unable to couple to the sqs .
more in general , the property that two objects whose optical separation is an integer multiple of @xmath86 behave _ as if they were at the same place _ will be exploited repeatedly in this work .
in addition to @xmath2 and the perfect mirror , the setup now comprises a further sq , dubbed @xmath98 , located on the left of 1 at a distance @xmath99 from it as shown in fig .
[ fig2](b ) .
hence , the spin - spin coupling term in @xmath100 now reads @xmath101 where @xmath102 and @xmath103 .
we aim to implement either an @xmath0-@xmath2 or an @xmath0-@xmath98 swap operation , i.e. either the unitary @xmath104 or @xmath105 , respectively ( note that in any case we require one of the two sqs to be unaffected ) .
analogously to the single - sq case , the mirror suppresses the transmission channel and thereby one can define a unitary reflection operator @xmath106 within the 8-dimensional ( 8d ) overall spin space that fully describes the interaction process output . in the spirit of scattering matrices combination via sum over different feynman paths @xcite , the scattering operator @xmath106 results from a superposition of all possible paths , the first of which are sketched in fig .
[ fig4 ] . ) . ]
the overall sum is obtained in terms of a geometric series as @xmath107 where although not shown by our notation , despite it involves qubits @xmath0 and @xmath108 , each reflection or transmission operator on the right - hand side is intended as the extension to the present 8d spin space . also , note that @xmath109 is a function of @xmath110 .
the present setup ensures qst between @xmath0-@xmath2 and @xmath0-@xmath98 , respectively , in the regimes @xmath111 where @xmath112 , while @xmath113/2 $ ] is a periodic function of period @xmath86 plotted in fig .
[ fig3](b ) . condition ( [ 2qstf2 ] ) is easily understood : we have already discussed ( see the previous section ) that when @xmath114 the optical distance between @xmath2 and the mirror is effectively zero , hence it is as if the mirror lied at @xmath115 so as to inhibit the @xmath0-@xmath2 coupling .
we are thus left basically with the same setup as the one in the previous section , which shows that if condition @xmath116 is fulfilled [ cf .
( [ g(kd0 ) ] ) ] then @xmath117 . to prove eq . ( [ 2qstf1 ] ) , which is key to the central findings in this paper , it is convenient to introduce the coupled spin basis arising from the coupling of @xmath118 , @xmath119 and @xmath120 .
we define @xmath121 ( @xmath122 ) and the total spin @xmath123 .
it is then straightforward to check that eq .
( [ v2 ] ) can be expressed as @xmath124 and thus @xmath125= 0 $ ] ( owing to @xmath126= 0 $ ] ) . also , @xmath127= 0 $ ] .
note , however , that neither @xmath128 nor @xmath129 is conserved since @xmath130\neq0 $ ] .
using now the coupling scheme where @xmath118 is first summed to @xmath119 @xcite , the coupled basis reads @xmath131 , where @xmath132 , @xmath133 and @xmath134 are the quantum numbers associated with @xmath128 , @xmath135 and @xmath136 , respectively . as @xmath137 ( singlet and triplet , respectively ) the possible values for @xmath133 are @xmath138 . in the subspace @xmath139 only @xmath140 occurs , while for @xmath141 , @xmath132 can be both 0 and 1
. it should be clear now that given that @xmath133 and @xmath142 are good quantum numbers ( @xmath135 and @xmath136 are conserved ) @xmath106 is block diagonal in the basis @xmath143 : four blocks are 1d , each identified by one of the vectors @xmath144 ; two blocks are instead 2d , each spanned by @xmath145 and labeled by @xmath146 . due to symmetry reasons , for fixed @xmath133
the effective form of @xmath106 in each block is independent of @xmath142 .
let us first begin with the two @xmath141 blocks . in the light of the previous section , for both of them , independently of the value of @xmath142 , we can write @xmath147 , where we have introduced the concise notation @xmath148 . as for @xmath149 ,
one has to solve an effective scattering problem in a 2d spin space in the presence of the spin - dependent potential barrier @xmath150 , where @xmath151 is the quantum number associated with @xmath152 and we have introduced the discrete function @xmath153 ( here , although @xmath154 , we leave such quantum number unspecified for reasons that will become clear later on ) .
such task can be carried out easily , as we show in the appendix .
next , by requiring condition ( [ g(kd0 ) ] ) , which ensures that @xmath155 implements a qst between @xmath0 and @xmath2 by setting @xmath156 , and plugging @xmath109 and @xmath157 into eq .
( [ rf12 ] ) , the matrix elements of @xmath158 in the @xmath141 block @xmath159 are calculated as @xmath160/\delta,\label{r00}\\ r_{11}&=&{- [ } i\tilde{g } ( 2+iq_{s_2}\tilde{g } ) -2(2+i\tilde{g})r_{\rm s}^{(\rm m)}e^{2 i k d_2 } + q_{s_2 } \tilde{g}^2 r_{\rm s}^{(\rm m)2}e^{4 i k d_2 } ] /\delta\label{r11 } , \\
r_{01}&=&r_{10}= 2 i\sqrt{q_{s_2}}\,\tilde{g}(1- r_{\rm s}^{(\rm m)2}e^{4
i k d_2})/\delta , \label{r01}\end{aligned}\ ] ] with @xmath161 , \ ] ] ( for compactness of notation the dependance of @xmath162 on @xmath83 is not shown ) .
to realize an @xmath0-@xmath2 swap , i.e. @xmath163 , @xmath164 and @xmath165 must be eigenstates of @xmath106 with opposite eigenvalues , namely @xmath166 must hold .
thereby , off - diagonal entries @xmath167 must vanish , which yields the condition @xmath168 , i.e. @xmath169/2= h(kd_1)$]/2 $ ] for @xmath89 ( @xmath170 integer ) . all these solutions
are physically equivalent .
lower values of @xmath170 , i.e. @xmath171 , are to be discarded since they would make @xmath172 negative . ] according to our definition of the @xmath173 function ( see above ) . by replacing this into eqs .
( [ r00 ] ) and ( [ r11 ] ) we immediately end up with @xmath174 . since for the 1d blocks @xmath139 ,
as mentioned , @xmath132 can only take value 1 and the same occurs for @xmath175 as is easily seen .
hence , @xmath176 and the interaction hamiltonian is given by @xmath177 , i.e. it is effectively spinless .
it should be clear then that the corresponding entry of @xmath158 , denoted by @xmath178 , can be found from eq .
( [ rf12 ] ) through the replacements @xmath179 and @xmath180 ( see the previous section ) .
the formerly introduced condition @xmath181 immediately yields @xmath182 ( matching the value found for @xmath183 as it must be given that they both correspond to @xmath140 ) .
this demonstrates that , up to an irrelevant global phase factor , the @xmath0-@xmath2 swap indeed occurs under condition ( [ 2qstf1 ] ) .
it is important to stress that this result is _ independent _ of the value taken by @xmath184 . in other words ,
the same result is achieved by replacing @xmath185 with @xmath186 with arbitrary @xmath187 .
we now address the case where an arbitrary number @xmath188 of sqs are present , the @xmath189th one lying at @xmath190 in a way that @xmath191 is the distance between the @xmath189th and @xmath192th ones [ see fig . [ fig2](c ) ] .
hence , now @xmath193 again , we aim at implementing a selective swap between @xmath0 and @xmath194 ( @xmath195 ) .
selective qst is achieved for @xmath196 where @xmath197 can be any positive integer .
regime ( [ n ] ) is immediately explained since it entails that @xmath198 , namely the distance between @xmath199 and the mirror , is a multiple integer of @xmath86 , hence the mirror behaves as if it lied at @xmath200 .
all the static qubits from @xmath2 to @xmath199 are thus decoupled from @xmath0 .
we in fact retrieve the case of one sq at a distance @xmath201 from the mirror , where qst is ensured by condition ( [ g(kd0 ) ] ) ( with the replacement @xmath202 ) .
the case in eq .
( [ alpha ] ) is explained as follows .
the mirror is effectively positioned at @xmath203 since each @xmath204 is a multiple integer of @xmath86 . on the other hand , @xmath205 holds as well : the static qubits indexed by @xmath18 such that @xmath206 behave as if they were _ all _ located at @xmath207 .
thereby , effectively @xmath208 ( subject to a hard - wall boundary condition at @xmath209 .
let @xmath210 be the total spin of the @xmath211 sqs effectively located at @xmath212 and @xmath213 the quantum number associated with @xmath214 . for @xmath215 even , @xmath216 ,
while for @xmath215 odd @xmath217 .
as , clearly , @xmath213 is a good quantum number , in each subspace of fixed @xmath213 an effective static spin-@xmath213 particle lies at @xmath212 particle , in our case a given value of @xmath213 can exhibit degeneracies ( e.g. for @xmath218 the value @xmath219 is two - fold degenerate ) . yet , such degeneracies do not play any role here and can in fact be ignored . ] . by coupling this spin to @xmath0 and @xmath194 ,
we find that the total quantum number can take values @xmath220 ( we can assume @xmath221 since the case @xmath222 has been analyzed in the previous section ) . among these ,
only @xmath223 is degenerate since in the corresponding eigenspace either @xmath224 or @xmath225 can take _ two _ possible values , i.e. @xmath226 and @xmath227 ( @xmath228 is the quantum number associated with @xmath229 ) .
the reflection matrix for the system is thus block - diagonal , where each block corresponding to either @xmath230 or @xmath231 is 1d , while a block corresponding to @xmath232 is 2d . in the latter case ,
the corresponding reflection amplitudes in the basis @xmath233 can then be worked out in full analogy with the @xmath141 subspace in the case of two sqs ( see the previous section ) .
hence , they are given by eqs .
( [ r00])([r01 ] ) under the simple replacements @xmath234 , @xmath235 and @xmath236 .
thereby , @xmath0-@xmath189 qst occurs for @xmath237 , which holds provided that @xmath238 and @xmath239 .
on the other hand , for @xmath240 ( @xmath241 ) we have @xmath242 ( @xmath243 ) , while @xmath244 .
hence , similarly to the @xmath139 case in the previous section , in either of these subspaces the interaction hamiltonian has the spinless effective form @xmath245 .
the condition @xmath246 then ensures that in each case the corresponding overall reflection amplitude equals @xmath247 ( see the comment at the end of the previous section ) . a swap operation between @xmath0 and @xmath194
is therefore implemented .
based on the above findings , in particular eq .
( [ alpha ] ) , the following working conditions to achieve selective writing / read - out of the static register can be devised .
firstly , one fixes once for all the desired coupling strength @xmath248 [ provided that it exceeds the threshold value @xmath249 , equivalent to @xmath250 ; see fig .
[ fig3](a ) ] .
next , we choose one of the two different distances ( in unit of @xmath251 ) that correspond to @xmath248 according to the function @xmath252 [ see fig .
[ fig3](a ) ] .
let us call such a distance @xmath253 , which therefore fulfills @xmath254 .
a further distance @xmath255 [ cf .
[ fig3](b ) ] is then univocally identified .
all the nearest - neighbour distances are set equal to an integer multiple of @xmath86 ( in unit of @xmath251 ) but the @xmath189th and @xmath256th ones , which are set to @xmath253 and @xmath257 , respectively . in a practical implementation , such tunable setting of nearest - neighbor distances
could be achieved by fabricating the setup in such a way that the fq can propagate along three possible paths instead of a single one ( similarly to the geometry of the well - known aharonov - bohm rings ) .
if the paths have different lengths , the actual path followed by the fq can be chosen by means of tunable beam splitters , in fact setting the effective sq - sq distance . in practice
, unavoidable static disorder will affect the ideal pattern of nearest - neighbor sq distances . through a proof - of - principle resilience analysis we have assessed
that , by assuming gaussian noise and in the case of a single sq , an uncertainty in its position of order of about @xmath258 yields a process fidelity above the 95%-threshold .
this witnesses an excellent level of tolerance , in line with similar tests @xcite .
preliminary studies for the cases of two and three sqs have been carried out as well , confirming comparable performances .
a comprehensive conclusive characterization of the effects of static disorder in the case of an arbitrary number of sqs , though , requires a rather involved analysis and thus goes beyond the scopes of this paper .
we have considered a typical scenario envisaged in distributed quantum information , where writing and read - out of a register of sqs is performed through a fq . in a fully quantum theory ,
the motional degrees of freedom ( mdofs ) of the fq should be treated as quantum , which is expected to substantially complicate the dynamics . by taking a paradigmatic hamiltonian
, we have discovered that , as long as the @xmath0-@xmath259 coupling is above a certain threshold value ( i.e. @xmath260 with @xmath1 being the input momentum of the fq ) , for an arbitrary number of sqs _ selective _ qst can be achieved on demand by tuning only two sq distances . throughout ,
as is customary in scattering - based theories , we have assumed to deal with a perfectly monochromatic plane wave for the flying qubit . in practice , clearly , this is a narrow - bandwidth wavepacket centered at a carrier wave vector @xmath261 .
a detailed resilience study of the performances of our protocol in such conditions is beyond the scope of the present paper . yet , similarly to refs .
@xcite , it is reasonable to expect the gate fidelity to be only mildly affected owing to the smoothness of functions @xmath252 and @xmath262 ( cf .
[ fig3 ] ) . in our model
, we assumed a heisenberg - type spin - spin interaction .
as already stressed , our attitude here was to take this well - known coupling as a paradigmatic model to show the possibility that selective writing / read - out is in principle achievable . however , there exist setups where the heisenberg - type coupling occurs so as to make them potential candidates for realizing our protocol
. for instance ( see also refs .
@xcite ) this is the case of an electron propagating along a semiconducting carbon nanotube @xcite and scattered from single - electron quantum dots or molecular spin systems featuring unpaired electrons , such as sc@c82 @xcite .
alternatively , one can envisage a photon propagating in a 1d waveguide to embody the fq in a way that its spin is encoded in the polarization dofs .
a three - level @xmath263-type atom could then work as the static qubit , where the @xmath264 basis is encoded in the ground doublet , while each transition to the excited state requires orthogonal photonic polarizations ; see refs .
although similar , the corresponding ( pseudo ) spin - spin coupling , yet , is not equivalent to a heisenberg - type one .
we found some numerical evidence that this alternative coupling model could work as well , at least in the few - sq case .
an analytical treatment , however , is quite involved and thus no definite answer can be given .
this is connected to the question whether some specific symmetry is a necessary prerequisite for such remarkable effects to take place ( in passing , note that the heisenberg model conserves the squared total spin , which was crucial to carry out our proofs ) .
all these issues are the focus of ongoing investigations .
it is worth mentioning that in a recent work @xcite , ping _
et al . _
proposed a protocol for imprinting the quantum state of a writing " fq on an array of sqs and retrieving it through a reading " fq at a next stage @xcite .
there , information is intentionally encoded over the _
register , which has some advantages , while mdofs are in fact treated as classical . significantly enough , here we have shown that the inclusion of quantum mdofs can allow for control over _ local _ encoding / decoding .
in line with other works @xcite , such apparent complication appears instead a powerful resource to carry out refined qip tasks .
we thank g. cordourier - maruri , s. bose , d. browne , s. lloyd and h. schomerus for useful discussions .
we acknowledge support from miur through the firb - ideas project rbid08b3fm , the italian ministry of foreign affairs through the joint italian - japanese laboratory on quantum technologies : information , communication and computation and the ministry of education , culture , sports , science and technology , japan , through grant - in - aid for young scientists ( b ) ( no .
21740294 ) .
ky is also supported by a waseda university grant for special research projects ( 2012a-878 ) .
here , we derive the matrix elements of the operator @xmath267 in the degenerate subspace @xmath268 ( which is @xmath141 in the case of @xmath269 sqs ) , namely all the reflection coefficients @xmath270 in terms of the basis @xmath271 , where @xmath272 . in line with the main text , we give the proof without specifying @xmath151 ( which can thus be any positive integer or semi - integer number ) .
the hamiltonian reads @xmath273 with @xmath274 ( we have set @xmath275 since the result is evidently independent of @xmath276 ) .
the key task is to work out the matrix representation of @xmath129 in the basis of eigenstates of @xmath128 , @xmath271 .
we first observe that in the present @xmath268 subspace @xmath277 .
accordingly , the scheme where @xmath0 is first coupled to 2 leads to the alternative basis @xmath278 such that @xmath279 .
thereby , in the basis @xmath280 , @xmath129 has the diagonal matrix representation @xmath281 .
the transformation matrix between the two basis can be calculated through 6@xmath19 coefficients @xcite as @xmath282 using these then yields @xmath129 in the basis @xmath143 as @xmath283 next , in close analogy with refs .
@xcite , we search for a stationary state @xmath284 such that @xmath285 , where @xmath286 labels the initial spin state ( prior to the interaction process ) .
each function @xmath287 has the form @xmath288 the unknown coefficients , including @xmath289 , i.e. the entries of @xmath267 , can be found by imposing the continuity condition of @xmath290 and @xmath291 at @xmath292 and the two constraints @xmath293 where @xmath294 is the jump of the derivative at @xmath292 . with the help of eqs .
( [ s00 ] ) and ( [ s01 ] ) , eqs .
( [ boundcondacc1 ] ) and ( [ boundcondacc2 ] ) can be straightforwardly obtained from the schrdinger equation by integrating it across @xmath292 and then projecting onto @xmath295 and @xmath296 @xcite . by solving the linear system in the cases @xmath286
, we thus end up with @xmath297 where @xmath298 .
|
we propose a setup comprising an arbitrarily large array of static qubits ( sqs ) , which interact with a flying qubit ( fq ) .
the sqs work as a quantum register , which can be written or read - out by means of the fq through quantum state transfer ( qst ) .
the entire system , including the fq s motional degrees of freedom , behaves quantum mechanically .
we demonstrate a strategy allowing for selective qst between the fq and a single sq chosen from the register .
this is achieved through a perfect mirror located beyond the sqs and suitable modulation of the inter - sq distances .
|
some numbers can be written as a sum of consecutive integers , for example , @xmath1 but some can not , for example , @xmath2 .
the following is a beautiful characterization of this phenomenon : [ th : po2 ] a number is a sum of consecutive integers if and only if it is not a power of 2 .
proofs of theorem [ th : po2 ] can be found in @xcite . to simplify the subsequent discussion ,
let us call a sequence of consecutive natural numbers a * decomposition of @xmath3 * if its terms sum to @xmath4 .
the * length * of a decomposition is the number of terms in the decomposition and the * parity * of a decomposition is the parity of its length . a * trivial * decomposition is a decomposition of length @xmath5 .
clearly every number @xmath4 has a trivial decomposition , namely @xmath6 .
the following result in @xcite ( also in @xcite ) is fundamental and will be used frequently throughout this article .
[ th : corr ] let @xmath4 be a natural number and @xmath7 be an odd factor of @xmath4 . if @xmath8 , then the sequence @xmath9 is an odd decomposition of @xmath4 of length @xmath7 . on the other hand ,
if @xmath10 , then the sequence @xmath11 is an even decomposition of @xmath4 of length @xmath12 .
moreover , every decomposition of @xmath4 has one of these forms .
theorem [ th : corr ] explicitly demonstrates a 1-to-1 correspondence between the odd factors of @xmath4 and its decompositions .
since the powers of @xmath13 are the only numbers having no odd factors other than @xmath5 , their decompositions can only be trivial .
this establishes theorem [ th : po2 ] as a consequence of theorem [ th : corr ] .
the * length spectrum * ( or simply the * spectrum * ) of a number @xmath4 , denoted by @xmath14 , is the set of lengths of the decompositions of @xmath4 . according to theorem [ th : corr ] ,
the spectrum of @xmath4 is the set @xmath15 as an example , we list in table [ tb : lspec(45 ) ] the decompositions of the number @xmath16 along with their lengths , parities and associated odd factors .
.the spectrum of @xmath16 is @xmath17 [ cols="^,^,^,^",options="header " , ] we say that two numbers are * spectral equivalent * if they have the same length spectrum .
the * spectral class * of @xmath4 , denoted by @xmath18 , is the equivalence class of @xmath4 under spectral equivalence . with these notions ,
theorem [ th : po2 ] can be restated in the following way : the powers of 2 form a spectral class with @xmath19 as their common spectrum .
the odd primes form another interesting spectral class . according to theorem [
th : corr ] , they are the numbers having @xmath20 as spectrum .
these two examples motivate the following question : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given @xmath4 , can we compute its spectral class ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we will give an algorithm in section [ s : e - a ] which answers this question in an affirmative way
. we also derive from it another algorithm which solve the following problem . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given a finite set of numbers @xmath21 , compute the set of numbers with @xmath21 as their common spectrum . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
the following notations and conventions will be adopted throughout the rest of this article . * for a set @xmath22 , we write @xmath23 for its cardinality . * for a set of natural numbers @xmath22 , we write @xmath24 and @xmath25 for the set of even and odd elements of @xmath22 , respectively . * for a rational number @xmath26 and a set of natural numbers @xmath22 , we write @xmath27 for the set @xmath28 . * for a prime @xmath29 and a natural number @xmath30 , we write @xmath31 for the exponent of @xmath29 in the prime factorization of @xmath30 .
let @xmath4 be a fixed but arbitrary natural number , we use @xmath32 to denote the list of odd factors of @xmath4 in ascending order .
so @xmath33 and @xmath34 where @xmath35 .
we use @xmath36 to denote the largest index such that @xmath37 .
let us note that for @xmath38 , @xmath39 and @xmath40 are complementary factors of @xmath41 , i.e. @xmath42 hence the length spectrum of @xmath4 can be re - written as @xmath43 the first important observation about length spectra is : [ th : even<=odd ] the number of even decompositions of a number is at most the number of its odd decompositions .
suppose @xmath39 corresponds to an even decomposition of @xmath4 . by theorem
[ th : corr ] , @xmath44 and so @xmath45 .
therefore , again by theorem [ th : corr ] , @xmath40 corresponds to an odd decomposition of @xmath4 .
the theorem now follows since the map @xmath46 is 1-to-1 . in the light of theorem
[ th : even<=odd ] , we call a spectrum * * balanced * if it has an equal number of even and odd elements ; * * unmixed * if it has no even elements ; * * lopsided * if it is neither balanced nor unmixed .
our next result characterizes the numbers with an unmixed spectrum .
[ th : class - unmixed ] the set of numbers with an unmixed spectrum is @xmath47 it follows from that a number with no even decompositions if and only if it is of the form @xmath48 with @xmath49 , @xmath7 odd and @xmath50 , i.e. @xmath51 . numbers with a balanced spectrum are trickier to capture . for a natural number @xmath7 ,
let @xmath52 be the minimum of the ratios @xmath53 where @xmath54 are complementary factors of @xmath7 .
note that by definition @xmath55 .
moreover , for a pair @xmath56 of complementary factors of @xmath7 , @xmath57 if and only if no factor of @xmath7 is strictly in between @xmath30 and @xmath58 . also , it may be worth pointing out that @xmath52 measures how far is @xmath7 from being a prefect square : @xmath7 is a prefect square if and only if @xmath59 .
[ th : class - balanced ] the set of numbers with a balanced spectrum is @xmath60 suppose @xmath48 is a member of the set in display .
let @xmath61 be a pair of complementary factors of @xmath7 with @xmath57 .
then @xmath62 , hence @xmath63 and so @xmath64 since no factor of @xmath7 is strictly in between @xmath30 and @xmath58 , the inequalities in imply the odd ( resp . even )
decompositions of @xmath48 correspond precisely to those factors of @xmath7 that are @xmath65 ( resp .
@xmath66 ) .
thus the assignment @xmath67 induces an injective map from the odd decompositions of @xmath68 to its even decompositions .
therefore , by theorem [ th : even<=odd ] , @xmath48 has a balanced spectrum .
conversely , suppose @xmath4 has a balanced spectrum .
then we have ( in the notations introduced earlier ) @xmath69 .
hence @xmath70 and @xmath71 .
moreover , @xmath72 , therefore @xmath73 and so @xmath4 belongs to the set displayed in the statement of the theorem .
in this section , we determine the spectral class of a number according to the type of its spectrum .
we begin with an observation which is clear from the form of the spectrum given in : [ p : even ] if @xmath4 has an even decomposition , then the highest power of @xmath13 dividing @xmath4 is half the least even element of @xmath14 .
in fact , @xmath74 for any even element @xmath75 of @xmath14 .
let @xmath21 be a finite set of natural numbers . recall that @xmath76 is the set of elements of @xmath21 congruent to @xmath77 .
we define @xmath78 , the * difference set * of @xmath21 , to be @xmath79 with its least @xmath80 elements removed if @xmath81 , or the empty set otherwise .
[ p : largest ] if a number has more odd than even decompositions , then its greatest odd factor is the product of the maximum and minimum of the difference set of its spectrum .
suppose @xmath82 and @xmath83 . in this case
, @xmath78 is the non - empty set @xmath84 . since @xmath85 , the largest odd factor of @xmath4 , the proposition follows .
[ th : lsc - lopsided ] if @xmath86 is lopsided , then @xmath87 . if @xmath14 is lopsided , then by proposition
[ p : even ] and [ p : largest ] both @xmath88 and @xmath41 , and therefore @xmath4 , can be recovered from @xmath14 . to illustrate the results that we have just proved ,
let us decide whether the set @xmath89 is a spectrum .
first , let us note that @xmath90 .
so if @xmath21 is a spectrum , say @xmath82 , then by proposition [ p : largest ] the largest odd factor of @xmath4 is @xmath91 . by proposition [ p : even ] , @xmath92 .
therefore , @xmath4 can only be @xmath93 .
one then finds that @xmath21 is indeed a spectrum by verifying @xmath94 .
next we determine the spectral class of @xmath4 when its spectrum is unmixed .
[ th : lsc - unmixed ] if @xmath14 is unmixed , then @xmath95 where @xmath30 is the largest odd element of @xmath14 .
suppose @xmath14 is unmixed .
let @xmath30 be the largest odd element of @xmath14 , @xmath21 be the set of factors of @xmath30 and @xmath96 be the set on the right - hand - side of the equation .
first , since @xmath14 is unmixed , it follows easily from that @xmath97 . by theorem
[ th : corr ] , every element of @xmath96 has spectrum @xmath21 . by theorem
[ th : class - unmixed ] , the converse is true hence @xmath96 is precisely the set of number with spectrum @xmath21 and the theorem follows . for a finite set of natural numbers
@xmath21 , we define the * exceptional set * of @xmath21 to be the set @xmath98 where @xmath99 is the set @xmath100 , @xmath101 ( @xmath102 ) and @xmath103 denotes the set of factors of @xmath104 which are strictly less than @xmath7 .
the following simple facts about elements of exceptional sets will come in handy for our subsequent arguments : [ l : facts ] suppose @xmath21 is a balanced spectrum and @xmath105 , then a. [ i : notprime ] @xmath106 has a proper prime factor ; in particular , @xmath106 is not a prime .
b. [ i : proper ] every proper factor of @xmath106 is in @xmath79 . c. [ i : q ] @xmath107 . by definition , @xmath108 .
moreover , @xmath109 therefore @xmath106 can not be a prime and so follows .
every proper factor of @xmath106 is clearly a member of @xmath110 which is @xmath79 and so follows . to show , it suffices to show that no factor of @xmath111 is strictly in between @xmath112 and @xmath106 .
if not , then @xmath110 , i.e. @xmath79 will contain a number larger than @xmath112 , a contradiction .
the next lemma is crucial to our analysis of balanced spectra .
[ l : incl ] suppose @xmath82 is balanced , then @xmath113 where @xmath114 is the set of primes that are larger than @xmath115 . since @xmath82 is balanced , it is of the form @xmath116 where @xmath35 .
in particular , @xmath117 .
pick @xmath105 , and let @xmath118 be @xmath119 .
by lemma [ l : facts ] , @xmath107 and so the inequality @xmath120 implies @xmath121 .
therefore , @xmath122 is balanced according to theorem [ th : class - balanced ] .
note that the inequality @xmath123 also implies @xmath124 these inequalities together with theorem [ th : corr ] and the fact that @xmath125 imply the set of odd elements of @xmath122 is @xmath110 . but
@xmath110 is simply @xmath79 , as @xmath106 is a member of @xmath126 .
therefore , @xmath127 so we conclude that @xmath128 . to show the second inclusion , suppose @xmath128 , i.e. @xmath129 . thus @xmath112 is the largest odd factor of @xmath118 corresponding to an odd decomposition .
since @xmath122 is balanced , the largest odd factor of @xmath118 is of the form @xmath111 for some odd number @xmath130 such that @xmath131 . by proposition [ p : even ] , @xmath132 and so @xmath133 . by theorem
[ th : class - balanced ] , @xmath134 . therefore , @xmath135 .
since no factor of @xmath111 is strictly in between @xmath106 and @xmath112 , @xmath110 is the set of odd factors of @xmath118 not exceeding @xmath112 and that is @xmath79 . moreover , since every proper factor of @xmath106 belongs to @xmath136 , that means either @xmath137 or @xmath106 is a prime . in the former case , @xmath105 and in the latter case , @xmath138 . before stating our next result , which gives the spectral class of @xmath4 when its spectrum is balanced , we need to introduce one more concept .
suppose @xmath21 is a spectrum . by theorem
[ th : corr ] , @xmath79 contains the factors of its elements .
in particular , the sets of factors of @xmath112 is always a subset of @xmath79 . in the light of this observation
, we call a balanced spectrum @xmath21 * non - excessive * if @xmath79 is precisely the set of factors of @xmath112 ; otherwise we call @xmath21 * excessive*. note that we use the word ( non-)excessive to describe balanced spectra only .
[ th : lsc - balanced ] suppose @xmath82 is balanced
. then a. [ i : non - ex ] @xmath139 , if @xmath21 is non - excessive ; or b. [ i : ex ] @xmath140 , if @xmath21 is excessive . to proof , thanks to lemma [ l : incl ] , we only need to show that the set @xmath141 is a subset of @xmath18 .
take @xmath142 and let @xmath143 where @xmath35 . since @xmath29 is prime , the odd factors of @xmath118 are the factors of @xmath112 and their @xmath29 multiples .
the inequalities @xmath144 ensure @xmath145 therefore , the factors of @xmath112 correspond to the odd decompositions of @xmath118 while their @xmath29 multiplies correspond to the even decompositions of @xmath118 ( theorem [ th : corr ] ) .
thus , @xmath122 is balanced and the set of odd elements of @xmath122 coincides with the set of factors of @xmath112 .
since @xmath21 is non - excessive , the set of factors of @xmath112 equals @xmath79 .
so we actually have @xmath146 this finishes the proof of part .
incidentally , the argument above also shows that if @xmath82 is balanced then every element of @xmath141 has non - excessive spectrum .
consequently , @xmath18 and @xmath141 do not intersect if @xmath21 is excessive .
thus part follows from lemma [ l : incl ] as well .
we study of the structures of exceptional sets in this section . as a result
, we prove a rather curious fact : a spectral class can only have either @xmath147 or infinitely many elements .
we start by making the following conventions and definitions . throughout this section
, @xmath21 denotes a balanced spectrum and @xmath30 denotes the largest odd element of @xmath21 , moreover : * for a prime @xmath29 , let @xmath148 denote the largest integer such that @xmath149 and we write @xmath150 for @xmath31 .
the * excessive index of @xmath29 with respect to @xmath21 * is defined to be @xmath151 .
note that @xmath152 is always non - negative . *
a prime @xmath29 is called an * excessive prime * of @xmath21 if @xmath153 ; otherwise @xmath29 is called a * non - excessive prime * of @xmath21 .
we also say that @xmath29 is * excessive ( non - excessive ) with respect to @xmath21 * if it is an excessive ( a non - excessive ) prime of @xmath21 .
note that @xmath13 is always a non - excessive prime of @xmath21 .
also , every excessive prime of @xmath21 is in @xmath79 and every non - excessive prime of @xmath21 that is in @xmath79 divides @xmath30 . * the * excessive number * of @xmath21 is defined to be the product @xmath154 where @xmath29 runs through the primes .
note that @xmath21 is excessive if and only if @xmath21 has an excessive prime if and only if @xmath155 .
* every @xmath105 can be written as @xmath156 where @xmath157 ( @xmath158 ) is a product of ( non-)excessive primes .
the numbers @xmath157 and @xmath158 are called the * excessive part * and the * non - excessive part * of @xmath106 , respectively .
note that either @xmath157 or @xmath158 can be @xmath5 but not both since @xmath159 .
the next two lemmas tell us what kind of factors that an element of @xmath126 can / must have .
[ l : excessive ] for every prime @xmath29 , @xmath160 divides every element of @xmath126 . for any @xmath105 , since @xmath149 , @xmath161 and so @xmath162 .
[ l : non - excessive ] if @xmath163 is a non - excessive prime of @xmath21 dividing some element of @xmath126 then @xmath164 for any prime @xmath29 other than @xmath163 .
suppose @xmath163 is non - excessive and @xmath165 for some @xmath105 .
then @xmath166 divides @xmath167 ; moreover for any prime @xmath168 , if @xmath169 then @xmath170 but that means @xmath171 , contradicting the fact that @xmath163 is non - excessive .
therefore , we must have @xmath172 .
[ p : form ] for every @xmath105 , a. [ i : divide ] @xmath157 is divisible by @xmath173 .
b. [ i : either ] @xmath158 is either @xmath5 or a power of the largest non - excessive prime of @xmath21 in @xmath79 . in particular , if @xmath174 then @xmath79 contains a non - excessive prime of @xmath21 . c. [ i : n_a>1 ] if @xmath175 , then @xmath176 . by lemma [ l : excessive ] , @xmath177 .
since @xmath173 is a product of excessive primes , so in fact , @xmath173 divides @xmath157 . by lemma [ l : facts ] and , every prime factor of @xmath106 is in @xmath79 .
so @xmath158 is a product of non - excessive primes ( of @xmath21 ) in @xmath79 .
therefore , @xmath178 if @xmath79 contains no non - excessive primes of @xmath21 .
so suppose otherwise and let @xmath163 be the largest non - excessive prime of @xmath21 in @xmath79 . if @xmath179 is another non - excessive prime of @xmath21 dividing @xmath158 , then @xmath180 and so @xmath181 .
however , by lemma [ l : non - excessive ] @xmath182 and this leads to a contradiction since @xmath183 .
so we conclude that @xmath158 can have no prime factors other than @xmath163 , therefore part follows .
suppose @xmath174 then , by part , @xmath106 is of the form @xmath184 where @xmath163 is the largest non - excessive prime of @xmath21 in @xmath79 and @xmath185 .
for a prime @xmath29 , let us write @xmath186 for @xmath187 .
consider the factor @xmath188 of @xmath167 .
for @xmath168 , @xmath189 and by lemma [ l : non - excessive ] @xmath190 .
therefore , @xmath191 and so @xmath192 since @xmath193 . consequently , @xmath194 , i.e. @xmath195 .
but by part , @xmath196 .
therefore we conclude that @xmath197 for every prime @xmath168 and hence @xmath198 . we should point out that it is possible for a balanced spectrum ( other than @xmath20 ) to have no non - excessive odd primes ( see example [ ex : no - ne ] ) .
also , @xmath158 in the above proposition may still be @xmath5 even @xmath79 contains a non - excessive prime of @xmath21 .
next we give a characterization of non - excessive spectra in terms of exceptional sets .
[ th : e(s)non - excessive ] a balanced spectrum @xmath21 is non - excessive if and only if @xmath199 or @xmath200 for some prime @xmath163 . in particular , the size of the exceptional set of a non - excessive spectrum is at most one .
suppose @xmath21 is non - excessive and @xmath105 .
since every prime is non - excessive with respect to @xmath21 , @xmath201 and so @xmath174 . therefore , by proposition [ p : form ] , @xmath106 is a power of the largest non - excessive prime @xmath163 in @xmath79 ( since @xmath21 is non - excessive , so in fact @xmath163 is outright the largest prime in @xmath79 ) .
since @xmath202 , so on one hand @xmath203 ; on the other hand @xmath204 but @xmath205 thus @xmath206 . therefore , @xmath126 must be the singleton @xmath207 if it is non - empty . to show the other implication , let us note that if @xmath126 is empty , then @xmath21 is non - excessive by theorem [ th : lsc - balanced ] .
so let us assume @xmath200 for some prime @xmath163 .
then every @xmath208 divides @xmath209 and @xmath210 .
thus @xmath211 . moreover , for any prime @xmath104 other than @xmath163 , @xmath212
therefore , we conclude that @xmath213 and hence @xmath21 is non - excessive .
the next result was a surprise to us .
[ th : e(s)>1 ] @xmath214 .
suppose @xmath215 , then @xmath21 is excessive according to theorem [ th : e(s)non - excessive ] .
let @xmath29 be an excessive prime of @xmath21 .
by lemma [ l : excessive ] , @xmath160 divides every element of @xmath126 .
moreover , since @xmath216 , by lemma [ l : facts ] @xmath217 and hence @xmath218 is a subset of @xmath79 .
let @xmath219 be the list of elements of @xmath220 in ascending order . for @xmath221
, since @xmath222 , therefore @xmath223 must not divide @xmath224 . in other words
, @xmath225 does not divide @xmath226 .
however , since @xmath227 , @xmath228 , i.e. @xmath229 .
that means @xmath230 and for any prime @xmath104 other than @xmath29 , @xmath231 .
so the fact that @xmath225 does not divide @xmath226 implies @xmath232 .
therefore , we must have @xmath233 we claim that for @xmath221 , @xmath234 is not divisible by @xmath29 . if not , then @xmath235 and so @xmath236 would be a proper factor of @xmath223 and hence , by lemma [ l : facts ] , an element of @xmath79 , a contradiction .
therefore , we conclude that @xmath237 for some @xmath238 not divisible by @xmath29 . since @xmath239 , then again by lemma [ l : facts ] @xmath240 and so @xmath236 is the least element of @xmath126 .
as elements of @xmath79 the @xmath234 s all divide @xmath241 .
but for @xmath221 , @xmath234 and @xmath29 are relatively prime so @xmath234 must divide @xmath30 . therefore ,
if @xmath242 , then we would have @xmath243 and @xmath244 , contradicting @xmath245 .
so we conclude that @xmath214 .
[ th : size ] every spectral class has either @xmath0 or infinitely many elements .
moreover , for any @xmath4 , * @xmath246 if and only if @xmath14 is lopsided or excessive with an exceptional set of size @xmath5 .
* @xmath247 if and only if @xmath14 is excessive with an exceptional set of size @xmath13 .
* @xmath18 is infinite if and only if @xmath14 is unmixed or non - excessive .
suppose a spectral class @xmath18 is finite , then @xmath248 must be either lopsided or excessive ( theorem [ th : lsc - unmixed ] and [ th : lsc - balanced ] ) .
in the former case , @xmath249 ( theorem [ th : lsc - lopsided ] ) ; in the latter case , @xmath250 according to theorem [ th : lsc - balanced ] and theorem [ th : e(s)>1 ] .
so we establish the first statement .
the three equivalences are simply reorganizing what we have already proved in theorem [ th : lsc - lopsided ] , [ th : lsc - unmixed ] and [ th : lsc - balanced ] .
we conclude this section with a few more precise descriptions of the exceptional sets .
[ p : s1 ] if @xmath251 , then @xmath252 .
if @xmath251 , then in particular @xmath173 is greater than @xmath5 so @xmath21 is excessive and hence @xmath253 .
let @xmath105 , by lemma [ l : excessive ] , @xmath177 .
however , since @xmath251 , therefore by lemma [ l : facts ] , @xmath254 .
[ p : e(s)=1 ] suppose @xmath255 then the unique element of @xmath126 is either a product of excessive primes or of the form @xmath256 ( @xmath185 ) where @xmath163 is the largest non - excessive primes of @xmath21 in @xmath79 .
let @xmath106 be the unique element of @xmath126 . if @xmath106 is not a product of excessive primes then by proposition [ p : form ] and , @xmath106 is of the form @xmath256 with @xmath257 .
[ p : s2 ] suppose @xmath215 then @xmath21 has a unique excessive prime @xmath29 and @xmath126 is of the form @xmath258 ( @xmath185 ) where @xmath29 and @xmath163 are the two largest primes in @xmath79 . moreover , a. [ i : p > q ] if @xmath29 is the largest prime in @xmath79 , then @xmath259
. b. [ i : q > p ] if @xmath29 is not the largest prime in @xmath79 , then @xmath260 .
we have already proved ( theorem [ th : e(s)>1 ] ) that @xmath215 implies @xmath126 is of the form @xmath261 where @xmath29 is an excessive prime of @xmath21 , @xmath262 and @xmath263 is not divisible by @xmath29 . by lemma [ l : excessive ] , every excessive prime of @xmath21 divides @xmath236 , therefore @xmath29 is the only excessive prime of @xmath21 and so @xmath263 is the non - excessive part of @xmath264 .
since @xmath265 , by proposition [ p : form ] @xmath263 is a positive power of @xmath163 where @xmath163 is the largest non - excessive prime of @xmath21 in @xmath79 .
therefore , we conclude that @xmath126 must be of the form @xmath266 for some @xmath185 . since @xmath163 is the largest non - excessive prime in @xmath79 and @xmath29 is the unique excessive prime in @xmath79 , therefore if @xmath29 is the largest prime in @xmath79 then @xmath163 must be the second largest prime in @xmath79 . since @xmath163 divides @xmath267 , by lemma [ l : non - excessive ] , @xmath164 .
therefore , @xmath150 must be @xmath268 , i.e. @xmath269 .
this completes the proof of .
suppose @xmath29 is not the largest prime in @xmath79 , then the largest prime in @xmath79 is non - excessive and so must be @xmath163 .
we claim that in this case @xmath270 is actually @xmath5 .
first , note that @xmath271 and since @xmath272 divides @xmath273 , @xmath274 i.e. @xmath275 . but by lemma [
l : non - excessive ] , we also have @xmath190 . therefore , @xmath270 must be @xmath5 .
to finish the proof of , we argue that @xmath29 must be the second largest prime in @xmath79 . since @xmath276 and @xmath163 is non - excessive , @xmath277 . in other words , @xmath278 .
so if there were a prime @xmath279 strictly in between @xmath29 and @xmath163 then @xmath280 .
but since @xmath104 is non - excessive , it divides @xmath30 and therefore @xmath281 , a contradiction .
we give some examples here to illustrate the results in previous sections .
the smallest number with a balanced spectrum is @xmath283 .
the spectrum of @xmath283 is @xmath20 which is non - excessive and has an empty exceptional set . by theorem
[ th : lsc - balanced ] , @xmath284 is the set of odd primes .
the number @xmath288 is the smallest number with a spectrum that has a non - empty exceptional set .
the spectrum of @xmath288 is @xmath289 .
it is non - excessive with exceptional set @xmath290 . by theorem
[ th : lsc - balanced ] , @xmath291 the smallest number with two elements in the exceptional set of its spectrum is @xmath296 .
the spectrum of @xmath296 is @xmath297 .
the exceptional set of @xmath298 is @xmath299 ( c.f .
proposition [ p : s2 ] ) . by theorem
[ th : lsc - balanced ] , @xmath300 the proof of theorem [ th : e(s)>1 ] would be considerably simpler if every excessive spectrum contains an odd prime not dividing its largest odd element . however , it is not always the case . the smallest number with a spectrum witnessing this fact is @xmath301 .
the set of odd elements of the spectrum of @xmath302 is @xmath303 .
it contains two primes , @xmath283 and @xmath304 , both of them divide @xmath305 . [ ex : no - ne ] the number @xmath306 is the smallest number such that its spectrum contains an odd prime and every odd prime in its spectrum is excessive .
the set of odd elements of @xmath307 is @xmath308 .
it contains two primes , @xmath283 and @xmath309 , both of them have excessive index @xmath5 .
let us explain how we found this example .
suppose @xmath82 has the required property . then by proposition [ p : s2 ]
, @xmath126 must be a singleton .
also , @xmath310 can not be a prime power , otherwise the prime of which @xmath30 is a power will be non - excessive .
so @xmath30 has at least two prime factors , say @xmath311 . since @xmath312 is excessive ,
therefore @xmath313 .
but that means @xmath314 and so @xmath315 for some odd number @xmath316 . by lemma [ l : excessive ] ,
the unique element @xmath105 is of the form @xmath317 . since it must be greater than @xmath318 , @xmath319 . at this juncture
, we make a guess : suppose @xmath320 and @xmath321
. then @xmath322 and we need @xmath323 . minimizing @xmath4 subjected to the inequality yields @xmath324 and so @xmath325 .
now let @xmath326 be the smallest number such that its spectrum contains an odd prime and every odd prime in its spectrum is excessive .
the argument above shows that @xmath327 and is the product of two numbers of the forms @xmath328 and @xmath317 with @xmath319 . since @xmath329 , @xmath330
therefore , @xmath26 must be @xmath283 and @xmath331 . from this
, we see that @xmath332 . if @xmath333 , then @xmath334 , i.e. @xmath335 or @xmath336 . but none of these choices produces a balanced spectrum .
so @xmath337 , this forces @xmath338 and hence @xmath339 .
[ ex : beta>1 ] the exponent @xmath270 in proposition [ p : s2 ] can be greater than @xmath5 .
let us find the smallest witness , say @xmath326 , of this fact .
first , let @xmath4 be a number with @xmath82 witnessing @xmath340 .
let @xmath29 be the unique excessive prime of @xmath21 and @xmath163 be the largest non - excessive prime in @xmath79 . for simplicity ,
let us write @xmath341 for @xmath148 . by proposition [ p : s2 ] , @xmath342 , @xmath343 and @xmath29 does not divide @xmath310 . therefore , @xmath30 is of the form @xmath344 where @xmath345 and @xmath26 is not divisible by either @xmath29 or @xmath163 .
the number of factors of the two elements in @xmath346 are both equal to the size of @xmath21 ; by equating them , we get @xmath347 . clearly , one solution of this equation with @xmath348 is @xmath349 . with these values , @xmath350 .
so the choices of @xmath29 and @xmath163 have to meet the following inequalities : @xmath351 minimizing @xmath352 subjected to these inequalities yields @xmath353 , @xmath354 and @xmath355 .
so @xmath356 and the exceptional set of its spectrum is @xmath357 .
it is indeed an example with @xmath358 . to show that @xmath326 is @xmath359 ,
let us note that , from the argument above , @xmath326 is of the form @xmath360 with @xmath361 .
therefore , @xmath362 . since @xmath363 and @xmath364 , @xmath365 .
that means @xmath366 .
but one quickly rules out the possibility that @xmath367 , since no choice of @xmath29 would satisfy the inequalities in .
hence @xmath355 , and it follows that the minimal choices for @xmath26 and @xmath163 are @xmath5 and @xmath368 , respectively .
this completes the proof . to find an exceptional set with both @xmath341 and @xmath340 will lead one to the number @xmath369 .
the exceptional set of its spectrum is @xmath370 .
an analysis similar to the one given above shows that @xmath371 is indeed the smallest possible choice .
we will leave the verification to the reader this time .
finally , we give two algorithms answering the questions that we posed in the introduction . algorithm
[ al : spectral - class ] computes from a given number @xmath4 its spectral class @xmath18 .
its correctness is guaranteed by theorem [ th : lsc - lopsided ] , [ th : lsc - unmixed ] and [ th : lsc - balanced ] .
[ al : lspec ] computes , using algorithm [ al : spectral - class ] , from a finite set of numbers @xmath21 the set of numbers with spectrum @xmath21 . in particular , algorithm [ al : lspec ] returns the empty set if @xmath21 is not a spectrum .
here is the strategy : computes from @xmath21 a number @xmath4 such that @xmath82 if @xmath21 is a spectrum .
the algorithm then returns either @xmath18 or the empty set depending on whether the equality holds or not .
we have implemented both algorithms using pari / gp script .
a natural number @xmath4 @xmath18 the spectral class of @xmath4 @xmath248 ; @xmath372 ( @xmath102 ) .
@xmath373 ; @xmath374 the least integer such that @xmath375 .
@xmath380 . a finite set of natural numbers @xmath21 .
the set of natural numbers with length spectrum @xmath21 .
@xmath381 ( @xmath102 ) .
@xmath382 ; @xmath374 the least integer such that @xmath375 ; @xmath383 .
@xmath385 ; @xmath386 .
@xmath385 ; @xmath387 the least prime @xmath388 ; @xmath389 .
|
two numbers are spectral equivalent if they have the same length spectrum .
we show how to compute the equivalence classes of this relation .
moreover , we show that these classes can only have either @xmath0 or infinitely many elements .
|
classical information theory was laid down by shannon in the nineteen forties to characterize the ultimate rate at which one could hope to transmit classical information over a classical communication channel : the channel capacity .
surprisingly in retrospective , not only it achieved its purpose but the capacity of classical channels turned out to comply with all the properties that one could expect for such a quantity : it can be efficiently computed @xcite and it gauges the usefulness of the channel in the presence of any additional contextual channel .
it is a natural consequence of additivity and convexity of the capacity in the set of channels . with quantum channels complemented by various auxiliary resources ,
a whole new range of communication tasks became feasible .
notably , they allow for the transmission of quantum and private classical communication tasks beyond the reach of classical channels . for most of these tasks ,
the tools used to prove the capacity theorems in the classical case can be generalized .
however , computability , additivity , and convexity the three convenient properties of the classical capacity of classical channels do not necessarily translate to the quantum case . in table [ tableofcaps ] we summarize what is known about these properties for a set of relevant quantum channel capacities . with the exception of the entanglement - assisted capacity @xcite , there is no known algorithm to compute any of these capacities .
it is due to their characterization which in most cases is given by a regularized formula @xcite .
moreover , even non - regularized quantities are notoriously hard to compute .
for instance , the holevo information is known to be np - complete @xcite .
a capacity is non - additive as a function of a channel if for a given pair of channels the sum of their individual capacities is strictly smaller than the capacity of another channel which is constructed by using both channels in parallel .
hence , a non - additive capacity is contextual : the usefulness of a channel for communication depends on what other channels are available .
the private and quantum capacities are known to be non - additive @xcite .
this observation motivated authors in @xcite to define a new quantity the potential capacity which characterizes the usefulness of a channel used in parallel with the best possible contextual channel .
.main properties of quantum channel capacities : convexity , additivity , and computability .
we consider quantum capacity @xmath0 , private capacity @xmath1 and unassisted , entanglement - assisted and environment - assisted classical capacities , @xmath2 , @xmath3 and @xmath4 respectively .
[ cols="^,^,^,^,^ " , ] another important property of the capacities of quantum channels is convexity .
the capacity @xmath5 of a quantum channel @xmath6 is non - convex if there exists a pair of channels @xmath7 and @xmath8 and @xmath9 such that : @xmath10 in a same vein , non - convexity also implies that capacity is contextual . for a channel @xmath6 , a contextual channel @xmath11 and a mixing parameter @xmath9
we can define a non - convexity functional : @xmath12\ ] ] analogous to the one defined in @xcite for non - additivity .
this functional induces a ( new ) potential capacity given by the maximization of @xmath13 over all contextual channels @xmath11 and @xmath14 $ ] .
if @xmath5 is non - convex , then there exists a channel @xmath6 such that its potential capacity is strictly larger than @xmath15 or , equivalently , there exists a triple @xmath16 for which @xmath17 .
non - convexity is a surprising property in connection to two communication scenarios in which alice , the sender , has access to two channels that are used with probabilities @xmath18 and @xmath19 . in the first one
, alice uses both channels independently . in the second one ,
alice encodes jointly over the two channels but has no control over which of the channels is applied ; instead , a black box applies them at random with the same probabilities @xmath18 and @xmath19 .
the two scenarios are depicted in fig .
[ fig : prot ] .
in contrast to the classical capacity of classical channels , it was shown that the capacity of quantum channels for transmitting quantum information , i.e. the quantum capacity , is non - convex @xcite .
the question that we address in the following is whether non - convexity is limited to the transmission of quantum information or can be observed beyond the task of entanglement transmission .
we show that the private capacity and the classical environment - assisted capacities of a quantum channel are non - convex .
the action of a quantum channel can always be defined by an isometry @xmath20 that takes the input system @xmath21 to the output @xmath22 together with an auxiliary system called the environment @xmath23 : @xmath24 .
this isometry allows to define the action of the complementary channel : @xmath25 .
we denote the systems involved by a superscript , which we omit when they are clear from the context .
let @xmath26 be a quantum state , we denote by @xmath27 the von neumann entropy .
let @xmath28 be a bipartite quantum state , we denote by @xmath29 the mutual information between the systems @xmath30 and @xmath22 .
we are interested in the following communication tasks and the associated channel capacities .
the first task is the transmission of quantum information .
the quantum capacity characterizes the ability of a quantum channel for this task in the absence of additional resources @xcite @xmath31 where @xmath32 is the coherent information of a quantum channel .
the maximum is taken over all input states purified with a reference system @xmath30 .
the quantity @xmath33 , where @xmath34 are the von neumann entropies of @xmath35 , @xmath36 and @xmath37 denotes the identity channel . for some channels , the coherent information is additive and thus exactly characterizes their capacity . in these cases ,
it is possible to compute the capacity exactly @xcite .
however , there are examples when this is not the case @xcite : coherent information is superadditive .
not only the coherent information is superadditive , but also , the quantum capacity itself is superadditive @xcite there exist pairs of channels such that their joint capacity is strictly larger than the sum of their capacities .
the second task is the transmission of private classical information .
the capacity of a channel for this task without additional resources is called the private capacity @xcite .
we define the private information to be @xmath38 where @xmath39 and @xmath40 are evaluated on the states @xmath41 and @xmath42 .
the private capacity is given by the regularization of the private information @xmath43 both private information @xcite and the private capacity @xcite were found to be superadditive .
the third task is the transmission of classical information .
the classical capacity @xcite characterizes the capacity of a quantum channel for transmitting classical information without additional resources . to characterize the classical capacity
we first define the holevo information @xmath44 the classical capacity is given by the regularization of the holevo information @xmath45 holevo information is superadditive @xcite but it is a challenging open question whether or not the classical capacity verifies any of the three properties of convexity , additivity , and computability .
in some scenarios , sender and receiver may share additional resources which they can leverage to increase their communication rates .
the capacities of a channel for a communication task assisted by additional resources turn out to have completely different properties than their unassisted counterparts .
one such example is shared entanglement .
the entanglement - assisted classical capacity of a quantum channel @xmath46 is both convex and additive and can be computed efficiently @xcite .
alternatively , one may consider the environment of the channel as a friendly helper that ` assists ' the sender during information transmission @xcite .
this third party can input states independently of the sender or even interact with the sender by exchanging messages .
this gives rise to a host of environment - assisted classical capacities depending on whether we have active or passive environment assistance @xcite or whether the sender and environment are allowed to share entanglement or interact by means of local operations and classical communication . in our work ,
we focus on the weakest variant of assistance for classical communication when the helper is in the product state with the sender @xcite .
the corresponding capacity is given by @xmath47 where @xmath48 .
@xmath49 is an isometric extension of the channel such that : @xmath50 and @xmath51 is a state of the system @xmath23 over @xmath52 uses of the channel .
we first show that private capacity is non - convex .
let us first define two families of channels .
the first is the @xmath53-dimensional erasure channel @xmath54 .
its action is defined as follows : @xmath55 that is , @xmath54 takes the input to the output with probability @xmath19 and with probability @xmath18 it outputs an erasure flag . the private capacity of the erasure channel
is known to be @xcite : @xmath56 the second is the ` rocket channel ' @xmath57 .
it was introduced by smith and smolin in @xcite .
it takes two @xmath53-dimensional inputs that we label @xmath58 and @xmath59 .
the channel chooses two unitaries @xmath60 and @xmath20 at random and @xmath20 are chosen uniformly at random from a unitary 2-design . choosing the clifford group or any other finite unitary 2 design has the advantage that the output of the channel is finite dimensional . ] and applies them to @xmath58 and @xmath59 respectively , followed by the application of a joint dephasing operation @xmath61 .
the map is given by @xmath62 with @xmath63 being a primitive @xmath53-th root of unity .
finally , the first system is traced out and the second system together with a classical description of @xmath60 and @xmath20 is sent to the receiver .
given @xmath60 and @xmath20 the action of the channel can be written as @xmath64 where @xmath65 .
the total action of the channel is the average @xmath66 rocket channels have small classical capacity for @xmath67 @xcite : @xmath68 now let us consider a convex combination of a flagged erasure channel and a flagged rocket channel : @xmath69 where @xmath70 and @xmath71 . in the following we prove that for some ranges of @xmath53 , @xmath18 and @xmath72 @xmath73 the right - hand side of is bounded from above by @xmath74 we can bound @xmath75 from below by @xmath76 .
hence , we can argue that any achievable rate for quantum communication ( itself a lower bound on the quantum capacity ) is a lower bound on the private capacity .
let @xmath77 be some input for two uses of channel @xmath78 .
then : @xmath79 now , let the input be : @xmath80 where @xmath81 represents a maximally entangled state between systems @xmath30 and @xmath22 .
we use a subscript if the register corresponds to a concrete channel use and a superscript to number the subsystem : @xmath82 stands for the first subsystem of the register @xmath58 in the second use of the channel and @xmath83 the second subsystem of an auxiliary register @xmath30 .
the coherent information achieved by is : @xmath84 see the supplemental material for details .
consequently , the private capacity of @xmath78 is bounded from below by : @xmath85 it remains to compare the achievable bound in with the converse bound in . for any triple @xmath86 such that is strictly greater than the private capacity is non - convex .
figure [ fig : nonconvexity ] depicts the achievable region for which we exhibit non - convexity .
when @xmath53 goes to infinity .
a value larger than zero implies non - convexity of @xmath87 .
, width=302 ]
we now turn to non - convexity of classical capacity with the weakest environment assistance .
we start with providing two channels and a special entangled input state which we use to demonstrate this effect .
consider a flagged combination of the two channels used in @xcite to show superadditivity of @xmath4 .
the first channel is defined by a controlled unitary @xmath88 where @xmath89 , @xmath90 , @xmath91 and @xmath63 is again the primitive @xmath53-th root of unity .
the second channel is a @xmath92 channel : @xmath93 .
thus , our channels will have the form @xmath94 and @xmath95 .
fix @xmath96 . in the following
we prove that for some range of @xmath18 : @xmath97 it follows from @xcite that @xmath98 and @xmath99 .
hence , the right - hand side of is bounded from above by @xmath100 in order to bound from below the left - hand side of , consider two uses of the channel @xmath101 .
let the state of the environment be the maximally entangled state between @xmath102 and @xmath103 : @xmath104 and the input state to the channel : @xmath105 then , @xmath106 and since @xmath11 is flagged , we can also divide the mutual information into the sum of the mutual information associated with each channel action .
let us compute the corresponding output states : @xmath107 note that @xmath108 is just @xmath109 with @xmath110 and @xmath111 swapped .
the state obtained from the action of @xmath112 follows from the observation that @xmath113 .
it is easy to verify that @xmath114 vanishes when @xmath115 is applied and takes the value @xmath116 when either @xmath117 or @xmath118 is applied . in the case of
@xmath119 we can bound the mutual information by : @xmath120 let us justify .
the input state is a clasical - quantum state of the form : @xmath121 .
we can write explicitly the input states as : @xmath122 if we apply the channel to an input state , we can conclude from that the output does only depend on @xmath123 and it simplifies to : @xmath124 let @xmath125 and @xmath126 , we can check the orthogonality between two output states : @xmath127 is a geometric series . then ,
if @xmath128 : @xmath129 that is , @xmath130 and @xmath131 are orthogonal except if @xmath132 and then @xmath133 .
this is the case if @xmath53 divides @xmath134 which can only occur for @xmath135 .
hence if @xmath53 is even there are @xmath136 orthogonal states and if @xmath53 is odd there are @xmath53 orthogonal states .
we conclude that @xmath114 equals @xmath137 if @xmath53 is odd and @xmath138 if @xmath53 is even as claimed . adding all the contributions we obtain for odd @xmath53 : @xmath139 finally
, comparing the achievable bound in with the converse bound in one observes that for odd @xmath140 and @xmath141 the classical capacity with passive environment - assisted capacity is non - convex .
computability , additivity , and convexity are three fundamental properties of capacity which allow to characterize the usefulness of a quantum channel for a concrete communication task . here
, we focused our attention on non - convexity .
prior to our work , non - convexity had only been proven for the quantum capacity .
we exhibit non - convexity of communication tasks involving classical information via quantum channels .
hence , our results show that non - convexity is a generic feature of communications over quantum channels that is not merely restricted to the transmission of quantum information .
furthermore , non - convexity is not an effect which concerns only a zero - measure set of quantum channels : by perturbing the channels in our construction one finds that the result still holds .
however , it remains open how typical is non - convexity ( and non - additivity ) if one chooses two channels at random .
both our non - convexity proofs and that of the quantum capacity build on top of non - additivity proofs .
it is unclear if this is an artifact of the constructions or they hint to a deeper relation between both properties .
moreover , the non - convexity functional that we introduce here gives rise to a potential capacity analogous to the one induced by non - additivity .
it is tempting to conjecture that the the two potential capacities , and more broadly , non - convexity and non - additivity , are closely related .
hence , a better understanding of this relation might shed some light into how much do the different capacities really gauge the usefulness of quantum channels for communication tasks .
we thank kenneth goodenough , frdric grosshans , jonas helsen and stephanie wehner for useful discussions and feedback .
ss acknowledges the support of sidney sussex college and european union under project qalgo ( grant agreement no
. de has been partially supported by stw , the nwo vidi grant large quantum networks from small quantum devices " and by the project hyqunet ( grant no . tec2012 - 35673 ) , funded by ministerio de economa y competitividad ( mineco ) , spain .
30ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreva.55.1613 [ * * , ( ) ] , @noop * * , ( ) link:\doibase 10.1109/18.651037 [ * * , ( ) ] ( , ) @noop * * , ( ) link:\doibase 10.1109/tit.2004.839515 [ * * , ( ) ] , @noop * * , ( ) link:\doibase 10.1103/physrevx.4.011016 [ * * , ( ) ] link:\doibase 10.1109/tit.2016.2522192 [ * * , ( ) ] http://arxiv.org/abs/1602.02036 [ ( ) ] , link:\doibase 10.1109/tit.2016.2519920 [ * * , ( ) ] @noop ( ) link:\doibase 10.1126/science.1162242 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.103.120501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.120503 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physreva.57.830 [ * * , ( ) ] , http://dx.doi.org/10.1038/ncomms7739 [ * * ( ) ] link:\doibase 10.1103/physrevlett.108.040501 [ * * , ( ) ] , @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.115.040501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.010501 [ * * , ( ) ] link:\doibase 10.1038/nphys1224 [ * * , ( ) ] , @noop ( )
now we analyze the coherent information achieved by the input @xmath142 after sending @xmath143 through two copies of the channel @xmath144 , the resulting state is : @xmath145 since the channel is a flagged combination of @xmath146 and @xmath147 , the coherent information is just the weighted sum of four terms @xmath148 by symmetry of the input state , one has @xmath149 let us compute each of the three terms .
first , we consider two erasure channels . the resulting state is @xmath150 where @xmath151 stands for the maximally mixed state .
the four states of this mixture can be differentiated by checking the erasure flag .
this implies that the coherent information can also be divided into the sum of the coherent information of each term .
@xmath152 the resulting state in the case of one erasure channel and one rocket channel is @xmath153 which yields @xmath154 finally , the use of two rocket channels yields @xmath155 for justification of and see appendix [ ap : b ] .
we plug , , and back into @xmath156
the arguments follow from @xcite .
let us analyze the action of one rocket channel and one erasure .
this action can be decomposed into the action of the identity channel with probability @xmath157 and an erasure with probability @xmath18 as stated in .
let us compute the resulting state in both situations .
for @xmath158 we get : @xmath159 the key idea here is that the register @xmath160 is maximally entangled with the register @xmath161 , which is available to the receiver . hence
, the receiver can undo each unitary applied to @xmath160 by applying the inverse of the transpose of the corresponding unitary .
more precisely , for each choice of @xmath60 and @xmath20 from the channel : @xmath162 in the case of rocket channel and erasure we obtain : @xmath163 let us denote by @xmath164 , then : @xmath165 . if we focus our attention on the action of the rocket channel for some concrete @xmath60 and @xmath20 : @xmath166 where @xmath167 denotes the completely dephasing channel in the computational basis .
we can conclude that @xmath168 , @xmath169 and @xmath170 .
now , let us analyze the action of two rocket channels . from the data processing inequality for coherent information we have that @xmath171 .
now we will show that , @xmath172&\otimes\left[\id^{a^2}\otimes \deph\circ r_d^{wx}\right]\left(\phi^{a^1d_1}\otimes\phi^{c_1c_2}\otimes\phi^{a^2d_2}\right)\\ & = \left[\left(v^t\right)^{a^1}\otimes\deph\right]\otimes\left[\left(x^t\right)^{a^2}\otimes\deph\right]\left(\phi^{a^1d_1}\otimes\phi^{a^2d_2}\right),\end{aligned}\ ] ] where the action of @xmath173 and @xmath174 is described in . then , since coherent information is invariant under the application of local unitaries : @xmath175 we can write explicitly the form of the output after acting on the input state with @xmath176 : @xmath177 if we apply a dephasing channel at the output we obtain the following : @xmath178
|
the capacity of classical channels is convex .
this is not the case for the quantum capacity of a channel : the capacity of a mixture of different quantum channels exceeds the mixture of the individual capacities and thus is non - convex . here
we show that this effect goes beyond the quantum capacity and holds for the private and classical environment - assisted capacities of quantum channels .
|
a knowledge of the shell effects near the r - process path is important to discerning astrophysical scenario of nucleosynthesis @xcite .
the question whether the shell effects near the drip lines are strong or do quench has become crucial to understanding the nucleosynthesis of heavy nuclei .
the @xmath3 nuclei at the r - process path are assumed to play a significant role in providing nuclear abundances about @xmath4 . since nuclei contributing to this peak are extremely neutron - rich and are not accessible experimentally , it has not been possible to ascertain the nature of the shell effects in the vicinity of the r - process path .
due to the lack of experimental data , there prevail conflicting view points @xcite on the strength of the shell effects near the r - process path . in ref .
@xcite the shell effects in the region of the r - process path at @xmath3 were studied on the basis of the relativistic mean - field ( rmf ) theory using a successful nuclear interaction .
it was predicted @xcite that the shell effects at @xmath3 in zr nuclei are strong .
in contrast , the shell effects were suggested @xcite to be quenched on the basis of the hartree - fock - bogoliubov ( hfb ) calculations using the density - dependent skyrme force skp . in the absence of experimental data on the shell effects for @xmath3 nuclei near the r - process path , however , it is difficult to verify either of the two predictions . in view of this , we first attempt to reproduce the available data on the shell effects in the known regions and then extrapolate the method to the inaccessible region . in this paper , we examine how the shell effects evolve with isospin in the region of the astrophysically important magic number @xmath3 near the r - process path . using the experimental data available on the waiting - point nucleus @xmath2zn ( @xmath5 ) , we explore the shell effects near the r - process path in the framework of the relativistic hartree - bogoliubov ( rhb ) theory with the self - consistent finite - range pairing . in order to improve the description of the shell effects ,
we have also introduced the vector self - coupling of @xmath0 meson in addition to the non - linear scalar coupling of the @xmath1-meson in the rhb theory .
it has been shown that with this approach the high - precision experimental data on the shell effects about the stability line can be reproduced well @xcite .
this forms our basis to be able to predict as to how the shell effects evolve in the unknown region .
the shell effects are known to manifest strongly in terms of the magic numbers .
this is signified by a prominent kink about the major magic numbers as can be seen vividly in the 2-neutron separation energies ( @xmath6 ) all over the periodic table @xcite .
this is a demonstration of the existence of large shell gaps at the magic numbers in nuclei about the stability line . here
the spin - orbit interaction is pivotal in the creation of the magic numbers @xcite . in the rmf theory
@xcite the spin - orbit term arises naturally as a consequence of the dirac - lorentz structure of nucleons .
this has shown much usefulness in explaining properties which involve shell effects such as anomalous isotope shifts in stable nuclei @xcite .
the form of the spin - orbit interaction in the rmf theory has been found to be advantageous over that in the non - relativistic approaches @xcite . for nuclei near the drip line , a coupling to the continuum is required .
a self - consistent treatment of pairing is also desirable .
the framework of the rhb theory provides an appropriate tool to include both these features .
thus , in the rhb theory the advantage of a relativistic description of the rmf approach in the hartree channel is combined with that of a self - consistent finite - range pairing force .
we provide a short summary of the rhb theory in the next section . in section iii we give the details of the calculations .
the ensuing results are presented and discussed in section iv . here a detailed comparison of the shell effects using different forces and approaches is made .
the single - particle levels are used to illustrate the nature of the shell effects in the different approaches .
the last section concludes the results of the paper .
the rmf lagrangian which describes the nucleons as dirac spinors moving in meson fields is given by @xcite @xmath7 where @xmath8 is the bare nucleon mass and @xmath9 is its dirac spinor .
in addition , we have the scalar meson ( @xmath1 ) , isoscalar vector meson ( @xmath0 ) , isovector vector meson ( @xmath10 ) and the electromagnetic field @xmath11 , with the masses @xmath12 , @xmath13 and @xmath14 and the coupling constants @xmath15 , @xmath16 , and @xmath17 , respectively . the field tensors for the vector mesons are given as @xmath18 and by similar expressions for the @xmath10-meson and the photon .
for a realistic description of nuclear properties a nonlinear self - coupling @xmath19 for @xmath1-mesons has become standard .
we have added the non - linear vector self - coupling of @xmath0-meson @xcite , which is represented by the coupling constant @xmath20 .
using green s function techniques @xcite it has been shown in ref .
@xcite that a relativistic hartree - bogoliubov theory can be implemented using such a lagrangian . neglecting retardation effects
one obtains a relativistic dirac - hartree - bogoliubov ( rhb ) equations @xmath21 where @xmath22 are the quasiparticle energies and the coefficients @xmath23 and @xmath24 are four - dimensional dirac spinors normalized as @xmath25 the average field @xmath26 contains the chemical potential @xmath27 which is adjusted to the proper particle number .
the meson fields @xmath1 and @xmath0 are determined self - consistently from the klein gordon equations : @xmath28 with the scalar density @xmath29 and the baryon density @xmath30 .
the sum on @xmath31 runs only over all the particle states in the _ no - sea _ approximation .
the pairing field @xmath32 in eq .
( [ rhb ] ) is given by @xmath33 where @xmath34 are the matrix elements of a general two - body pairing interaction , and the pairing tensor is defined as @xmath35 the rhb equations ( [ rhb ] ) are a set of four coupled integro - differential equations for the dirac spinors @xmath36 and @xmath37 which are obtained self - consistently .
the solution of the dirac - hartree - bogoliubov eigenvalue equations and the meson - field equations determines the nuclear ground state .
nuclei which are known to show strong pairing correlations are treated appropriately within the framework of the rhb approach .
this is especially important when the pairing correlations in the middle of a shell become significant . on the other hand ,
the pairing correlations for nuclei 2 neutrons less or more than a magic number can be expected to be similar in the rhb and the bcs approach .
this is due to the reason that the pairing correlations in such nuclei are reduced to a minimal level .
however , in the present work , we have performed the rhb calculations for all nuclei . for the pairing channel
, we have taken the finite - range gogny force d1s .
it is known to represent the pairing properties of a large number of finite nuclei appropriately @xcite .
the rhb approach becomes important while dealing with nuclei far away from the stability line and in particular for those in the vicinity of the drip lines where the fermi level is usually very close to the continuum .
the rhb theory which is based upon the quasi - particle scheme employing bogoliubov transformations takes into account the coupling between the bound states and the states in the continuum .
the pairing and the coupling to the continuum can be important in nuclei with a pair of particles above the shell closure in determining the shell effects especially as the fermi surface is very close to the continuum .
therefore , the rhb theory with the self - consistent pairing serves as an ideal scheme for this purpose .
the rhb calculations have been performed for nuclei in the neighbourhood of the waiting - point zn nuclei and for nuclei from r - process path to the neutron drip line .
nuclei are treated within a spherically symmetric configuration .
the method of expansion of the wavefunctions into the harmonic oscillator basis is employed @xcite .
we have taken 20 shells both for the fermionic as well as bosonic wavefunctions for the expansion .
a comparative study is made between the lagrangian model with the non - linear scalar self - coupling of the @xmath1 meson and the lagrangian with both the non - linear scalar self - coupling of @xmath1 and the vector self - coupling of @xmath0 meson .
for the former , we have used the force nl - sh which is a typical representative for this lagrangian model . in a large number of studies
it has been shown that the force nl - sh @xcite is able to provide a very good description of the ground - state properties of nuclei all over the periodic table .
it has been found to be especially useful for exotic nuclei near the drip lines .
recently , the force nl3 @xcite has been brought out with a view to improve the description of the giant monopole resonance . however , as shown in ref .
@xcite ground - state properties of nuclei predicted by nl3 are very similar to those of nl - sh . for the lagrangian with the non - linear vector self - coupling of @xmath0 meson which is in addition to the usual non - linear scalar self - coupling of @xmath1 meson
, we have used the forces nl - sv1 and nl - sv2 .
these forces have been developed recently @xcite with a view to bring about an improvement in the ground state properties of nuclei vis - a - vis the scalar self - coupling .
consequently , it has been shown by sharma et al .
@xcite that the introduction of the vector self - coupling was a remedial measure in respect of the shell effects in nuclei about the stability line .
in addition , the high - density equation of state ( eos ) of the nuclear matter becomes softer as compared to that with the scalar self - coupling only .
this would make the lagrangian model with the vector self - coupling compatible with the neutron star masses .
the detailed properties of these forces will be provided elsewhere @xcite .
in our attempt to scrutinize the shell effects , we started our investigation @xcite of the shell effects in nuclei at the stability line , where we examined the role of @xmath1- and @xmath0-meson couplings on the shell effects in ni and sn isotopes .
it was observed that the existing nuclear forces based upon the nonlinear scalar self - coupling of @xmath1-meson exhibit shell effects which were stronger than suggested by the experimental data . in order to remedy this problem ,
the nonlinear vector self - coupling of @xmath0-meson in the rhb theory was introduced . as a result of this
, the experimental data on the shell effects in ni and sn nuclei on the stability line were reproduced well @xcite .
having established the functional basis , we consider nuclei around @xmath2zn ( @xmath5 ) in the present work .
the isotope @xmath2zn is assumed to be a waiting - point nucleus in the chain of r - process nucleosynthesis and comes closest to the @xmath38-stability line for nuclei with @xmath5 .
the experimental @xmath6 value for this nucleus is about 11 mev .
the nuclei @xmath2zn and @xmath39zn being very rich in neutrons would serve to test the theoretical models far away from the stability line . here , we present the results on the shell effects in the waiting - point zn nuclei about @xmath5 using different model lagrangians .
the @xmath6 values for the zn isotopes obtained from the rhb calculations are shown in fig . 1 .
the experimental data is taken from the 1995 mass tables @xcite .
because the mass of the nucleus @xmath39zn is based upon the systematics , there may be a slight uncertainty in the @xmath6 value of the nucleus @xmath39zn .
thus , the data point for @xmath39zn is shown in fig . 1 enclosed within a diamond . the rhb calculations [ fig .
1(a ) ] with nl - sh ( scalar self - coupling only ) show a very large kink at @xmath5 implying that the shell effects with nl - sh are much stronger than exhibited by the available data . in comparison ,
the shell gap with nl - sv2 ( both the scalar and vector self - couplings ) is reduced as compared to nl - sh .
it is , however , still larger than the data shown .
the reduction in the shell gap with the vector self - coupling is consistent with that observed in ni nuclei at the stability line vis - a - vis the scalar self - coupling only .
we show in fig .
1(b ) the @xmath6 values obtained with the other vector self - coupling force nl - sv1 .
the slope of the kink shows that the shell gap with nl - sv1 agrees well with the available data .
thus , rhb with the vector self - coupling lagrangian nl - sv1 is able to reproduce the available data on the shell effects in the waiting - point region satisfactorily .
we also show in fig .
1(b ) the non - relativistic hfb calculations with the skyrme force skp @xcite .
the skp results produce nearly a straight line passing through the @xmath5 point .
a comparison of these results with the data shows that the shell effects with skp are quenched strongly .
this is reminiscent of the behaviour of skp as observed in ref .
@xcite at the stability line , where the shell effects were found to be quenched strongly in ni isotopes at @xmath40 in contrast to the experimental data .
this feature of skp is evidently due to its high effective mass @xmath41 which increases the level density near the fermi surface and thus suppresses the shell gaps significantly @xcite .
nuclei in the vicinity of @xmath42 with the magic neutron number @xmath3 are expected to lie on or close to the r - process path .
we have chosen the isotopic chains @xmath43 with neutron numbers running across the magic number @xmath3 .
the r - process path is expected to pass through these nuclei especially those with the higher @xmath44 values .
some nuclei in these chains are close to the neutron drip - line as the neutron separation energy approaches a vanishing value .
how the shell effects behave in nuclei near the r - process path is an open question ?
this is due to the fact that the @xmath3 nuclei near the r - process path possess an extremely large neutron - to - proton ratio .
although a tremendous progress has been made in the last decade in synthesizing very neutron - rich nuclei in the laboratory , such nuclei are still very far from being realized .
the last such nucleus attained in this region in the laboratory is @xmath45cd ( @xmath46 and @xmath3 ) .
the mass of the nucleus @xmath47cd with @xmath48 is still not known .
the knowledge of the latter would allow to discern the shell effects in regions far away from the stability line . for @xmath3 nuclei with @xmath49 ,
i.e. , in the region of zr ( @xmath50 ) , the magnitude of the nuclear isospin becomes increasingly large and hence the difficulty in producing these nuclei . in the present paper
, we examine the shell effects in these nuclei , therefore , in the light of the data available in the other regions .
we also study how the shell effects evolve as one approaches the continuum in moving from the r - process path towards the neutron drip - line .
having described the available data on the waiting - point zn nuclei by the forces with the vector self - coupling as shown in fig . 1 , we extend our formalism to explore the inaccessible region of the r - process path and the neutron drip - line about @xmath3 . with a view to visualize the evolution of the shell effects with the isospin
, we show our rhb results on the @xmath6 values for the isotopic chains from the higher @xmath44 to the lower ones using the two lagrangian models .
2 shows the corresponding values for the pd ( @xmath51 ) and ru ( @xmath52 ) nuclei across the magic number @xmath3 .
the results [ fig . 2 ] with the non - linear scalar self - coupling ( nl - sh ) for the pd and ru isotopic chains show a shell gap at @xmath3 , which is largest amongst all the forces shown here . with the force nl - sv2 with the vector self - coupling of the @xmath0-meson ,
the shell effects are milder as compared to nl - sh .
this feature is similar to that observed at the stability line @xcite and also in fig . 1 .
furthermore , the results with the force nl - sv1 show that the shell gap at @xmath3 is reduced as compared to nl - sv2 .
this is again similar to that seen for the zn isotopes . for a comparison
, we also show the hfb results obtained with the interaction skp .
the shell gap with skp is much reduced as compared to the forces with the vector self - coupling .
thus , the shell effects with nl - sv1 are stronger as compared to the hfb+skp results shown in the figure .
it is interesting to note that there is a remarkable agreement in the results below and up to the magic number for all the forces except for nl - sh .
the neutron separation energy from these results for @xmath3 pd and ru nuclei is about 3 - 4 mev .
accordingly , these nuclei are expected to lie well on the r - process path as per the conventional wisdom .
we show the rhb results for the isotopic chains of mo ( @xmath53 ) and zr ( @xmath42 ) in fig . 3 .
from the argument of the neutron separation energy , the mo and zr nuclei about the magic number are also expected to lie along the r - process path .
thus , both the figs . 2 and 3
portray how the shell effects evolve in moving to nuclei with larger neutron excess .
the main feature that the shell effects with nl - sh are strong is also displayed by fig .
it is worth reminding that the force nl - sh was also used in an earlier work @xcite to calculate the shell effects at @xmath3 in zr nuclei about the drip line .
the shell effects with nl - sh were noted to be similarly strong in the previous results .
however , the previous results with nl - sh were based upon the bcs pairing and there was no mechanism in place to include the effects of the coupling to the continuum .
the qualitative behaviour that the shell effects with nl - sh are strong is reproduced here also in the present rhb calculations .
the reason for this overestimation of the shell effects as also noted in ref .
@xcite is that the properties which were considered mainly in constructing these forces such as nl - sh and others were the binding energies and charge radii of nuclei .
the shell effects as such were not included in such considerations heretofore . as shown in ref .
@xcite , the shell effects emerge as an observable which should be employed to constrain the nuclear interaction . with the force nl - sv2 with the vector self - coupling of the @xmath0-meson , the shell effects [ figs .
2 - 3 ] are milder as compared to nl - sh for all the chains .
this is similar to that observed at the stability line @xcite and also in fig .
the results with our benchmark force nl - sv1 show that the shell gap at @xmath3 is reduced as compared to nl - sv2 .
this is again similar to that observed for the zn isotopes .
the @xmath6 values obtained with the hfb calculations using the skyrme interaction skp are also shown for comparison .
the magnitude of the kink at n=82 shows that the shell effects with hfb+skp are much weaker as compared to nl - sv1 .
the quenching predicted by skp in this region is as expected .
as noted earlier , hfb+skp is known to exhibit a strong quenching at the stability line @xcite and the same has also been observed about the waiting - point nucleus @xmath2zn as shown in fig .
1(b ) above . in comparison , the rhb approach with nl - sv1 reproduces the shell effects about the stability line @xcite as well as in the waiting - point region at @xmath5 . for any given force
, the shell gap at @xmath3 shows a steady decrease in moving from pd ( @xmath51 ) to zr ( @xmath42 ) . as nuclei
become increasingly neutron rich , the gradual decrease in the @xmath6 values and also in the corresponding shell gap shows the evolution of the shell effects in going away from the r - process path towards the drip line .
this reduction in the shell gap is largest for skp which shows intrinsically weaker shell effects . as seen from fig .
3(b ) for the zr isotopes , the curve for @xmath6 has become nearly a straight line , implying a disappearance of magicity with skp .
this picture is at variance with that from nl - sv1 .
the persistence of the kink with nl - sv1 in figs .
2 and 3 shows that with nl - sv1 the shell effects for the r - process nuclei at @xmath3 are much stronger than with skp .
thus , notwithstanding the fact that nl - sv1 is commensurate with the available data along the stability line and in the waiting point region , it can be stated that the shell effects in @xmath3 nuclei on the r - process path still remain stronger .
nuclei become increasingly unbound and a coupling to the continuum arises in going to extremely large neutron to proton ratios .
this situation is expected to arise naturally in going to @xmath3 nuclei below @xmath42 .
for this situation , we show the results for sr ( @xmath54 ) and kr ( @xmath55 ) isotopes in fig . 4 .
the results show that the @xmath6 value for the @xmath3 nuclei decreases from @xmath56 2.5 mev for sr isotopes to @xmath56 1.5 mev . the corresponding @xmath57 value would amount to a merely 1 mev or less . on the basis of the @xmath57 , the @xmath3 kr nucleus represents the arrival of the drip line .
the shell effects with nl - sv1 are seen to become successively weaker ( which is true for all the forces ) as one moves to nuclei with higher isospin such as @xmath58sr [ fig .
4(a ) ] and @xmath59kr [ fig .
4(b ) ] . here ,
the case of the kr isotopes [ fig .
4(b ) ] deserves a mention . except with nl - sh ,
all the other forces show that the shell effects are washed out significantly .
this stems from the fact that for @xmath59kr the fermi energy is very close to the continuum and the nucleus is pushed to the very limit of binding .
the binding energy of an additional neutron is then close to zero and the shell gap ceases to exist .
thus , any semblance of the shell effects for @xmath59kr ( @xmath3 ) is completely lost .
this situation is destined to befall on a nucleus or nuclei for every magic number , leading to a termination of the corresponding shell gap .
all the forces nl - sv2 , nl - sv1 and skp predict such a behaviour except nl - sh .
this situation at the drip line at the magic number can be termed as _
terminus stratum_. the nuclei close to the drip line show no binding to an extra neutron to be able to form a higher mass isotope .
such nuclei at the drip line ( @xmath60 ) would not contribute to the r - process nucleosynthesis .
in view of this , it is therefore important to understand the nature of the shell effects for nuclei along the r - process path , which are assumed to play an important role in the nucleosynthesis , rather than those at the drip line
. we show in fig . 5 the neutron single - particle levels obtained with nl - sv1 for @xmath61 nuclei from a few chains .
our focus is the evolution of the shell gap at @xmath3 as one approaches the drip line ( @xmath60 ) from the r - process path . as discussed in the subsection above , the mo and zr nuclei
are expected to lie on the r - process path , whereas the sr and kr nuclei should be very close to the drip line .
the single - particle levels shown in this figure portray essentially the results of figs .
the fermi energy ( shown by the dashed lines ) approaches the continuum as one moves towards a larger neutron to proton ratio .
the shell gap ( @xmath3 ) shows a constant decrease as the last neutrons become more and more unbound in going from mo to kr . for @xmath59kr ( @xmath3 )
the fermi energy is close to zero and the shell gap merges into the continuum .
however , for nuclei on the r - process path such as mo and zr shell gaps do remain large . for the corresponding nuclei of pd and ru , the shell gaps are ( not shown here ) even larger than that for mo .
the difference in the response of the various forces near the drip line is illustrated in the single - particle levels shown for @xmath58zr in fig .
6 . since this nucleus has 2 neutron less than the magic numbers , all the occupied levels are bound .
moreover , we do not see any indication of a partial occupancy of any level above the fermi energy , which might arise out of the bogoliubov pairing .
the figure shows that the @xmath3 shell gap is largest with nl - sh and it shows a reduction in going to nl - sv1 .
this is consistent with the general behaviour of the shell effects from nl - sh to nl - sv1 as seen in figs . 2 - 4 .
on the other hand , the shell gap with skp is reduced significantly as compared to nl - sv1 .
the reduced shell strength with skp seems to be generic as shown above . in fig .
7 we have shown the neutron single - particle levels for the nucleus @xmath62zr with 2 neutrons more than the magic number @xmath3 . here we have made a comparison of the levels for nl - sh , nl - sv1 and skp .
the shell gap at @xmath3 shows a decreasing trend from nl - sh to skp as seen also in fig .
however , it can be noticed that the extra 2 neutrons above the magic number go to the levels in the continuum and therefore with the force nl - sh and nl - sv1 this nucleus is unbound .
this should imply that @xmath62zr is already a drip line nucleus , whereas the same is not the case for @xmath58zr ( see fig .
the single - particle levels with skp show a remarkable difference from those of the rmf forces . with skp
the @xmath63 level is below the continuum and the extra 2 neutrons occupy this level predominantly .
thus , the nucleus @xmath62zr is bound within the hfb+skp calculations in contrast to the rmf forces .
the spin - orbit potential is the key ingredient of the nuclear interaction responsible for creation of the magic numbers and the shell gaps . as mentioned earlier , the spin - orbit interaction arises naturally in the rmf theory as a result of the dirac - lorentz structure of nucleons .
the @xmath1 and @xmath0 potentials contribute constructively to generate the required spin - orbit strength @xcite . this is able to reproduce the experimental data on spin - orbit splittings in nuclei such as @xmath64o and @xmath65ca
. it has been shown by several authors @xcite that the effective mass in the rmf theory can be constrained by the spin - orbit splittings in light nuclei to a value close to 0.60 .
however , the shell gaps and thus the shell effects in heavier nuclei can not be addressed adequately by fixing the spin - orbit splittings in @xmath64o only .
this becomes clear by examining the successful rmf forces such as nl - sh and nl3 which show stronger shell effects as compared to the experimental data in heavy nuclei @xcite .
this implies that the shell effects in heavier nuclei need to be taken into account in order to make a successful nuclear interaction . in this respect ,
the experimental data on @xmath6 values around the shell closure provide an additional useful ground - state observable .
the advantage which the rmf theory derives from its intrinsic spin - orbit interaction is not enjoyed by the non - relativistic approaches such as the hartree - fock skyrme theories .
the spin - orbit potential is usually added on an ad - hoc basis and the strength of the spin - orbit potential is fitted in order to reproduce the spin - orbit splittings in light nuclei in these approaches .
this may , however , not suffice for the shell effects in heavy nuclei as illustrated for the case of the rmf theory .
the spin - orbit ( s.o . ) potential plays a central role in determining the magnitude of the spin - orbit splittings in nuclei .
we show in fig .
8 the spin - orbit potential obtained in the rhb theory for the lagrangian parameters nl - sh , nl - sv2 and nl - sv1 .
the contribution of the @xmath1 and @xmath0 terms add together to give rise to the spin - orbit potential which is predominant in the surface region of the nucleus @xcite . the s.o .
potential is largest for the non - linear scalar force nl - sh .
this is consistent with the strong shell effects observed with nl - sh at the shell closures in nuclei .
the depth of the s.o .
potential is reduced by about 10 % for the force nl - sv2 .
it is reduced further with nl - sv1 . this reduction from nl - sh to nl - sv2 and to nl - sv1 is consistent with all the results we have discussed on the shell effects with these forces in this work .
this underlines the point that the spin - orbit interaction plays a key role in determining the strength of the shell effects .
this would require a an adjustment of both the @xmath1 and the @xmath0 fields in order to produce a suitable spin - orbit field .
we have investigated the evolution of the shell effects in nuclei from the r - process path to the neutron drip line .
the theoretical approach that is employed in the present work is that of the relativistic hartree - bogoliubov theory based upon the quasi - particle scheme with the self - consistent pairing .
the rhb approach with the inclusion of the vector self - coupling of @xmath0 meson was developed with a view to remedy the problem of the strong shell effects with the lagrangian with the scalar self - coupling only .
having reproduced the shell effects with the vector self - coupling at the stability line in our earlier work @xcite , we have attempted to reproduce the available data on the shell effects in the waiting - point region of @xmath2zn .
the available data on zn isotopes are reproduced well within the rhb approach with the vector self - coupling .
this establishes further the reliability of the working basis we employ . with the above approach we have carried out the rhb calculations for nuclei along the r - process path and near the neutron drip - line using the two different lagrangian models .
it is shown that the shell effects with the vector self - coupling are in general softer than those with the scalar self - coupling .
this is found to be the case over a broad range of nuclei considered in our work .
this implies that the shell structure near the fermi surface with the vector self - coupling forces is denser than with the lagrangian with the scalar self - coupling .
this is consistent with what was found to be the case also for nuclei about the stability line .
the shell gap at @xmath3 shows a gradual decrease in going from the r - process path towards the neutron drip line for all the forces considered .
this is the natural behaviour for the shell gaps to evolve with the isospin .
however , we show that for nuclei on the r - process path , the shell effects at @xmath3 still remain strong vis - a - vis a quenching exhibited by hfb+skp .
this quenching with skp has been discussed extensively in the literature @xcite .
the origin of the shell quenching stems primarily from the very large effective mass of skp which leads to a significant compression of the shell structure near the fermi surface .
a quenching has been requested for an improved fit to the global r - process abundances @xcite .
however , a quenching does not seem not to be compatible with the data at the stability line and in the waiting - point region . in moving away from the r - process path where nuclei are expected to have @xmath66 mev to nuclei where @xmath57 is vanishingly low , the shell gap at n=82 merges into the continuum .
consequently , the ensuing shell structure above n=82 is washed out significantly . since a nucleus can not take on any more neutrons , nuclei in the vicinity of the drip line are , therefore , not expected to contribute to the r - process nucleosynthesis .
kratz j.p .
bitouzet , f.k .
thielemann , p. mller and b. pfeiffer , astrophys . j. * 403 * , 216 ( 1993 ) .
sharma , g.a .
lalazissis , w. hillebrandt , and p. ring , phys .
* 72 * , 1431 ( 1994 ) ; * 73 * , 1870 ( 1994 ) .
j. dobaczewski i. hamamoto , w. nazarewicz and j.a .
sheikh , phys .
* 72 * , 981 ( 1994 ) .
sharma , a.r .
farhan and s. mythili , phys .
c * 61 * 054306 ( 2000 ) . c. borcea
_ et al . _ ,
phys . a * 565 * , 158 ( 1993 ) .
mayer and j.h.d .
jensen , _ elementary theory of nuclear structure _
( wiley , new york , 1955 ) .
serot and j.d .
walecka , adv .
* 16 * , 1 ( 1986 ) .
sharma , g.a .
lalazissis , and p. ring , phys .
b * 317 * , 9 ( 1993 ) .
sharma g.a .
lalazissis , j. knig , and p. ring , phys . rev . lett .
* 74 * , 3744 ( 1994 ) .
bodmer , nucl .
a * 526 * , 703 ( 1991 ) .
gorkov , sov .
jetp * 7 * , 505 ( 1958 ) . h. kucharek and p. ring , z. phys .
a * 339 * , 23 ( 1991 ) .
berger , m. girod , and d. gogny , nucl .
a * 428 * , 32 ( 1984 ) .
gambir , p. ring and a. thimet , ann .
( n.y . ) * 198 * , 132 ( 1990 ) .
sharma , m.a .
nagarajan and p. ring , phys .
b * 312 * , 377 ( 1993 ) . g.a .
lalazissis , j. knig , and p. ring , phys .
c * 55 * , 540 ( 1997 ) .
sharma , in preparation ( 2001 ) .
g. audi and a.h .
wapstra , nucl .
a * 595 * , 409 ( 1995 ) .
j. dobaczewski , h. flocard , and j. treiner , nucl .
a * 422 * , 103 ( 1984 ) .
reinhard , rep .
* 52 * ( 1989 ) 439 .
sharma , m.a .
nagarajan , and p. ring , ann .
( n.y . ) * 231 * , 110 ( 1994 )
. b. chen , j. dobaczewski , k .-
kratz , k. langanke , b. pffeifer , f .- k .
thielemann , and p. vogel , phys .
b * 355 * , 37 ( 1995 ) . * the @xmath6 values for zn isotopes in the waiting - point region obtained with rhb using ( a ) the scalar self - coupling force nl - sh and the vector self - coupling force nl - sv2 and ( b ) with the vector self - coupling force nl - sv1 along with the nonrelativistic hfb+skp results .
the experimental data is shown for comparison .
the datum on @xmath39zn is adopted from the mass systematics @xcite and is shown enclosed within a diamond . * the rhb results on the shell effects at @xmath3 in the r - process nuclei of ( a ) pd and ( b ) ru .
the hfb+skp results are also shown for a comparison . * the same as fig .
2 for the r - process nuclei ( a ) mo and ( b ) zr . the @xmath6 values and shell gaps show a gradual decline in moving from pd to zr . * the same as fig . 2 for the drip line nuclei near @xmath3 for the sr and kr isotopic chains .
the shell gap and the @xmath6 values show a vanishing trend for the kr isotopes about @xmath3 . *
the neutron single - particle levels for the mo , zr , sr and kr nuclei ( @xmath61 ) showing an evolution of the shell gap in moving from the r - process path towards the drip line , with the force nl - sv1 .
the fermi energy is shown by the dashed lines .
the fermi energy for the kr nucleus near the drip line is seen to merge with the continuum . * the shell gap @xmath3 for the nucleus @xmath58zr using the various interactions . * the neutron single - particle levels for the nucleus @xmath62zr with 2 neutrons above the closed shell .
the reduction of the @xmath3 shell gap from nl - sh to nl - sv1 to skp is clearly evident .
in contrast to the rmf forces , the @xmath63 level is bound with skp . * the neutron spin - orbit potential for @xmath58zn obtained in the rhb calculations with the various rmf forces .
the predominance of the spin - orbit potential in the surface region is clearly evident .
|
we have investigated the evolution of the shell structure of nuclei in going from the r - process path to the neutron drip line within the framework of the relativistic hartree - bogoliubov ( rhb ) theory . by introducing the quartic self - coupling of @xmath0 meson in the rhb theory in addition to the non - linear scalar coupling of @xmath1 meson ,
we reproduce the available data on the shell effects about the waiting - point nucleus @xmath2zn . with this approach
, it is shown that the shell effects at @xmath3 in the inaccessible region of the r - process path become milder as compared to the lagrangian with the scalar self - coupling only .
however , the shell effects remain stronger as compared to the quenching exhibited by the hfb+skp approach .
it is also shown that in reaching out to the extreme point at the neutron drip line , a terminal situation arises where the shell structure at the magic number is washed out significantly .
= 10000
|
the acis operations team , combining elements from the cxc and the acis mit@xmath2ipi team , has been developing and testing the new squeegee modes since the spring of 2000 . in squeegee mode ,
charge is collected in the top few rows of the ccd and then swept across the imaging array once per readout , thus filling some of the radiation - induced electron traps that cause degraded performance .
the development of a novel method of reading out the acis ccds was first developed and tested on ground ccds that are similar to flight ccds ( see prigozhin , _ et al . _
2000 for a characterization of the radiation damage of acis ccds ) .
this new mode ameliorates some of the effects caused by the radiation damage suffered early in the mission by the fi ccds . for a more comprehensive discussion on the design and clocking method of squeegee mode , see bautz and kissel ( 2000 ) .
a description of the different squeegee modes is included in an internal mit@xmath2acis memo by bautz and grant ( 2000 ) .
that memo also presents an analysis of the squeegee modes and describes the trade - offs in choosing between the various squeegee modes . in this paper , we present the results of the various squeegee measurements on the i0 ccd .
.*list of i0 squeegee tests * [ cols="<,<,<,<,<",options="header " , ] this analysis utilized the standard level 0 and level 1 data products produced by the cxc data system .
the datasets included in this analysis are listed in table 1 .
the control run ( obsid 62895 ) was a long charge transfer inefficiency ( cti ) measurement .
there are five `` flavors '' of squeegee discussed in this paper : l1 `` vanilla '' , l2 `` maximum spectral resolution '' , l4 `` maximum angular resolution '' , l7 `` maximum observing efficiency '' , and l9 `` sub - array near the aimpoint '' .
table 1 summarizes the important parameters which distinguish one squeegee mode from another : static integration time , number of squeegee rows ( 2 or 16 ) , number of rows in the exclusion window ( 8 or 24 or 32 ) , on - chip summing ( yes or no ) , parallel transfer time ( @xmath3s or @xmath4s ) , number of frame flushes ( always 24 for these squeegees ) , and number of reverse clocks ( 1010 or 1026 + 32 ) .
in addition , the table includes the average exposure time for these measurements . with all of these squeegee modes ,
the exposure time varies from row - to - row .
the number listed is the correct exposure time for the middle of the ccd or for the middle of the sub - array for squeegee l9 . in the analysis presented in this paper ,
the row - to - row variation in exposure time has been included by computing the exposure for each of the 32 row elements .
data from each i0 measurement was separated by node and then further reduced to thirty - one , 32-row region files .
spectra files were generated from the cxc level 1 events files .
events were selected in the asca g02346 grade set and were filtered on node and chipy coordinates .
the gain was computed using the so - called `` local - gain '' method , meaning the fitted peak in each 32 row element is used to compute a conversion from adus to ev for that element .
the prominent lines of the internal calibration source were then fit using gaussians ; the figure of merit employed was the c - statistic .
the results of this analysis are presented in figures 1 and 2 .
figure 1 shows the fitted values of the fwhm in ev for the i0 ccd at 1.5 kev ( al - k@xmath0 ) at a focal plane temperature of -110 c , -120 c , and at -120 c using the `` vanilla '' squeegee mode . for comparative purposes , the s3 ccd fwhm at -120 c
is also overlaid .
figure 2 shows the fitted values of the fwhm in ev for the i0 ccd at 5.9 kev ( mn - k@xmath0 ) at a focal plane temperature of -110 c , -120 c , and at -120 c using the `` vanilla '' squeegee mode . for comparative purposes , the s3 ccd fwhm at -120 c
is also overlaid .
what is clear from both plots is that squeegee mode improves the spectral resolution of i0 ccd compared to standard clocking .
figures 3 and 4 displays the quantities asca g0234/all grades , asca g02346/all grades , asca g7/all grades versus row number , where `` all grades '' means all events which pass the on - board filtering and make it into telemetry , for the al - k@xmath0 and mn - k@xmath0 lines .
all of these measurements were executed using an upper event amplitude cutoff of 3750 adus and a grade filter which rejected flight grades 24 , 66 , 107 , 214 , and 255 .
figure 5 is a plot of the detection efficiency for each line as a function of row number .
we have included a statistical error bar for the count rate plots at the location of the first data element which is representative of the uncertainties in these measurements . in producing figure 5 ,
we have corrected for the decrease in intensity of the radioactive source as the data span 6 months .
see 3.4 for further discussion .
= 4.0 in = 4.0 in = 4.0 in = 4.0 in = 3.0 in = 3.0 in figure 3 shows the grade distribution as a function of row number for the squeegee measurements and the control run .
the enhanced cti of the fi ccds causes events to migrate from asca grades g0234 to higher grades .
this can be seen in the control run and the squeegee runs .
the effect is quite small at al - k@xmath0 , never more than 3% for the grade sets listed here .
squeegees l2 and l4 show a very different behavior with regards to grade migration when the individual grades are examined .
analysis of the g0 events as a function of row number show that the percentage of grade g0 events is dropping from @xmath5 in the first resolution element to @xmath6 by the top of the ccd .
asca grade 2 is the beneficiary of these counts as it increases from its nominal value of 5% to over 80% of the events at the top of the ccd .
asca grade 2 is composed of flight grades 2 ( down splits ) and 64 ( up splits )
. a further examination of the data reveals that all of the migrating events are going into flight grade 64 .
the one characteristic which the squeegees l2 and l4 have in common is the slow parallel transfer .
lastly , l2 and l7 both have much higher g0234@xmath2all grades ratios than the other modes because both l2 and l7 utilize 2x2 summing .
we suggest that this severe grade migration provides a clue to the cause of the lower detection efficiency seen in figure 5 .
a common cause of reduced efficiency in a given grade combination is that events are migrating to grades outside of the chosen set .
we will defer a full discussion of this until section 3.4 in which the decay in the intensity of the radioactive source is included .
= 3.0 in = 3.0 in the grade distribution versus number row number data are included in figure 4 .
the squeegee runs now show greater variability with respect to each other and a larger difference with respect to the control run .
particularly interesting is the grade migration exhibited by squeegee l4 .
the fraction of g02346 events shows a linear decrease from row 100 to row 400 , at which point it flattens out and is much lower than any other measurement .
this behavior is also seen in the migration to g7 events .
the same effect is seen in l2 but at a reduced level .
this is presumably due to the fact that l2 utilizes on - chip summing , which should retain more of the valid x - ray events in the g02346 grade set .
the fraction of g02346 and g7 events in the sub - array squeegee l9 lies between l2 and l4 .
the shorter frame - time of l9 compared to l4 appears to reduce some of this migration indicating that the effect can not be attributed solely to the short timescale traps . the grade migration effect seen at al - k@xmath0 for squeegees l2 and
l4 is more pronounced at mn-@xmath7 .
analysis of the individual grade `` branching ratios '' shows an even stronger migration from grades g0 , g3 , and g4 to g2 than for al-@xmath7 .
squeegee l4 also shows a steep increase in the g7 events from row 100 to row 500 .
the flight grade analysis shows that the migrating events are going into flight grade 64 .
it is interesting to note that the percentage of g2 events is larger in l2 than l4 .
the difference between these two squeegees is that l2 uses on - chip summing while l4 does not .
this suggests that a significant amount of charge is trailing the center of the event by 2 - 3 or perhaps even more pixels .
lastly , l2 and l7 both have much higher g0234@xmath2all grades ratios than the other modes because both l2 and l7 utilize 2x2 summing .
the results for the ti-@xmath7 line are not presented in this paper due to space restrictions .
however , the ti-@xmath7 data for i0 resemble the mn-@xmath7 data .
the grade distribution results for ti are also similar to the mn results as one would expect given that the performance characteristics of the ccd vary little from 4.5 kev to 5.9 kev .
these squeegee tests were run between 3 and 6 months after the control run .
the half - life of the @xmath8 source is 2.7 yr .
therefore , the intensity of the source decays by @xmath9 in three months and by @xmath10 after six months . in figure 5
we have corrected for this decrease by simply normalizing the detected count rates of the squeegee measurements to the control run . after applying this correction ,
squeegees l1 and l7 have the same or higher detection efficiency than the control run across the ccd , with the possible exception of l7 at al - k@xmath0 .
the upturn in the detection efficiency for the control run for al - k@xmath0 is probably an artifact of the fitting process .
it is interesting that the squeegee runs show a decrease of the detection efficiency with row number relative to the control run at al - k@xmath0 , but squeegees l1 and l7 match or exceed the control run at mn - k@xmath0 and ti - k@xmath0 .
l2 appears to be still lower than the control run near the frame - store and drops more rapidly with row number than the control run .
l4 still shows the largest drop with row number , but may now be consistent with the control run near the frame - store .
the most important conclusion to be drawn from these data is that squeegees l1 and l7 _ have the same detection efficiency as the standard clocking _ at mn - k@xmath0 and ti - k@xmath0 .
= 3.0 in = 3.0 in the question still remains as to the lower detection efficiency of l2 and l4 .
the detection efficiency of l4 is @xmath6 lower than the standard clocking at the top of the ccd .
only about 6% of this difference can be explained by grade migration to asca g7 .
these measurements were executed with the now - standard on - board rejection of flight grades 24 , 66 , 107 , 214 , and 255 .
of these , flight grade 66 may be the most interesting because it is the closest to flight grade 64 which is exhibiting the tremendous increase .
perhaps there is also a significant migration to grade 66 , which would lead to a lower detection efficiency since these events are never telemetered .
it is also possible that flight grade 255 is enhanced by this effect .
the variation with row number for l2 and l4 confirm what our analysis of the s0 data indicated ( virani and plucinsky 2000 ) .
the detection efficiency is changing with row number and the percentage decreases are the same for the squeegee runs which utilize the slow parallel transfer .
the severe grade migration effect appears to lower the effective detection efficiency around row 200 for mn - k@xmath0 and row 700 for ti - k@xmath0 .
the al - k@xmath0 data are not effected until about row 900 . with the improved statistical precision of the ti - k@xmath0 data
, we can start to see the energy dependence of this effect .
clearly , the mn - k@xmath0 photons are effected sooner than the ti - k@xmath0 .
this is suggestive of a strong dependence on energy in a rather narrow range from 4.5 to 5.9 kev .
one suggested explanation for the lower detection efficiency is that the analysis is confused by the blending of the k@xmath0 and k@xmath11 lines as the spectral resolution degrades .
we note that this effect should vary in magnitude with row number since the resolution is degrading with row number .
however , squeegees l2 and l4 show a discrepancy with respect to the control run in the first 200 rows .
line blending is not an issue in the first 200 rows on i0 since the spectral resolution is still close to pre - launch values and is more than sufficient to resolve the k@xmath0 and k@xmath11 lines for mn and ti .
line blending may be part of the explanation near the top of the ccd as the discrepancy between the squeegees and the control runs increases with row number .
nevertheless , the lower detection efficiency near the frame - store is a puzzling effect which warrants more investigation .
we confirm that all of the tested squeegee modes improve the spectral resolution of the i0 ccd compared to the standard clocking .
our analysis of the detection uniformity indicates that squeegees l1 and l7 have the same detection efficiency as the standard clocking after correcting for the decay in the intensity of the radioactive source , while the squeegees l2 , l4 , and l9 still exhibit a lower detection efficiency .
the discrepancy is as large as 20% for squeegee
l4 at the top of the ccd . l2 and l4 also produce a highly spatially - dependent grade distribution .
we suggest that the slow parallel transfer of both these modes is the likely explanation for this effect .
we suggest that this effect should be investigated further with the hope that a squeegee mode can be developed which optimizes the spectral resolution , the detection efficiency , and the _ uniformity _ of the detection efficiency .
we thank our acis and cxc colleagues , particularly mark bautz , dan schwartz , and peter ford , for many useful ideas over the course of this analysis .
snv and ppp acknowledge support for this research from nasa contract nas8 - 39073 ; ceg and bl acknowledge support for this research from nasa contracts nas8 - 37716 and nas8 - 38252 .
|
the mit and cxc acis teams have explored a number of measures to ameliorate the effects of radiation damage suffered by the acis fi ccds .
one of these measures is a novel ccd read - out method called `` squeegee mode '' .
a variety of different implementations of the squeegee mode have now been tested on the i0 ccd . our results for
the fitted fwhm at al - k@xmath0 and mn - k@xmath0 clearly demonstrate that all the squeegee modes provide improved performance in terms of reducing cti and improving spectral resolution .
our analysis of the detection efficiency shows that the so - called squeegee modes `` vanilla '' and `` maximum observing efficiency '' provide the same detection efficiency as the standard clocking , once the decay in the intensity of the radioactive source has been taken into account .
the squeegee modes which utilize the slow parallel transfer ( `` maximum spectral resolution '' , `` maximum angular resolution '' , and `` sub - array '' ) show a significantly lower detection efficiency than the standard clocking . the slow parallel transfer squeegee modes exhibit severe grade migration from flight grade 0 to flight grade 64 and a smaller migration into asca g7 .
the latter effect can explain some of the drop in detection efficiency .
there are a few observational penalities to consider in using a squeegee mode . utilizing
any squeegee mode causes a loss of fov near the aimpoint ( 4 to 16@xmath1 strips along the full length of the ccds ) , as well as the attendant dead - time increase .
secondly , the cost of the software implementation and its testing will be significant .
lastly , each squeegee mode `` flavor '' would require lengthy , mode - specific calibration observations .
therefore , since an efficacious , ground - based cti corrector algorithm is now available ( see paper by plucinsky , townsley , _ et al . _ in this proceedings ) , a scientific judgment will have to be made to determine which , if any , squeegee modes should be developed and calibrated for use by chandra observers .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
|
molecular dynamics ( md ) simulations allow for analysis and understanding of the dynamical behaviour of molecular systems .
however realistic simulations on timescales beyond microseconds are still infeasible even on the most powerful general purpose computers , which renders the md - based analysis of many biological equilibrium processes , that are often rare compared to the characteristic time scale of the system and hence require prohibitively long simulations , impossible .
the hallmark of these rare events is that the average waiting time between the events is orders of magnitude longer than the timescale of the switching event itself .
thus rare event simulation and estimation are among the most challenging topics in molecular dynamics .
the molecular dynamics literature on rare event simulations is rich . since direct numerical equilibrium simulation
is infeasible , all available techniques try to sample from the rare event statistics by biasing the system in one or the other way .
roughly speaking , we can distinguish between two major classes of sampling techniques : class @xmath0 consists of splitting methods that decompose state space , but are still essentially based on an equilibrium distribution , whereas methods from class @xmath1 proceed by driving the system under consideration into a nonequilibrium regime that changes the rare events statistics . for a general overview of monte - carlo methods for rare events in other application fields ,
we refer to the textbook @xcite .
the list of methods in class @xmath0 range from reaction - coordinate based techniques via path - space oriented techniques to approaches based on interface sampling or generalized dynamics .
reaction - coordinate based techniques consider the marginal of the equilibrium distribution in some low - dimensional collective variables like in direct free energy calculations @xcite ; they suffer from the fact that appropriate reaction coordinates are often not available .
path - space oriented techniques approximate the most important reaction paths that govern the rare event statistics either by sampling distribution of reactive paths like in transition path sampling ( tps ) @xcite or by optimizing an appropriate path functional like in the string method @xcite ; they become problematic if the path space distribution is multi - modal or generally too complex ( e.g. , involving bifurcations ) .
interface sampling techniques like milestoning @xcite or forward flux sampling ( ffs ) @xcite place a set of suitably chosen interfaces in state space between the initial and final state and use them to follow the transition of the system in an iterative manner using equilibrium trajectories that connect neighbouring interfaces . the idea of _ generalized dynamics _ such as hyperdynamics @xcite , metadynamics @xcite , conformational flooding @xcite , or the adaptive biasing force ( abf ) method @xcite is to bias the system on - the - fly ( e.g. , by filling in certain energy wells in which the system got trapped during a simulation ) so as to enhance rare transitions between metastable states .
although seemingly different , generalized dynamics belong to class @xmath0 , in that they only alter the underlying equilibrium distribution along a predefined set of low - dimensional collective variables .
although these methods have proven to be very efficient , they require that the interesting processes can be described by a few collective coordinates that have to be known in advance .
class @xmath1 consists of methods based on the jarzynski and crooks formulae @xcite that relate the equilibrium helmholtz free energy to the nonequilibrium work exerted under external forcing .
instances of nonequilibrium simulations that mimic experiments on controlling and manipulating single molecules ( see , e.g. , @xcite ) are single - molecule pulling @xcite , steered molecular dynamics @xcite or bridge sampling @xcite , to mention just a few .
the corresponding path functionals have the form of cumulant - generating functions for the exerted work @xcite which poses immense challenges to monte - carlo simulations and limits the usability of the formulae in practice . roughly speaking ,
the usability is limited by the fact that the likelihood ratio between equilibrium and nonequilibrium trajectories is highly degenerate , for the overwhelming majority of nonequilibrium forcings generate trajectories that have almost zero weight with respect to the equilibrium distribution that is relevant for the rare event ; cf .
also the discussion in @xcite .
nevertheless the underlying idea is appealing and a cleverly designed external force may speed up the sampling of the rare events by biasing the equilibrium distribution of the system towards a distribution under which the rare events is no longer rare , while giving numerical estimators that are useful in terms of variance and convergence properties .
the method presented in this article belongs to the latter class , but shares somes ideas with ideas from class @xmath0 .
it takes up the idea that external forcings can speed up the rare event but avoids sampling issues related to nonequilibrium processes . instead it uses optimal nonequilibrium forcing in connection with splitting methods such as ffs or milestoning , in the sense that the new method uses interfaces to follow the transition of an _ optimally driven _
system where the external forcing that drives the system from one interface to the next results in a considerable speed - up compared to ffs or milestoning .
specifically , the new method replaces the path sampling problem using an exponential change of measure that can be explicitly computed by minimizing a cross - entropy - like functional , which then yields the optimal forcing .
although the minimization involves solving an optimal control problem , the numerical effort can be drastically reduced when the minimization is done in a clever way ; one reason is that the path functional becomes linear after the change of measure whereas it was exponential in the original cumulant - generating function .
transformations based on exponential change of measures have a rich tradition in the ( risk - sensitive ) optimal control literature @xcite and the theory of large deviations @xcite , and are regularly rediscovered mostly aiming at turning certain optimal control problems into linearly solvable sampling problems @xcite ; cf .
also @xcite . here
we pursue the reversed strategy and turn a difficult rare event estimation problem into an optimal control problem that can be solved by minimizing a suitable functional .
thus the basic outline of the new method is : iteratively determine the optimal nonequilibrium forcing by an optimization procedure based on milestoning ideas that avoid path - space sampling and compute the equilibrium rare event statistics from the optimal nonequilibrium forcing . besides introducing the new method
the purpose of this article is to explain the basic ideas of how to use optimal control for the estimation and simulation of rare events .
therefore we present only the simplest possible scenario ( a particle following an overdamped langevin dynamics in a conservative force field ) , without paying too much attention to complete generality or mathematical rigour .
the first issue in section [ sec : free ] then is to introduce the variational characterization of ( generalized ) free enregy and the exponential change of measures that are the basis of our optimal control approach
. the precise formulation of the optimal control problem , a stochastic control problem with quadratic control costs and an indefinite time horizon , is given in section [ sec : ocp ] . in section
[ sec : milestoning ] we describe the numerical method for computing the optimal control , based on an inexact gradient descent in connection with a milestoning algorithm , and apply it to the controlled first passage between metastable sets .
we briefly summarize the results in section [ sec : conclusion ] and sketch possible generalization that have been omitted for the sake of brevity .
we consider a particle with position @xmath2 at time @xmath3 which moves in an energy landscape @xmath4 according to the equation @xmath5 here @xmath6 denotes standard @xmath7-dimensional brownian motion , and @xmath8 is the temperature of the system . under mild conditions on the energy landscape function
@xmath9 we have ergodicity , and the law of @xmath10 converges to a unique equilibrium distribution with density @xmath11 we assume throughout that the temperature is small , relative to the largest energy barriers , i.e. , @xmath12 . as a consequence ,
the relaxation of the dynamics towards equilibrium is dominated by the rare transitions over the largest energy barriers .
let @xmath13 be a random variable that depends on the sample paths @xmath14 up to a stopping time @xmath15 .
we will call @xmath13 _ work _ in the following . given some continuous function @xmath16
, we suppose that it can be expressed as ) and path functionals like ( [ work ] ) and can be easily can be easily generalized to , e.g. , non - gradient systems with multiplicative and/or degenerate noise or observables @xmath17 that are explicitly time - dependent . ] @xmath18 let us further denote by @xmath19 the probability measure on the space of continuous trajectories that is generated by the brownian motion in ( [ sde ] ) , and let @xmath20={{\mathbf e}}[\cdot|x_{0}=x]$ ] be the expectation with respect to @xmath19 , i.e. , the average over all realizations of @xmath10 starting at @xmath21 .
we call the quantity @xmath22\,.\ ] ] the _ ( conditional ) free energy _ of @xmath13 with respect to @xmath19
. clearly , the functions and the expectation on the right hand side of ( [ cgf ] ) do not commute , and it follows by jensen s inequality that @xmath23 $ ] , in accordance with the second law of thermodynamics . but @xmath24 encodes information about the cumulants of the work @xmath13 ( assuming they exist ) , namely , @xmath25 + \frac{1}{2{\epsilon}}{{\mathbf e}}^{x}\big[(w-{{\mathbf e}}^{x}[w])^{2}\big ] + \ldots\,.\ ] ] the similarity between ( [ cgf ] ) and jarzynski s formula @xcite is no coincidence .
if @xmath26 is a deterministic stopping time and @xmath13 is the nonequilibrium work done on a system during a transition between two equilibrium states @xmath27 and @xmath28 , then @xmath29 equals the equilibrium free energy difference between @xmath27 and @xmath28 .
the phrases `` work '' for the quantity @xmath13 defined in ( [ work ] ) and `` free energy '' for @xmath24 as of ( [ cgf ] ) are just used to relate to jarzynski s formula .
the framework is much more general as the following example will show .
[ [ guiding - example . ] ] guiding example .
+ + + + + + + + + + + + + + + + + one example , of which we will consider variants below , is the first hitting time of a subset of state space . to this end
let @xmath30 a set and define @xmath31 to be the first time at which @xmath10 hits @xmath32 .
choosing the constant function @xmath33 in ( [ work ] ) , the free energy @xmath34\,.\ ] ] considered as a function of @xmath35 is the scaled cumulant - generating function of @xmath15 when @xmath10 is started at @xmath21 .
in particular , we can compute the mean first hitting time by @xmath36\,.\ ] ] the strict convexity of the exponential function implies that equality @xmath37 $ ] is only attained if @xmath13 is @xmath19-almost surely constant ; one such case is the adiabatic limit @xmath38 we will restore ( [ cgf ] ) to an expression that becomes linear in @xmath13 after a suitable change of measure . to this end
let @xmath39 denote a probability measure on the space of continuous trajectories that is absolutely continuous with respect to @xmath19 ( i.e. , @xmath40 exists ) .
we define the relative entropy of @xmath39 with respect to @xmath19 as @xmath41 ( this is also called the _ kullback - leibler divergence_. ) we declare that @xmath42 if @xmath39 is not absolutely continuous with respect to @xmath19 . then , by jensen s inequality , @xmath43\\ & = -{\epsilon}\log { { \mathbf e}}^{x}_{q}[\exp(-w/{\epsilon}- \log\varphi)]\\ & \le { { \mathbf e}}^{x}_{q}[w ] + { \epsilon}i(q\|p)\,,}\ ] ] where we have used the notation @xmath44 $ ] to denote the expectation with respect to @xmath39 .
the last inequality that appears in the literature in various forms as _ second - law - like identity _ or _ generalized jarzysnki inequality _
@xcite ) suggests that the free energy and the relative entropy are related by a legendre - type transformation , viz .
, @xmath45 + { \epsilon}i(q\|p)\right\},\ ] ] and a result in @xcite implies that the infimum exists and is attained when @xmath39 runs over all path measures that are absolutely continuous with respect to @xmath19 . by the strict convexity of the exponential function ,
the latter implies that @xmath46 is @xmath39-almost surely constant .
the idea of the approach sketched below then is to represent @xmath39 in terms of suitable ( parametric ) control variables and minimize the right hand side of ( [ equivcgf ] ) over all admissible controls .
the aim of this section is to derive necessary and sufficient conditions for the optimal change of measure that turns ( [ equivcgf ] ) into an equality . to this end
we follow ideas by fleming and co - workers @xcite and consider the exponential cost functional : @xmath47.\ ] ] for a stopping time @xmath15 that is the first hitting time of a set @xmath48 , the feynman - kac formula @xcite implies that @xmath49 solves the elliptic boundary value problem @xmath50 where @xmath51 is the infinitesimal generator of @xmath10 , defined on a suitable subspace of @xmath52 .
we want to transform the boundary value problem ( [ fk ] ) into an equation for the unknown control variable in ( [ equivcgf ] ) .
for this we proceed in two steps .
[ [ step-1 ] ] step 1 : + + + + + + + we can safely assume that @xmath15 is almost surely finite
. as a consequence , the function @xmath49 in ( [ mgf ] ) admits a formal representation of the form @xmath53 we seek an equation for the free - energy @xmath24 . by chain rule
, it follows that @xmath54 which entails that ( [ fk ] ) is equivalent to @xmath55 the last equation is known as the hamilton - jacobi - bellmann ( hjb ) equation of optimal control @xcite ; its solution is called _ value function _ or _ optimal cost - to - go_. [ [ step-2 ] ] step 2 : + + + + + + + to reveal the stochastic optimal control problem that corresponds to the hjb equation ( [ hjb ] ) , we first note that @xmath56 from which we recognize that ( [ hjb ] ) is equivalent to @xmath57 with the shorthands @xmath58 and @xmath59 equation ( [ hjbmin ] ) is the hamilton - jacobi - bellman equation of the following optimal control problem that should be compared to the right hand side of ( [ equivcgf ] ) : minimize @xmath60\ ] ] over an admissible set @xmath61 of control laws @xmath62 with values in @xmath63 and subject to the tilted dynamics @xmath64 that is , the expectation in ( [ indefinite1 ] ) has to be taken wrt the path measure @xmath39 generated by the dynamics given by ( [ indefinite2 ] ) . the dynamics that generates the new path measure @xmath39 is again of gradient form if @xmath65 is the optimal markovian feedback control , i.e. when @xmath66 . as a consequence ,
the optimally controlled process satisfies detailed balance @xcite . indeed ,
since ( [ indefinite1 ] ) is quadratic and ( [ indefinite2 ] ) is affine in the control , the minimizer @xmath67 in ( [ hjbmin ] ) is unique ( provided that @xmath24 is sufficiently smooth ) .
the optimal feedback law is then given by @xmath68 and gives rise to the tilted dynamics @xmath69 with the tilted potential @xmath70 [ [ guiding - example - contd . ] ] guiding example , contd .
+ + + + + + + + + + + + + + + + + + + + + + + + in some cases it is helpful to pursue a reverse strategy and transform the nonlinear hjb equations of an optimal control problem into a linear equation that may be easier to solve ( cf .
@xcite ) . consider a brownian particle under a microscope with a moveable object holder .
let @xmath71 denote the microscope s focal disc , @xmath72 the particle position at time @xmath3 , relative to the position of the object holder , and @xmath73 the motor force .
the control task is to move the object holder such that the particle stays in the focus as long as possible .
hence the control objective is the maximization of the mean first exit time from @xmath74 which amounts to minimizing the cost functional @xmath75,\ ] ] subject to @xmath76 let @xmath77,\ ] ] be the value function ( free energy ) of the problem and @xmath78 \,.\ ] ] then the linear boundary value problem for @xmath79 is a helmholtz equation with dirichlet boundary conditions , @xmath80 which can be solved by standard means .
at first sight it seems that we have not gained much , for we have transformed the original path sampling problem into a complicated nonlinear optimal control problem .
however the optimal control formulation opens up other options for the numerical treatment of the rare event sampling in terms of a minimization problem .
another advantage is that it is relatively easy to construct unbiased estimators of the control functional , avoiding both bias and variance issues when estimating exponential observables such as ( [ cgf ] ) .
[ [ discretization ] ] discretization + + + + + + + + + + + + + + together with the information that the optimal markov control is of feedback form our minimization problem ( [ indefinite1])([indefinite2 ] ) takes the form @xmath81\label{final}\end{aligned}\ ] ] with @xmath39 denoting the path measure generated by the dynamics given by ( [ indefinite2 ] ) .
we discretize this optimization problem by choosing a finite dimensional ansatz space for the space of admissable feedback functions @xmath82 : we choose sufficiently smooth and integrable vector fields @xmath83 , @xmath84 , so that @xmath85 or , respectively , we choose scalar ansatz function @xmath86 , @xmath84 , so that @xmath87 the minimization problem then amounts to minimizing the cost functional @xmath88\ ] ] over the unknown coefficients @xmath89 where @xmath90 , the path measure of the controlled diffusion ( [ indefinite2 ] ) also depends on the coefficients ; for the moment we remain with the imprecise statement that the measure @xmath39 has a density @xmath91 with respect to a ( fictitious ) uniform measure on the space of all continuous paths in @xmath63 , which is a function of the unknown coefficients . is the probability to find paths @xmath92 in a small tube around a smooth curve @xmath93\to{{\mathbb r}}^{n}$ ] , i.e. , @xmath94 . by the girsanov theorem , @xmath95 has a density @xmath96 with respect to the gaussian measure induced by the brownian motion @xmath97 , where @xmath98 is the onsager - machlup functional @xcite . ]
[ [ gradient - descent ] ] gradient descent + + + + + + + + + + + + + + + + we minimize the cost functional @xmath99 by a doing a gradient descent in the coefficient vector @xmath89 . specifically , we iterate the map @xmath100 where @xmath101 is the iteration index and @xmath102 is a bounded sequence of stepsizes for the gradient search .
for instance , we can do a line search in the descent direction and determine @xmath103 so that it satisfies the wolfe condition @xcite .
details of the iteration that is based on an euler - maruyama discretization of the path measure @xmath39 will be given below in the appendix .
the overall algorithm thus has the following steps : * choose scalar - valued ansatz functions @xmath104 with support in the interesting region of state space and related vector fields @xmath105 .
* choose initial coefficients @xmath106 such that the free energy or value function @xmath107 fills up the main wells in the energy landscape @xmath9 .
* iterate the following steps in @xmath101 , starting with @xmath108 , until a prescribed termination criterium is satisfied : 1 . sample the path measure @xmath109 and evaluate @xmath110 ( see formula ( [ gradi ] ) in the appendix ) .
2 . perform a gradient descent @xmath111 .
the gradient search algorithm can be regarded as a variant of the cross - entropy method that is a relatively new monte - carlo technique for the sampling of rare events which goes back to rubinstein and others @xcite .
it is based on the idea that an optimal change of measure can be found by minimizing the kullback - leibler divergence ( [ kldiv ] ) over a family of probability measures @xmath39 in terms of the tilting parameter @xmath82 .
compared to equilibrium rare event simulation algorithms used in molecular dynamics using the optimal change of measure has the advantage that the likelihood ratio @xmath112 stays of order one , while rare events under the original dynamics ( here : diffusion in an energy landscape @xmath9 ) are no longer rare under the forced dynamics ( [ indefinite2 ] ) . as a consequence , sampling the path measure @xmath39 is significantly more efficient than sampling the original path measure @xmath19 since the trajectories to be sampled from @xmath39 are much shorter on average ( i.e. , the expected hitting time is considerably shorter ) . [ [ milestoning - algorithm ] ] milestoning algorithm + + + + + + + + + + + + + + + + + + + + + for problems with a large state space or for strongly metastable systems , the above algorithm may still be inefficient since sampling the path measure @xmath39 may involve many rather long trajectories . in this case
the computation can be broken down to transitions between neighbouring interfaces as in milestoning @xcite or in ffs @xcite .
we explain the basic steps of this procedure : let @xmath113\ ] ] denote the semi - discretized value function of the problem , with the shorthand @xmath114 suppose that @xmath115 is the set of interest and @xmath116 is the first hitting time of @xmath117 ; we now choose nested sets or _ milestones _ @xmath118 ( cf .
figure [ f : illu ] ) .
we first compute @xmath119 in @xmath120 by finding the optimal control policy @xmath82 in @xmath120 .
that is , our ansatz functions in the above gradient descent algorithm only have to be non - vanishing in @xmath120 . in particular this gives @xmath119 on @xmath121 , the outer boundary of @xmath120 .
we can repeat the same algorithm in the set @xmath122 ; then letting @xmath123 and letting @xmath124 denote the first entry time into @xmath125 , we have @xmath126,\ ] ] where @xmath127 for which @xmath119 has been computed in the previous step . by iterating the algorithm
we eventually obtain @xmath119 on all set boundaries @xmath128 , @xmath129 .
thus , the milestoning iteration can be implemented as an outer loop which contains the above gradient descent algorithm in every of its iterations .
the milestoning variant of the gradient descent algorithm only requires the computation of an ensemble of _ short _ trajectories of the controlled system ( [ indefinite2 ] ) . here
`` short '' means that they are orders of magnitude shorter than those in typical path - space sampling algorithms like tps , and equilibrium milestoning or ffs .
we consider the uncontrolled dynamics ( [ sde ] ) with the one - dimensional potential shown in figure [ fig : pot ] .
suppose we are interested in computing the mean first passage time to the set @xmath130 $ ] in terms of the free energy ( [ cgf ] ) .
let @xmath131 be the first hitting time of @xmath32 , consider the constant function @xmath33 , and the scaled moment generating function @xmath132,\ ] ] considered as a function of @xmath35 .
the quantity of interest is the mean first passage time of the uncontrolled dynamics , @xmath133,\ ] ] for @xmath134 . in order to obtain a reference solution with high accuracy we first compute @xmath135 by discretizing the elliptic boundary value problem ( [ fk ] ) based on a standard finite element discretization on a fine grid
this is possible because the state space dimension in this guiding example is small but will not be possible in realistically high dimensions .
the resulting reference solution for @xmath136 $ ] is shown in the left panel of figure [ fig : fmptunc ] below , along with the associated free energy @xmath137 in the right panel .
an approximation of the free energy was then computed by the greedy milestoning / gradient descent algorithm described above that minimizes the cost functional ( [ semicost ] ) in the coefficients @xmath89 . as scalar ansatz functions
@xmath104 we chose @xmath138 gaussians with width @xmath139 whose centers where uniformly spaced in the complement of @xmath32 .
once the minimization had been converged , the value function ( free energy ) and the resulting optimal control law were given by @xmath140 with @xmath141 .
the result agree with the reference solution shown in fig .
[ fig : fmptunc ] ( deviations are of the order of the accurarcy threshold used in the gradient descent algorithm ) . figure [ f : u ] shows the resulting optimally tilted potential @xmath142 , together with first few iteration steps of the gradient search .
the mean first passage time of the tilted system @xmath143 i.e. , with @xmath9 in ( [ sde ] ) replaced by the new potential @xmath144 , is shown in figure [ f : fmptc ] . as has been outlined above the algorithm only requires the computation of rather short trajectories since for all iterative potentials the mean first passage time is orders of magnitude smaller than for the original dynamics ; the mean first passage time of the optimally tilted potential , e.g. , is around 100 times smaller than originally .
we have developed a simulation scheme for rare events that is based on an optimal change of measure that boils down to a logarithmic transformation of the path functional under consideration .
the measure transformation turns the original exponential path functional into the functional of an optimal control problem that is linear in the observable and quadratic in the control variables .
although analytic solutions to the optimal control problem are available only in simple situations and computing the optimal change of measure may require to solve a possibly high - dimensional optimal control problem numerically , there is a considerable speed - up coming from ( a ) the fact that the functional is linear - quadratic and allows for the design of robust unbiased monte - carlo estimators and ( b ) the fact that events that were rare originally are no longer rare under the new probability measure .
the gain in the numerical complexity requires that the optimal control problem can be solved efficiently , and , with the equivalence between path sampling and optimal control in hand , we have sketched a numerical algorithm for computing the optimal control that is based on an easy - to - implement inexact gradient descent that can be solved rather efficiently using milestoning .
the algorithm was tested , computing the optimal feedback for the controlled passage between metastable sets in a double - well potential .
even though the numerical example that we presented is tiny on the scale of typical molecular dynamics applications , we emphasize that the minimization algorithm is independent of the dimension of the system and hence admits an easy generalization to more complicated systems ; we refer to the rich literature on machine learning and queuing networks where various strategies for treating high dimensional systems have been developed ( e.g. , see @xcite ) .
finally we note that all ideas presented in this article can be readily extended to more complicated dynamics ( e.g. , degenerate diffusions with dissipation ) and time - dependent path functionals ( e.g. , to simulate single - molecule experiments ) ; it is even possible to consider situations where the exponential path functional involves additional control variables , in which case a logarithmic transformation leads to a game rather than an optimal control problem ( cf .
further open issues are the deterministic limit of the stochastic control problem , the convergence analysis of the gradient descent and the rigorous analysis of fluctuations in systems under feedback control ( cf . @xcite ) .
in order to compute the gradient of ( [ semicost ] ) with respect to to the unknown coefficients @xmath89 , it is convenient to discretize the path measure @xmath90 . to this end , let @xmath145 be a set of time nodes with @xmath146 where we assume for the moment that @xmath147 is deterministic .
euler s method applied to @xmath148 gives @xmath149 where the @xmath150 are i.i.d .
random variables that are normally distributed with mean zero and unit covariance . since the @xmath151 are gaussian , the density of the distribution @xmath152 of discrete paths @xmath153 conditional on @xmath154 is readily shown to be @xmath155 with the discrete action @xmath156 and the normalization constant @xmath157 computing the gradient of the discretized functional @xmath158\ ] ] with @xmath159 is now straightforward . assuming that @xmath160 is independent of the control , we have @xmath161,\ ] ] where both @xmath162 and @xmath163 are evaluated at @xmath164 .
specifically , @xmath165.\end{aligned}\ ] ] together with the projection property of the conditional expectation this gives @xmath166 + \mathbf{c}^{\tilde{x}_{0}}_{q_{h}}\left[\tilde{g}_{h},\,\frac{\partial s_{h}}{\partial a_{j } } \right ] , \ ] ] where @xmath167 denotes the covariance operator @xmath168={{\mathbf e}}_{q_{h}}[uv]-{{\mathbf e}}_{q_{h}}[u]{{\mathbf e}}_{q_{h}}[v].\ ] ] [ [ inexact - gradient ] ] inexact gradient + + + + + + + + + + + + + + + + we are interested in the situation when @xmath15 in ( [ indefinite1 ] ) is a random stopping time rather than a fixed time ; otherwise the optimal control policy would be a function of time , i.e. , @xmath169 . but in case that @xmath15 is a first entry time of a set @xmath48 , this stopping time @xmath170 will be a function of the control .
hence the derivative of the cost functional with respect to the unknown control coefficients @xmath171 would involve additional derivatives of @xmath15 or its time - discrete counterpart @xmath172 ; for example , for the discretized running cost this would result in an expression like @xmath173 in principle the dependence of the stopping time on the control variable can be made explicit in terms of the solution to an elliptic boundary value problem for @xmath15 , yet it is unclear how terms such as @xmath174 can be handled numerically efficiently . in many cases
the gradient descent will also converge even though the gradient @xmath175 is not exact , and it turns out that the boundary cost in the last equations is typically small compared to the accumulated cost . ignoring the contribution from the boundary terms in the derivatives hence gives a gradient descent method with inaccurate gradient . in our numerical example where @xmath33 is constant
, the inexact gradient reads @xmath176\\ & & -\frac{h^{3/2}}{{\epsilon}\sqrt{2 } } \mathbf{c}^{\tilde{x}_{0}}_{q_{h}}\left[\sum_{k=0}^{n_\tau-1}(\sigma+\frac{1}{2}|c(x_k)|^{2}),\,\sum_{k=0}^{n_\tau-1 } \eta_{k+1 } b_{j}(x_k)\right]\\ & = & h { { \mathbf e}}^{\tilde{x}_{0}}_{q_{h}}\left[\sum_{k=0}^{n_\tau-1}c(x_k ) b_j(x_k)\right]\\ & & -\frac{h^{3/2}}{{\epsilon}\sqrt{2 } } { { \mathbf e}}^{\tilde{x}_{0}}_{q_{h}}\left[\left(\sum_{k=0}^{n_\tau-1}(\sigma+\frac{1}{2}|c(x_k)|^{2})\right)\sum_{k=0}^{n_\tau-1 } \eta_{k+1 } b_{j}(x_k)\right]\,,\end{aligned}\ ] ] where @xmath177 is the discrete analog of the first hitting time ( here @xmath178 is the nearest integer larger than @xmath179 ) , and we used the fact that @xmath180 and @xmath181 are independent .
10 d. chandler . finding transition pathways : throwing ropes over rough mountain passes , in the dark . in g.
ciccotti b. j. berne and d. f. coker , editors , _ computer simulation of rare events and dynamics of classical and quantum condensed - phase systems classical and quantum dynamics in condensed phase simulations _ , pages 5166 .
world scientific , 1998 .
k. dvijotham and e. todorov .
linearly - solvable optimal control . in f.l .
lewis and liu d. , editors , _ to appear in : reinforcement learning and approximate dynamic programming for feedback control _ , chapter 6 .
wiley & sons , 2012 .
k. rawlik , m. toussaint , and s. vijayakumar . on stochastic optimal control and reinforcement learning by approximate inference . in _ proc .
robotics : science and systems conference ( r : ss 12 ) _ , 2012 ( in press ) .
|
rare event simulation and estimation for systems in equilibrium are among the most challenging topics in molecular dynamics . as was shown by jarzynski and others , nonequilibrium forcing can theoretically be used to obtain equilibrium rare event statistics .
the advantage seems to be that the external force can speed up the sampling of the rare events by biasing the equilibrium distribution towards a distribution under which the rare events is no longer rare . yet algorithmic methods based on jarzynski s and related results often fail to be efficient because they are based on sampling in path space .
we present a new method that replaces the path sampling problem by minimization of a cross - entropy - like functional which boils down to finding the optimal nonequilibrium forcing .
we show how to solve the related optimization problem in an efficient way by using an iterative strategy based on milestoning .
|
glendenning@xcite had realized the essentially different character of a first - order transition in simple system possessing a single conserved quantity and complex one having more than one conserved charge .
one of the most remarkable features of a simple system is the constancy of the pressure during the transition from one homogeneous phase to the other .
such phase transition is the typical description of first - order one in textbook .
the properties of the transition are quite different in complex .
the pressure varies continuously with the proportion of two pure phases in equilibrium .
some quantities are obviously nonlinear functions in the proportion in the mixed phase . as well - known
, the occurrence of first - order transition will eventually accompany the release of energy .
the energy is called latent heat related to the release of entropy in the example of gas - liquid transition .
the phase transition at zero temperature showed the liberation of binding energy .
we still use the concept of `` latent heat '' for the release of energy associated with such the first - order phase transition .
the calculations of `` latent heat '' from hadronic matter to strange quark matter were firstly made by haensel and zdunik@xcite .
they considered hadrons be absorbed into strange quark matter in accreting strange stars and had expected that the heat per absorbed neutron is in the range 10@xmath030mev .
generally the energy release for constant - pressure phase transition at zero temperature takes place nearby a transition point where an over - compressed state , metastable , is probably the first to form and then go to a stable state@xcite .
however , the phase transition in a multicomponent system takes place in a varying pressure region .
an appearance of over - compressed state at every pressure is possible .
so we need to face up to a series of metastable states if the assumed system has a durative transfer from one pure phase to the other . relatively speaking ,
the treatment of such energy release is a more difficult task than the constant - pressure case .
we will focus on solving the problem in present paper .
many phase transitions in nuclear matter are the expected ones having multicomponent mixture in chemistry , the nuclear gas - liquid transition relevant to accelerator experiments , where the neutron and proton number are conserved , pion , koan condensates and deconfinement of quarks in high density , where the conserved charges are baryon and electric .
typically , we will treat the problem on the converted energy taking the confined - deconfined phase transition as an example .
ideally the two phase should be described selfconsistently in the same theory , but we can not solve qcd through confined - deconfined phase transition based on our present knowledge . hence we will use separate models for each phase .
we describe the equation of state of hadronic matter in the framework of the relativistic mean - field theory and depict the deconfined quark matter in the mit bag model .
we will study the baryon number density dependence of energy per baryon for different phases , hadronic , quark and their mixed phase .
the plan of the paper is as follows . in sec .
ii we state our views on `` latent heat '' and formulate the problem . in sec.iii
we solve numerically the equations in the confined - deconfined phase transition for the given equations of state .
our discussion and conclusion will be made in sec.iv .
for superdense nuclear matter where a first - order phase transition would take place , the pressure @xmath1 versus baryon number density could generally be described in figure 1 if we follow the same philosophy and method used by glendenning .
correspondingly , the number density dependence of the energy per baryon could also be revealed in the figure(low panel ) .
the body experiences from low density phase , mixed phase to high density phase with the number density increase .
we find the suppression of the increase binding energy in the mixed phase as compared with the low density phase when the body gradually compressed .
the starting point @xmath2 in the mixed phase is so conspicuous that the body may go beyond @xmath2 in low density phase to an over - compressed state and then have a transition to a stable state in mixed phase .
analogous situation would appear at every point in mixed phase region .
the process depicted by @xmath3 and @xmath4 should repeatedly occur through the duration when the body were transferred from @xmath2 to @xmath5 .
we can consider that the transfer is nearly along the line @xmath6 , denoting the equation of state , if the over - compressed states have very small deviations from the corresponding state in mixed phase region .
as well - known , the total energy and baryon number densities for the mixed phase , given by glendenning@xcite , read @xmath7 @xmath8with the volume fraction @xmath9 where @xmath10 and @xmath11 represent the volumes occupied by low density and high density phase respectively , @xmath12 and @xmath13 denote the corresponding energy densities , @xmath14 and @xmath15 , the baryon number densities .
we further introduce the baryon number fraction @xmath16 to rewrite the equations ( 1 ) and ( 2 ) for convenience of the following treatments , where @xmath17 and @xmath18 present high density phase and total baryon numbers .
substitute @xmath19 into eq(3 ) , we get @xmath20 @xmath21 therefore eqs(1 ) and ( 2 ) become @xmath22 @xmath23 the energy per baryon(binding energy ) is expressed as @xmath24 at zero temperature the energy @xmath25 only depends on the density . the change in @xmath25 arises from the density increases by means of compression and deconfinement.the energy @xmath25 can therefore be assumed to be of the function of form @xmath26 .
we find the derivative of @xmath25 with respect to @xmath27 , which reads @xmath28 furthermore @xmath29 where @xmath30 eq(10 ) shows the fact that the increase of binding energy in mixed phase amounts to the composition of two processes , the simple compressional change without phase transition and the compressional one due to the occurrence of phase transition .
we define @xmath31 and hence substitute it into eq(9 ) , we immediately have @xmath32\delta\rho=\left [ ( e+\left({{\partial}e\over{\partial}\rho}\right ) _
\eta\delta\rho ) - ( e+{{\delta}e\over{\delta } \rho}\delta\rho)\right]\ ] ] in equation(11 ) , @xmath33 has been supposed invariable factor with increase density .
@xmath34 indicates the the binding energy increase by a simple compression regardless of phase transition since no particle is converted to the other phase from one phase .
the second term of the right - hand side in eq(9 ) or eq(12 ) is obviously associated with phase transition because it arises from the change of the baryon number fraction .
we further change eq(13 ) to reveal the physical meanings of quantity @xmath35 . from an initial state(see @xmath36 in figure 1 ) to the adjacent stable state , the system suffering phase transition(see @xmath37 ) , or the adjacent metastable one , the system experiencing simple compression ( see @xmath4 ) , through a equivalent density change , the binding energies for @xmath37 and @xmath4 can be calculated as @xmath38 and @xmath39 .
the @xmath35 is immediately the energy difference between the metastable state and the stable one , i.e.@xmath40 , where @xmath41 represent total energies of the system for the two states .
evidently @xmath35 indicates the binding energy release , just the same mechanism as the physical situation in @xcite , when the body change from the metastable state @xmath4 into the stable state@xmath37 .
a very small change of density considered here is the only difference from ref@xcite and hence the energy release , @xmath35 , tends to zero .
but the accumulation of a series of infinitesimal released energies is able to be nonvanishing because the system certainly experiences the finite change during real phase transition , which can be expressed by an definite integral .
therefor , when we consider the assumed system with @xmath18 baryon number from pure low density phase(@xmath2 in figure 1 ) to pure high density one(@xmath5 in the figure ) , the energy release can be expressed as @xmath42\delta\rho,\ ] ] where @xmath43 is mean energy release per baryon .
we can also calculate the energy release per converted baryon for different densities .
the total energy release for given density @xmath27 in mixed phase region equals @xmath44 .
meanwhile , the converted baryon numbers should be @xmath45 at infinitesimal compression @xmath46 .
thus the energy release per converted baryon is calculated by @xmath47 the meanings of the formula will be displayed thoroughly when eq(12 ) is substituted into eq(15 ) , @xmath48 clearly the energy release per converted baryon is different for various baryon number density during phase transition in varying pressure .
we turn back to the understandings of eq(10 ) .
we easily find that the simple compression causes the binding energy increase but the phase transition leads to the reduction .
the reduced energy is just about the energy release .
we will numerically solve the integral(12 ) together with equations(1)and(2 ) for the transition from hadronic matter to quark matter .
the low density phase represents hadronic matter , called hp for short , and the high density one indicates quark matter , qp for short .
we use the equations of state in relativistic mean - field theory(rmf ) description for hadronic matter and the equations of state in mit bag model for quark matter . @xmath49 where the spinor for the baryon species b is denoted by @xmath50 , @xmath51
represent meson fields .
the energy density and pressure can be obtained in the familiar manner as @xmath52 @xmath53 where the coupling constants @xmath54,@xmath55,@xmath56,@xmath57 and @xmath58 can be determined by the nuclear saturation density @xmath59 , the binding energy at saturation @xmath60 , the symmetry energy@xmath61 , the compression modulus@xmath62 and the effective nucleon mass @xmath63 . in mit bag model ,
the energy density and pressure of quark matter at zero temperature are given by @xmath64.\ ] ] @xmath65.\ ] ] we calculate the mixed phase of hadronic and quark matter with gibbs construction which first used by glendenning@xcite .
for simplicity , we neglect coulom and surface effects in mp . the gibbs condition for mechanical and chemical equilibrium at zero temperature in mp
is written as @xmath66 we will obtain eqs(1 ) and ( 2 ) by applying the condition ( 20 ) together with eqs(16)@xmath0 ( 19 ) and imposing the condition of global charge neutrality in mp where @xmath68 and @xmath69 respectively denote negative charged density of quark matter and positive one of hadronic matter . when @xmath70=0 and 1 , the global charge neutrality spontaneously come back local charge neutralities for pure hadronic matter and pure quark matter .
the equilibrium chemical potentials can be determined in mp from eq(20 ) while @xmath70 can be solved from eq(21 ) .
we choose the representative parameters for soft , moderate and stiff hadronic matter equations of state listed in table 1 .
the bag constant is taken as @xmath71=170mev , 180mev and 190mev .
we construct a series of the equations of state with mp denoted by rmfn+@xmath71(n=1,2,3 ) under those considerations . as an example , rmf1 +
180 is depicted in figure 2 .
|
we consider the energy release associated with first - order transition by gibbs construction and present such energy release as an accumulation of a series of tiny binding energy differences between over - compressed states and stable ones .
universal formulae for the energy release from one homogeneous phase to the other is given .
we find the energy release per converted particle varies with number density . as an example , the deconfinement phase transition at supranuclear densities is discussed in detail .
the mean energy release per converted baryon is of order 0.1mev in rmf theory and mit bag descriptions for hadronic matter and strange quark matter for a wider parameter region .
pacs numbers : 97.60.jd , 05.70.fh , 12.38.mh , 64.60.-i
|
the type ii linear supernova ( sn ii - l ) sn 1980k in ngc 6946 reached a peak brightness of v = 11.4 in november 1980 ( see barbon , ciatti , & rosino 1982 ; @xcite and references therein ) . despite a steadily declining flux through 1982 ( @xcite ) , faint h@xmath1 emission from sn 1980k was detected in 1987 through narrow passband imaging ( @xcite , hereafter fb90 ) .
low - disperson optical spectra obtained in 1988 and 1989 showed broad , 6000 km s@xmath0 h@xmath1 and [ o i ] 6300,6364 emission , along with weaker line emission from [ ca ii ] 7291,7324 and/or [ o ii ] 7319,7330 , [ o iii ] 4959,5007 , and [ fe ii ] 7155 ( fb90 ; @xcite ; leibundgut et al .
monitoring of its optical flux from 1988 through 1992 indicated a nearly constant luminosity ( leibundgut , kirshner , & porter 1993 ; fesen , & matonick 1994 , hereafter fm94 ) .
following the optical recovery of sn 1980k , a handful of other sne ii - l have been optically detected 7 25 yr after maximum light .
these include sn 1986e in ngc 4302 ( @xcite ) , sn 1979c in m100 ( @xcite ) , and sn 1970
g in m101 ( @xcite ) .
the late - time optical spectra of all these sne show broad h@xmath1 and forbidden oxygen emission lines , and these characteristics have been exploited in searching for fainter , late - time sne
ii - l ( e.g. , sn 1985l in ngc 5033 ; fesen 1998 ) . of
the few sne ii - l recovered at late times , sn 1979c in m100 ( ngc 4321 ) is particularly noteworthy .
this sn was exceptionally bright at outburst for a type ii event , with a peak @xmath10 of 20 mag ( @xcite ; @xcite ) , and became bright in the radio as well , eventually reaching a 6 cm luminosity more than 200 times that of cas a ( weiler et al . 1989 ) .
it also exhibits among the highest , late - time h@xmath1 luminosity ( @xmath11 @xmath12 10@xmath13 erg s@xmath0 ) and shows radio emission variability which may be due to a periodic modulation of the progenitor s pre - sn mass loss ( @xcite ) .
late - time optical sn ii emission lines are thought to arise from interactions with surrounding circumstellar mass loss material ( csm ) .
these interactions lead to the formation of a reverse shock moving back into the expanding ejecta which then subsequently ionizes , either by far - uv or x - ray emission , a broad inner ejecta region from which the optical lines are produced ( chevalier & fransson 1994 , hereafter cf94 ) .
this model is also consistent with the presence of accompanying nonthermal radio emission in all optically recovered sne ii - l , with radio light curves like those predicted from the `` mini - shell '' model involving shock generated synchrotron emission from the forward shock ( chevalier 1982 ; weiler et al .
1993 ) .
late - time sn ii emission lines are important for the information they can provide on sne ejecta abundances , late - time shock emission processes , and the mass - loss history and evolutionary status of sn progenitors ( leibundgut et al . 1991 ; @xcite , montes et al .
1998 ) . unfortunately ,
optical emission lines from old sne ii - l are quite faint ( @xmath14 10@xmath15 erg @xmath5 s@xmath0 @xmath0 ) , making accurate measurements difficult even using 4-m class telescopes .
consequently , much is still uncertain about their spectra , even for sn 1980k , the best and longest studied object .
for example , virtually nothing is known about sn ii - l spectra below 5000 , and uncertainties remain as to whether [ ca ii ] or [ o ii ] is chiefly responsible for the strong emission commonly seen near 7300 .
sn 1980k s current optical properties are of special interest due to a recently reported sharp decline in nonthermal radio flux in the interval 19941997 ( montes et al .
1998 ) . here
we present a 1995 keck spectrum of sn 1980k which shows the region below 5000 for the first time , and a november 1997 mdm spectrum useful for investigating its recent h@xmath1 flux .
we then compare these sn 1980k data to a 1993 mmt spectrum of sn 1979c at a similar age to the keck sn 1980k spectrum .
we also present a 1997 _ hst _ uv spectrum of sn 1979c which reveals several strong uv and near - uv lines which help clarify the nature of the observed emission .
finally , we outline some general observed properties of late - time sne ii - l and compare them with young galactic supernova remnants ( snrs ) .
two consecutive low - dispersion spectra of sn 1980k were obtained on 28 november 1995 with the low resolution imaging spectrometer ( lris ; @xcite ) at the cassegrain focus of the keck - i telescope .
the duration of each exposure was 30 minutes .
conditions were not photometric , and the seeing was about 1@xmath16 .
the position angle of the slit ( of width @xmath17 ) was 128.5 degrees , through two stars that are known to be colinear with the supernova position ; this was also close to the parallactic angle at the time of observation .
we used a tektronix 2048 x 2048 pixel ccd with a scale of @xmath18 per binned pixel in the spatial direction , a gain of 1.1 e@xmath19/count , and a readout noise of 6 e@xmath0/pixel .
cosmic rays were eliminated from the two - dimensional spectra through a comparison of the pair of exposures .
the background sky was measured in regions adjacent to the supernova , and subtracted from the extracted spectrum .
spectra of hg - kr - ne - ar comparison lamps were used to determine the wavelength scale and resolution ( @xmath20 ) .
flux calibration and removal of telluric absorption lines were achieved with a spectrum of the sdf star bd+174708 ( @xcite .
a spectrum of sn 1980k was also obtained on 3 november 1997 using the 2.4 m hiltner telescope at the mdm observatory on kitt peak . a modular spectrograph like that in use at las campanas observatory
was employed with a n - s aligned @xmath21 slit , a 600 line mm@xmath0 5000 blaze grating , and a 1024 x 1024 tektronics ccd detector . a single 6000 s exposure covering the spectral region 4000 8500
was taken with a spectral resolution of about 5 .
cosmic rays were removed using standard iraf routines for pixel rejection and replacement . observing conditions were photometric but due to slit light losses caused by variable seeing during the long exposure , absolute fluxes are reliable only to @xmath22% .
a spectrum of sn 1979c was obtained on 21 may 1993 using the red channel long - slit ccd spectrograph ( schmidt , weymann , & foltz 1989 ) on the 4.5-m multiple mirror telescope ( mmt ) .
three 1200 s exposures were taken using a @xmath23 slit and covering a spectral range 3800 9900 with 12 resolution .
a strong blue continuum , possibly from o and b stars near sn 1979c s location in ngc 4321 ( m100 ) , was removed in the final reduction and an arbitrary zero flux level set .
because of the greatly increased noise level redward of 8000 , we show here only the region 4000 to 8000 . an _ hst _ spectrum of sn 1979c was obtained using the faint object spectrograph ( fos ) on 30 january 1997 . three g270h exposures ( 2200 3275 ) were taken for a total time of 7170 s , plus one g400h spectrum ( 3250 4750 ) with an exposure of 2410 s. both the g270h and g400h data shown here have been smoothed by a 5 point average .
the keck and mdm spectra of sn 1980k are shown in figures 1 and 2 respectively .
though the keck spectrum has significantly higher s / n , both show broad h@xmath1 and [ o i ] 6300,6364 lines plus emission near 7100 and 7300 .
these features have been seen in earlier sn 1980k spectra with roughly the same relative strengths and widths as seen here .
however , the line profiles are better defined in the keck data and the spectrum reveals other fainter emission lines not previously seen .
* h@xmath1 : * sn 1980k s late - time optical flux was first detected in 1987 via its strong h@xmath1 emission and it remains among the strongest optical lines observed a decade later . the 1995 keck spectrum shows the h@xmath1 line profile with an expansion velocity of 5700 to + 5500 km s@xmath0 with a strong asymmetry toward the blue , in good agreement with earlier 1992 and 1994 measurements ( fm94 ; fesen , hurford , & matonick 1995 , hereafter fhm94 ) .
a comparison of sn 1980k s h@xmath1 line profile and strength changes over the last ten years is shown in figure 3 using the lick june 1988 spectrum ( fb90 ) and our november 1997 mdm data .
besides illustrating a drop in h@xmath1 flux over this time period ( see section 4 below ) , one also sees a substantial decrease in the line width , most noticeable toward the blue edge . in 1988 ,
the fwhm of h@xmath1 was around 220 , compared to 190 in 1997 .
this decrease is consistent with earlier measurements ( fm94 ) and is predicted by sn
csm interaction models ( cf94 )
. however , the asymmetric profile of h@xmath1 appears not to have changed much during the last ten years , with a sharp blue emission edge and an emission peak near 6500 visible in both the 1988 and 1997 line profiles .
a faint , unresolved line is seen on top of the broader h@xmath1 emission around zero radial velocity .
a narrow h@xmath1 emission feature was evident in the 1988 lick spectrum , but much less so in subsequent data ( fm94 ; fhm95 ) . well detected again here , this narrow h@xmath1 emission may be due to a small h ii region near the sn site or ionized wind material associated with the progenitor ( fb90 ; fhm95 ) . from the 1997 mdm spectrum
, we estimate a broad emission h@xmath1 flux of @xmath4 erg @xmath5 s@xmath0 .
this is close to the 1.4 @xmath24 10@xmath25 erg @xmath5 s@xmath0 value reported by fhm95 for a 1994 spectrum and supports the conclusion of a @xmath7 25% fading from flux levels observed in the late 1980s and early 1990s .
* h@xmath26 : * there is broad emission near 4850 , some of which may be due to h@xmath26 4861 . its line strength , however , is difficult to disentangle from other possible emission lines between 4800 and 5000 ( see below ) .
nonetheless , the observed h@xmath1/h@xmath26 line ratio must be @xmath27 6 , suggesting , after correcting for a foreground reddening of e(b v ) = 0.40 mag ( @xcite ) , an intrinsic value @xmath27 4 .
* h@xmath28 : * weak emission near 4350 may be attributable in part to h@xmath28 emission , blended with [ o iii ] 4363 . *
[ o i ] : * the broad [ o i ] 6300,6364 emission is asymmetric towards the blue with a steep blue edge and a more gradual decline toward the red . the line s full velocity range is 6000 km s@xmath0 if attributed to 6300 on the blue side , and @xmath27 + 1700 km s@xmath0 if attributed to 6363 on the red .
the blueward velocity is larger than the 4800 km s@xmath0 estimated by fm94 , probably due in part to weaker emission detected due to the higher s / n in the keck spectrum .
in addition , a prominent emission peak at 6280 suggests strong clumping of the o - emitting material in the sn s facing expanding hemisphere . * [ o ii ] : * one of the strongest emission features in the spectrum , rivaling h@xmath1 and [ o i ] , lies near 7300 , and extends from 7180 to 7420 .
it has been alternatively identified as [ ca ii ] 7291,7324 ( @xcite ; fhm95 ) , [ o ii ] 7319,7330 (in sn1979c ; fm93 ) , or a combination of the two ( fb90 ; fm94 ) .
cf94 suggest it is likely to be solely [ ca ii ] since these [ o ii ] lines are weak in a variety of sn csm models and in a variety of photoionized objects .
however , these models also predict appreciable ca ii infrared triplet emission ( 8498 , 8542 , 8662 ) which is absent in sn 1980k s near - ir spectrum ( fm94 ; fhm95 ) .
figure 4a shows sn 1980k s [ o i ] 6300,6364 line profile in terms of expansion velocity relative to 6300 compared to the 7300 feature if identified as the [ o ii ] line blend at 7325 ( top panel ) , or [ ca ii ] using the stronger
7291 [ ca ii ] line ( bottom panel ) . the [ o ii ] blend interpretation is seen to provide a better match to [ o i ] s line width and its emission substructure than a [ ca ii ] interpretation .
for instance , the 6280 emission peak in the [ o i ] line profile matches well with the 7300 feature s emission feature at 7300 if interpretated as an [ o ii ] blend . in addition , the velocity of the 7300 feature s fwhm blueward edge agrees within measurement uncertainties to those of [ o i ] and h@xmath1 using the 7319 [ o ii ] line : namely , 4750 , 4500 , and 4600 km s@xmath0 for [ o i ] , h@xmath1 , and 7300 respectively , compared to 3450 km s@xmath0 using [ ca ii ] 7291 . a possible complication is an atmospheric absorption band at 7190 which might have affected the 7300 feature s blue emission .
this telluric absorption band is not especially strong , however , and though it could help to explain the poor [ ca ii ] profile match to that of [ o i ] at large negative velocities ( see fig .
4 ) , there still would remain the discrepancy along the line s red end .
we therefore conclude that [ o ii ] 7319,7330 emission is the dominant cause for the 7300 emission feature . * [ o iii ] : * previous sn 1980k spectra detected broad emission near 5000 with an emission peak near 4955 .
the broad emission was initially identified as [ o iii ] 4959,5007 emission ( fb90 ; @xcite ; fm94 ) although @xcite suggested it might be a blend of permitted fe ii lines . as seen in figure 1 ,
broad emission extends from 4750 to 5020 and from 5100 to 5400 , with little emission between 5020 and 5100 .
absence of strong emission just redward of 5007 is hard to reconcile with a broad , 5000 km s@xmath0 [ o iii ] line profile like seen for [ o i ] , [ o ii ] , and h@xmath1 unless the [ o iii ] emission is weak and has a strong blue asymmetry .
in fact , the sharp cutoff of emission at 5020 places an expansion limit on strong [ o iii ] 5007 emission of @xmath29 km s@xmath0 . nonetheless , [ o iii ] appears a much more likely identification than either permitted or forbidden fe ii and fe iii lines ( cf .
fhm95 ) , especially in view of the clear [ o iii ] emission in other sne ii - l ( e.g. , sn 1979c ; see section 3.2 ) .
the far blue side of the feature is probably due to h@xmath26 4861 .
finally , weak emission near 4350 may be a blend of [ o iii ] 4363 and h@xmath28 .
the keck spectrum also shows two narrow lines at 4957 and 5005 .
these are probably related to the narrow h@xmath1 emission observed .
however , in many earlier spectra , only one emission spike , around 4950 4960 , could be clearly seen ( fb90 ; @xcite ; fm94 ) and was interpreted as possible evidence of emission substructure in the 4959,5007 [ o
iii ] blend .
an alternative explanation is that the narrow 4959 line emission happens to coincide with a @xmath30 km s@xmath0 emission feature in the broad [ o iii ] line , like that seen in the [ o i ] and [ o ii ] line profiles . *
[ fe ii ] : * emission near 7100 which merges into the red wing of the 7300 line is likely [ fe ii ] 7155 ( fb90 ; fm94 ) .
adopting a line center of 7100 gives an emission centroid velocity of 2300 km s@xmath0 .
emission extends toward the blue as far as @xmath7 7010 suggesting a maximum velocity of 6000 km s@xmath0 . due to the blend with the [ o ii ] 7300 feature , little can be determined regarding its maximum recession velocity .
however , the line extends at least to 7175 ( + 800 km s@xmath0 ) . broad weak emission from 5020 to 5400 can be understood in terms of blends of several [ fe ii ] lines specifically the 5112 ( 19f ) , 5158 ( 18f ) , 5159 ( 19f ) , 5262 ( 19f ) , 5269 ( 18f ) , 5273 ( 18f ) , and 5334 ( 19f ) lines .
these represent all of the stronger [ fe ii ] lines from the multiplets 18f and 19f typically seen in shocked supernova remnant emissions ( cf . fesen & hurford 1996 ) .
adopting a line center shift of 2300 km s@xmath0 and a velocity width of 6000 to + 3500 km s@xmath0 like that seen for the [ fe ii ] 7155 line , a blend of [ fe ii ] lines can explain the 50205400 emission : specifically , its width , peak intensity near the strong 5158 and 5159 [ fe ii ] lines , and the lack of emission longward of 5400 and shortward of 5050 .
* [ s ii ] : * the weak emission feature near the blue end of the keck spectrum appears to be [ s ii ] 4069,4076 . adopting this identification ,
sulfur has a velocity range of @xmath31 to + 4000 km s@xmath0 , comparable to the broad h and o lines . as with other lines , the [ s ii ] profile also shows a strong blue asymmetry , with a line center shifted to the blue at least 2000 km s@xmath0 .
finally , we note there is no obvious emission from the nebular [ s ii ] 6716,6731 lines in either the keck or mdm spectra , and previous sn 1980k studies have failed to detect any [ s iii ] 9069,9531 emission ( fhm95 ) . a 1993 mmt spectrum of sn 1979c taken 14 yr after maximum light , or roughly the same age as sn 1980k when the 1995 keck spectrum was obtained , is shown in figure 5 .
the spectrum is similar to those presented by fm93 but with much improved s / n especially in the blue .
* h@xmath1 : * broad h@xmath1 emission is detected with an expansion velocity of @xmath32 km s@xmath0 .
a strong , narrow h@xmath1 component is also seen at + 120 velocity in m100 s rest frame ( v = + 1570 km s@xmath0 ) .
the broad h@xmath1 emission has a flux of @xmath33 erg @xmath5 s@xmath0 , or slightly larger than the @xmath34 erg @xmath5 s@xmath0 reported by fm93 .
s peak flux is somewhat weaker than [ o i ] 6300 and the profile shows a strong blue asymmetry .
narrow [ s ii ] 6716,6731 emission is seen along its red edge , presumably associated with the narrow h@xmath1 emission . * [ o i ] : * sn 1979c s late - time [ o i ] 6300,6364 emission profile is double - peaked much like that reported by fm93 but with different peak velocities . from the mmt spectrum ,
we measure peaks at 6230 and 6320 ( @xmath35 and @xmath36 km
s@xmath0 if attributed to just the 6300 line ) , whereas fm93 found peaks at 6214 and 6329 ( @xmath37 and @xmath38 km s@xmath0 ) .
however , [ o i ] s full expansion velocity in the mmt spectrum is @xmath39 to @xmath40 km s@xmath0 and in good agreement with fm93 .
a june 1998 2.4 m mdm [ o i ] image of sn 1979c taken using the same filter used by fm93 indicates a 10% increase in [ o i ] luminosity between 1991.5 and 1998.5 . *
[ o ii ] : * fm93 identified the emission near 7300 as [ o ii ] 7319,7330 based upon the excellent match of the feature s double - peaked velocity profile with the [ o i ] emission peaks .
two strong emission peaks can also be seen in the mmt spectrum at 7240 and 7347 .
these correspond to @xmath41 and @xmath42 km s@xmath0 in m100 s rest frame , in agreement with the above measured [ o i ] peak velocities .
this is graphically shown in figure 4b and helps validate the [ o ii ] identification made for the sn 1980k spectrum .
we note that the [ o ii ] 7319,7330 emission appears stronger in 1993 relative to [ o i ] when compared to 1991/1992 data ( fm93 ) , making it now the strongest feature in sn 1979c s optical spectrum . * [ o iii ] : * the mmt spectrum of sn 1979c also shows broad , strong [ o iii ] emission centered at 4965 and spanning 4880 to 5060 , which translate to @xmath39 to + 1600 km s@xmath0 .
no emission peaks like those seen for [ o i ] and [ o ii ] are evident from these data .
we measure an observed [ o iii ] flux of @xmath43 erg @xmath5 s@xmath0 . broad emission at 5750 : weak but very broad emission
is also seen near 5750 and extending from 5550 to 6000 .
though faint , this emission is consistent with an earlier report of broad emission at 57005900 by fm93 .
possible identifications include blends of [ n ii ] 5755 , na i 5890,5896 , and he i 5876 .
figure 6 shows the combined 1997 fos spectra ( g270h and g400h ) of sn 1979c covering the wavelength region 2200 to 4750 .
the continuum levels at the g270h / g400h 3250 crossover point agreed quite well and no zero - point correction was applied .
three prominent emission lines were detected below 3000 : c ii ] 2324,2325 , [ o ii ] 2470,2470 , and mg ii 2796,2803 .
these lines show clumpy and strongly asymmetric profiles with the blueward peaks about twice as bright as those along the redward side , suggestive of internal dust extinction of the receding emission regions .
sharp blue emission peaks in these lines match well those seen in the optical lines from the mmt spectrum ; specifically , the c ii ] , [ o ii ] and mg ii emission peaks at 2297 , 2440 , and 2767 correspond to the @xmath44 km s@xmath0 peaks in the [ o i ] 6300 and [ o ii ] 7325 optical lines .
the mg ii 2800 line appears broader than the two other uv lines , and this might indicate the presence of weak mg i 2852 along its red wing .
it is also possible that weak si ii ] 2335 emission might also be blended with the c ii ] 2324,2325 feature .
the [ o ii ] 2470 lines arise from the same upper level ( @xmath45 @xmath46 ) as the near - ir [ o ii ] 7319,7330 lines .
the relative strength of the 2470/7325 ratio is determined by radiative transition probabilities ( 0.78 ; mendoza 1983 ) and thus can be used to derive an estimate of the foreground extinction . assuming no contribution from [ ca ii ] 7291,7330
, the observed 2470/7325 ratio of @xmath47 suggests that e(b
v ) = @xmath48 mag .
however , this includes dust extinction within the sn s ejecta as indicated by the weaker red emission peaks in these uv lines compared to the near infrared lines .
therefore , measuring only the blue emission peaks in the [ o ii ] 2470 and 7325 lines , we find a 2470/7325 ratio of @xmath49 which implies e(b v ) = @xmath50 assuming r = a@xmath51/e(b v ) = 3.1 .
this should be a reasonable estimate of the foreground extinction ( both galactic and in m100 ) for sn 1979c .
however , it is larger than the previous total reddening value of @xmath52 estimated toward sn 1979c ( @xcite ) and may indicate that some weak [ ca ii ] 7291,7330 emission contributes to the 7300 feature s strength .
this would make the 2470/7325 ratio appear smaller and hence the reddening larger .
alternatively , there might have been changes in [ o ii ] line strengths during the four year interval between when the optical ( 1993 ) and uv ( 1997 ) data were taken .
one can also estimate the internal dust extinction in the young , developing remnant by comparing the red emission spikes in the [ o ii ] 2470 and 7325 lines .
the observed 2470/7325 ratio of 0.15 suggests e(b v ) = @xmath53 implying an internal reddening of e(b v ) = @xmath6 in addition to the foreground e(b v ) of @xmath54 .
although the evidence for internal extinction is strong due to the consistently weaker red emission spikes in the uv line profiles , our measurement is not very accurate from this single measurement .
although the red emission peak is comparable or slightly stronger than the blue in the optical lines of [ o i ] 6300,6364 and [ o ii ] 7319,7330 , much greater variation in blue / red peak strengths is seen for the uv lines in the _ hst _ data .
for example , the red emission peak in the [ o ii ] 2470 line is about 45% as strong as the blue emission peak , around 65% in the c ii ] 2324 line , and less than 35% for the mg ii 2800 line .
emission features detected between 3000 and 4600 in the short g400h exposure are considerably weaker and consequently have less certain identifications but appear to be [ ne iii ] 3869 , [ s ii ] 4068,4076 , and [ o iii ] 4363 .
the strength of the temperature sensitive [ o iii ] 4363 line is surprisingly large relative to the [ o iii ] 4959,5007 lines and suggests electron densities @xmath55 10@xmath56 @xmath3 .
the only other viable line identification for the 4335 emission is h@xmath28 .
this seems unlikely , however , due to the weakness of the h@xmath1 emission .
finally , our detection of [ s ii ] 4068,4076 is tentative but would be consistent with the sn 1980k spectrum .
to date , five papers have been published on the late - time optical spectrum of sn 1980k plus one each on sn 1970 g , sn 1979c , and sn 1986e .
none of these , however , presented spectra of sufficient s / n or covered a wide enough wavelength range to constrain several basic physical parameters such as densities or internal extinction , or test current late - time emission models .
the new sn 1979c and sn 1980k spectra do provide some of this information and give a much clearer picture for late - time optical and uv properties of sne ii - l .
in addition , because sn 1979c and sn 1980k were observed at comparable ages , they can be used to gauge the spread of sn ii - l late - time emission properties .
table 1 lists the observed emission properties of four type ii - l sne having detected late - time optical emission .
they can be seen to exhibit several spectral similarities . from @xmath11 8
yr and extending at least to 17 yr , the strongest optical emission lines in sne
ii - l are h@xmath1 , [ o i ] 6300,6364 , [ o ii ] 7319,7330 , and [ o iii ] 4959,5007 .
these lines show similar expansion velocities , @xmath11 5000 6000 km s@xmath0 , often with roughly equal line strengths .
several of these lines also have emission profiles showing evidence for dust formation by virtue of diminished flux from the receding emission portions .
the presence of this dust extinction creates both blueshifted line centers and asymmetric blue line profiles .
in addition , all four sne show narrow emission lines ( e.g. , h@xmath1 and [ o iii ] ) which could imply similar interstellar environments for the progenitors such as local h ii regions . however , significant spectral differences also exist .
luminosities cover more than an order of magnitude , with a factor of six difference between objects of similar age ( e.g. , 79c and 80k ) .
sn 1970 g and sn 1986e exhibit little if any broad [ o iii ] emission , sn 1980k some , while in sn 1979c [ o iii ] is stronger than h@xmath1 .
conversely , sn 1980k has fairly strong [ fe ii ] emission , both at 7155 and the blend at 51005300 , whereas sn 1979c and the other two do not .
in addition , the very broad , unidentified emission near 5750 in sn 1979c is not seen in sn 1980k .
bright optical emission appears correlated with the presence of strong late - time radio nonthermal emission and inferred high mass loss rates ( see table 1 ) in the sense that h@xmath1 luminosity tracks fairly well the peak 6 cm radio luminosity and the derived mass loss rate from late - time radio data .
for example ,
sn 1979c is the brightest radio and optical sn and has the largest estimated mass loss rate while sn 1970 g is optically the faintest with the lowest estimated mass loss rate .
however , this correlation is not a particularly tight one .
sn 1986e is nearly as bright optically as sn 1979c , yet it is quite faint in the radio .
moreover , strong radio emission is also not always a predictor of bright optical emission .
two of us ( fesen & filippenko ) have been unsuccessful in detecting late - time optical emission from sn 1968d in ngc 6946 ( the same parent galaxy as sn 1980k ) as a follow - up to its late - time radio emission recovery ( hyman et al .
the lack of detectable optical emission may be related to sn 1968d s location closer to ngc 6946 s center and in a dustier region ( see @xcite and @xcite ) which may be attenuating faint , late - time optical emission below the detection threshold .
late - time , optical emission lines from sne ii - l appear surprisingly steady over time spans of many years .
our 1993 spectrum of sn 1979c indicated no decrease in h@xmath1 flux compared to 1990/1991 data ( fm93 ) and actually showed a slight increase as does a comparison of 1991/1998 [ o i ] images .
recent radio measurements indicate a flattening of its radio light curve ( van dyk et al .
1998b ) and new optical measurements should be undertaken .
sn 1980k s h@xmath1 flux remained nearly constant from 1987 through 1994 , decreasing only recently . in figure 7 , we show the h@xmath1 light curve over one decade : from 1987 through 1997 .
the data indicate no change between 1987 and 1992.5 .
spectra taken in 1991.4 and 1992.5 , though indicating possibly a slow fading , were of lower quality than one obtained in 1992.6 which indicated no significant drop in h@xmath1 strength ( fm94 ) . moreover , broadband photometric studies covering the years 1990 1992 indicated sn 1980k s optical flux level remained virtually unchanged from that seen in 1987 , especially in the r band which is sensitive to h@xmath1 emission ( r = 21.9 @xmath57 0.1 mag ; leibundgut et al .
1993 ) . a more recent spectrum taken in 1994 , however , showed an h@xmath1 emission flux of @xmath58 erg @xmath5 s@xmath0 ( fhm95 ) , about 25% less than the @xmath59 erg @xmath5 s@xmath0 found earlier .
our 1997 measurement of @xmath60 erg @xmath5 s@xmath0 supports a fading h@xmath1 strength which may have started during the last several years , possibly in 1994 with the fhm95 reported decline . while this optical fading would roughly coincide with a drop in radio emission ( montes et al .
1998 ) , we can not completely rule out a small , steady decline in h@xmath1 over the 19871997 period from the spectroscopic data . during t = 10 20 yr ,
sn 1980k s broad h@xmath1 line profile gradually narrowed , changing from a fwhm = 220 in 1988.6 to 190 in 1997.9 ( see fig .
this trend has been noted earlier ( fm94 ) and is predicted by sn
csm interaction models ( cf94 ) . however , h@xmath1 appears to be the line most affected , with the [ o i ] 6300,6364 line profile changing little during this time interval . with the notable exception of h@xmath1 , all the broad emission lines
exhibit a spiky profile suggestive of clumpy emission regions at particular velocities .
this is perhaps most clearly seen in sn 1979c where emission peaks at @xmath41 and @xmath61 km s@xmath0 can be seen in the [ o i ] 6300 , [ o ii ] 2470 , [ o ii ] 7325 , c ii ] 2324 , and mg ii lines .
the less blueshifted emission peaks are always weaker in the uv lines compared to the red or near infrared lines , consistent with the presence of internal extinction . in a similar fashion ,
sn 1980k also exhibits strong emission peaks in the [ o i ] 6300 and [ o ii ] 7325 lines .
the fact that the h@xmath1 profile in both sn 1979c and sn 1980k does not show such emission peaks suggests that the o , c , mg - rich emitting material is coming mostly from ejecta and is physically separate from the dominant h@xmath1 emitting material which includes the swept - up shell .
our new spectra also allow one to get a handle on the electron density and temperature .
in sn 1980k , we detected [ s ii ] 4068,4076 lines but not [ s ii ] 6716,6731 emission .
the [ s ii ] 4068,4072 lines have a critical density , n@xmath62 , of @xmath63 @xmath3 , whereas the 6731 line has n@xmath62 = @xmath64 @xmath3 .
this suggests collision de - excitation of the [ s ii ] 6716,6731 lines due to electron densities near 10@xmath2 @xmath3 .
an upper limit of @xmath11 10@xmath65 @xmath3 is suggested by the presence of strong [ o i ] 6300,6364 emission ( n@xmath62 = @xmath66 @xmath3 ) and [ o ii ] 7319,7330 (n@xmath62 = @xmath67
@xmath3 ) , although [ o i ] 6300,6364 emission can be strong even at high densities because it is an important coolant for low ionization gas . for example , [ o i ] emission is an important for the cool shell in the cf94 models even though @xmath68 @xmath3 . if sn 1979c s densities were similar to those of sn 1980k , we would predict little [ o ii ] 3726,3729 emission which has n@xmath62 = @xmath69 @xmath3 , and this agrees with observation .
indeed , the fos spectrum of sn 1979c reveals strong [ o ii ] 2470 emission but virtually no [ o ii ] 3727 emission .
as shown in table 2 , only those emission lines with critical densities above 10@xmath70 @xmath3 are detected in both 79c and 80k .
consequently , we conclude that all lines with critical densities below 10@xmath70 @xmath3 are collisionally suppressed relative to lines having higher critical densities
. electron densities of @xmath71 @xmath3 would also explain the observed low ratio of [ o iii ] ( 4959 + 5007)/4363 .
the [ o iii ] 4959,5007 lines have n@xmath62 = @xmath72 @xmath3 . if the electron densities were a bit above @xmath73 @xmath3 , then the normally strong [ o iii ] lines would be suppressed relative to [ o iii ] 4363 line intensity ( n@xmath62 = @xmath74 10@xmath65 @xmath3 ) .
furthermore , the [ o iii ] electron temperature must be greater than @xmath11 15,000 k ( i.e. , i(4959 + 5007)/i(4363 ) @xmath7 4 assuming n = @xmath75 @xmath3 ) in order to generate sufficient 4363 emission to be detected given the s / n of the data . a very broad emission feature seen in sn 1979c s spectrum near 5800 may be partially due to the temperature sensitive [ n ii ] 5755 line .
if correct , this would also indicate a [ n ii ] temperature above 10@xmath76 k. the overall weaker [ o iii ] emission seen in the sn 1980k spectrum might be due to low ejecta ionization as indicated by the presence of of strong [ fe ii ] 50505400 and 7155 emission .
late - time optical observations of sne ii - l can be used to test emission predictions in cf94 s recent sn csm models . in table 1 , we list cf94 s predicted relative line intensities for t = 10 and 17.5 yr arising from sn csm interactions assuming either a power law or red supergiant ( rsg ) density gradient .
there is a general agreement with the observed line list , although there are significant differences with the relative line intensities .
the observed line widths are approximately those expected and the general picture of emission from circumstellar interaction is supported , but cf94 present two detailed models that appear not to be well matched to the conditions in sn 1980k and sn 1979c .
the models are based on the explosion of a @xmath77 red supergiant with two variations for the outer density profile .
although not ideal , these models do provide a starting point for discussing the current observations .
the rsg model predicts about the right h@xmath1 luminosity but weaker [ o i ] and stronger [ o iii ] ( relative to h@xmath1 ) than observed in all four cases . on the other hand , the 10 and 17.5 yr old power law models predict about the right h@xmath1/[o i ] ratio , but less [ o ii ] 7319,7330 emission , too much [ o iii ] 4959,5007 , and much weaker h@xmath1 luminosity than observed .
larger discrepancies between observations and models are seen for several uv lines .
both rsg and power law models predict mg ii 2800 to be the dominant uv / optical line ( after ly@xmath1 ) at a strength @xmath78 times more than h@xmath1 . however , the sn 1979c spectrum shows mg ii to be slightly weaker than h@xmath1 , with the strongest line actually being c ii ] 2324,2325 .
likewise , mg i ] 4571 is predicted to be about half the strength of h@xmath1 yet is not seen in either 79c or 80k .
large differences between observed and predicted mg ii intensities are a problem common in young , o - rich snrs ( blair et al .
1994 ; sutherland & dopita 1995 ) and may simply be indicating an incorrect assumed mg abundance .
however , large differences exist for other uv lines , such as c ii ] 2324,2325 which is about twice the predicted strength , and this difference grows somewhat larger if internal sn dust extinction is considered .
other notable differences involve the [ o ii ] lines .
we find strong [ o ii ] 7319,7330 emission in both sn 1979c and sn 1980k and suspect its presence in sn 1986e where it may have been mis - identified as [ ca ii ] 7291,7330 ( @xcite ) . but [ o ii ] 7319,7330 emission is weak in all cf94 models , due to the high excitation energy of the upper [ o ii ] level and its small collision strength . because of this cf94 argued that [ ca ii ] was the chief source of line emission observed near 7300 .
however , in view of the new spectra , it now appears quite likely that the majority of the 7300 emission is due to [ o ii ] .
the presence of strong [ o ii ] 2470 supports this conclusion .
since cf94 do not predict a 2470 line strength we can not do a direct comparison . but based upon the theoretical [ o ii ] 2470/7325 ratio of 0.78 , the model prediction should be about an order of magnitude too low .
although [ o ii ] 7319,7330 lines are weak in most photoionized nebulae , the presence of strong [ o ii ] emission in young snrs is not uncommon . in almost all young , o - rich snrs such as cas
a ( sn 1680 ) , as well as in sn 1957d ( cappellaro et al .
1995 ) and sn 1986j ( leibundgut
et al . 1991 ) , one finds strong [ o ii ] 7325 emission which is equal to or stronger than the other forbidden oxygen line emissions .
strong [ o ii ] emission has been successfully modeled by sutherland and dopita ( 1995 ) using a mixture of shock and photoionized emissions zones in o - rich ejecta .
additional model - observation differences involve the the ca ii ir triplet and the [ s iii ] 9069,9531 lines .
these emission lines are predicted to be strong in the power law models but are not detected in the published near - ir spectrum of sn 1980k ( fhm95 ) .
finally , the favored power law models predict a steady decline of the h@xmath1 emission and a rapid increase of [ o iii ] 4959,5007 ( cf94 ) .
many model - observation differences appear to be related to higher electron and gas densities than the cf94 models assume .
as noted above , the presence of strong [ o iii ] 4363 emission indicates densities of @xmath79 @xmath3 .
high oxygen gas densities would also account for the lack of any [ o iii ] 4959 and 5007 line discrimination in the [ o iii ] profile for sn 1979c . at densities above 10@xmath56 @xmath3
, several lines predicted to be observable in the models , like [ o ii ] 3727 , [ n ii ] 6548,6583 and [ s ii ] 6716,6731 , will be strongly de - excited ( see table 2 ) .
the differences indicate that the cf94 stellar model is not the best one to model sn 1979c or sn 1980k .
their assumed circumstellar density may also be a factor .
models for the light curves of these supernovae suggest a h envelope mass of @xmath80 ( swartz , wheeler , & harkness 1991 ; blinnikov & bartunov 1993 ; arnett 1996 ) , much lower than the mass assumed in cf94 . with the low envelope mass , processed core material
may not be strongly decelerated and could be moving at @xmath81 km s@xmath0 .
some deceleration of the core material could lead to hydrodynamic instabilities and clumping of the gas , which is consistent with the spiky profiles seen in the spectral lines of o , c , and mg . if the gas is h - depleted core material , strong differences can be expected with the models of cf94 who assumed cosmic abundances . the models presented in cf94 also assumed a circumstellar density determined by a mass loss rate @xmath82 for a wind velocity of @xmath83 .
however , the analysis of the radio turn - on by lundqvist & fransson ( 1988 ) leads to @xmath84 for sn 1979c for the same wind velocity .
there is new evidence for a high circumstellar density around sn 1979c .
immler , pietsch , & aschenbach ( 1998 ) have detected x - ray emission from the supernova with _
rosat _ and determined a @xmath85 kev x - ray luminosity of @xmath86 .
the summed luminosity of the observed lines listed in table 1 is @xmath87 and this figure is likely to be substantially increased when unobserved lines such as l@xmath1 are included ( cf94 ) . in the circumstellar interaction model ,
the optical and ultraviolet lines are excited by the x - ray radiation , so the implication is that the x - rays from the reverse shock wave are being absorbed by the cool circumstellar shell .
this would explain the observed constancy of the line emission because the x - ray emission at the reverse shock should decline slowly if the reverse shock is a cooling shock ( cf94 ) .
application of eq .
( 2.17 ) of cf94 shows that the circumstellar shell can still be optically thick at 1 kev at the time of the observations if the density power law for the supernova is @xmath88 , where @xmath89 , or if the circumstellar density is about twice that quoted above . in the cf94 power law density model at an age of 10 years , the supernova h density just inside the reverse shock front is @xmath90 @xmath3 . for sn 1979c
, the density may be several times higher , but it still falls short of the density implied by the [ o iii ] line ratio .
the implication is that the [ o iii ] lines are formed inside of the reverse shock , which is consistent with the lack of narrowing of the lines with time .
from keck and mdm optical spectra of sn 1980k and optical mmt and uv _ hst _ spectra of sn 1979c , we find the following : \1 ) the optical spectrum of sn 1980k taken at 15 and 17 yr shows continued strong and broad 5500 km s@xmath0 emission lines of h@xmath1 , [ o i ] 6300,6364 , and [ o ii ] 7319,7330 , with weaker but similarly broad lines of [ s ii ] 4068,4072 , h@xmath26 , [ fe ii ] 7155 , and a [ fe ii ] blend at 50505400 .
the presence of [ s ii ] 4068,4072 but a lack of [ s ii ] 6716,6731 emission suggests collisional de - excitation of the [ s ii ] 6716 and 6731 lines due to electron densities of 10@xmath2 @xmath3 .
the 1997 mdm spectra indicates a h@xmath1 flux of @xmath4 erg @xmath5 s@xmath0 , suggesting a 25% drop from 19871992 levels sometime during the period 1994 to 1997 , possibly simultaneously with an observed decrease in nonthermal radio emission .
\2 ) like sn 1980k , sn 1979c s optical spectrum at t = 14.0 yr shows @xmath11 6000 km s@xmath0 wide emission lines but weaker h@xmath1 , strong [ o iii ] 4959,5007 , clumpy [ o i ] and [ o ii ] line profiles , no detectable [ fe ii ] 7155 emission , and a faint but very broad emission feature near 5750 .
a 1997 _ hst _ faint object spectrograph spectrum covering the range 2200 4500 shows strong lines of c ii ] 2324,2325 , [ o ii ] 2470 , and mg ii 2796,2803 along with weaker [ ne iii ] 3969 , [ s ii ] 4068,4072 , and [ o iii ] 4363 .
a lack of [ o ii ] 3726,3729 emission together with a [ o iii ] ( 4959 + 5007)/4363 @xmath7 4 indicates electron densities @xmath11 10@xmath8 @xmath3 .
furthermore , the [ o iii ] temperature must be greater than @xmath11 15,000 k in order to generate sufficient 4363 emission to be so easily detected ( i.e. , i(4959 + 5007)/i(4363 ) @xmath7 4 assuming n = @xmath75 @xmath3 ) . a very broad emission feature seen in sn 1979c s spectrum near 5800
may be partially due to the [ n ii ] line at 5755 .
\3 ) in both sn 1979c and sn 1980k , several lines show one or more sharp emission peaks .
the blueward peak(s ) are substantially stronger than those toward the red indicating internal dust extinction with the expanding ejecta .
the amount of internal extinction in sn1979c is estimated to be e(b v ) = @xmath91 mag .
the line profile differences exhibited between h@xmath1 and the oxygen lines suggests that we are seeing emission from two or more separate regions , possibly the shell ( h@xmath1 ) and the inner sn ejecta ( o lines ) .
\4 ) comparison of these observations to late - time sn model predictions indicates several areas of significant differences , many of which can be attributed to model electron densities being several orders of magnitude too low .
for example , [ o ii ] 3727 emission is predicted to be moderately strong in the models but is not seen due to densities well above the [ o ii ] critical density of @xmath92 10@xmath93 @xmath3 . in other cases , such as mg ii 2800 which arises chiefly from the swept - up shell , the observed emission is far weaker than predicted .
the parameters for future models will have to be more closely adapted to the conditions in these supernovae . it should be noted that other sne ii besides type
ii - l can exhibit bright , late - time optical emission , most notably the type iin objects sn 1986j ( @xcite ) and sn 1988z ( @xcite ; @xcite ; and @xcite ) .
these objects show quite a different optical spectrum from the sne ii - l discussed above and are believed to be encountering dense , clumpy csm ( @xcite ; @xcite ) .
the current , single type ii - p detection of sn 1923a in m83 in the radio ( @xcite , @xcite ) together with the lack of any reported late - time optical detections of sne ii - p suggests significant mass loss differences between sne ii - l and ii - p .
it will be interesting to see if sn 1923a can be recovered optically and how its uv and optical emission properties compare to those of sne
ii - l . into
what kind of young remnants will these sn ii - l emission nebulae evolve ? except for the higher densities and the presence of strong hydrogen emission , the observed late - time spectra of sne ii - l bear some resemblance to galactic and lmc o - rich snrs such as cas a and 1e0102 - 7219 ( @xcite ; @xcite ) .
their o , s , and c emission - line dominated spectrum with expansion velocities around 5000 km s@xmath0 is not unlike that seen in 10@xmath94 yr old o - rich snrs
. the weakness of [ o iii ] 4959,5007 emission and the absence of lines like [ o ii ] 3727 , [ ne iv ] 2425 , and [ ne v ] 3426 can be attributed to higher filament densities and lower ionization levels .
however , it is not at all clear if the progenitors of these older , metal - rich remnants are related to sne ii - l events . in the case of cas
a , the progenitor may have been a wn star , quite different from the rsg progenitor usually assumed for sne ii - l .
moreover , the cas a supernova appears to have been subluminous and thus very different from the super - luminous sn ii - l events sn 1979c and sn 1980k .
this all raises the issue of whether sne of different types can yet leave similar looking young remnants at ages 10 to 100 yr .
if sn 1979c and sn 1980k remain luminous for several more decades , we may be able to address this question directly .
we are grateful to the keck , mmt , and mdm observatory staffs for their excellent observing assistance , and k. weiler and m. montes for generously communicating their radio study results on sn 1980k prior to its publication .
financial support for this work was provided by nasa through grant nos .
go-6043 and go-6584 from the space telescope science institute , which is operated by aura , inc .
, under nasa contract no .
nas 5 - 26555 .
we also acknowledge nsf grants ast-9529232 to r.a.f . and ast-9417213 to a.v.f .
lccccccc age at observation ( yr ) & 8 & 14 - 15 & 1214 & 22 & 10 & 10 & 17.5 + distance ( mpc ) & 16.8 & 7.5 & 17 & 7 & & & + e(b v ) ( mag ) & 0.02 & 0.40 & 0.23 & 0.15 & & & + mass loss rate@xmath95 & 4.7 & 2.0 & 19 & 2 & & & + h@xmath1 luminosity@xmath96 & 8.9 & 2.5 & 15 & 1 & 4.8 & 0.9 & 0.5 + peak radio luminosity@xmath97 & 1.1 & 1.0 & 26 & 1.5 & & & + relative line fluxes & & & & & & & + h@xmath1 6563 & 10 10 & 10 10 & 10 10 & 10 10 & 10 & 10 & 10 + [ o i ] 6300,6364 & 9 9 & 7 7 & 9 9 & 6 6 & 1 & 5 & 7 + [ o ii ] 2470,2470 & & & 2 6 & & & & + [ o ii ] 7319,7330 & 3 3 & 10 9 & 12 11 & & & 1 & 1 + [ o iii ] 4363 & & 1 2 & 2 3 & & & & + [ o iii ] 4959,5007 & @xmath983 @xmath985 & 2 3 & 9 11 & @xmath986 @xmath987 & 34 & 23 & 46 + c ii 2324,2325 & & & 6 25 & & 11 & 15 & 10 + mg ii 2796,2803 & & & 4 9 & & 49 & 79 & 57 + [ ne iii ] 3868,3869 & & & 2 3 & & 3 & 4 & 5 + [ s ii ] 4068,4076 & & 2 4 & 1 2 & & & 2 & 3 + [ fe ii ] 7155 & @xmath983 @xmath983 & 3 3 & @xmath985 @xmath985 & & & & + & & & & & & & + detection upper limits & & & & & & & + [ o ii ] 3726,3729 & & & @xmath99 @xmath982 & & @xmath981 & 2 & 5 + mg i ] 4571 & & @xmath982 @xmath984 & @xmath982 @xmath984 & & 3 & 6 & 4 + [ na i ] 5890,5896 & @xmath983
@xmath984 & @xmath982 @xmath982 & ? & @xmath985 @xmath986 & 2 & 10 & 13 + [ n ii ] 6548,6583 & @xmath983 @xmath983 & @xmath985 @xmath985 & @xmath985 @xmath985 & @xmath985 @xmath985 & @xmath981 & 5 & 7 + [ c i ] 8729 & & @xmath982 @xmath982 & & & 0 & 2 & 4 + [ s iii ] 9069,9531 & & @xmath983 @xmath983 & & & 2 & 11 & 13 + & & & & & & & + expan .
velocity ( km s@xmath0 ) & & & & & & & + h@xmath1 6563 ( blue)&@xmath100 & @xmath37&@xmath39 & @xmath101 & & & + ( red ) & @xmath102 & @xmath103&@xmath104 & @xmath105 & & & + [ o i ] 6300,6364 ( blue)&@xmath106 & @xmath100&@xmath39 & @xmath107 & & & + [ o ii ] 7319,7330 ( blue ) & & @xmath41&@xmath39 & & & & + ( red)&&@xmath108&@xmath109 & & & & + [ s ii ] 4068,4076 ( blue ) & & @xmath31 & & & & & + ( red ) & & @xmath110 & & & & & + h@xmath1 line asymmetry & red ?
& blue & blue & blue & & & + clumpy [ o i ] emission ?
& yes ? & yes & yes & yes & & & + narrow nebular lines ? & yes & yes & yes & yes & & & + llccc n@xmath111 @xmath112 10@xmath70 @xmath3 & & & & + [
n ii ] 6548,6583 & @xmath113 & absent & absent & strong + [ o ii ] 3726,3729 & @xmath114 , @xmath115 & absent & absent & moderate + [ s ii ] 6716,6731 & @xmath116 , @xmath64 & absent & absent & weak + [ ne iv ] 2422,2425 & @xmath117 , @xmath118 & & absent & weak + & & & & + n@xmath111 = 10@xmath2 @xmath3 & & & & + [ o iii ] 4959,5007 & @xmath119 & weak & moderate & v. strong + [ s iii ] 9069,9531 & @xmath120 & absent ? & & strong + & & & & + n@xmath111 @xmath27 10@xmath56 @xmath3 & & & & + [ o i ] 6300,6364 & @xmath121 & strong & strong & strong + [ o ii ] 2470,2470 & @xmath122 , @xmath123 & & strong & + [ o ii ] 7319,7330 & @xmath123 & strong & strong & weak + [ o iii ] 4363 & @xmath124 & weak ? & weak & + [ s ii ] 4069,4076 & @xmath125 , @xmath126 & weak & weak & moderate + [ ne iii ] 3868,3869 & @xmath127 & & moderate & moderate + [ ne
v ] 3346,3426 & @xmath128 & & absent & moderate + c ii ] 2324,2325 & @xmath129 , @xmath130 & & strong & strong + arnett , w. d. 1996 , supernovae and nucleosynthesis , ( princeton : princeton university press ) barbieri , c. , bonoli , c. , & cristiani , s. 1982 , , 114 , 216 barbon , r. , ciatti , f. , & rosino , l. 1979 , , 72 , 287 barbon , r. , cappellaro , e. , & turatto , m. 1989 , , 81 , 421 blair , w. p. , raymond , j. c. , danziger , j. , & matteucci , f. 1989 , , 338 , 812 blair , w. p. , raymond , j. c. , & long , k. s. 1994 , , 334 blinnikov , s. i. , & bartunov , o. s. 1993 , , 273 , 106 burstein , d. , & heiles , c. h. 1982 , , 54 , 33 cappellaro , e. , danziger , i. j. , & turatto , m. 1995 , , 277 , 106 chevalier , r. a. 1982 , , 259 , 302 chevalier , r. a. , & fransson , c. 1994 , , 420 , 268 ( cf94 ) chugai , n. n. 1993 , , 414 , 401 chugai , n. n. , & danziger , i. j. 1994 , , 268 , 173 cowan , j. j. , goss , w. m. , & sramek , r. a. 1991 , , 379 , l49 de vaucouleurs , g. , de vaucouleurs , a. butta , r. , ables , h. d. , & hewitt , a. v. 1981 , , 93 36 eck , c. r. , cowan , j. j. , & branch , d. 1997 , , 191 , 3901 fesen , r. a. 1993 , , 413 , l109 fesen , r. a. 1998 , , 115 , 1107 fesen , r. a. , & becker , r. h. 1990 , , 351 , 437 ( fb90 ) fesen , r. a. , & hurford , a. p. 1996
, , 106 , 563 fesen , r. a. , hurford , a. p. , & matonick , d. m. , 1995 , , 109 , 2608 ( fhm95 ) fesen , r. a. , & matonick , d. m. 1993 , , 07 , 1104 ( fm93 ) fesen , r. a. , & matonick , d. m. 1994 , , 428 , 157 ( fm94 ) filippenko , a. v. in eso worshop and conf .
37 , `` sn 1987a and other supernovae '' eds .
i.j . danziger & k. kjar , ( eso , garching ) p343 .
gaskell , c. m. 1992 , , 389 , l17 hurst , g. m. , & taylor , m. d. 1986 , j. brit .
assoc . , 96 , 102 hyman , s. d. , van dyk , s. d. , weiler , k. w. , & sramek , r. a. 1995 , , 443 , l77 iben , i. , & renzini , a. 1983 , , 21 , 271 immler , s. , pietsch , w. , & aschenbach , b. 1998 , , 331 , 601 kirshner , r. p. , & chevalier , r. a. 1977 , , 218 , 142 leibundgut , b. , kirshner , r. p. , & porter , a. c. 1994 , , 182 , 2910 leibundgut , b. , kirshner , r. p. , pinto , p. a. , rupen , m. p. , smith , r. c. , gunn , j. e. , & schneider , d. p. 1991
, , 372 , 531 lundqvist , p. , & fransson , c. 1988 , , 192 , 221 358 , 344 mendoza , c. 1983 , in iau symposium 103 , planetary nebulae , ed . d.r .
fowler ( dordrecht : reidel ) , 143 montes , m. j. , van dyk , s. d. , weiler , k. w. , sramek , r. a. , & panagia , n. 1997 , , 482 , l61 montes , m. j. , van dyk , s. d. , weiler , k. w. , sramek , r. a. , & panagia , n. 1998 , preprint oke , j. b. , & gunn , j. e. 1983 , , 266 , 713 oke , j. b. , et al .
1995 , , 107 , 375 schmidt , g. d. , weymann , r. j. , & foltz , c. b. 1993 , , 101 , 713 sparks , w. 1997 in `` the extragalactic distance scale '' stsci symposium series # 10 , eds . m. livio , m. donahue , and n. panagia , ( cambridge university press ) p.281 stathakis , r. a. , & sadler , e. m. 1991 , , 250 , 786 sutherland , r. s. , & dopita , m. a. 1995 , , 439 , 381 swartz , d. a. , wheeler , j. c. , & harkness , r. p. 1991
, , 374 , 266 trewhella , m. 1998 , , 297 , 807 turatto , m. cappellaro , e. , danziger , i. j. , benetti , s. , gouiffes , c. , & della valle , m. 1993 , , 262 , 128 uomoto , a. 1991 , , 101 , 1275 uomoto , a. , & kirshner , r. p. 1986
, , 308 , 685 van dyk , s. d. , hamuy , m. , & filippenko , a. v. , 1996 , , 111 , 2017 weiler , k. w. , van dyk , s. d. , peng , c. y. , barth , a. j. , & filippenko , a. v. , 1998a , , submitted van dyk , s. d. , sramek , r. a. , weiler , k. w. , montes , m. j. , & panagia , n. 1998 , asp conference series , 144 , p.357 weaver , t. a. , & woosley , s. e. 1980 , ann .
, 336 , 335 weiler , k. w. , panagia , n. , sramek , r. a. , van der hulst , j. m. , roberts , m. s. , & nguyen , l. 1989 , , 336 , 421 weiler , k. w. , sramek , r. a. , panagia , n. , van der hulst , j. m. , & salvati , m. 1986 , , 301 790 weiler , k. w. , van dyk , s. d. , panagia , n. , sramek , r. a. , & discenna , j. l. 1991 , , 380 , 161 weiler , k. w. , van dyk , s. d. , panagia , n. , & sramek , r. a. 1993 , in massive stars : their lives in the interstellar medium , ed .
cassinelli & e.b .
churchwell , asp conf .
series , 104 , p.436 young , t. r. , & branch , d. 1989 , , 342 , l79 weiler , k. w. , van dyk , s. d. , montes , m. , panagia , n. , & sramek , r. a. 1998 , 500 , 51
|
a low - dispersion keck i spectrum of sn 1980k taken in august 1995 ( t = 14.8 yr after explosion ) and a november 1997 mdm spectrum ( t = 17.0 yr ) show broad 5500 km s@xmath0 emission lines of h@xmath1 , [ o i ] 6300,6364 , and [ o ii ] 7319,7330 .
weaker but similarly broad lines detected include [ fe ii ] 7155 , [ s ii ] 4068,4072 , and a blend of [ fe ii ] lines at 50505400 .
the presence of strong [ s ii ] 4068,4072 emission but a lack of [ s ii ] 6716,6731 emission suggests electron densities of 10@xmath2 @xmath3 . from the 1997 spectra
, we estimate an h@xmath1 flux of @xmath4 erg @xmath5 s@xmath0 indicating a 25% decline from 19871992 levels during the period 1994 to 1997 , possibly related to a reported decrease in its nonthermal radio emission .
a may 1993 mmt spectrum of sn 1979c ( t = 14.0 yr ) shows a somewhat different spectrum from that of sn 1980k .
broad , 6000 km s@xmath0 emission lines are also seen but with weaker h@xmath1 , stronger [ o iii ] 4959,5007 , more highly clumped [ o i ] and [ o ii ] line profiles , no detectable [ fe ii ] 7155 emission , and a faint but very broad emission feature near 5750 .
a 1997 _ hst _ faint object spectrograph , near - uv spectrum ( 2200 4500 ) shows strong lines of c ii ] 2324,2325 , [ o ii ] 2470 , and mg ii 2796,2803 along with weak [ ne iii ] 3969 , [ s ii ] 4068,4072 and [ o iii ] 4363 emissions .
the uv emission lines show a double - peak profile with the blueward peak substantially stronger than the red , suggesting dust extinction within the expanding ejecta ( e[b v ] = @xmath6 mag ) .
the lack of detectable [ o ii ] 3726,3729 emission together with [ o iii ] ( 4959 + 5007)/4363 @xmath7 4 imply electron densities 10@xmath8 @xmath3 .
these new sne ii - l spectra show general agreement with the lines expected in a circumstellar interaction model , but the specific models that are available show several differences with the observations .
high electron densities ( 10@xmath9 @xmath3 ) result in stronger collisional de - excitation than assumed in the models , thereby explaining the absence of several moderate to strong predicted lines such as [ o ii ] 3726,3729 , [ n ii ] 6548,6583 , and [ s ii ] 6716,6731 .
interaction models are needed that are specifically suited to these supernovae .
we review the overall observed range of late - time sne ii - l properties and briefly discuss their properties relative to young , ejecta dominated galactic snrs .
|
star counts in the direction of the andromeda galaxy ( m31 ) have shown that it is embedded in a large , complex stellar halo with significant substructure ( ibata et al .
2001 ; mcconnachie et al .
such halos are thought to be comprised of the debris of shredded satellite galaxies ( searle & zinn 1978 ; newberg et al . 2002 ; mcconnachie et al .
2009 ) , and their existence around luminous spiral galaxies is a central prediction of galaxy formation models ( bullock & johnston 2005 ; abadi , navarro , & steinmetz 2006 ; purcell , bullock , & zentner 2008 ; johnston et al .
2008 ; cooper et al . 2010
; martnez - delgado et al . 2010 ; cooper et al .
2013 ) .
testing these predictions for large samples of galaxies is difficult , as star count studies are limited to relatively small distances ( barker et al .
2009 ; tanaka et al .
a solution to this problem is to study tidal features and stellar halos in integrated light rather than star counts , as the integrated surface brightness is independent of distance cosmological surface brightness dimming is unimportant . ]
( see , e.g. , mihos et al . 2005 ; van dokkum 2005 ; tal et al . 2009 ; martnez - delgado et al . 2010 ; atkinson , abraham , & ferguson 2013 ) .
however , conventional reflecting telescopes can not reliably observe low surface brightness emission below @xmath12magarcsec@xmath1 due to systematic errors in flat fielding and the complex point spread functions of stars ( see slater , harding , & mihos 2009 , for an in - depth discussion of these issues ) .
the dragonfly telephoto array ( abraham & van dokkum 2013 ) is a new class of telescope that is optimized for detecting spatially - extended low surface brightness emission .
the prototype dragonfly telescope consists of eight canon ef 400 mm f/2.8l is ii usm telephoto lenses on a common mount .
telephoto lenses have no central obstruction and are optimally baffled .
furthermore , the specific lenses used in the array are of superb optical quality , partially owing to nano - fabricated sub - wavelength corrugations on their anti - reflecting coatings . as we show in abraham & van dokkum ( 2013 ) the dragonfly point spread function
has an order of magnitude less scattered light than the best reflecting telescopes . by combining eight lenses that image the same area of sky we built a `` compound eye '' that acts as a 40 cm f/1.0 refractor . in its default configuration ,
four of the lenses are equipped with sdss @xmath13-band filters and four with sdss @xmath14 filters .
the pixel scale is @xmath15 and the angular field covered by each camera is @xmath16 . in this _
we present the first results from dragonfly : a study of the stellar halo of the well - known galaxy m101 ( a.k.a . the pinwheel galaxy ) . with a distance of 7mpc ( lee & jang 2012 ) , an absolute magnitude @xmath17 , and a stellar mass of @xmath18(see 4 ) ,
m101 is one of the nearest massive spiral galaxies .
m101 has been the subject of many detailed studies ( e.g. , van der hulst & sancisi 1988 ; kenney , scoville , & wilson 1991 ; kennicutt , bresolin , & garnett 2003 ; mihos et al .
2012 ) , including one of the deepest photometric investigations of the outskirts of galaxies beyond the local group done so far ( mihos et al .
2013 ) .
dragonfly is located at the new mexico skies observatory in mayhill , nm .
it is robotic and operates in a semi - autonomous way .
a total of 35 hours of observations were obtained in 13 nights in may and june 2013 . a typical m101 observing sequence consisted of nine dithered 600s exposures with all eight cameras .
the size of the dither box was typically @xmath19 . as we have four slightly offset cameras per filter this provides 36 independent lines of sight per filter for each 1.5 hr of observing time .
sky flats were taken at the beginning of each night and darks were taken throughout each night .
the data reduction followed standard procedures for imaging data , taking care to preserve the large scale faint structure in the images .
after initial dark subtraction and flat fielding with night - specific calibration frames a low - order illumination correction was applied , created from a large number of dithered observations over many nights .
each frame was also corrected for the 1%
2% gradient in the night sky emission across the field - of - view ( see garstang 1989 ) , by subtracting a tilted plane . combined @xmath13 and @xmath14 images
were created for each night , using optimal weighting .
the 13 images of all nights were combined for each filter , again using optimal weighting .
the reduced , combined @xmath13-band image is shown in fig .
[ full.fig]a .
the background in the images shows large scale variation at a level of @xmath20% ( peak - to - peak ) over the @xmath21 field ( see fig.[full.fig]a ) . as
this background is independent of camera orientation and variations in the dither pattern it is most likely galactic cirrus emission , m image of the m101 field , which shows a broadly similar morphology ( e.g. , zagury , boulanger , & banchet 1999 ) .
] at levels of @xmath22magarcsec@xmath1 .
this large scale background was removed by fitting a third - order polynomial to a background image determined with sextractor , aggressively masking m101 and other objects in the field . in the analysis of the surface brightness profile of m101
the average subtracted background value in a particular radial bin was added in quadrature to the uncertainty in the measured surface brightness in that bin .
stars were removed by modeling their spatially - varying point spread function ( psf ) .
first , large numbers of bright but unsaturated stars were used to construct average psfs in image sections .
these psfs were then interpolated so that a psf can be constructed for any location in the image .
next , the wings of the psf were modeled by averaging saturated stars over the entire image .
care was taken to mask neighboring stars in an iterative way when doing the averaging , both when determining the spatially - dependent inner parts of the psf and when constructing its wings .
the background- and star - subtracted @xmath13-band image is shown in fig .
[ full.fig]b .
the central @xmath23 of the star - subtracted @xmath13-band image , binned to a scale of @xmath24 per pixel , is shown in fig .
[ full.fig]c .
the image shows the outer spiral arms of m101 , including the faint extensions to the east that were first identified by mihos et al .
this emission , which corresponds to spiral structure seen in neutral hydrogen emission ( walter et al .
2008 ; mihos et al .
2012 , 2013 ) , has a surface brightness of @xmath25magarcsec@xmath1
. we do not see coherent faint emission at larger radii .
the surface brightness of m101 falls off rapidly outside of the area defined by the spiral arms , and we see no evidence for an extended stellar halo or features such as m31 s `` giant stream '' ( ibata et al .
we quantify this visual impression with the projected surface brightness profile , shown in fig .
[ sbprof.fig]a .
the profile was determined by averaging the flux in circular annuli at increasing distance from the center of the galaxy .
the profile reaches @xmath0magarcsec@xmath1 at @xmath26kpc , and there is no evidence for an upturn that might have indicated a regime where light from the stellar halo dominates over that of the disk . the @xmath2 color profile is shown in fig .
[ sbprof.fig]b .
the galaxy becomes progressively bluer at larger radii .
= 8.2 cm the grey line in fig .
[ sbprof.fig]a shows the surface brightness profile measured by mihos et al .
( 2013 ) , converted from @xmath27 to @xmath28 using their @xmath29 profile and eq .
23 in fukugita et al .
( 1996 ) . within @xmath30kpc
the datasets agree to @xmath31mag . at @xmath32kpc
there is a discrepancy , which is caused by a difference in methodology : mihos et al .
( 2013 ) determined the median flux in each radial bin ( c. mihos , priv .
comm . ) , whereas we use the mean .
the mean and median are different at radial distances where the ne spiral arm is prominent . the grey line in the bottom panel of fig .
[ sbprof.fig ] shows the @xmath29 profile of mihos et al .
( 2013 ) , converted to @xmath2 ( fukugita et al .
the profiles are offset by @xmath33 , where the error bar reflects the uncertainty in our zeropoint determinations only . assuming a similar uncertainty in the mihos et al .
( 2013 ) zeropoints and/or their conversions to standard filters , the difference is not significant .
we quantify the contribution of the stellar halo to the total mass of m101 by fitting the radial profile of m101 . we first convert the observed surface brightness profile to a radial mass density profile .
this step is important as m101 has a strong color gradient ( fig .
2b ) , and the mass - to - light ( @xmath34 ) ratio correlates with color ( bell & de jong 2001 ) .
we use the following relation between surface brightness and stellar surface density density : @xmath35 with @xmath28 in magarcsec@xmath1 and @xmath36 in @xmath37kpc@xmath1 .
equation [ mlcol.eq ] was determined from the observed relation between rest - frame @xmath2 color and @xmath38 ratio for galaxies with @xmath39 , @xmath40 , and @xmath41 in the sloan digital sky survey dr7 ( as provided by the mpa - jhu release ; brinchmann et al .
the mpa - jhu relation between @xmath38 and @xmath2 has a scatter of 0.12dex and assumes a chabrier ( 2003 ) imf . instead of the observed colors , which have large uncertainties at @xmath42kpc
, we used a fit of the form @xmath43 this fit is indicated by the broken green line in fig .
2b . the mass density profile is shown in fig . [ massdens.fig ] .
the form is similar to the surface brightness profile , except for the central regions as those are more prominent in mass than in blue light .
the data reach surface densities of @xmath44kpc@xmath1 , and a non - parametric limit on the stellar halo of m101 is that it has a stellar mass density lower than this limit at @xmath45kpc .
the profile is well - described by an exponential disk and a sersic ( 1968 ) bulge .
this fit has the form @xmath46,\ ] ] with @xmath47kpc@xmath1 , @xmath48kpc , @xmath49kpc@xmath1 , @xmath50kpc , and @xmath51 .
the fit is shown by the red solid line in fig .
[ massdens.fig]a .
we note that the `` bulge '' may in fact be more appropriately called an inner disk ; as is well known m101 has a very low central velocity dispersion and spiral arms that continue into the central few arcsec ( kormendy et al .
2010 ) .
the residuals from the fit are shown below panel a of fig.[massdens.fig ] .
they are @xmath52 dex at @xmath53kpc and within the @xmath54 error bars at larger radii , confirming that there is no significant upturn in the profile .
we quantify the contribution of a halo component by fitting the residuals . to parameterize the halo we adopt model `` u '' in the courteau et al .
( 2011 ) analysis of the m31 light profile ( their preferred model ) .
this model is a power law : @xmath55^{\alpha}.\ ] ] the values of @xmath56 and @xmath57 are fixed to the best fits for m31 , @xmath58kpc and @xmath59 ( see table 4 of courteau et al .
2011 ) . fitting the normalization ( i.e. , the halo surface density at 30kpc ) gives @xmath60kpc@xmath1 .
the combined disk + bulge + halo model is shown by the red line in fig .
[ massdens.fig]b .
the total mass implied by this model of m101 , integrated to @xmath61kpc , is @xmath62 .
the halo mass is @xmath63 , and we infer that the fraction of mass in the halo is @xmath64 .
this fraction is significantly lower than the halo fraction of m31 : courteau et al .
( 2011 ) find @xmath8 using the same decomposition method .
we illustrate this difference in fig .
[ massdens.fig]c , where we show what m101 s profile would look like if the galaxy had a 4% , m31-like halo .
such halos are clearly inconsistent with the data .
we have measured the surface brightness profile of m101 to @xmath65 disk scale lengths and to surface brightness levels @xmath0magarcsec@xmath1 .
we do not find evidence for the presence of a stellar halo , or more precisely for a photometric component at large radii that can be distinguished from the disk . taking the halo profile of m31 as a model and fitting the normalization
, we find a halo fraction of @xmath66 . this fraction is lower than that of m31 ( @xmath67% ; courteau et al .
2011 ) and also the milky way ( @xmath68% ; carollo et al . 2010 ; courteau et al .
2011 ) . in fig .
[ halofrac.fig ] we show the relation between stellar halo fraction and galaxy stellar mass .
the stellar masses of the milky way and m31 were taken from mcmillan ( 2011 ) and tamm et al .
( 2012 ) respectively . for comparison
, we show the relation between the accreted fraction of stars and galaxy stellar mass as predicted by numerical models that trace the light of accreted satellites in dark matter halos .
this relation was derived from the data in fig .
12b of cooper et al . (
2013 ) for bulge - to - total ratios @xmath69 ; the relation for other @xmath70 limits is very similar .
interestingly , the stellar halo masses of all three galaxies are below the predicted relation , with m101 a factor of @xmath11 below the median expectation .
we caution , however , that stellar halo masses are not measured in a self - consistent way in such comparisons .
there is no universal definition of a stellar halo , and it is unclear whether it even makes sense to model it as a single component ( see , e.g. , carollo et al .
2010 ) . from a practical perspective
it is perhaps most fruitful to test the model predictions by comparing the predicted and observed radial surface density profiles directly , or by fitting a model such as eq .
[ halo.eq ] with only the normalization as a free parameter . in this context
it is interesting to note that the cooper et al .
( 2010 , 2013 ) models predict that stellar halos begin to dominate at @xmath71kpc and surface densities @xmath72kpc@xmath1 again inconsistent with the m101 observations ( but not with m31 ) .
given the stochastic nature of accretion events and , as a consequence , the large scatter predicted in the @xmath73 relation ( purcell et al .
2008 ; cooper et al . 2013 ; fig.4 ) , it is important to build up a sample of galaxies with radial profiles reaching surface densities of @xmath74kpc@xmath1 .
star counts with the gemini telescope ( bland - hawthorn et al .
2005 ) , the subaru telescope ( tanaka et al . 2011 ) and the hubble space telescope ( barker et al . 2009 ; radburn - smith et al .
2011 ; monachesi et al .
2013 ) have reached depths of @xmath22magarcsec@xmath1 but only for very nearby , low mass galaxies .
reaching those limits at distances beyond @xmath75mpc is exceedingly difficult as the apparent brightness of stars decreases with the square of their distance .
by contrast , the integrated - light surface brightness is independent of distance , and low surface brightness - optimized telescopes such as dragonfly can study galaxies out to the virgo cluster and beyond .
this makes it possible to construct complete samples and search for correlations of @xmath9 with other galaxy parameters . as an example , elliptical galaxies and spiral galaxies likely had different accretion histories ( e.g. , guedes et al .
2011 ; tal & van dokkum 2011 ; cooper et al . 2013
; van dokkum et al .
2013 ) .
finally , we note that halos can also be identified by their substructure , as it reflects the detailed accretion history of a galaxy ( see , e.g. , fardal et al .
2008 , for a discussion on m31 ) .
this history is not very informative for an individual galaxy but a large sample can provide very strong constraints on galaxy formation models ( johnston et al .
2008 ) . to illustrate the capabilities of the dragonfly telephoto array in this context
, we created a simulated dragonfly image of m31 by redshifting this galaxy to the distance of m101 ( fig .
[ m31.fig ] ) .
the m31 observations that were used are a combination of a dragonfly image taken on 26 june 2013 and star count data from pandas ( mcconnachie et al .
2009 ; carlberg et al .
the star counts go out very far from the center of m31 but they are incomplete at small radii due to crowding . following a similar procedure to that described in irwin et al .
( 2005 ) dragonfly data at @xmath76 were used to tie the star count data to the integrated - light data .
the combined dragonfly + pandas image was redshifted to 7.0mpc and placed in a relatively empty region of the full - field m101 image ( see fig .
[ full.fig]b ) .
this last step ensures that the noise characteristics and artifacts from the reduction are identical to the m101 data . as the scaling is identical to fig .
[ full.fig]c the actual dragonfly image of m101 can be compared directly to the simulated dragonfly image of m31 .
prominent features in the m31 halo ( see fig . 1 in ferguson et al .
2005 ) are labeled in fig .
[ m31.fig ] .
remarkably , these features are all clearly visible , including the famous giant stream first identified by ibata et al .
it has been known for many decades that dramatic tidal features can be detected in integrated light ; telescopes such as dragonfly are now enabling us to detect the subtle relics of galaxy formation that should be present around every @xmath77 galaxy .
we are grateful to the pandas team for sharing their m31 star count data in digital form , to andrew cooper for providing us with the data to construct the model curves in fig . 4 , and to the staff at new mexico skies for their dedication and support .
the anonymous referee is thanked for excellent comments that substantially improved the manuscript .
we thank the nsf ( grant ast-1312376 ) and nserc for financial support .
|
we use a new telescope concept , the dragonfly telephoto array , to study the low surface brightness outskirts of the spiral galaxy m101 .
the radial surface brightness profile is measured down to @xmath0magarcsec@xmath1 , a depth that approaches the sensitivity of star count studies in the local group .
we convert surface brightness to surface mass density using the radial @xmath2 color profile .
the mass density profile shows no significant upturn at large radius and is well - approximated by a simple bulge + disk model out to @xmath3kpc , corresponding to 18 disk scale lengths . fitting a bulge + disk + halo model we find that the best - fitting halo mass @xmath4 .
the total stellar mass of m101 is @xmath5 , and we infer that the halo mass fraction @xmath6 . this mass fraction is lower than that of the milky way ( @xmath7 ) and m31 ( @xmath8 ) .
all three galaxies fall below the @xmath9 @xmath10 relation predicted by recent cosmological simulations that trace the light of disrupted satellites , with m101 s halo mass a factor of @xmath11 below the median expectation . however , the predicted scatter in this relation is large , and more galaxies are needed to better quantify this possible tension with galaxy formation models .
dragonfly is well suited for this project : as integrated - light surface brightness is independent of distance , large numbers of galaxies can be studied in a uniform way .
|
the thermal hall effect ( the ) is a heat analog of the hall effect , namely , the heat current flows perpendicular to a temperature gradient when the time - reversal symmetry is broken .
recently , the the has been experimentally studied in ferromagnetic metals @xcite , ferromagnetic insulators @xcite , and frustrated magnets in a magnetic field @xcite . in ferromagnetic metals @xcite ,
the the is utilized in order to investigate effects of inelastic scattering on the anomalous hall effect ( ahe ) . according to the wiedemann - franz law , the lorenz ratio @xmath1 coincides with the universal lorenz number @xmath2 at @xmath3 , but decreases as temperature increases and inelastic scattering dominates elastic scattering . here
@xmath4 and @xmath5 are the electric and thermal ( hall ) conductivities , respectively , and @xmath0 is temperature . therefore the lorenz ratio is a quantitative indicator of inelastic scattering . in ferromagnetic insulators @xcite and frustrated magnets @xcite ,
the the is a powerful probe of chargeless spin excitations .
further experimental progress in the the is expected in the future . theoretically , it is difficult to calculate the thermal hall conductivity owing to the fundamental reasons below .
first , the kubo formula is justified by introducing luttinger s gravitational potential since a temperature gradient is not mechanical but statistical force @xcite .
second , the heat magnetization correction to the kubo formula is necessary because the heat current as well as the density matrix is perturbed by a gravitational potential @xcite . once the berry - phase formula has been established @xcite , we can easily calculate the thermal hall conductivity in the clean noninteracting limit .
however , as mentioned above , it is highly desired to calculate the thermal hall conductivity in disordered or interacting systems .
the formula derived in the celebrated paper by smrka and steda , who first pointed out the necessity of the magnetization correction , can be applied to generic disordered systems in principle @xcite , but not easily in practice .
practically , we take into account disorder or interactions by a perturbation theory , e.g. , feynman diagrams . in this context , the ordinary hall effect in disordered metals was studied with feynman diagrams @xcite .
recently , we found a gravitational vector potential by gauging the time translation symmetry and derived the keldysh formalism in curved spacetime @xcite . with this formalism
, we can calculate the kubo formula and the heat magnetization in disordered or interacting systems within a perturbation theory . in this paper ,
we first apply the keldysh formalism in curved spacetime to a three - dimensional disordered weyl ferromagnet , which is known to exhibit the ahe .
the main purpose is to show the wiedemann - franz law by a practical perturbation theory with respect to the impurity strength .
we take nonmagnetic impurities into account within the self - consistent @xmath0-matrix approximation and calculate @xmath4 and @xmath6 on an equal footing .
we begin with brief reviews of gauging the time translation symmetry and the keldysh formalism in curved spacetime .
the details of calculation are given in appendices .
this paper serves as the first step towards a unified theory of the ahe at finite temperature , in which we should take into account both disorder and interactions . before going to the main part ,
let us review some theoretical aspects of the ahe @xcite .
the ahe occurs in ferromagnets with spin - orbit coupling . in the first microscopic theory by karplus and luttinger @xcite ,
the ahe was attributed to the anomalous velocity , which is now understood as the berry curvature @xcite .
since this mechanism works in the absence of disorder and depends only on the band structure , leading to @xmath7 , it is called the intrinsic mechanism . soon later , it was revealed that the skew scattering , i.e. , the antisymmetric part of the transition probability due to disorder , also gives rise to the ahe @xcite .
an important consequence of the skew - scattering mechanism is @xmath8 .
another mechanism due to disorder was proposed by berger @xcite .
this side - jump mechanism results in @xmath7 , which is the same as the intrinsic mechanism .
the most systematic way to deal with all these mechanisms on an equal footing is based on the keldysh formalism @xcite .
one important advantage of this formalism is that we can go beyond a finite - order perturbation theory with respect to the impurity strength with preserving the gauge invariance . by using this formalism , some authors investigated a two - dimensional disordered rashba ferromagnet within the self - consistent @xmath0-matrix approximation and found the crossover between the superclean and moderately dirty regimes , where the skew - scattering and intrinsic contributions are dominant , respectively @xcite .
we also note that the nontrivial scaling relation @xmath9 was found in the dirty regime @xcite .
let us also review three - dimensional weyl semimetals @xcite which we investigate in this paper .
weyl semimetals have pairs of weyl nodes at which two bands touch each other .
typically , weyl semimetals are realized from dirac semimetals by breaking the time - reversal @xcite or inversion symmetry @xcite .
in particular , weyl semimetals without the time - reversal symmetry exhibit the ahe @xcite . when the fermi level is sufficiently close to weyl nodes , the ahe is purely intrinsic @xcite , which results from chiral anomaly @xcite . on the other hand ,
when the fermi level is away from weyl nodes , the extrinsic mechanisms due to disorder can be relevant . in this paper , we mean such doped weyl semimetals without the time - reversal symmetry by weyl ferromagnets because we are interested in both the extrinsic and intrinsic mechanisms of the ahe and the .
although the the in weyl ferromagnets was studied with the semiclassical boltzmann theory @xcite , we go beyond the relaxation time approximation used in the literature .
hereafter we assign the latin ( @xmath10 ) and greek ( @xmath11 ) alphabets to locally flat and global coordinates , respectively .
we follow the einstein convention , which implies summation over the spacetime dimension @xmath12 when an index appears twice in a single term .
the minkowski metric is taken as @xmath13 .
the planck constant and the charge are denoted by @xmath14 and @xmath15 , respectively , and the speed of light and the boltzmann constant are put to @xmath16 . the upper or lower signs in equations correspond to boson or fermion .
first , we review the idea of gravitational vector potential by gauging the time translation symmetry @xcite .
it is well known that electromagnetic scalar and vector potentials are introduced by gauging the u(@xmath17 ) symmetry . by gauging the time translation symmetry as well as the u(@xmath17 ) symmetry
, we can calculate the electric and thermal ( hall ) conductivities on an equal footing .
furthermore , the heat magnetization is defined by the magnetic component of the field strength induced from a gravitational vector potential . according to noether
s first theorem , a global continuous symmetry is associated with the conservation law of a current . in the presence of the global time translation symmetry , namely , when the matter action is invariant under global time translation @xmath18 , @xmath19 , and @xmath20 , the conserved energy current density @xmath21 is constructed as @xmath22 with @xmath23 $ ] being the matter lagrangian density in flat spacetime .
we impose the local time translation symmetry by following the gauge principle .
indeed , the local symmetries of time translation , space translation , and lorentz transformations are imposed by the general covariance principle . under local time translation @xmath24 , @xmath19 , and @xmath20 ,
the matter action is no longer invariant because @xmath25 transforms as @xmath26 . to preserve the local time translation symmetry of the total action
, we introduce a gauge field called vielbein @xmath27 , which transforms as @xmath28 , by replacing @xmath29 with the covariant derivative @xmath30 . here
@xmath31 is the inverse of @xmath32 .
we also introduce a u(@xmath17 ) gauge field @xmath33 by gauging the u(@xmath17 ) symmetry . according to noether s second theorem ,
the energy current density is also expressed by @xmath34 note that @xmath35 $ ] is the matter lagrangian density in curved spacetime and contains the determinant of vielbein @xmath36 , which comes from the covariant volume element @xmath37 . from eq . , @xmath38 is coupled to the hamiltonian density and is identified as luttinger s gravitational potential which describes nonuniform temperature @xcite .
@xmath39 is coupled to the energy current density and is called gravitational vector potential @xcite .
vielbein induces a field strength called torsion .
in analogy to electromagnetic fields induced from a u(@xmath17 ) gauge field , we define torsional electromagnetic fields by @xmath40 [ eq : torsion1 ] a torsional electric field describes a temperature gradient , and a torsional magnetic field is conjugate to the heat magnetization introduced above .
thus , the kubo formula for the thermal ( hall ) conductivity and the heat magnetization are calculated by @xcite @xmath41 [ eq : def ] in which @xmath42 is the grand potential .
note that eq .
is defined by @xmath43 but not by @xmath44 , because @xmath45 is invariant under local time translation @xmath24 and @xmath19 .
this is obvious since @xmath31 and @xmath44 transform as @xmath46 and @xmath47 , respectively .
nonetheless , we explicitly show the invariance of @xmath45 in appendix [ app : inv ] .
we also note that the spatial component of vielbein @xmath48 is introduced by gauging the space translation symmetry , and the current responses to torsion @xmath49 were discussed in refs . .
next , we review the keldysh formalism in curved spacetime @xcite , which is a natural extension of that in electromagnetic fields @xcite . the keldysh formalism is a powerful quantum - mechanical tool to treat nonlinear and nonequilibrium phenomena @xcite . in the wigner representation in terms of the center - of - mass
coordinate @xmath50 and the relative canonical momentum @xmath51 , convolution in the dyson equation is represented by the moyal product . in the presence of electromagnetic fields and torsion ,
the keldysh green s function depends on @xmath50 through the mechanical momentum @xmath52 $ ] . by changing the variables @xmath53 ,
the moyal product is replaced by the star product which contains electromagnetic fields @xmath54 @xcite and torsion @xmath45 @xcite .
thus we can develop a systematic perturbation theory with respect to electromagnetic fields and torsion .
we start from the keldysh formalism in flat spacetime @xcite .
the keldysh green s function @xmath55 is defined on the closed time path and satisfies the dyson equation . in terms of real time , it is expressed in the matrix form by @xmath56 in which @xmath57 , @xmath58 , and @xmath59 are the retarded , advanced , and lesser green s functions , @xmath60_{\mp } \right\rangle , \label{eq : green2a } \\
g^{\rm a}(x_1 , x_2 ) \equiv & + i/\hbar \theta(t_2 - t_1 ) \left\langle [ \psi(x_1 ) , \psi^{\dag}(x_2)]_{\mp } \right\rangle , \label{eq : green2b } \\ g^<(x_1 , x_2 ) \equiv & \mp i/\hbar \left\langle \psi^{\dag}(x_2 ) \psi(x_1 ) \right\rangle , \label{eq : green2c}\end{aligned}\ ] ] [ eq : green2 ] respectively . the dyson equation is expressed by @xmath61 where @xmath62 and @xmath63 are the lagrangian density and the self - energy , respectively , and @xmath64 indicates convolution . in the presence of vielbein ,
the volume element @xmath65 is replaced by the covariant volume element @xmath37 .
correspondingly , the dyson equation is modified as @xcite @xmath66 in which the modified convolution and @xmath67 function are given by @xmath68 [ eq : det1 ] respectively .
although eq . looks complicated , by introducing a weight - one tensor density @xmath69 from a tensor @xmath70 , it is symbolically reduced to the dyson equation in flat spacetime as @xmath71 do not confuse a tensor density with a vector in three dimensions , which is indicated by an arrow below . in order to deal with convolution , we move to the wigner representation @xcite @xmath72 in terms of the center - of - mass coordinate @xmath73 and the relative canonical momentum @xmath51 conjugate to @xmath74 . in this representation , convolution is represented by the noncommutative moyal product as @xmath75 .
@xmath76 is the poisson bracket in flat spacetime known in analytical mechanics .
note that the partial derivative on the left ( right ) side of @xmath77 acts on the left ( right ) only .
see appendix [ app : moyal ] for derivation . as a result , the dyson equation eq .
is simply represented by @xmath78 now we take into account electromagnetic fields @xcite and torsion @xcite .
these gauge fields are introduced through the covariant derivative @xmath79 , which corresponds to the so - called peierls substitution @xmath52 $ ] in the wigner representation .
thus the dyson equation in curved spacetime is given by @xmath80 however , since @xmath81 is a complicated function of @xmath50 , it is better to change variables @xmath53 , which leads to replacing the moyal product @xmath82 by the star product @xmath83 . by using the chain rule of partial derivatives , the poisson bracket in curved spacetime
is given by @xmath84 and the dyson equation in curved spacetime eq .
is modified as @xmath85 see appendix [ app : moyal ] for calculation details .
although we do not derive the exact form of the star product , eqs . and
complete the first - order perturbation theory of green s functions with respect to static uniform electromagnetic fields and torsion .
this is the final section for reviews . in this section ,
we derive the first - order perturbation theory with respect to static uniform electromagnetic fields @xcite and torsion @xcite .
we can neglect the @xmath50 dependence in the green s functions and the first term in eq . .
we also expand the keldysh green s function and the self - energy with respect to @xmath54 and @xmath86 as @xmath87 \partial_{\pi_a } \otimes \partial_{\pi_b}/2 , \label{eq : gem1a } \\
{ { \hat { \bm g } } } = & { \hat g}_0 + \hbar f_{ab } { \hat g}_{f_{ab}}/2 + \hbar ( -t^{\hat 0}_{\phantom{\hat 0 } ab } ) { \hat g}_{-t^{\hat 0}_{\phantom{\hat 0 } ab}}/2 , \label{eq : gem1b } \\ { { \hat { \bm \sigma } } } = & { \hat \sigma}_0 + \hbar f_{ab } { \hat \sigma}_{f_{ab}}/2 + \hbar ( -t^{\hat 0}_{\phantom{\hat 0 } ab } ) { \hat \sigma}_{-t^{\hat 0}_{\phantom{\hat 0 } ab}}/2 .
\label{eq : gem1c}\end{aligned}\ ] ] [ eq : gem1 ] by substituting these equations into eq . , we obtain @xmath88 and @xmath89/2 i , \label{eq : gem2a } \\ { \hat g}_{-t^{\hat 0}_{\phantom{\hat 0 } ab } } = & { \hat g}_0 { \hat \sigma}_{-t^{\hat 0}_{\phantom{\hat 0 } ab } } { \hat g}_0 - ( -\pi_{\hat 0 } ) [ { \hat g}_0 \partial_{\pi_a } { \hat g}_0^{-1 } { \hat g}_0 \partial_{\pi_b } { \hat g}_0^{-1 }
{ \hat g}_0 - ( c \leftrightarrow d)]/2 i. \label{eq : gem2b}\end{aligned}\ ] ] [ eq : gem2 ] obviously , the only difference in eq . is the charge ; @xmath90 in eq . and @xmath91 in eq .
below , we explicitly write down the results for electromagnetic fields @xcite only , because those for torsion @xcite can be obtained by replacing @xmath92 . in order to calculate expectation values
, we have to extract the lesser green s function . in the absence of electromagnetic fields or torsion , namely , in equilibrium , we know @xmath93 and @xmath94 , where @xmath95 is the distribution function . by introducing @xmath96 [ eq : gem3 ]
we obtain @xmath97/2 i , \label{eq : gem4a } \\
g_{f_{ab}}^ { < ( 0 ) } = & \pm ( g_{f_{ab}}^{\rm r } - g_{f_{ab}}^{\rm a } ) , \label{eq : gem4b } \\ \sigma_{f_{ab}}^ { < ( 0 ) } = & \pm ( \sigma_{f_{ab}}^{\rm r } - \sigma_{f_{ab}}^{\rm a } ) , \label{eq : gem4c } \\ g_{f_{{\hat \jmath } { \hat 0}}}^ { < ( 1 ) } = & g_0^{\rm r } \sigma_{f_{{\hat \jmath } { \hat 0}}}^ { < ( 1 ) } g_0^{\rm a } -q [ g_0^{\rm r } \partial_{\pi_{\hat \jmath } } g_0^{{\rm r } -1 } ( g_0^{\rm r } - g_0^{\rm a } ) - ( g_0^{\rm r } - g_0^{\rm a } ) \partial_{\pi_{\hat \jmath } } g_0^{{\rm a } -1 } g_0^{\rm a}]/2 i. \label{eq : gem4d}\end{aligned}\ ] ] [ eq : gem4 ] the electric component of the self - energy @xmath98 is determined self - consistently together with @xmath99 , and the magnetic component is found to satisfy @xmath100 .
this is because systems remain in equilibrium even in the presence of a static uniform magnetic field .
first , we derive the greens - function formula for the heat magnetization @xcite .
although we define the heat magnetization with the grand potential in eq . , we can also define the auxiliary heat magnetization with the total energy @xmath101 . in the wigner representation
, the total energy is represented by @xmath102^ < + { { \rm c.c.}}\label{eq : en3}\ ] ] the auxiliary heat magnetization is defined and calculated as @xmath103 [ eq : auxhm1 ] which is obtained by using eq . and the cyclic property of trace .
the proper and auxiliary heat magnetizations are related as @xcite @xmath104 second , we write down the kubo formula for the electric and thermal ( hall ) conductivities @xcite .
the charge and heat currents are represented by @xmath105^ < + { { \rm c.c . } } , \label{eq : j1a } \\ j_{\rm q}^{\hat \imath } = & \pm \frac{i \hbar}{2 } \int \frac{{\rm d}^d \pi}{(2 \pi \hbar)^d } \operatorname{tr}[(-\partial_{\pi_{\hat \imath } } { { \hat { \bm g}}}^{-1 } ) \star { { \hat { \bm g } } } \star ( -\pi_{\hat 0})]^ < + { { \rm c.c . } } , \label{eq : j1b}\end{aligned}\ ] ] [ eq : j1 ] in which @xmath106 is the renormalized velocity to guarantee the ward identity . the electric ( hall )
conductivity is defined and calculated as @xmath107 [ eq : sigma1 ] these terms are called `` int ii '' , `` int i '' , and `` ext '' , respectively .
obviously , the fermi - sea term denoted by `` int ii '' appears only in the hall conductivity .
the kubo formula for the thermal ( hall ) conductivity is essentially the same as that for the electric ( hall ) conductivity and is given by @xmath108 [ eq : kappa1 ] note that these expressions were already obtained by the keldysh formalism in electromagnetic and gravitational fields , and the electric and thermal transport properties in a disordered interacting fermi liquid were deeply studied @xcite . as introduced above
, the heat magnetization is necessary for calculating the proper thermal hall conductivity , whose expression was first obtained by gauging the time translation symmetry and developing the keldysh formalism in curved spacetime @xcite .
equations , , and can be applied to disordered or interacting systems , because these are obtained by a perturbation theory with respect to field strengths but not with respect to the impurity or interaction strength . below
we restrict ourselves to the self - consistent @xmath0-matrix approximation to deal with nonmagnetic impurities .
for @xmath67-function nonmagnetic impurities , the @xmath0 matrix is defined by @xmath109 and is expressed by @xmath110^{-1}. \label{eq : tma1}\ ] ] note that the @xmath0 matrix is defined by @xmath111 for generic disordered systems . in terms of feynman diagrams ,
the self - consistent @xmath0-matrix approximation is expressed in fig .
[ fig : diag ] .
the unperturbed and perturbed @xmath0-matrices are given by @xcite @xmath112^{-1 } , \label{eq : tma2a } \\
t_{f_{cd}}^{\rm r}(-\pi_{\hat 0 } ) = & v_{\rm i } t_0^{\rm r}(-\pi_{\hat 0 } ) \int \frac{{\rm d}^d \pi}{(2 \pi \hbar)^d } g_{f_{cd}}^{\rm r } t_0^{\rm r}(-\pi_{\hat 0 } ) , \label{eq : tma2b } \\ t_{f_{j0}}^ { < ( 1)}(-\pi_{\hat 0 } ) = & v_{\rm i } t_0^{\rm r}(-\pi_{\hat 0 } ) \int \frac{{\rm d}^d \pi}{(2 \pi \hbar)^d }
g_{f_{j0}}^ { < ( 1 ) } t_0^{\rm a}(-\pi_{\hat 0 } ) . \label{eq : tma2c}\end{aligned}\ ] ] [ eq : tma2 ]
now that we have prepared all the greens - function formulas to calculate the electric and thermal ( hall ) conductivities , let us apply them to a disordered weyl ferromagnet .
we study the linearized weyl hamiltonian without the time - reversal symmetry , @xmath113 in which @xmath114 , @xmath115 , @xmath116 , and @xmath117 indicate the fermi velocity , the zeeman interaction , chirality , and the pauli matrices for spin , respectively . the distance between weyl nodes in the wave number space
is denoted by @xmath118 , and the intrinsic contribution is given by @xmath119 @xcite .
we deal with @xmath67-function nonmagnetic impurities within the self - consistent @xmath0-matrix approximation .
calculation is summarized as follows . for simplicity , we omit the hat symbols for the locally flat coordinates and write @xmath120 . we expand the green s function and the self - energy with respect to the pauli matrices , i.e. , @xmath121 .
all the momentum integrals can be carried out analytically for each value of the energy @xmath122 , and the self - consistent calculation of the self - energy and the energy integrals are done numerically .
we introduce two cutoffs for the momentum integrals ; @xmath123 and @xmath124 . here
we emphasize that the intrinsic contribution strongly depends on ultraviolet regularizations @xcite .
in particular , if we take the limit of @xmath125 first , the intrinsic contribution for a massive dirac ferromagnet with mass @xmath126 is given by @xmath127 @xcite , although the system is not a weyl semimetal but a trivial insulator . in our case ,
the intrinsic contribution consists of the topological term @xmath128 and this additional term due to the incorrect regularization and hence vanishes .
we should keep @xmath129 to obtain the proper intrinsic contribution .
the chemical potential is determined self - consistently to fix the particle number without impurities at @xmath3 . as a representative
, we calculate the unperturbed retarded self - energy @xmath130 .
see appendices [ app : gbr]-[app : sigma ] for the first - order self - energies with respect to electromagnetic fields and the physical quantities .
we introduce @xmath131 and @xmath132 and obtain @xmath133 \cdot { \vec \sigma}_{\perp } - [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \sigma_z , \label{eq : g0r1a } \\ g_0^{\rm r}(\xi , { \vec p } ) = & \gamma_0^{\rm r}(\xi , { \vec p } ) \{z^{\rm r}(\xi ) + [ v \rho_x { \vec p}_{\perp } + { \vec \sigma}_{0 \perp}^{\rm r}(\xi ) ] \cdot { \vec \sigma}_{\perp } + [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \sigma_z\ } , \label{eq : g0r1b } \\ \gamma_0^{\rm r}(\xi , { \vec p } ) \equiv & \det g_0^{{\rm r } -1}(\xi , { \vec p } ) = \{z^{{\rm r } 2}(\xi ) - [ v \rho_x { \vec p}_{\perp } + { \vec \sigma}_{0 \perp}^{\rm r}(\xi)]^2 - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}^{-1}. \label{eq : g0r1c}\end{aligned}\ ] ] [ eq : g0r1 ] in order to calculate the self - consistent @xmath0 matrix , we define @xmath134 by the momentum integral of @xmath135 .
explicitly , it is given by @xmath136 , \label{eq : g0r2c } \\ g_{0z}^{\rm r}(\xi ) = & \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{\rm r}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : g0r2d}\end{aligned}\ ] ] [ eq : g0r2 ] from which we obtain @xmath137^{-1 } = \tau_0^{\rm r}(\xi ) \{[1 - v_{\rm i } g_{00}^{\rm r}(\xi ) ] + v_{\rm i } { \vec g}_0^{\rm r}(\xi ) \cdot { \vec \sigma}\ } , \label{eq : g0r3a } \\ \tau_0^{\rm r}(\xi ) \equiv & \det t_0^{{\rm r } -1}(\xi ) = \{[1 - v_{\rm i } g_{00}^{\rm r}(\xi)]^2 - [ v_{\rm i } { \vec g}_0^{\rm r}(\xi)]^2\}^{-1}. \label{eq : g0r3b}\end{aligned}\ ] ] [ eq : g0r3 ] if we put @xmath138 , @xmath139 is even with respect to @xmath140 , which leads to @xmath141 in eq . and
@xmath142 in eq .
. therefore we obtain @xmath138 and @xmath143 \sigma_z , \label{eq : g0r4a } \\ g_0^{\rm r}(\xi , { \vec p } ) = & \gamma_0^{\rm r}(\xi , { \vec p } ) \{z^{\rm r}(\xi ) + v \rho_x { \vec p}_{\perp } \cdot { \vec \sigma}_{\perp } + [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \sigma_z\ } , \label{eq : g0r4b } \\ \gamma_0^{\rm r}(\xi , { \vec p } ) = & \{z^{{\rm r } 2}(\xi ) - ( v { \vec p}_{\perp})^2 - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}^{-1}. \label{eq : g0r4c}\end{aligned}\ ] ] [ eq : g0r4 ] here we define the integrals by @xmath144^m}{\{z^{{\rm r } 2}(\xi ) - x - [ y + \eta^{\rm r}(\xi)]^2\}^n } , \label{eq : irmn}\ ] ] with @xmath145 and @xmath146 .
their explicit forms are given in appendix [ app : int ] . by introducing @xmath147 , the self - consistent equations which we solve
are given by @xmath148 ^ 2 - [ ( v_{\rm i}/8 \pi^2 \hbar^3 v^3 ) { \tilde g}_{0z}^{\rm r}(\xi)]^2\}^{-1 } , \label{eq : g0r5c } \\ t_{00}^{\rm r}(\xi ) = & \tau_0^{\rm r}(\xi ) [ 1 - ( v_{\rm i}/8 \pi^2 \hbar^3 v^3 ) { \tilde g}_{00}^{\rm r}(\xi ) ] , & t_{0z}^{\rm r}(\xi ) = & \tau_0^{\rm r}(\xi ) ( v_{\rm i}/8 \pi^2 \hbar^3 v^3 ) { \tilde g}_{0z}^{\rm r}(\xi ) , \label{eq : g0r5d } \\
\sigma_{00}^{\rm r}(\xi ) = & n_{\rm i } v_{\rm i } t_{00}^{\rm r}(\xi ) , & \sigma_{0z}^{\rm r}(\xi ) = & n_{\rm i } v_{\rm i } t_{0z}^{\rm r}(\xi ) . \label{eq : g0r5e}\end{aligned}\ ] ] [ eq : g0r5 ] the particle number to be fixed is calculated by @xmath149 \notag \\ = & \frac{1}{4 \pi^3 \hbar^3 v^3 } \int { \rm d } \xi
f(\xi ) [ i { \tilde g}_{00}^{\rm r}(\xi ) + { { \rm c.c . } } ] .
\label{eq : g0r6}\end{aligned}\ ] ] first , we show the @xmath150 dependences of the electric and thermal resistivities and the lorenz ratio in fig .
[ fig : yy_nivi2 ] .
the momentum cutoffs are given by @xmath151 and @xmath152 , and the energy interval @xmath153 is divided into @xmath154 subintervals .
we use @xmath155 , @xmath156 , and @xmath157 for fig .
[ fig : yy_nivi2 ] . at the lowest temperature ,
@xmath158 remains unity , and the wiedemann - franz law holds for each component .
next , we show the @xmath150 dependences of the electric and thermal hall conductivities and the hall lorenz ratio in fig . [
fig : xy_nivi2 ] . note that @xmath159 is defined by sum of @xmath160 and @xmath161 .
the `` ext '' can be identified as the skew - scattering contribution since it is enhanced as @xmath150 decreases .
the `` int ii '' remains constant as a function of @xmath150 , which is specific for the linearized weyl hamiltonian .
in fact , the intrinsic contribution is known to remain constant by changing @xmath162 or @xmath0 @xcite , because the density of berry curvature is finite only at @xmath163 . as a result
, we do not find the nontrivial scaling relation @xmath9 which was found in a two - dimensional disordered rashba ferromagnet in the dirty regime @xcite .
we also note that @xmath164 and @xmath165 are less than unity owing to the finite momentum cutoffs . at the lowest temperature ,
@xmath166 remains unity , which is consistent with the wiedemann - franz law .
thus , the keldysh formalism in curved spacetime provides a practical perturbation theory for calculating the electric and thermal ( hall ) conductivities on an equal footing .
we also show the @xmath0 dependences of the electric and thermal resistivities and the lorenz ratio in fig .
[ fig : yy_tt ] and those of the electric and thermal hall conductivities and the hall lorenz ratio in fig .
[ fig : xy_tt ] .
we use @xmath155 , @xmath156 , and @xmath167 .
the lorenz ratio @xmath158 remains unity as widely believed in the absence of inelastic scattering , while the hall lorenz ratio @xmath166 does not .
below , we discuss this temperature dependence more seriously .
here we discuss the temperature dependences of the fermi - surface terms denoted by `` int i '' and `` ext '' . to do this , we use the sommerfeld expansion @xmath168 g(\epsilon ) \simeq & g(\mu ) + ( 1/6 ) ( \pi t)^2 g^{(2)}(\mu ) \notag \\ & + ( 7/360 ) ( \pi t)^4 g^{(4)}(\mu ) , \label{eq : sommer1}\end{aligned}\ ] ] for an infinitely differentiable function @xmath169 .
the fermi - surface terms @xmath170 for the electric and thermal ( hall ) conductivities can be approximated by @xmath171 s^{ij}_{\alpha}(\epsilon ) \notag \\
\simeq & s^{ij}_{\alpha}(\mu ) + ( 1/6 ) ( \pi t)^2 s^{ij(2)}_{\alpha}(\mu ) , \label{eq : fermi - surf1a } \\ \kappa^{ij}_{\alpha}/l_0 t = & \frac{3}{(\pi t)^2 } \int { \rm d } \epsilon [ -f^{\prime}(\epsilon - \mu ) ] ( \epsilon - \mu)^2 s^{ij}_{\alpha}(\epsilon ) \notag \\
\simeq & s^{ij}_{\alpha}(\mu ) + ( 7/10 ) ( \pi t)^2 s^{ij ( 2)}_{\alpha}(\mu ) , \label{eq : fermi - surf1b } \\ l^{ij}_{\alpha}/l_0 = & \kappa^{ij}_{\alpha}/l_0 t \sigma^{ij}_{\alpha } \notag \\
\simeq & 1 + ( 8/15 ) ( \pi t)^2 s^{ij ( 2)}_{\alpha}(\mu)/s^{ij}_{\alpha}(\mu ) .
\label{eq : fermi - surf1c}\end{aligned}\ ] ] [ eq : fermi - surf1 ] at finite temperature , the ( hall ) lorenz ratio @xmath172 can be more or less than unity depending on @xmath173 .
however , when the chemical potential is fixed , the convexities of @xmath174 and @xmath175 depend only on the sign of @xmath176 and hence coincide . as seen in fig .
[ fig : xy_tt ] , @xmath177 is convex upward , while @xmath178 is convex downward .
the temperature dependence of the skew - scattering contribution can not be explained as it is . in our calculation , the chemical potential depends on temperature so as to fix the particle number .
again by using eq . , we obtain @xmath179 with @xmath180 being the density of states .
we employ the taylor expansion in eq . to obtain @xmath181 , \label{eq : fermi - surf2a } \\
\kappa^{ij}_{\alpha}/l_0 t \simeq & s^{ij}_{\alpha}(\mu_0 ) + ( 1/6 ) ( \pi t)^2 \notag \\ & \times \left[\frac{21}{5 } s^{ij ( 2)}_{\alpha}(\mu_0 ) - s^{ij \prime}_{\alpha}(\mu_0 ) \frac{d^{\prime}(\mu_0)}{d(\mu_0)}\right ] , \label{eq : fermi - surf2b } \\ l^{ij}_{\alpha}/l_0 \simeq & 1 + ( 8/15 ) ( \pi t)^2 s^{ij ( 2)}_{\alpha}(\mu_0)/s^{ij}_{\alpha}(\mu_0 ) .
\label{eq : fermi - surf2c}\end{aligned}\ ] ] [ eq : fermi - surf2 ] thus , the ( hall ) lorenz ratio @xmath172 can be more or less than unity depending on @xmath182 , and the convexities of @xmath174 and @xmath175 do not always coincide . in particular , for @xmath183/5 $ ]
, @xmath177 is convex upward , while @xmath178 is convex downward .
we show the energy dependences of @xmath184 and @xmath185 in fig .
[ fig : dos ] .
we find @xmath186 and @xmath187 with @xmath188 , which satisfy the above inequalities .
the lorenz ratio @xmath189 remains unity because @xmath190 is almost linear , and hence @xmath191 is negligible .
in realistic systems , we do not expect the above temperature dependences because of inelastic scattering .
experimentally , the hall lorenz ratio in ni and co is almost constant while that in fe rapidly decreases @xcite .
since the intrinsic contribution is dominant in ni and co while the skew - scattering contribution is dominant in fe @xcite , it is concluded that the intrinsic mechanism is robust against inelastic scattering , but not the extrinsic mechanisms .
this conclusion was also supported experimentally by the ( hall ) resistivity measurements @xcite and theoretically @xcite .
it is an important future problem to calculate the electric and thermal ( hall ) conductivities by taking into account both disorder and interactions . to summarize , we investigated the transport properties for a doped weyl ferromagnet with nonmagnetic impurities which exhibits the ahe .
we used the keldysh formalism in curved spacetime which recently we developed @xcite and calculated the electric and thermal ( hall ) conductivities on an equal footing within the self - consistent @xmath0-matrix approximation .
we successfully reproduced the wiedemann - franz law at low temperature and found that the lorenz ratio of the skew - scattering contribution @xmath192 deviates from unity as temperature increases , which comes from the higher - order sommerfeld expansion .
note that this temperature dependnece is not expected in the presence of inelastic scattering caused by phonons and magnons .
our results can be understood by the formula derived by smrka and steda , which works at any temperature as far as elastic scattering is concerned @xcite .
however , by using our practical perturbation theory , we can go beyond smrka and steda and deal with both disorder and interactions .
we discussed our numerical calculation with g. tatara and k. yamamoto and learned the cutoff dependence of the intrinsic contribution from j. fujimoto .
we thank n. nagaosa for reading our manuscript and giving us comments .
here we explicitly show the invariance of @xmath45 under local time translation @xmath24 and @xmath19 . by using @xmath193 [ eq : transf1 ] vielbein , its inverse , and torsion
transform as @xmath194 [ eq : transf2 ] @xmath195 [ eq : transf3 ] and @xmath196 [ ( 1 + \partial_0 \epsilon^0)^{-1 } e^{\hat 0}_{\phantom{\hat 0 } 0 } ] - ( 1 + \partial_0 \epsilon^0)^{-1 } \partial_0 [ e^{\hat 0}_{\phantom{\hat 0 } j } - ( 1 + \partial_0 \epsilon^0)^{-1 } \partial_j \epsilon^0 e^{\hat 0}_{\phantom{\hat 0 } 0 } ] \notag \\ = & ( 1 + \partial_0 \epsilon^0)^{-1 }
t^{\hat 0}_{\phantom{\hat 0 } j0 } , \label{eq : transf4a } \\ t^{\prime { \hat 0}}_{\phantom{\prime { \hat 0 } } ij}(x^{\prime } ) = & \partial_i^{\prime } e^{\prime { \hat 0}}_{\phantom{\prime { \hat 0 } } j}(x^{\prime } ) - \partial_j^{\prime } e^{\prime { \hat 0}}_{\phantom{\prime { \hat 0 } } i}(x^{\prime } ) \notag \\
= & [ \partial_i - ( 1 + \partial_0 \epsilon^0)^{-1 } \partial_i \epsilon^0 \partial_0 ] [ e^{\hat 0}_{\phantom{\hat 0 } j } - ( 1 + \partial_0 \epsilon^0)^{-1 } \partial_j \epsilon^0 e^{\hat 0}_{\phantom{\hat 0 } 0 } ] - ( i \leftrightarrow j ) \notag \\ = & t^{\hat 0}_{\phantom{\hat 0 } ij } + ( 1 + \partial_0 \epsilon^0)^{-1 } ( \partial_i \epsilon^0 t^{\hat 0}_{\phantom{\hat 0 } j0 } - \partial_j \epsilon^0 t^{\hat 0}_{\phantom{\hat 0 } i0 } ) , \label{eq : transf4b}\end{aligned}\ ] ] [ eq : transf4 ] respectively .
thus , @xmath44 is not invariant under local time translation . on the other hand , @xmath45 is invariant as @xmath197 ( 1 + \partial_0 \epsilon^0)^{-1 } t^{\hat 0}_{\phantom{\hat 0 } k0 } + [ ( 1 + \partial_0 \epsilon^0 ) e_a^{\phantom{a } 0 } + \partial_k \epsilon^0 e_a^{\phantom{a } k } ] e_b^{\phantom{b } l } ( 1 + \partial_0 \epsilon^0)^{-1 } t^{\hat 0}_{\phantom{\hat 0 } 0l } \notag \\ & + e_a^{\phantom{a } k } e_b^{\phantom{b } l } [ t^{\hat 0}_{\phantom{\hat 0 } kl } + ( 1 + \partial_0 \epsilon^0)^{-1 } ( \partial_k \epsilon^0 t^{\hat 0}_{\phantom{\hat 0 } l0 } - \partial_l \epsilon^0 t^{\hat 0}_{\phantom{\hat 0 } k0 } ) ] \notag \\ = & e_a^{\phantom{a } k } e_b^{\phantom{b } 0 } t^{\hat 0}_{\phantom{\hat 0 } k0 } + e_a^{\phantom{a } 0 } e_b^{\phantom{b } l } t^{\hat 0}_{\phantom{\hat 0 } 0l } + e_a^{\phantom{a } k } e_b^{\phantom{b } l } t^{\hat 0}_{\phantom{\hat 0 } kl } = t^{\hat 0}_{\phantom{\hat 0 } ab}. \label{eq : transf5}\end{aligned}\ ] ] we also note that eq
. indicates the exact gauge transformation of a gravitational ( vector ) potential @xmath198 , which is different from the u(@xmath17 ) gauge transformation .
however , when @xmath199 and @xmath200 , eq .
is approximately expressed by the u(@xmath17 ) gauge transformation @xmath201 and @xmath202 .
in this appendix , we derive the moyal product in flat spacetime and the star product in curved spacetime .
the moyal product @xmath64 is defined by the wigner representation of convolution and is calculated as @xmath203 \notag \\ & \times [ e^{i p_{2 a } ( x_1^a - i \hbar \partial_{p_a})/\hbar } e^{x_2^a \partial_{x^a}/2 } { \hat b}(x , p ) ] \notag \\
= & { \hat a}(x , p ) e^{i \hbar { \cal f}_0/2 } { \hat b}(x , p ) .
\label{eq : moyal}\end{aligned}\ ] ] in the second line , we employ the inverse fourier transformation . in the third line , we change variables @xmath204 and introduce @xmath205 . as defined in the main text , @xmath76 is the poisson bracket in flat spacetime .
the star product @xmath206 is obtained by the peierls substitution @xmath207 $ ] .
strictly , we can not employ the peierls substitution for the moyal product since the moyal product is given by an exponential function of the poisson bracket .
instead , we employ it for the poisson bracket , from which weconstruct the star product .
this procedure is known as deformation quantization in mathematics @xcite and was carried out to obtain the keldysh formalism in dynamical or nonuniform electromagnetic fields @xcite . according to the chain rule ,
partial derivatives are modified as @xmath208 and @xmath209 , and the poisson bracket is modified as @xmath210
here we calculate the first - order retarded self - energy with respect to a magnetic field @xmath211 from eq . .
we obtain @xmath212 ^ 2\ } + 2 z^{\rm r}(\xi ) { \vec \sigma}_{f_{xy } \perp}^{\rm r}(\xi ) \cdot \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x { \vec p}_{\perp } \notag \\ & + 2 z^{\rm r}(\xi ) \sigma_{f_{xy } z}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] - q v^2 \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : gbr1a } \\
g_{f_{xy } x}^{\rm r}(\xi , { \vec p } ) = & 2 z^{\rm r}(\xi ) \sigma_{f_{xy } 0}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x p_x \notag \\ & + \sigma_{f_{xy } x}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) + v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } + 2 \sigma_{f_{xy } y}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y \notag \\ & + 2 \sigma_{f_{xy } z}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x p_x [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : gbr1b } \\
g_{f_{xy } y}^{\rm r}(\xi , { \vec p } ) = & 2 z^{\rm r}(\xi ) \sigma_{f_{xy } 0}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x p_y \notag \\ & + 2 \sigma_{f_{xy } x}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y + \sigma_{f_{xy } y}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } \notag \\ & + 2 \sigma_{f_{xy } z}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x p_y
[ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : gbr1c } \\ g_{f_{xy } z}^{\rm r}(\xi , { \vec p } ) = & 2 z^{\rm r}(\xi ) \sigma_{f_{xy } 0}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] + 2 { \vec \sigma}_{f_{xy } \perp}^{\rm r}(\xi ) \cdot \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x { \vec p}_{\perp } [ v \rho_x p_z + \eta^{\rm
r}(\xi ) ] \notag \\ & + \sigma_{f_{xy } z}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - ( v { \vec p}_{\perp})^2 + [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } - q v^2 z^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) , \label{eq : gbr1d}\end{aligned}\ ] ] [ eq : gbr1 ] and their momentum integrals denoted by @xmath213 are given by @xmath214 ^ 2\ } \notag \\ & + 2 z^{\rm r}(\xi ) \sigma_{f_{xy } z}^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \notag \\ & - q v^2 \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : gbr2a } \\
{ \vec g}_{f_{xy } \perp}^{\rm r}(\xi ) = & { \vec \sigma}_{f_{xy } \perp}^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } , \label{eq : gbr2b } \\ g_{f_{xy } z}^{\rm r}(\xi ) = & 2 z^{\rm r}(\xi ) \sigma_{f_{xy } 0}^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \notag \\ & + \sigma_{f_{xy } z}^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - ( v { \vec p}_{\perp})^2 + [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } \notag \\ & - q v^2 z^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) .
\label{eq : gbr2c}\end{aligned}\ ] ] [ eq : gbr2 ] we also obtain the self - consistent @xmath0 matrix eq . as @xmath215 g_{f_{xy } 0}^{\rm r}(\xi ) + 2 v_{\rm i } t_{00}^{\rm r}(\xi ) t_{0z}^{\rm r}(\xi ) g_{f_{xy } z}^{\rm r}(\xi ) , \label{eq : gbr3b } \\ { \vec t}_{f_{xy } \perp}^{\rm r}(\xi )
= & v_{\rm i } [ t_{00}^{{\rm r } 2}(\xi ) - t_{0z}^{{\rm r } 2}(\xi ) ] { \vec g}_{f_{xy } \perp}^{\rm r}(\xi ) , \label{eq : gbr3c } \\ t_{f_{xy } z}^{\rm r}(\xi ) = & 2 v_{\rm i } t_{00}^{\rm r}(\xi ) t_{0z}^{\rm r}(\xi ) g_{f_{xy } 0}^{\rm r}(\xi ) + v_{\rm i } [ t_{00}^{{\rm r } 2}(\xi ) + t_{0z}^{{\rm r } 2}(\xi ) ] g_{f_{xy } z}^{\rm r}(\xi ) , \label{eq : gbr3d}\end{aligned}\ ] ] [ eq : gbr3 ] equations and lead to @xmath216 .
the remaining components are obtained by introducing @xmath217 and @xmath218 and solving @xmath219 [ eq : gbr4 ] the first - order self - energy with respect to a torsional magnetic field is obtained just by @xmath220 .
next , we calculate the first - order lesser self - energy with respect to an electric field @xmath221 from eq . .
we obtain @xmath222 \notag \\ & + \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{[z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] v \rho_x p_x + i [ \eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{-i [ \eta^{\rm r}(\xi ) -
\eta^{\rm a}(\xi ) ] v \rho_x p_x + [ z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + q v \rho_x |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{[\eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] v \rho_x p_x + i [ z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & - i q v \rho_x [ z^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x p_y + { { \rm c.c . } } ] , \label{eq : ge<11a } \\
g_{f_{y0 } x}^ { < ( 1)}(\xi , { \vec p } ) = & \sigma_{f_{y0 } 0}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2
\{[z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] v \rho_x p_x - i [ \eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 + v^2 ( p_x^2 - p_y^2 ) - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2 v^2 p_x p_y
+ i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{v \rho_x p_x [ 2 v \rho_x p_z + \eta^{\rm r}(\xi ) + \eta^{\rm a}(\xi ) ] - i [ z^{\rm r}(\xi ) - z^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + q v \rho_x |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2
i v^2 p_x p_y - z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & - i q v \rho_x [ \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y + { { \rm c.c . } } ] , \label{eq : ge<11b } \\
g_{f_{y0 } y}^ { < ( 1)}(\xi , { \vec p } ) = & \sigma_{f_{y0 } 0}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i [ \eta^{\rm r}(\xi ) -
\eta^{\rm a}(\xi ) ] v \rho_x p_x + [ z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2 v^2 p_x p_y - i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - v^2 ( p_x^2 - p_y^2 )
- |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i [ z^{\rm r}(\xi ) - z^{\rm a}(\xi ) ] v \rho_x p_x + v \rho_x p_y [ 2 v \rho_x p_z + \eta^{\rm r}(\xi ) + \eta^{\rm a}(\xi)]\ } \notag \\ & + i q v \rho_x |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - v^2 ( p_x^2 - p_y^2 ) -
|v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & - i q v \rho_x/2 ( \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } + { { \rm c.c . } } ) , \label{eq : ge<11c } \\
g_{f_{y0 } z}^ { < ( 1)}(\xi , { \vec p } ) = & \sigma_{f_{y0 } 0}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2
\{z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{[2 v \rho_x p_z + \eta^{\rm r}(\xi ) + \eta^{\rm a}(\xi ) ] v \rho_x p_x + i [ z^{\rm r}(\xi ) - z^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{- i [ z^{\rm r}(\xi ) - z^{\rm a}(\xi ) ] v \rho_x p_x + [ 2 v \rho_x p_z + \eta^{\rm r}(\xi ) + \eta^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - ( v { \vec p}_{\perp})^2 + |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & + q v \rho_x |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{- i [ z^{\rm r}(\xi ) - z^{\rm a}(\xi ) ] v \rho_x p_x + [ 2 v \rho_x p_z + \eta^{\rm r}(\xi ) + \eta^{\rm a}(\xi ) ] v \rho_x p_y\ } \notag \\ & - i q v \rho_x \{\gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v \rho_x
p_y [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] + { { \rm c.c.}}\ } , \label{eq : ge<11d}\end{aligned}\ ] ] [ eq : ge<11 ] and their momentum integrals denoted by @xmath223 are given by @xmath224 \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } , \label{eq : ge<12a } \\
g_{f_{y0 } x}^ { < ( 1)}(\xi ) = & \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + i q v \rho_x \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } , \label{eq : ge<12b } \\ g_{f_{y0 } y}^ { < ( 1)}(\xi ) = & -\sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] -
i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & + i q v \rho_x \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] \notag \\ & - \frac{1}{2 } i q v \rho_x \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \left(\gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } + { { \rm c.c.}}\right ) , \label{eq : ge<12c } \\ g_{f_{y0 } z}^ { < ( 1)}(\xi ) = & \sigma_{f_{y0 } 0}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \sigma_{f_{y0 } z}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - ( v { \vec p}_{\perp})^2 + |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] .
\label{eq : ge<12d}\end{aligned}\ ] ] [ eq : ge<12 ] the self - consistent @xmath0 matrix eq .
is expressed by @xmath225 g_{f_{y0 } 0}^ { < ( 1)}(\xi ) + v_{\rm i } [ t_{00}^{\rm r}(\xi ) t_{0z}^{\rm a}(\xi ) + t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) ] g_{f_{y0 } z}^ { < ( 1)}(\xi ) , \label{eq : ge<13b } \\ t_{f_{y0 } x}^ { < ( 1)}(\xi ) = & v_{\rm i } [ |t_{00}^{\rm r}(\xi)|^2 - |t_{0z}^{\rm r}(\xi)|^2 ] g_{f_{y0 } x}^ { < ( 1)}(\xi ) + v_{\rm i } [ i t_{00}^{\rm r}(\xi ) t_{0z}^{\rm a}(\xi ) - i t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) ] g_{f_{y0 } y}^ { < ( 1)}(\xi ) , \label{eq : ge<13c } \\ t_{f_{y0 } y}^ { < ( 1)}(\xi ) = & - v_{\rm i } [ i t_{00}^{\rm
r}(\xi ) t_{0z}^{\rm a}(\xi ) - i t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) ] g_{f_{y0 } x}^ { < ( 1)}(\xi ) + v_{\rm i } [ |t_{00}^{\rm r}(\xi)|^2 - |t_{0z}^{\rm r}(\xi)|^2 ] g_{f_{y0 } y}^ { < ( 1)}(\xi ) , \label{eq : ge<13d } \\ t_{f_{y0 } z}^ { < ( 1)}(\xi ) = & v_{\rm i } [ t_{00}^{\rm r}(\xi ) t_{0z}^{\rm a}(\xi ) + t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) ] g_{f_{y0 } 0}^ { <
( 1)}(\xi ) + v_{\rm i } [ |t_{00}^{\rm r}(\xi)|^2 + |t_{0z}^{\rm r}(\xi)|^2 ] g_{f_{y0 } z}^ { < ( 1)}(\xi ) , \label{eq : ge<13e}\end{aligned}\ ] ] [ eq : ge<13 ] leading to @xmath226 .
the remaining components are obtained by introducing @xmath227 and @xmath228 and solving @xmath229/2 - { { \rm c.c.}}\end{bmatrix } , \label{eq : ge<14a } \\
\begin{bmatrix } { \tilde \sigma}_{q f_{y0 } x}^ { < ( 1)}(\xi ) \\ { \tilde \sigma}_{q f_{y0 } y}^ { < ( 1)}(\xi ) \end{bmatrix } = & \frac{n_{\rm i } v_{\rm i}^2}{8 \pi^2 \hbar^3 v^3 } \begin{bmatrix } |t_{00}^{\rm r}(\xi)|^2 - |t_{0z}^{\rm r}(\xi)|^2 & i t_{00}^{\rm
r}(\xi ) t_{0z}^{\rm a}(\xi ) - i t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) \\
-i t_{00}^{\rm r}(\xi ) t_{0z}^{\rm a}(\xi ) + i t_{00}^{\rm a}(\xi ) t_{0z}^{\rm r}(\xi ) & |t_{00}^{\rm r}(\xi)|^2 - |t_{0z}^{\rm r}(\xi)|^2 \end{bmatrix } \begin{bmatrix } { \tilde g}_{q f_{y0 } x}^ { < ( 1)}(\xi ) \\ { \tilde g}_{q f_{y0 } y}^ { < ( 1)}(\xi ) \end{bmatrix } , \label{eq : ge<14b } \\ j_1^ { < ( 1)}(\xi ) \equiv & \int_0^{\lambda_{\perp}^2 } { \rm d } x \int_{-\lambda_z}^{\lambda_z } { \rm d } y \frac{|z^{\rm r}(\xi)|^2 - |y + \eta^{\rm r}(\xi)|^2}{|z^{{\rm r } 2}(\xi ) - x - [ y + \eta^{\rm r}(\xi)]^2|^2 } , \label{eq : ge<14c } \\
j_2^ { < ( 1)}(\xi ) \equiv & \int_0^{\lambda_{\perp}^2 } { \rm d } x \int_{-\lambda_z}^{\lambda_z } { \rm d } y \frac{i z^{\rm r}(\xi ) [ y + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ y + \eta^{\rm r}(\xi)]}{|z^{{\rm r } 2}(\xi ) - x - [ y + \eta^{\rm r}(\xi)]^2|^2 } , \label{eq : ge<14d}\end{aligned}\ ] ] [ eq : ge<14 ] the explicit forms of @xmath230 and @xmath231 are given in appendix [ app : int ] .
again , the first - order self - energy with respect to a torsional electric field is obtained by @xmath232 .
now we calculate the auxiliary heat magnetization eq . and the proper one with eq .
. by using the results in appendix [ app : gbr ] , the trace parts in eq
. are given by @xmath233 , \label{eq : mz1a } \\ & \operatorname{tr}g_0^{\rm r } \sigma_{-t^0_{\phantom{0 } xy}}^{\rm r } g_0^{\rm r } \notag \\
= & 2 \sigma_{-t^0_{\phantom{0 } xy } 0}^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) + ( v { \vec p}_{\perp})^2 + [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } + 4 \sigma_{-t^0_{\phantom{0 } xy } z}^{\rm r}(\xi ) z^{\rm r}(\xi ) \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] , \label{eq : mz1b}\end{aligned}\ ] ] [ eq : mz1 ] respectively .
since the heat magnetization depends on temperature through the distribution function @xmath234 in the absence of interactions , eq . can be easily solved .
in fact , it is achieved by replacing @xmath235 . here
@xmath236 is a dilogarithm function defined by @xmath237 and hence we can check @xmath238/\partial \beta = \beta f(\xi)$ ] . therefore , the proper heat magnetization is obtained by @xmath239 + { { \rm c.c.}}\notag
\\ = & \frac{i}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi
f^{(-2)}(\xi ) \xi^2 i_{12}^{\rm r}(\xi ) + { { \rm c.c . } } , \label{eq : mz2a } \\ m_{{\rm q } z}^{\rm ext } = & -\frac{i \hbar}{\pi q } \sum_{\rho_x = \pm 1 } \int { \rm d } \xi f^{(-2)}(\xi ) \xi^2 \sigma_{f_{xy } 0}^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) + ( v { \vec p}_{\perp})^2 + [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } \notag \\ & - \frac{2 i \hbar}{\pi q } \sum_{\rho_x = \pm 1 } \int { \rm d } \xi f^{(-2)}(\xi ) \xi^2 \sigma_{f_{xy } z}^{\rm r}(\xi ) z^{\rm r}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) [ v \rho_x p_z + \eta^{\rm r}(\xi ) ] \notag \\ = & -\frac{i}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi f^{(-2)}(\xi ) \xi^2 \{{\tilde \sigma}_{q f_{xy } 0}^{\rm r}(\xi ) [ 2 z^{{\rm r } 2}(\xi ) i_{02}^{\rm
r}(\xi ) - i_{01}^{\rm r}(\xi ) ] + { \tilde \sigma}_{q f_{xy } z}^{\rm r}(\xi ) 2 z^{\rm r}(\xi ) i_{12}^{\rm r}(\xi)\ } + { { \rm c.c.}}\label{eq : mz2b}\end{aligned}\ ] ] [ eq : mz2 ]
we also calculate the kubo formulas for the electric and thermal ( hall ) conductivities eqs . and .
for the hall conductivity , by using the results in appendix [ app : ge<1 ] , the trace parts in eq .
are given by @xmath240 - \eta^{\prime { \rm r}}(\xi ) z^{\rm r}(\xi)\ } , \label{eq : sigmaxy1a } \\ & \operatorname{tr}(g_0^{\rm r } - g_0^{\rm a } ) \partial_{p_x } ( g_0^{{\rm r } -1 } + g_0^{{\rm a } -1 } ) g_0^{\rm r } \partial_{p_y } g_0^{{\rm r } -1 } \notag \\
= & -4 v^2 |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2 v^2 p_x p_y + i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } + 8 v^2 \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y , \label{eq : sigmaxy1b } \\ & \operatorname{tr}\sigma_{f_{y0}}^ { < ( 1 ) } ( g_0^{\rm r } - g_0^{\rm a } ) \partial_{p_x } g_0^{{\rm r } -1 } g_0^{\rm r } \notag \\ = & 2 v \rho_x \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \notag \\ & \times \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 + v^2 ( p_x^2 - p_y^2 ) - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] - \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) + v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}\right ) \notag \\ & + 2 v \rho_x \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \notag \\ & \times \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2 v^2 p_x p_y + i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } - 2 \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y\right ) , \label{eq : sigmaxy1c}\end{aligned}\ ] ] [ eq : sigmaxy1 ] respectively . by integrating over @xmath122 and @xmath241 and summing over @xmath116 ,
we obtain @xmath242 - z^{\rm r}(\xi ) \eta^{\prime { \rm r}}(\xi)\ } + { { \rm c.c.}}\notag \\ = & -\frac{i q^2}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi f(\xi ) \{[1 - \sigma_{00}^{\prime { \rm r}}(\xi ) ] i_{12}^{\rm r}(\xi ) - z^{\rm r}(\xi ) \sigma_{0z}^{\prime { \rm r}}(\xi ) i_{02}^{\rm r}(\xi)\ } + { { \rm c.c . } } , \label{eq : sigmaxy2a } \\
\sigma^{xy}_{\rm int~i } = & -\frac{\hbar q^2 v^2}{\pi } \sum_{\rho_x = \pm 1 } \int { \rm d } \xi f^{\prime}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x + \eta^{\rm r}(\xi)]\ } \notag \\
= & \frac{q^2}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi [ -f^{\prime}(\xi ) ] j_2^ { < ( 1)}(\xi ) , \label{eq : sigmaxy2b } \\ \sigma^{xy}_{\rm ext } = & \frac{i \hbar q v}{\pi } \sum_{\rho_x = \pm 1 } \rho_x \int { \rm d } \xi f^{\prime}(\xi ) \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \notag \\ & \times \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] - \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}/2 - { { \rm c.c.}}\right ) \notag \\ & + \frac{i \hbar q v}{\pi } \sum_{\rho_x = \pm 1 } \rho_x \int { \rm d } \xi f^{\prime}(\xi ) \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\
= & \frac{q^2}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi [ -f^{\prime}(\xi ) ] \left({\tilde \sigma}_{q f_{y0 } x}^ { < ( 1)}(\xi ) \{j_1^ { < ( 1)}(\xi ) - [ z^{{\rm r } 2}(\xi ) i_{02}^{\rm r}(\xi ) - i_{22}^{\rm r}(\xi)]/2 - { { \rm c.c.}}\ } + { \tilde \sigma}_{q f_{y0 } y}^ { < ( 1)}(\xi ) j_2^ { < ( 1)}(\xi)\right ) \label{eq : sigmaxy2c}\end{aligned}\ ] ] [ eq : sigmaxy2 ] for the electric conductivity , the trace parts are given by @xmath243 \notag \\ & + 4 v^2 \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\ } , \label{eq : sigmayy1b } \\ & \operatorname{tr}\sigma_{f_{y0}}^ { < ( 1 ) } ( g_0^{\rm r } - g_0^{\rm a } ) \partial_{p_y } g_0^{{\rm r } -1 } g_0^{\rm r } \notag \\ = & 2 v \rho_x \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{2 v^2 p_x p_y - i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] + i z^{\rm
a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } - 2 \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) v^2 p_x p_y\right ) \notag \\ & + 2 v \rho_x \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - v^2 ( p_x^2 - p_y^2 ) - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2]\right . \notag \\ & \left.- \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - v^2 ( p_x^2 - p_y^2 ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}\right ) , \label{eq : sigmayy1c}\end{aligned}\ ] ] [ eq : sigmayy1 ] and we get @xmath244 - \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}/2 - { { \rm c.c.}}\right ) \notag
\\ = & \frac{q^2}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi [ -f^{\prime}(\xi ) ] \{j_1^ { < ( 1)}(\xi ) - [ z^{{\rm r } 2}(\xi ) i_{02}^{\rm r}(\xi ) - i_{22}^{\rm r}(\xi)]/2 - { { \rm c.c.}}\ } , \label{eq : sigmayy2b } \\
\sigma^{yy}_{\rm ext } = & -\frac{i \hbar q v}{\pi } \sum_{\rho_x = \pm 1 } \rho_x \int { \rm d } \xi f^{\prime}(\xi ) \sigma_{f_{y0 } x}^ { < ( 1)}(\xi ) \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } |\gamma_0^{\rm r}(\xi , { \vec p})|^2 \{i z^{\rm r}(\xi ) [ v \rho_x p_z + \eta^{\rm a}(\xi ) ] - i z^{\rm a}(\xi ) [ v \rho_x p_z + \eta^{\rm r}(\xi)]\ } \notag \\ & + \frac{i \hbar q v}{\pi } \sum_{\rho_x = \pm 1 } \rho_x \int { \rm d } \xi
f^{\prime}(\xi ) \sigma_{f_{y0 } y}^ { < ( 1)}(\xi ) \notag \\ & \times \int \frac{{\rm d}^3 p}{(2 \pi \hbar)^3 } \left(|\gamma_0^{\rm r}(\xi , { \vec p})|^2 [ |z^{\rm r}(\xi)|^2 - |v \rho_x p_z + \eta^{\rm r}(\xi)|^2 ] - \gamma_0^{{\rm r } 2}(\xi , { \vec p } ) \{z^{{\rm r } 2}(\xi ) - [ v \rho_x p_z + \eta^{\rm r}(\xi)]^2\}/2 - { { \rm c.c.}}\right ) \notag \\
= & \frac{q^2}{4 \pi^3 \hbar^2 v } \int { \rm d } \xi [ -f^{\prime}(\xi ) ] \left(-{\tilde \sigma}_{q f_{y0 } x}^ { < ( 1)}(\xi ) j_2^ { < ( 1)}(\xi ) + { \tilde \sigma}_{q f_{y0 } y}^ { < ( 1)}(\xi ) \{j_1^ { < ( 1)}(\xi ) - [ z^{{\rm r } 2}(\xi ) i_{02}^{\rm r}(\xi ) - i_{22}^{\rm r}(\xi)]/2 - { { \rm c.c.}}\}\right ) .
\label{eq : sigmayy2c}\end{aligned}\ ] ] [ eq : sigmayy2 ] the kubo formula for the thermal ( hall ) conductivity is obtained by replacing @xmath245 in the above expressions .
we give the explicit forms of the momentum integrals defined above .
. is expressed by @xmath246 \ln [ \sqrt{z^{{\rm r } 2}(\xi ) - x } + y + \eta^{\rm r}(\xi ) ] \notag \\ & \left.\left.+ [ \sqrt{z^{{\rm r } 2}(\xi ) - x } - y - \eta^{\rm r}(\xi ) ] \ln [ \sqrt{z^{{\rm r } 2}(\xi ) - x } - y - \eta^{\rm r}(\xi ) ] \right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : i01r } \\ i_{11}^{\rm r}(\xi ) = & \left.\left .
-\frac{1}{2 } \{[y + \eta^{\rm r}(\xi)]^2 - z^{{\rm r } 2}(\xi ) + x\ } \ln \{[y + \eta^{\rm r}(\xi)]^2 - z^{{\rm r } 2}(\xi ) + x\ } \right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : i11r } \\
i_{02}^{\rm r}(\xi ) = & \left.\left .
\frac{1}{2 \sqrt{z^{{\rm r } 2}(\xi ) - x } } \ln \frac{y + \eta^{\rm r}(\xi ) + \sqrt{z^{{\rm r } 2}(\xi ) - x}}{y + \eta^{\rm r}(\xi ) - \sqrt{z^{{\rm r } 2}(\xi ) - x } } \right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : i02r } \\
i_{12}^{\rm r}(\xi ) = & \left.\left .
-\frac{1}{2 } \ln \{[y + \eta^{\rm r}(\xi)]^2 - z^{{\rm r } 2}(\xi ) + x\ } \right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : i12r } \\
i_{22}^{\rm r}(\xi ) = & \left.\left .
\frac{1}{2 } \sqrt{z^{{\rm r } 2}(\xi ) - x } \ln \frac{y + \eta^{\rm r}(\xi ) + \sqrt{z^{{\rm r } 2}(\xi ) - x}}{y + \eta^{\rm r}(\xi ) - \sqrt{z^{{\rm r } 2}(\xi ) - x } } \right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z}. \label{eq : i22r}\end{aligned}\ ] ] @xmath248 b(\xi , x , y ) + c_1(\xi ) c(\xi , x , y)\right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : varxi1<1 } \\
j_2^ { < ( 1)}(\xi ) = & \left.\left.a_2(\xi ) a(\xi , x , y ) + b_2(\xi ) b(\xi , x , y)\right|_{x = 0}^{x = \lambda_{\perp}^2}\right|_{y = -\lambda_z}^{y = \lambda_z } , \label{eq : varxi2<1 } \\
a(\xi , x , y ) \equiv & -\frac{1}{2 i } \left[\operatorname{li}_2\left(\frac{r(\xi ) - y}{r(\xi ) - q_+^{\rm r}(\xi , x)}\right ) + \operatorname{li}_2\left(\frac{r(\xi ) - y}{r(\xi ) - q_-^{\rm r}(\xi , x)}\right)\right ] + { { \rm c.c . } } , \label{eq : a<1 } \\
b(\xi , x , y ) \equiv & \{[y - q_+^{\rm r}(\xi , x ) ] \ln [ y - q_+^{\rm r}(\xi , x ) ] + [ y - q_-^{\rm r}(\xi , x ) ] \ln [ y - q_-^{\rm r}(\xi , x)]\}/2 i + { { \rm c.c . } } , \label{eq : b<1 } \\ c(\xi , x , y ) \equiv & \{[y^2 - q_+^{{\rm r } 2}(\xi , x ) ] \ln [ y - q_+^{\rm r}(\xi , x ) ] + [ y^2 - q_-^{{\rm r } 2}(\xi , x ) ] \ln [ y - q_-^{\rm r}(\xi , x)]\}/4 i + { { \rm c.c . } } , \label{eq : c<1 } \\
q_{\pm}^{\rm r}(\xi , x ) \equiv & -\eta^{\rm r}(\xi ) \pm \sqrt{z^{{\rm r } 2}(\xi ) - x } , \label{eq : q<1 } \\
r(\xi ) \equiv & [ z^{{\rm r } 2}(\xi ) - \eta^{{\rm r } 2}(\xi ) - z^{{\rm a } 2}(\xi ) + \eta^{{\rm a } 2}(\xi)]/2 [ \eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] , \label{eq : r<1 } \\
a_1(\xi ) \equiv & -\{[z^{\rm r}(\xi ) + z^{\rm a}(\xi)]^2 + [ \eta^{\rm r}(\xi ) -
\eta^{\rm a}(\xi)]^2\ } \notag \\ & \times [ z^{\rm r}(\xi ) + \eta^{\rm r}(\xi ) - z^{\rm a}(\xi ) - \eta^{\rm a}(\xi ) ] [ z^{\rm r}(\xi ) - \eta^{\rm r}(\xi ) - z^{\rm a}(\xi ) + \eta^{\rm a}(\xi ) ] /4 i [ \eta^{\rm r}(\xi ) -
\eta^{\rm a}(\xi)]^3 , \label{eq : a1<1 } \\
b_1(\xi ) \equiv & -[z^{{\rm r } 2}(\xi ) - z^{{\rm a } 2}(\xi)]/i [ \eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi)]^2 , \label{eq : b1<1 } \\ c_1(\xi ) \equiv & -1/i [ \eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] , \label{eq : c1<1 } \\
a_2(\xi ) \equiv & [ z^{\rm r}(\xi ) + z^{\rm a}(\xi ) ] \notag \\ & \times [ z^{\rm r}(\xi ) + \eta^{\rm r}(\xi ) - z^{\rm a}(\xi ) - \eta^{\rm a}(\xi ) ] [ z^{\rm r}(\xi ) - \eta^{\rm r}(\xi ) - z^{\rm a}(\xi ) + \eta^{\rm a}(\xi ) ] /2 [ \eta^{\rm r}(\xi )
- \eta^{\rm a}(\xi)]^2 , \label{eq : a2<1 } \\
b_2(\xi ) \equiv & [ z^{\rm r}(\xi )
- z^{\rm a}(\xi)]/[\eta^{\rm r}(\xi ) - \eta^{\rm a}(\xi ) ] , \label{eq : b2<1 } \end{aligned}\ ] ] 59ifxundefined [ 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.100.016601
[ * * , ( ) ] link:\doibase 10.1103/physrevb.81.054414 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.100404 [ * * , ( ) ] link:\doibase 10.1126/science.1188260 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.134411 [ * * , ( ) ] link:\doibase 10.1126/science.1257340 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.106603 [ * * , ( ) ] link:\doibase 10.1073/pnas.1524076113 [ * * , ( ) ] link:\doibase 10.1103/physrev.135.a1505 [ * * , ( ) ] link:\doibase 10.1088/0022 - 3719/10/12/021 [ * * , ( ) ] link:\doibase 10.1103/physrevb.55.2344 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.197202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.184406 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.236601 [ * * , ( ) ] link:\doibase 10.1093/ptep / ptu162 [ * * , ( ) ] link:\doibase 10.1143/ptp.42.494 [ * * , ( ) ] link:\doibase 10.1143/ptp.42.1284 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1539 [ * * , ( ) ] link:\doibase 10.1103/physrev.95.1154 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1959 [ * * , ( ) ] link:\doibase 10.1016/s0031 - 8914(55)92596 - 9 [ * * , ( ) ] link:\doibase 10.1016/s0031 - 8914(58)93541 - 9 [ * * , ( ) ] link:\doibase 10.1103/physrevb.2.4559 [ * * , ( ) ] link:\doibase 10.1103/physrevb.5.1862 [ * * , ( ) ] link:\doibase 10.1143/ptp.116.61 [ * * , ( ) ] link:\doibase 10.1143/ptp.117.415 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.126602 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.165103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.195129 [ * * , ( ) ] link:\doibase 10.1016/j.crhy.2013.10.010 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/27/11/113201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.205101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.127205 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/9/9/356 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.035103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.115133 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.245107 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.187202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.165115 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.035116 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.075502 [ * * , ( ) ] link:\doibase 10.1093/ptep / pts063 [ * * , ( ) ] link:\doibase 10.1103/physrevd.88.025040 [ * * , ( ) ] link:\doibase 10.1103/physrevd.90.105004 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.134510 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.116.166601 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.80.115111 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.214516 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.214418 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.086602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.087206 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.220403 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/24/48/482001 [ * * , ( ) ] link:\doibase 10.1143/jpsj.81.083704 [ * * , ( ) ] \doibase 10.1023/b : math.0000027508.00421.bf [ * * , ( ) ]
|
we investigate the electric and thermal transport properties in a disordered weyl ferromagnet on an equal footing by using the keldysh formalism in curved spacetime .
in particular , we calculate the anomalous thermal hall conductivity , which consists of the kubo formula and the heat magnetization , without relying on the wiedemann - franz law .
we take nonmagnetic impurities into account within the self - consistent @xmath0-matrix approximation and reproduce the wiedemann - franz law for the extrinsic fermi - surface and intrinsic fermi - sea terms , respectively .
this is the first step towards a unified theory of the anomalous hall effect at finite temperature , where we should take into account both disorder and interactions .
|
the main motivation to study solar wind composition is to obtain information on the isotopic composition of elements in the sun .
this is important because the sun constitutes @xmath6 of solar system matter , and for most elements the solar composition could provide the most reliable information on the composition of the primordial solar nebula .
because nowhere in the sun have temperatures ever been high enough to alter isotopic abundances of heavy elements by nuclear reactions , solar ca is thought to reflect the original isotopic composition in the solar nebula . yet , most of what is known of solar isotopic abundances is inferred from terrestrial and from meteoritic abundances with a few exceptions : recently , the isotopic composition of solar wind magnesium , the first analysis of the isotopic abundances of a refractive element in solar matter measured in situ with the spacecraft borne mass spectrometer wind / mass , has been reported ( @xcite ) .
the solar isotopic composition of the volatile noble gases helium and neon have been determined in situ with the apollo foil experiments ( @xcite ) . in situ measurements with celias / mtof ( @xcite ) on board soho
have confirmed that the solar neon isotopic composition differs significantly from the terrestrial and meteoritic abundances .
isotopic abundances also have been determined for several elements ( c , n , o , he , ne , mg ) in the higher energy ( above 10 mev / amu ) solar energetic particles ( @xcite ) . from similar measurements it has been possible to obtain reliable information on coronal elemental abundances including the abundance of calcium ( @xcite ) . however , the isotopic composition of solar energetic particles ( sep ) may not be representative of the solar composition because fractionation processes occur during the particle acceleration and transport and these processes may vary from event to event .
the far more fluent solar wind is the most authentic sample of the solar source composition . from recent theoretical models ( e.g. @xcite )
it is expected that the isotopic fractionation in the solar wind flow due to differences in coulomb drag depletes the heavier isotopes by at most a few percent .
therefore measurements of the solar wind isotopic abundances of the heavy element calcium together with previous measurements of magnesium abundances provide additional evidence that the solar isotopic composition of refractive elements agrees with terrestrial and meteoritic values .
the mtof sensor ( figure [ fig1 ] ) of celias ( charge , element , and isotope analysis system ) on board the soho ( solar and heliospheric observatory ) spacecraft is an isochronous time - of - flight mass spectrometer ( @xcite ) with a resolution m/@xmath7 m of better than 100 .
this provides the possibility of resolving the different isotopes of almost all solar wind elements in the range from @xmath8 to @xmath9 atomic mass units ( amu ) .
the instrument detects ions at solar wind bulk velocities of 300 to 1000 km / s , corresponding to energies of about 0.3 to 3 kev / amu .
details on the principle of operation , the calibration of the instrument functions , and the format of the data transferred by the telemetry can be found in the initial publication on celias / mtof isotope abundance measurements of solar wind neon ( @xcite ) .
here we only briefly describe the instrument characteristics which are necessary to understand the quantitative evaluation of time - of - flight ( tof ) spectra .
highly charged solar wind ions enter the instrument ( figure [ fig1 ] ) through the wave ( wide angle variable energy ) entrance system that has an energy - per - charge ( @xmath10 ) acceptance bandwidth of about half a decade , and a conic field of view of @xmath11 width ( @xcite ) .
all calcium isotopes have approximately the same bulk velocity , the same charge state , and the same width of the drifting maxwellian velocity distribution so that the heavier isotopes have a higher center @xmath10 .
the strongest instrumental fractionation occurs when the center @xmath10 of either isotope coincides with one of the edges of the @xmath10-acceptance of the wave .
the acceptance function including the ion optical effects of the postacceleration voltage to the time - of - flight section has been well analyzed in the calibration system for mass spectrometers ( @xcite ) of the university of bern ( uob ) .
therefore the flight data can reliably be filtered with respect to the ion optical instrument discrimination for any element or isotope detected in the solar wind .
a much weaker instrumental fractionation is due to the element specific detection efficiencies of the vmass subsystem which is a v - shaped isochronous time - of - flight spectrometer ( @xcite ) .
the ratio of the double coincidence rate over the front secondary electron detection assembly rate has been measured as a function of ion energy per mass with a @xmath0ca@xmath12 beam of the mefisto laboratory at uob ( @xcite ) .
the model functions reproduce the calibration data very well so that they can be applied to derive the differences in the detection efficiencies of the isotopes @xmath13ca ( figure [ figca2 ] ) . with very good confidence the uncertainty of these detection efficiencies
can be assumed to be lower than the statistical error evaluated ( see table [ fit ] ) .
figure [ figca3 ] shows a spectrum where all pulse height analysis words of the full year 1996 have been collected in the mass range 39 to 45 amu .
tof channel contents have been put into bins six channels wide and the uncertainties of the data points have been estimated by the linear - optimization algorithm of rauch - tung - striebel ( @xcite ) where the filter width corresponds to half the tof resolution of the instrument ; this filter introduces a minor reduction of the peak resolution of @xmath14 .
this algorithm has been applied to improve the signal - to - noise ratio by a factor of four so that @xmath1ca@xmath15 and @xmath4ca@xmath15 can be identified with high significance .
note that most incident ions leave the carbon foil as neutrals or singly charged .
the @xmath0ca@xmath15 distribution is asymmetrically broadened .
it exhibits a different shape than @xmath16ne@xmath15 ( @xcite ) for one main reason : singly charged calcium has generally much higher @xmath10 inside the vmass than singly charged neon so that calcium follows a longer trajectory in the field free region above the ion stop microchannel plate ( imcp ) detector ( @xcite ) .
longer trajectories in the field - free region correspond to longer times of flight .
this leads to the tails of higher channels in the tof distribution of all the mass peaks in figure [ figca3 ] . as in the case of @xmath17ne@xmath15 ( @xcite )
the electronic ringing peaks are visible at about 30 tof channels to the left of the @xmath13ca@xmath15 signatures ; they arise from reflected start signals .
the tof counts have been weighted by the detection efficiencies for the different calcium isotopes within the window of @xmath18 instrument fractionation between @xmath4ca@xmath15 and @xmath0ca@xmath15 .
all three detectable calcium peaks , @xmath0ca@xmath15 , @xmath1ca@xmath15 , and @xmath4ca@xmath15 , are described by the same model function with identical asymmetric shape . also , the marginally detectable or non - detectable peaks @xmath19k@xmath15 and @xmath20ca@xmath15 are included into the fit function , assuming terrestrial isotopic abundances : @xmath21 } } & \nonumber\\ { + { \sum_{j=1}^{5}}{{b_j}\over{1+{{{(t - p_{i}-jl)}^{2}}\over{w^{2 } } } } } } & \nonumber\\ { + { { c}\times{exp[-{{{(t - p_{i}+r)}^{2}}\over{2u^{2}}}]}})}. \nonumber \\ % \eqnum{1 } \label{eqn}\end{aligned}\ ] ] @xmath22 is the time - of - flight channel number and the free fit parameters are @xmath23 representing the amplitudes of the @xmath13ca@xmath15 peaks , @xmath24 the main gaussian peak position of @xmath0ca@xmath15 , @xmath25 the shift of the lorentzians describing the asymmetry against the main gaussians for all three peaks of @xmath13ca@xmath15 , @xmath26 the shift of the ringing peaks against the main gaussians , the @xmath27 the relative amplitudes of the shifted lorentzians compared to the main gaussians , @xmath28 the relative amplitude of the ringing peaks , @xmath29 the width of the main gaussian peaks , @xmath30 the width of the lorentzian peaks , and @xmath31 the width of the ringing peaks .
the gaussian peak positions @xmath32 of @xmath19k@xmath15 and @xmath33ca@xmath15 are kept fixed at their calibrated tof channels .
the total amplitudes @xmath23 , the widths @xmath30 and @xmath31 , and the relative amplitudes @xmath27 give the abundances for @xmath13ca@xmath15 . this analysis has been done twice assuming that the solar wind calcium is either tenfold or elevenfold charged .
table [ fit ] lists the integrated counts for @xmath13ca@xmath15 assuming tenfold charged solar wind calcium .
this gives the count ratios @xmath0ca@xmath15/@xmath1ca@xmath15 @xmath2 ( @xmath34 ) and @xmath0ca@xmath15/@xmath4ca@xmath15 @xmath2 ( @xmath35 ) .
analogously , assuming elevenfold charged solar wind calcium we find @xmath0ca@xmath15/@xmath1ca@xmath15 @xmath2 ( @xmath36 ) and @xmath0ca@xmath15/@xmath4ca@xmath15 @xmath2 ( @xmath37 ) .
we further correct these ratios for the effect of heavy elements such as si and fe having different mean velocities in the solar wind compared to the proton velocity @xmath38 derived from the proton monitor data . for ca the mean velocity can be estimated to be 1.027 x @xmath38 - 14.6 km / s ( @xcite ) .
this correction in the mean velocity and therefore in the mean e / q leads to the values @xmath0ca@xmath15/@xmath1ca@xmath15 @xmath2 ( @xmath39 ) and @xmath0ca@xmath15/@xmath4ca@xmath15 @xmath2 ( @xmath40 ) for tenfold charged solar wind calcium and @xmath0ca@xmath15/@xmath1ca@xmath15 @xmath2 ( @xmath41 ) and @xmath0ca@xmath15/@xmath4ca@xmath15 @xmath2 ( @xmath42 ) for elevenfold charged solar wind calcium . according to @xcite and @xcite the coronal freeze - in temperatures for the time period considered in this paper are such that the average solar wind calcium charge state should be somewhere between ten and eleven ;
the average number is approximately 10.5 .
therefore the final values for the solar wind calcium isotopic abundance ratios are taken as the average of the two values derived for the two assumptions of tenfold and elevenfold calcium .
the error of the mean value is chosen sufficiently large to be compatible with both assumptions .
for all ratios an additional uncertainty of @xmath43 is due to uncertainties in the absolute calibration of the instrument acceptance .
this results in @xmath0ca@xmath15/@xmath1ca@xmath15 @xmath2 ( @xmath3 ) and @xmath0ca@xmath15/@xmath4ca@xmath15 @xmath2 ( @xmath5 ) .
more precise values can be determined , once the charge distribution of solar wind calcium is better known for the year 1996 with other instrumentation on board the soho spacecraft .
the sensors wind / mass and celias / ctof will provide sufficient information to reevaluate the calcium data .
based on the assumption that the solar wind calcium charge state varies between ten and eleven as a long term average , we find that the isotopic abundance ratios of solar wind calcium are @xmath0ca/@xmath1ca @xmath2 ( @xmath3 ) and @xmath0ca/@xmath4ca @xmath2 ( @xmath5 ) . both ratios are consistent with the terrestrial ratios @xmath0ca/@xmath1ca @xmath2 ( @xmath44 ) and @xmath0ca/@xmath4ca @xmath2 ( @xmath45 ) ( @xcite ) .
more precise values can be obtained once a time series during the year 1996 of the calcium charge states in the solar wind at the soho location is available .
from astrophysical and geochemical considerations it can be concluded that the isotopic composition of photospheric ca must be within small fractions of per mills identical to the terrestrial composition .
the similarity of the solar wind results with the terrestrial values suggests the preliminary conclusion that isotopic fractionation within the solar wind plays a minor role .
this is supported by the model calculations of @xcite that consider differences in the coulomb drag for isotopes of various elements .
fractionation effects not larger than a few percent are expected for ne and mg . for the heavy element
ca even weaker isotopic fractionation has to be expected .
although there are as yet no detailed evaluations of a more extensive time series with calcium isotopic ratios to be determined in different solar wind regimes , we suggest that the ca abundance discussed in this work is within a few percent of the true solar composition .
table [ comparison ] shows a comparison between the terrestrial , solar and presolar grain calcium isotopic composition .
there is no evidence for large variations in these isotopic compositions except for the case of the presolar x - grains that originate from supernovae explosions ( @xcite ) .
the large enrichment in @xmath4ca in the x - grains is a consequence of the radioactive decay of @xmath4ti in supernovae material .
this work was supported by the swiss national science foundation , by the prodex program of esa , by nasa grant nag5 - 2754 , and by dara , germany , with grants 50 oc 89056 and 50 oc 9605 .
the flight - spare unit of mtof has been recalibrated with the electron cyclotron resonance ion source in the mefisto laboratory at the university of bern and with the support of adrian marti and reto schletti . crrrrrrrrrrr @xmath0ca@xmath15 @xmath46 & @xmath47 & @xmath48 & @xmath49 & @xmath50 & @xmath51@xmath1ca@xmath15 @xmath46 & @xmath52 & @xmath53 & @xmath54 & @xmath55 & @xmath56@xmath20ca@xmath15 @xmath46 & @xmath57 & @xmath58 & @xmath59 & @xmath60 & @xmath61@xmath4ca@xmath15 @xmath46 & @xmath62 & @xmath63 & @xmath64 & @xmath65 & @xmath66@xmath67k@xmath15 @xmath46 & @xmath68 & @xmath69 & @xmath70 & @xmath71 & @xmath72@xmath73k@xmath15 @xmath46 & @xmath58 & @xmath74 & @xmath75 & @xmath76 & @xmath77 aellig , m. , et al .
1998 , submitted to j. geophys .
amari , s. , zinner , e. , lewis , r. s. 1996 , lunar planet .
, 27 , 23 breneman , h. h. , and stone , e. c. 1985 , astrophys .
j. , 299 , l57 bochsler , p. , gonin , m. , sheldon , r. b. , zurbuchen , t. , gloeckler , g. , hamilton , d. c. , collier , m. r. , and hovestadt , d. 1995 , proc .
solar wind conf . , d. winterhalter , j. t. gosling , s. r. habbal , w. s. kurth , m. neugebauer , woodbury : aip press , 1995 , 199 bodmer , r. , and bochsler , p. , paper presented at meeting , eur .
soc . , the hague , netherlands , 1996 geiss , j. , bhler , f. , cerutti , h. , eberhardt , p. , and filleux , c. 1972 , apollo 16 preliminary scientific report , nasa spec .
sp-315 , 14.1 gelb , a. , kasper , j. f. , nash , r. a. , price , c. f. , and sutherland , a. a. 1974 , applied optimal estimation , a. gelb , cambridge : mit press , 1974 .
gonin , m. , kallenbach , r. , bochsler , p. , and brgi , a. 1995 , nucl .
methods phys .
res . , b101 , 313 hefti , s. 1997 , phd thesis , university of bern , switzerland hoppe , p. , strebel , r. , eberhardt , p. , amari , s. , and lewis , r. s. 1996 , science , 272 , 1314 hovestadt , d. , et al .
1995 , solar physics , 162 , 441 kallenbach , r. , et al .
1997 , j. geophys .
res . , 102 , 26895 kern , o. 1997 , private communication , university of bern , switzerland leske , r. a. , mewaldt , r. a. , cummings , a. c. , cummings , j. r. , stone , e. c. , and von rosenvinge , t. t. 1996 , space sci .
, 78 , 149 marti , a. 1997 , phd thesis , university of bern , switzerland mason , g. m. , mazur , j. e. , and hamilton , d. c. 1994 , astrophys .
j. , 425 , 843 mewaldt , r. a. , spalding , j. d. , and stone , e. c. 1984 , astrophys . j. , 280 , 892 nittler , l. r. , amari , s. , zinner , e. , woosley , s. e. , and lewis , r. s. 1996 , astrophys . j. , 462 , l31 russell , w. a. , papanastassiou , d. a. , tombrello , t. a. , and epstein , s. 1977 , proc .
lunar sci .
8th , 3791 selesnick , r. s. , cummings , a. c. , cummings , j. r. , leske , r. a. , mewaldt , r. a. , stone , e. c. , and von rosevinge , t. t. 1993 , astrophys .
j. , 418 , l45 steinacher , m. , jost , f. , and schwab , u. 1995 , rev .
, 66 , 4180 zurbuchen , t. , bochsler , p. , and scholze , f. 1995 , opt .
eng . , 34 , 1303
|
we present first results on the ca isotopic abundances derived from the high resolution mass time - of - flight ( mtof ) spectrometer of the charge , element , and isotope analysis system ( celias ) experiment on board the solar and heliospheric observatory ( soho ) .
we obtain isotopic ratios @xmath0ca/@xmath1ca @xmath2 ( @xmath3 ) and @xmath0ca/@xmath4ca @xmath2 ( @xmath5 ) , consistent with terrestrial values .
this is the first in situ determination of the solar wind calcium isotopic composition and is important for studies of stellar modeling and solar system formation since the present - day solar ca isotopic abundances are unchanged from their original isotopic composition in the solar nebula .
|
[ intro_sec ] a frequently found motif in biological networks , in particular in genetic networks , is the combination of a positive feedback loop in which one species ( @xmath0 ) activates itself , and a negative feedback loop , in which the first species activates its own repressor , the second species ( @xmath1 ) . in connection with genetic systems
such motifs are realized in the camp signalling system in the slime mold dictyosthelium discoideum @xcite , in the embryonic division control system @xcite , the mapk - cascade @xcite , or in the circadian clock @xcite . from the viewpoint of theoretical physics one
is interested in dynamical features that are in common to the many different dynamical realizations of this motif .
one common feature is certainly the occurrence of regular oscillations and the possibility of excitable behavior for an appropriate choice of parameters , and these features are captured already by a deterministic description in form of the bistable frustrated unit , considered in @xcite and references therein
. it should be emphasized , however , that the different realizations do not only differ by the biological systems in which they are realized , but also by the degree in which this representation as two coupled loops as shown in fig .
[ fig1 ] , activating also its own repressor @xmath1 . pointed arrows denote activation , blunt arrow denotes repression.,width=264 ] amounts only to an effective description .
the effective description should be seen in contrast to a one - to - one mapping of the participating ingredients . in principle
a number of intermediate steps may be included in these loops , and these intermediate steps need not be on the same level in case of a hierarchical organization .
they could amount to reactions between genes leading to production rates on the level of proteins , or on the same level , just to reactions between proteins , and it may make a big difference if protein @xmath0 is repressed via the binding of a transcription factor of type @xmath1 to gene @xmath2 ( design i ) , or via a direct repression of protein @xmath0 via @xmath1 ( design ii ) , as it was emphasized in @xcite .
there it was shown that the distinct set of equations , corresponding to the two designs , lead to different oscillatory features and different behavior with respect to noise and external periodic signals , so that the motif in the first design acts as an integrator " of external stimuli , in the second design as a resonator " . for a coarse - grained description that should reflect the essential dynamical implications of this motif
, the question therefore arises when it is actually justified to subsume the intermediate steps to effectively effective single steps and to represent the whole system by a set of two ordinary differential equations for the concentrations of @xmath0 and @xmath1 .
in particular such a description implies some kind of large - volume limit , in which the stochastic fluctuations of demographic origin , that is in the number of individuals , are ignored .
in @xcite we have analyzed the effect of demographic fluctuations and of fluctuations in the reaction times in simple realizations of this motif , that is , without any intermediate steps .
the stochastic fluctuations in the number of reactants and in the reaction times led to the occurrence of so - called quasi - cycles in parameter regimes for which the system would be deeply in the fixed - point regimes in the deterministic limit . on the other hand , the three regimes of the deterministic limit could be clearly recognized : as function of one bifurcation parameter @xmath3 , the maximal rate of production of @xmath0 for full activation and no repression , the dynamics of the coupled species @xmath0 and @xmath1 converges to a fixed point and shows excitable behavior in a first phase . in a second phase , regular limit - cycle behavior is observed , and in a third phase , again a fixed point with excitable behavior is seen . the quasi - cycles which are only due to the finite system size ,
are more regular in the transition regions than far off in the fixed point phases , they can be distinguished from regular oscillations by the decay of their autocorrelation functions .
when this motif is supposed to describe a genetic circuit , a possible zoom into the dynamical details would amount to a description in terms of proteins @xmath0 and @xmath1 , their associated mrna at an intermediate level and two types of genes a and b , respectively .
protein production of type @xmath0 would result from transcription factors of type @xmath0 , activating gene @xmath4 which is transcribed to the corresponding mrna that leads to the translation to proteins @xmath0 . at the same time
, protein @xmath0 leads to an activation of gene @xmath2 via binding of the transcription factor @xmath0 to gene @xmath2 , which is then transcribed to mrna of type @xmath2 , and when translated to proteins leads to a repression of the production of @xmath0 . in general , such intermediate steps as
the production of mrna and the binding and unbinding of transcription factors to the promoter region of genes induce additional time scales : delay of the protein production , and the switching rate between gene states . here
we distinguish the following gene states : states with bound activating transcription factor ( the gene being in the on - state " ) , states with bound repressing transcription factor ( the gene being in the off - state " ) , or no transcription factor bound ( which we call the gene being in the bare - state " ) . in principle
we could consider further situations such as the simultaneous binding of an activator and a repressor to the promoter region of the gene @xmath4 , but we leave out this option for simplicity . whether the delay time and the switching rate can be ignored compared to the time scales defined by the protein decay rates of @xmath0 and @xmath1 is a matter of relative size . in addition , the excitable behavior with large excursions in phase space of the original motif is related to the pronounced difference in the decay times of the proteins : @xmath0 is considered as the fast variable and @xmath1 as the slow variable , similarly to the distinction of fast and slow variables in a fitzhugh - nagumo element @xcite .
( due to this inherent difference , the oscillations become more spiky , the excursions in phase space from a fixed point more pronounced , and the description more realistic if we think of neurons realizing these circuits . )
therefore , when comparing the switching rate of gene states , we should specify with respect to which decay rate ( that of @xmath0 or of @xmath1 ) it is fast or slow . in the following we call the switching rates fast " if they are fast in comparison to both decay rates , of @xmath0 and of @xmath1 , slow " if the switching rate is of the order of the fast decay rate of @xmath0 , and ultra - slow " if it is of the order of the slow decay rate of @xmath1 . in former related studies of this genetic circuit
@xcite the switching rates of genes were assumed to be so high that their effect could be assumed to average out in the sense that mrna and proteins see only average values of gene activation or repression . in this paper
we want to analyze the effect that the inherent time scale of binding and unbinding rates of transcription factors has on the coarse - grained modeling . in order to project on this effect ,
we neglect the intermediate step of mrna production , but vary the binding / unbinding rates to values which are no longer high , but of the same order as the decay times of either the fast " protein @xmath0 or of the slow " protein @xmath1 . from results for genetic switches ,
in particular for the toggle switch @xcite , a simpler system than ours with only two mutually repressing genes , it is known that additional fixed points may show up for ( what we call ) slow genes .
more precisely , the joint probability distribution of the number of proteins in the cell and the dna - binding site state ( being on or off ) have additional peaks , corresponding to additional fixed points in the deterministic limit , if the switches are studied in the nonadiabatic limit .
if the switching rate between the genes is slow enough , the remainder of the system can follow the corresponding gene state of being either on or off .
the dynamics of our system is more versatile .
in addition to the fixed - point regime in the deterministic limit , we have the regime of regular limit - cycles , and the decay time of proteins differs by a factor of the order of hundred .
so independently of the possibility to realize our system in synthetic genetic circuits or to find it in natural systems , we want to focus on the impact of slow " and ultra - slow " components on the system s dynamics , in particular on the phase structure .
we start from a fully stochastic description in terms of master equations describing reactions directly between genes and proteins , skipping the intermediate mrna - production steps .
we simulate these equations via the gillespie algorithm @xcite . in view of a coarse - grained description in the form of deterministic equations
we take appropriate averages of the master equations which are adapted to the various limiting cases .
these averaging procedures correspond to taking the zeroth order of a van kampen expansion @xcite of the master equations , where the expansion parameter is the inverse system size .
the phase structure then differs for the resulting sets of deterministic equations , it depends on the competing time scales of genes and proteins . for fast genes
we recover the previously observed structure of three regimes , two fixed - point and one limit - cycle regime .
for this case we provide a detailed bifurcation analysis . for slow and ultra - slow genes ,
the regime of regular limit cycles is absent .
instead , the fast proteins @xmath0 see ( certain combinations of ) distinct states of gene expression separately rather than their averages : in the deterministic limit they therefore approach the fixed points corresponding to these states and stay there , unless the gene states change ; in the stochastic realization they switch between these states , so that the probability density to find @xmath5 proteins of type @xmath0 and @xmath6 proteins of type @xmath1 shows as many peaks as there are gene states distinguished .
in contrast , the inherent dynamics of the proteins @xmath1 is still so slow that these proteins can not reach the vicinity of these fixed points in the stochastic realization which they would approach in the deterministic description , leading to broad peaks or almost uniform probability densities .
the subsequent sections of the paper are organized as follows . in sect .
[ secii ] we present the model of interacting genes and proteins for the motif of a positive self - activating feedback loop coupled to a negative feedback loop with a second species repressing the first one .
the model is described in terms of chemical reactions and master equations . in sect .
[ seciii ] we derive deterministic models along with a stability analysis and a comparison with gillespie simulations for three limiting cases : fast genes ( sect . [ seciii1 ] ) , slow genes ( sect .
[ seciii2 ] ) and ultra - slow genes ( sect .
[ seciii3 ] ) . the summary and conclusions are given in sect .
[ seciv ] , followed by an appendix in which we analyze in detail the two transition regions between fixed - point behavior and limit cycles .
in terms of biochemical reactions we consider the following realization of the motif of fig .
[ fig1 ] displayed in fig .
[ fig2 ] . with a realization via genes
@xmath4 and @xmath2 leading to the production of proteins @xmath0 and @xmath1 with rates @xmath7 and @xmath8 , respectively , depending on the bound transcription factors to the promoter region of @xmath4 and @xmath2 .
transcription factors @xmath9 bind with rate @xmath10 to the promoter region of @xmath4 and unbind with rates @xmath11 , respectively .
transcription factor @xmath0 also binds to the promoter region of @xmath2 with rate @xmath12 and unbinds with rate @xmath13 .
proteins @xmath0 and @xmath1 decay with rates @xmath14 and @xmath15 , respectively . for further explanations
see the main text.,width=377 ] [ fig2 ] proteins @xmath0 and @xmath1 are produced under different conditions on the expression level of genes , but we assume in all cases that the production is proportional to the system size .
the system size is parameterized by a factor @xmath16 . for protein
@xmath0 we distinguish three situations : ( i ) an activating transcription factor @xmath0 is bound to the promoter region of gene @xmath4 , so that gene @xmath4 is in the on - state and produces proteins @xmath0 accordingly with rate @xmath17 ; here the superscript @xmath4 and subscript @xmath18 indicate that gene @xmath4 is responsible for the production of prbbotein @xmath0 and is itself in the on - state due to the binding of the ( self-)activating transcription factor @xmath0 . (
ii ) no transcription factor is bound to the promoter region of gene @xmath4 , leading to a production of @xmath0 in the so - called bare state of the gene with rate @xmath19 .
( iii ) a repressing transcription factor @xmath1 is bound to the promoter region of gene @xmath4 and turns the gene to the lower expression level , so that we term the gene state to be off " and the production rate of protein @xmath0 proceeds with rate @xmath20 , accordingly . for simplicity
we leave out a further possible situation that an activating and a repressing transcription factor @xmath0 and @xmath1 , respectively , are simultaneously bound to the respective promoter regions of gene @xmath4 , leading to a conflicting input of activation and repression . since protein @xmath1 is only activated , but not repressed via @xmath0 , we distinguish here only two states : ( i ) an activating transcription factor @xmath0 is bound to the promoter region of gene @xmath2 , leading to a production of @xmath1 with rate @xmath21 , or , ( ii ) no transcription factor is bound , leading to a production rate of @xmath22 of protein @xmath1 .
moreover , protein @xmath0 decays with rate @xmath14 and protein @xmath1 with rate @xmath15 .
choosing @xmath15 much smaller than @xmath14 and @xmath23 much smaller than @xmath24 is a way to implement the different inherent time scales in the protein dynamics : that of @xmath0 is much faster than that of @xmath1 , the reason for why we call @xmath0 the fast variable and @xmath1 the slow variable also in this realization of the motif
. the decay rate of the fast protein @xmath0 sets our time scale , @xmath25 . throughout this paper
we choose @xmath26 , so that the slow protein @xmath1 lives by a factor of 100 longer than @xmath0 .
the reactions referring to production and decay of proteins are summarized in the following equations : @xmath27 next we come to the binding / unbinding reactions of transcription factors to the promoter region of genes .
we distinguish between dimer binding of @xmath0 and @xmath1 to gene @xmath4 ( corresponding to a hill coefficient of @xmath28 ) and monomer binding of @xmath0 to the promoter region of gene @xmath2 ( corresponding to a hill coefficient of @xmath29 ) .
( the hill coefficient provides a quantitative measure for the binding cooperativity . )
the corresponding binding reactions along with the unbinding ones are listed in eq .
( [ eq3 ] ) .
@xmath30 the notation is chosen as follows : the superscript in the binding / unbinding coefficients @xmath31 , @xmath32 indicate the gene whose promoter region is affected in the binding / unbinding event , the subscripts refer to the monomer ( one index ) or dimer ( two indices ) binding / unbinding of the transcription factors . in our gillespie simulations
we choose the effective binding rates and the corresponding unbinding rates to be of the same order such that @xmath33 we normalize the binding rates which are proportional to the number of proteins @xmath5 with the appropriate power in @xmath16 to make them approximately independent of the system size , assuming that @xmath34 .
our simulations have shown that in case of a ratio of binding and unbinding rates different from one , it is the smaller value of the binding and unbinding rates that determines the dynamics in the following sense : the implied changes affect only the location of the fixed points and the limit - cycle regime , but lead to no qualitative change . given now the common values for the binding and unbinding rates of transcription factors , we call the switching of gene states , induced by the binding and unbinding events , fast if these rates are much higher than the decay rate of the fast protein ( set to one ) , slow if it is of the order of the fast protein @xmath0 , and ultra - slow if it is of the order of the slow protein @xmath1 . as we shall argue in sect .
[ seciii1 ] , the @xmath0-protein production rate in the on - state , @xmath35 , is chosen as the bifurcation parameter , while the production rates in the other states of gene @xmath4 are kept fixed , such that @xmath36 is by an order of magnitude smaller than the usual values of @xmath35 , and @xmath37 is set to zero .
the production rates of the protein @xmath1 are also kept fixed and chosen by two orders of magnitude smaller than the corresponding production rates of protein @xmath0 to compensate for the hundred times longer lifetime of protein @xmath1 in the competing gain and loss terms in eq .
( [ eqdetfastgenes ] ) , see sect .
[ seciii1 ] .
our choice of parameters is summarized in tables [ table1 ] and [ table2 ] .
.fixed parameters . [ cols="^,^,^,^,^,^",options="header " , ] the reactions described by eqs.([eq4 ] ) and ( [ eq3 ] ) correspond to a set of master equations which tell us the change in time of the probability to find at time @xmath38 @xmath5 proteins of type @xmath0 and @xmath6 proteins of type @xmath1 , given that gene @xmath4 is in either of the three states @xmath39 and at the same time gene @xmath2 is either in the on - state @xmath40 , or in the bare - state @xmath41 .
this probability is denoted as @xmath42 .
the six master equations for @xmath42 , resulting from the six combinations of indices , can be summarized in matrix notation according to : @xmath43 if we introduce the following definitions : @xmath44 for the dimer binding factors , and @xmath45 for the dimer unbinding , where the argument in the bracket refers to the state of either gene @xmath4 or gene @xmath2 , referred to via @xmath46 or @xmath47 in the master equations , respectively . for monomer
binding / unbinding factors we define @xmath48 the first eight terms of the master equation for @xmath42 for all values of @xmath49 result from the production or decay of proteins , leading to gain or loss terms as follows : loss terms contributing to the change of @xmath42 are due to the production of protein @xmath0 with constant rate @xmath50 , or the decay of @xmath0 proportional to @xmath5 , and the production of protein @xmath1 with rate @xmath51 , or the decay of @xmath1 proportional to @xmath6 .
gain terms , on the other hand , are due to the production of @xmath0 from a state with @xmath52 proteins @xmath0 , and the production of @xmath1 from a state with @xmath53 proteins @xmath1 , or the decay of one protein @xmath0 from a state with @xmath54 proteins @xmath0 , or the decay of one protein @xmath1 from a state with @xmath55 proteins @xmath1 . the last six terms in eq .
( [ eqmaster ] ) , of which between four and six are different from zero , describe changes in the probability due to the binding and unbinding of proteins to the promoter regions of gene @xmath4 and @xmath2 .
for example , a positive contribution to the probability @xmath56 results from a binding of a dimer of proteins @xmath0 with rate @xmath57 , given that the system before the binding event has @xmath58 proteins with gene @xmath4 being in the bare state ; a negative contribution results from an unbinding of a dimer of @xmath0-proteins with rate @xmath59 , given that the system contains @xmath58 proteins while gene @xmath4 being in the on - state before the unbinding event .
the other terms are derived similarly .
the coarse - graining that we intend to achieve refers primarily to a coarse - graining in time , averaging over events on a time scale much shorter than the scale on which the effective description should hold .
( usually , along with the coarse - graining in time goes a coarse - graining in space , as it is familiar from other realms of physics , but this is not in our focus here , as we do not arrange the genetic circuit and its constituents in any spatial ordering . here
the coarse - graining in time leads in general to a reduction of variables that are needed to describe the system s dynamics , but the amount of reduction depends on the limits which are taken , as we shall see below . ) in order to derive coarse - grained descriptions in the form of deterministic equations for the protein concentrations @xmath60 and @xmath61 , we apply appropriate averaging procedures to the master equation ( [ eqmaster ] ) .
we assume that the states of the genes are independent of each other and of the number of proteins @xmath5 and @xmath6 .
therefore we factorize the probability @xmath42 according to @xmath62 with @xmath63 the probability of finding gene @xmath4 in the @xmath46-state and @xmath64 the probability of finding gene @xmath2 in the @xmath47-state , while @xmath65 is the probability of finding the respective protein numbers whatever states the genes are in .
we then consider expectation values @xmath66 where @xmath67 stands for the species @xmath0 or @xmath1 and the average represents a summation over all @xmath5 , @xmath6 values .
we have to postulate @xmath68 so that @xmath69 and @xmath70 moreover , using ( [ eq8 ] ) and ( [ eqns ] ) , we have @xmath71 and @xmath72 now we multiply the master equation ( [ eqmaster ] ) with @xmath5 and @xmath6 , respectively , and sum over all @xmath58 ; next we sum the average value of @xmath5 , @xmath73 only over all states of gene @xmath2 , since the change in @xmath5 is assumed to be independent on the states of gene @xmath2 , and in analogy @xmath74 only over all states of gene @xmath4 , respectively .
we then obtain @xmath75 where @xmath76 similarly , @xmath77 where @xmath78 from eq .
( [ eqaibj ] ) we obtain @xmath79 these equations determine the time dependence of the probability to find the system in any of the five different gene states . in a stationary state of the genes , the left - hand side of eqs .
( [ eq18 ] ) vanishes .
this leads to @xmath80 using @xmath81 .
if the genes change their state fast enough as compared to other time scales in the system , they will reach the stationary values of eq .
( [ eq19 ] ) before the other processes are completed ; therefore they may be used in equations like ( [ eq13][eq17 ] ) .
now we are prepared to discuss the different limiting cases of fast , slow and ultra - slow genes . in the limit of fast genes
we choose all binding and unbinding rates a hundred times larger than the decay rate of the fast protein @xmath0 , that is @xmath14 . in this limit
the proteins see only average values of gene expression patterns , averaged over the five gene states
. therefore , in this limit we sum eqs .
( [ eq13 ] ) , ( [ eq14 ] ) , ( [ eq15 ] ) to predict the time evolution of @xmath5 , and eqs .
( [ eq16 ] ) and ( [ eq17 ] ) for the time evolution of @xmath6 to obtain @xmath82 where we have defined @xmath83 , @xmath84 , and @xmath85 with @xmath86 referring to the respective gene and @xmath87 indicating the monomer or dimer binding of the transcription factors according to the chosen hill coefficient in the biochemical reactions .
let us first briefly compare the equations ( [ eqdetfastgenes1],[eqdetfastgenes ] ) with the deterministic equations formerly used to describe the bistable frustrated unit in @xcite @xmath88 where @xmath89 is the ratio of the half - life of @xmath90 to that of @xmath91 , that is @xmath92 . in these units , the parameter @xmath93 sets the strength of repression of @xmath90 by @xmath91 .
the parameter @xmath2 determines the basal expression level of @xmath0 , @xmath94 .
the parameter @xmath3 is the maximal rate of production of @xmath0 for full activation ( @xmath95 ) and no repression ( @xmath96 ) . if we divide eqs .
( [ eqdetfastgenes1 ] [ eqdetfastgenes ] ) by @xmath14 and define @xmath97 , @xmath98 , similarly for the other @xmath99-parameters , we have @xmath100 in the previous model we used @xmath3 as a bifurcation parameter which multiplies @xmath101 in the gain term of eq .
( [ eqfirst ] ) , a similar role in eq .
( [ eq1a ] ) is played by @xmath102 which we therefore use here as the bifurcation parameter . differently from our former parametrization , apart from the common prefactor @xmath92 , that sets the slow time scale of @xmath91 , the gain term in the second equation ( [ eqdetfastgenesrescaled ] ) implicitly depends on @xmath103 , compared to the loss term .
therefore to align the scale of production with the slow decay , we have to adjust the production by choosing @xmath23 each by two orders of magnitude smaller than the corresponding production rates of @xmath104 which explains our choice in table [ table1 ] .
so the slow dynamics of protein @xmath1 is realized by both slow decay and slow production rate on the genetic level .
moreover , it should be noticed that we have changed the power of the hill coefficient in the binding term of the repressor concentration @xmath91 from one in eq .
( [ eqfirst ] ) to two in eq .
( [ eq1a ] ) corresponding to the choice of @xmath105 in the master equation ( [ eqmaster ] ) .
this appears as a minor difference in the equations .
the effect , however , is a considerable broadening of the intermediate limit cycle regime in case of a hill coefficient of @xmath28 .
since we are interested in the fate of the regular oscillations in case of slow and ultra - slow genes , it is important not to need a finetuning for seeing oscillations for fast genes .
furthermore , we would like to compare our equations ( [ eq1a],[eqdetfastgenesrescaled ] ) with the deterministic equations as they were derived for
design i " in @xcite . in common with those equations of @xcite is the limit of fast genes and the realization of the repression operating on the transcriptional level .
the main differences between both sets of equations are firstly the power one of @xmath90 in eq .
( [ eqdetfastgenesrescaled ] ) , which can be traced back to the monomer ( rather than dimer ) binding of the transcription factor @xmath0 to the promoter region of gene @xmath2 in eq .
[ eqmaster ] , and secondly , the relative weights between gain and loss terms . in particular
the bifurcation parameter affects in our case only the first equation directly and the second equation indirectly via @xmath90 , while it affects both equations directly in @xcite .
the combination of these apparently minor differences leads to different bifurcation patterns : a saddle - node bifurcation in @xcite and hopf bifurcations in our case . when increasing our bifurcation parameter @xmath35 , we see two fixed - point regimes for low and high values of @xmath35 , separated by an intermediate limit - cycle regime due to two hopf bifurcations . therefore , the deterministic equations ( [ eq1a],[eqdetfastgenesrescaled ] ) reproduce the phase structure of the bistable frustrated unit
. naively one may expect that the actual bifurcation scenarios in the deterministic limit are irrelevant for the final stochastic systems .
it is , however , known from the work of @xcite in the context of neural networks and also emphasized in @xcite that the very bifurcation scenario may have a strong impact on the final biological function of the motif , as the very onset of oscillations and the embedding in phase space have an impact on amplitude , frequency , noise resistance and other stability properties .
therefore we present a detailed bifurcation analysis of eqs .
( [ eq1a],[eqdetfastgenesrescaled ] ) in the appendix .
as it is seen there , the analysis requires a further zoom into the two transition regions , that is , a high resolution and finetuning of the bifurcation parameter .
it would be interesting to search for manifestations or remnants of these scenarios in a fully stochastic description , of which we studied so far the gross features only : noisy " fixed points and noisy " limit cycles . ,
@xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 ,
respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure
we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245]@xmath113@xmath113 , @xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 , respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245 ] , @xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 , respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure
we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245]@xmath113@xmath113 , @xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 , respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure
we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245 ] , @xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 ,
respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure
we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245]@xmath113@xmath113 , @xmath106 , @xmath107 .
phase portraits of the number of proteins @xmath6 versus @xmath5 within gillespie time @xmath108 ( left column ) and corresponding probability density functions ( pdf ) ( right column ) of @xmath5 ( full line black ) and @xmath6 ( dashed line black ) , while the gray full and dashed vertical lines indicate the position of the fixed points in the first and third row . in the first and third row @xmath109 and @xmath110 , respectively , in the left panels we see the stochastic pendant of the fixed points observed in the deterministic case . for clarity of the figure
we do not plot every gillespie step , but only @xmath111 of them .
the maxima of the pdfs agree well with the fixed points in the deterministic description . in the second row
we see the stochastic version of limit cycles for @xmath112 .
the vertical lines here mark the maximal and minimal extension of the limit cycles in @xmath90 and @xmath91 when integrated as solutions of the deterministic equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) .
these plots confirm our former model ( [ eqfirst]),([eqsan ] ) as a suitable coarse - grained description.,title="fig:",width=245 ] we present results of gillespie simulations of the reactions , listed in eqs .
( [ eq4 ] ) and ( [ eq3 ] ) for parameter values as in table [ table1 ] and the first row of table [ table2 ]
. figure [ fig11 ] , left column , shows phase portraits of @xmath6 versus @xmath5 for three values of the bifurcation parameter , @xmath109 ( first row ) , @xmath112 ( second row ) and @xmath110 ( third row ) , which are typical for the lower ( @xmath109 ) and higher ( @xmath110 ) fixed - point regime " , and for the limit - cycle regime " ( @xmath112 ) .
the phase portraits are made within a gillespie time of @xmath114 , where @xmath115 refers to the time interval , randomly chosen out of a poisson distribution in the @xmath46-th gillespie update .
the regimes are termed after their deterministic pendants : in the deterministic limit ( @xmath116 ) the clouds of @xmath58 values in the first and third row would shrink to the lower and higher fixed points as predicted from eqs .
( [ eqdetfastgenes1],[eqdetfastgenes ] ) , while the clouds in the second row would contract to a limit cycle .
( the higher density of @xmath117-values for small and large values of @xmath5 is due to the fact that also the stochastic version of a limit cycle spends more time in regions , where @xmath6 drastically changes , since @xmath1 is the slow variable , while large changes in @xmath5 happen rapidly , since @xmath0 is the fast variable . )
the probability density functions in the right column of the figures reflect the probability for finding concrete combinations of @xmath117-values in the phase portraits .
they are displayed for a quantitative comparison of their maxima with the prediction of the location of the fixed points and the extension of the limit cycle from the deterministic equations .
these locations are indicated via the vertical lines .
the vertical lines hit the maxima quite well where they correspond to fixed points ; the lines also match the typical extension of the cloud in case of the limit cycle .
figure [ fig10 ] ( upper part ) shows the time series of the number of proteins @xmath0 , @xmath5 ( gray ) and of proteins @xmath1 , @xmath6 ( black ) , which are fluctuating about constant values @xmath118 .
the fluctuations between the different gene states , in particular between the on- and bare - states , is so fast that it appears as a gray and black band ( fig .
[ fig10 ] lower part ) , so that the protein values , shown in the upper part of the figure , only fluctuate about the fixed point values .
, @xmath119 ( gray ) , @xmath120 ( black ) ( upper part of the figure ) and the gene states for gene @xmath4 ( gray ) and gene @xmath2 ( black ) ( lower part of the figure ) , recorded during the gillespie steps .
the bifurcation parameter is chosen as @xmath110 .
states of gene @xmath4 switch between @xmath18 , [ @xmath121 ( gray ) ] , @xmath122 , [ @xmath123 ( white ) ] , @xmath124 , [ @xmath125 ( gray ) ] , states of gene @xmath2 switch between @xmath18 , [ @xmath126 ( black ) ] and @xmath122 , [ @xmath127 ( white ) ] .
the other parameters are chosen as @xmath128 , @xmath129 , @xmath130 , @xmath131 , @xmath132 = @xmath133 , @xmath134 , @xmath135 @xmath136 @xmath137 @xmath138 , @xmath139 , @xmath140.,width=302 ] in the gillespie simulations of our former realization of the genetic circuit @xcite we identified quasi - cycles deeply in the fixed - point regimes .
such cycles , caused by large demographic fluctuations , are also found in the present realization of the genetic circuit if we wait sufficiently long for such fluctuations to happen .
figure [ fig12 ] shows a few such excursions below ( upper left figure for @xmath141 ) and above ( lower figures for @xmath142 and @xmath143 ) the limit - cycle regime , while the data in the upper right figure for @xmath144 are compatible with more frequent and smaller excursions into phase space due to the vicinity of the transition region to the region of regular limit cycles . for fast genes for four values of the bifurcation parameter .
@xmath141 ( upper left ) , @xmath144 ( upper right ) , @xmath142 ( lower left ) , @xmath145 ( lower right ) .
in the lower ( @xmath146 ) and higher ( @xmath147 ) fixed point regimes we see clear indications of quasi - cycles in the stochastic descriptions.,title="fig:",width=245]@xmath113@xmath113 for fast genes for four values of the bifurcation parameter . @xmath141 ( upper left ) , @xmath144 ( upper right ) , @xmath142 ( lower left ) , @xmath145 ( lower right ) . in the lower ( @xmath146 ) and higher ( @xmath147 ) fixed point
regimes we see clear indications of quasi - cycles in the stochastic descriptions.,title="fig:",width=245 ] for fast genes for four values of the bifurcation parameter . @xmath141
( upper left ) , @xmath144 ( upper right ) , @xmath142 ( lower left ) , @xmath145 ( lower right ) . in the lower ( @xmath146 ) and higher ( @xmath147 ) fixed point
regimes we see clear indications of quasi - cycles in the stochastic descriptions.,title="fig:",width=245]@xmath113@xmath113 for fast genes for four values of the bifurcation parameter . @xmath141
( upper left ) , @xmath144 ( upper right ) , @xmath142 ( lower left ) , @xmath145 ( lower right ) . in the lower ( @xmath146 ) and higher ( @xmath147 ) fixed point
regimes we see clear indications of quasi - cycles in the stochastic descriptions.,title="fig:",width=245 ] , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 . here
the switching between @xmath18 and @xmath122 states is much faster than the indirect switching between @xmath18 and @xmath124 states .
the other parameters are as in fig .
[ fig10 ] . ] in the limiting case of slow genes , the binding and unbinding rates are chosen to be of the order of the decay rate of the fast protein , so that they are still fast as compared to the decay rate of the slow protein @xmath1 .
therefore we sum the moments of @xmath6 from eqs .
( [ eq16]),([eq17 ] ) also over the states of gene @xmath2 .
moreover we have three equations ( [ eq13]),([eq14]),([eq15 ] ) that include already a summation over the states of gene @xmath2 , but in order to further sum over states of gene @xmath4 , we consider fig .
[ fig13 ] , which shows the switching events of gene @xmath4 between the three states @xmath151 and of gene @xmath2 between the two states @xmath18 and @xmath122 .
the switching events between the on- and the bare - state , determined by @xmath152 , @xmath153 , as well as the switching between the @xmath122- and the @xmath124-state , determined by @xmath154 , @xmath155 , are still frequent as compared to an indirect switching between an @xmath18- and an @xmath124-state of gene @xmath4 , since it is due to different binding events .
therefore the proteins @xmath0 and @xmath1 effectively see two gene levels , an average over the on- and bare - state , or over the bare- and off - state .
these two different levels are seen in the time evolution of @xmath5 : @xmath5 switches between the upper gray band in fig .
[ fig13 ] , that is between @xmath156 and @xmath157 , and the lower gray band close to zero .
the adaptation of @xmath6 to the abrupt changes in @xmath5 is more smooth and happens with delay .
( as we shall later see , the fast proteins @xmath0 are able to adapt to these two states , while the slow ones @xmath1 are not . )
accordingly we average eq .
( [ eq13 ] ) with eq .
( [ eq15 ] ) and eq .
( [ eq14 ] ) with eq .
( [ eq15 ] ) to obtain two sets of differential equations @xmath158 and @xmath159 which describe the two alternatives of the time evolution of the system , either following the first or the second set of equations .
while the time dependence of @xmath74 is the same in both equations ( [ eq24b ] ) and ( [ eq25b ] ) , the time dependence of @xmath73 depends on whether it is due to the activating binding of a dimer of @xmath0 or the repressing binding of @xmath1 .
this apparently minor detail leads to a very different bifurcation pattern , as it is revealed by a linear stability analysis . for eqs .
( [ eq25a]),([eq25b ] ) this analysis shows that the determinant of the jacobian , evaluated at the single fixed point , is positive for all combinations of positive values of parameters and @xmath160 and @xmath161 , while the trace of the jacobian , evaluated at the fixed point , is negative , so the fixed point is always stable .
in contrast , for the first set of equations , eqs .
( [ eq24a]),([eq24b ] ) , we have for our usual choice of parameters ( used in the gillespie simulations ) a single fixed point , but there is a combination of parameters for a very low value of @xmath36 , for which we obtain three positive real fixed points ; for example choosing @xmath112 , @xmath162 and the other parameters as before , we obtain two stable nodes with positive determinant , negative trace and negative eigenvalues at @xmath163 , and at @xmath164 , and a saddle with negative determinant , positive trace , and @xmath165 , @xmath166 at @xmath167 . depending on the initial conditions ,
the system will evolve in one of the stable nodes .
gillespie simulations for this parameter set are not displayed in fig .
[ fig14 ] below .
the basin of attraction of the lower fixed point is so small that the stochastic system practically never evolves to this fixed point due to the demographic fluctuations .
while for fast ( ultra - slow ) genes , gene switches are fast ( slow ) with respect to both proteins , for slow genes protein @xmath0 sees slow switches between the @xmath168 and @xmath169 average values , so that it can follow these different states of gene @xmath4 , while protein @xmath1 is intrinsically so slow that it still sees averages over all states of the genes . therefore inserting the separate average values of @xmath73 in eqs .
( [ eq24b]),([eq25b ] ) fails as an effective description .
the phase portraits of @xmath58 for slow genes in the left column of fig .
[ fig14 ] shows the pendant of one fixed point of eq .
( [ eq25a ] ) ( upper row ) , while the fixed point of eq .
( [ eq24a ] ) is not visible ( two fixed points each in the middle and lower row of the figure ) , where the left cloud corresponds to the solution of eq .
( [ eq25a ] ) and the right cloud to eq .
( [ eq24a ] ) . in particular we interpret the clouds of events in the second row also as a stochastic switching between two fixed points rather than a noisy version of limit cycles , since in contrast to the corresponding phase portrait in fig .
[ fig11 ] there is no empty space between large and small @xmath5 values , the events jump from the left - to the right side rather than performing full cycles , in agreement with the time series of @xmath5 in fig .
[ fig13 ] .
the vertical lines in fig .
[ fig14 ] indicate the two stable fixed points of eqs .
( [ eq24a]),([eq25a ] ) in all three cases .
the second fixed point of eq .
( [ eq24a ] ) is not visible for @xmath109 ( first row ) , but it is seen for @xmath112 and @xmath110 , where both fixed point locations fit well the maxima of the pdfs for @xmath5 , though not for @xmath6 . the failure is due to the inherent difficulty that protein @xmath1 sees averages over all @xmath2-states , leading to values of @xmath6 which are not solutions of eqs .
( [ eq24b]),([eq25b ] ) ( derived under the condition on protein @xmath0 to see @xmath5 states as two distinct averages over ( @xmath168 ) and @xmath169 ) states ) .
@xmath1 is still too inert to follow the two distinct @xmath73-values . , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle .
for further explanations we refer to the main text.,title="fig:",width=245]@xmath113@xmath113 , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle .
for further explanations we refer to the main text.,title="fig:",width=245 ] , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle . for further explanations we refer to the main text.,title="fig:",width=245]@xmath113@xmath113 , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle .
for further explanations we refer to the main text.,title="fig:",width=245 ] , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle . for further explanations we refer to the main text.,title="fig:",width=245]@xmath113@xmath113 , but for slow genes , that is @xmath148 , @xmath149 , @xmath150 .
the two vertical lines in each species in the pdfs indicate the prediction of two stable fixed points .
this is in agreement with the stochastic results only for @xmath110 ( third row ) . for @xmath109 ( first row )
we see the stochastic pendant of a single fixed point only in disagreement with the prediction from the deterministic model .
the interpretation of the phase portrait in the second row remains ambiguous : the second local maximum in the pdf seems to be a precursor of the second fixed point , and the interpretation of the phase portrait in terms of a limit cycle is less likely due to the absence of a white area which should not be visited by the gillespie trajectory in case of a limit cycle . for further explanations we refer to the main text.,title="fig:",width=245 ] independently of the possibility to realize this limit in natural or synthetic genetic circuits
, it is of interest from the dynamical point of view what the effective coarse - grained description in the deterministic limit amounts to in case of ultra - slow genes .
this limit refers to a situation , in which the binding / unbinding rates of transcription factors are of the order of the slow protein @xmath1 .
so the time which genes @xmath4 and @xmath2 spend in one of their possible states is long as compared to @xmath170 , the lifetime of protein @xmath0 . , but for ultra - slow genes , that is @xmath171 , @xmath172 , @xmath173 .
protein @xmath0 sees gene @xmath4 in three different states , while there are periods during which protein @xmath1 still sees averages over @xmath18 and @xmath122- states , here visible in the first two blocks , where gene @xmath2 switches frequently between @xmath18 and @xmath122 . , width=340 ] it then does no longer make sense to further sum eqs .
( [ eq13])([eq15 ] ) over any states of gene @xmath4 and eqs .
( [ eq16]),([eq17 ] ) over any states of gene @xmath2 , but we are left with these five equations , which can be summarized as @xmath174 with @xmath84 , and @xmath175 for @xmath119 and @xmath176 for @xmath120 , respectively . if we insert for @xmath177 eq . ( [ eq18 ] ) and solve ( [ eq18a ] ) for @xmath178 , we arrive at the following set of uncoupled differential equations for @xmath179 @xmath180 with solutions that for @xmath181 exponentially decay to the fixed points @xmath182 with @xmath183 , @xmath184 for @xmath185 , and @xmath186 for @xmath187 , leading to six fixed points . in the stochastic realization of this limit of ultra - slow genes
we expect the system to switch between three possible states with respect to @xmath5 and two with respect to @xmath6 , so between six fixed points in the deterministic limit .
the former oscillations in the limit cycle regime are clearly gone .
for @xmath5 we see both in the phase portraits and in the probability density functions remnants of three distinct fixed - point values of @xmath5 , while the remnants of two possible fixed - point values of @xmath6 are only vaguely visible as two maxima in the probability distribution .
obviously the ultra - slow genes are still not slow enough to allow protein @xmath1 to adjust to the different states of gene @xmath2 . , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows ,
left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170]@xmath113@xmath113 , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows , left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170 ] , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows ,
left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170]@xmath113@xmath113 , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows ,
left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170 ] , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows ,
left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170]@xmath113@xmath113 , but for ultra - slow genes , that is that is @xmath171 , @xmath172 , @xmath173 . in all three rows ,
left column , we see remnants of three fixed points in the deterministic limit with respect to values of @xmath5 , while the values of @xmath6 are broadly spread , since @xmath1 is too slow to follow the different states of gene @xmath2 .
the vertical lines from the deterministic prediction of the fixed point values ( right column ) match well the maxima of the pdfs for @xmath5 and only roughly for @xmath6 .
the insets zoom into the @xmath117 values of the two fixed points for lower @xmath5-values.,title="fig:",width=245,height=170 ]
the motif of the self - activating species which activates its own repressor is found in many realizations in biological systems . from the physics point of view
one would like to identify universal features of the dynamical behavior .
certainly regimes of excitable and oscillatory behavior are common features found over a wide range of parameters . also in our current realization we have recovered three regimes of excitable , oscillatory and excitable behavior as function of one bifurcation parameter , which we have chosen as @xmath35 , the production rate of protein @xmath0 if gene @xmath4 is in the on - state . however , as we have pointed out , these three distinct regimes are only obtained in our realization if the binding and unbinding rates of genes are fast as compared to the other inherent time scales , here the decay rates of the fast @xmath188 and the slow @xmath189 proteins . for this case
we derived deterministic equations equivalent to the former ones of @xcite .
as soon as the binding / unbinding rates are no longer small , but of the same order as the decay time of either proteins , the averaging procedure for deriving a deterministic limit has to be changed ; the proteins see no longer average values of the gene states , but tend to follow the distinct states , unless their own production is too slow to reach the appropriate state in time " .
these features were demonstrated by our gillespie simulations of the six master equations .
for the ultra - slow genes they were also visible in the derived coarse - grained description in terms of five deterministic equations . for the intermediate case of binding rates of the order of the decay of the fast protein ,
the intrinsic difficulty to derive deterministic equations lies in the fact that one and the same binding and unbinding events are differently seen from the proteins : protein @xmath0 sees already distinct events and follows the according gene state , protein @xmath1 still sees an average over the gene states . in both cases , slow and ultra - slow genes ,
the intermediate regime of stable regular oscillations is absent .
one should keep in mind that we have to deal with systems of nonlinear dynamics
. therefore one should be aware that apparently minor changes such as the value of the hill coefficient may have pronounced effects .
even if the qualitative picture after such a minor change remains the same , the quantitative features like the extension of the limit - cycle regime can drastically change , as we have seen .
such a change can finally decide on the relevance of the model for the real biological system . for real systems stable oscillations
would probably not be observed if they were restricted to a tiny interval of the bifurcation parameter .
although we have chosen our parameters independently of their possible realization in concrete genetic circuits , the conclusion from our analysis is generic and applies also to real biological systems : it is not only the gross features of the topology of the motifs and the couplings that determine the dynamical performance .
if different time scales are inherent , the gross bifurcation patterns may depend on their ratios .
we would like to thank ashok garai ( uc san diego ) for his participation in the beginning of this work .
two of us ( dl and hmo ) would also like to thank michael zaks ( humboldt university berlin ) for stimulating discussions .
financial support from the dfg ( ( dl , grant no .
me-1332/17 - 1 ) , ( hn , grant no .
ja 483/27 - 1 ) ) is gratefully acknowledged .
the appendix is devoted to a detailed bifurcation analysis of eqs .
( [ eqdetfastgenes1 ] ) , ( [ eqdetfastgenes ] ) . in fig .
[ fig3 ] we plot the real - and imaginary part of the two eigenvalues of the jacobian @xmath190 , evaluated at the respective fixed points of these equations .
we display only the negative imaginary part to keep the figure clear . in the following
we describe the stationary states as a function of increasing parameter @xmath35 .
the labels assigned to the specific values of @xmath35 shall remind to the change in the dynamical performance occurring at these values ( so that spirals are created at sp " , hopf bifurcations at h " , two limit cycles collide at coll " and str " marks the boundaries of the tiny strange " interval in which two limit cycles coexist ) . .
to keep the figure clear we present only the negative imaginary part . for the meaning of the various indicated special values of @xmath35 we refer to the main text.,width=377 ] *
@xmath191 : one stable node .
+ in this interval we observe one fixed point in the form of a stable node . * @xmath192 : one stable spiral .
+ at @xmath193 the two real eigenvalues become complex conjugate of each other , and the stable node turns into a stable spiral , shown in fig .
+ , @xmath129 , @xmath130 , @xmath131 , @xmath132 = @xmath133 , @xmath134 , @xmath135 @xmath136 @xmath137 @xmath138 , @xmath139 , @xmath140 .
gray dashed and dashed - dotted lines show the nullclines of @xmath5 and @xmath6.,width=377 ] * @xmath194 : one unstable spiral and one small stable limit cycle . + at the hopf bifurcation point @xmath195 the real part of the eigenvalues crosses the abscissa and becomes positive , so that the stable spiral becomes unstable . from that value on , as long as @xmath196 , there is a small stable limit cycle with a radius growing with increasing @xmath35 , surrounding the unstable spiral .
trajectories spiral out of the fixed point towards the limit cycle and spiral extremely slowly when approaching the limit cycle , cf . fig .
[ fig5 ] . the first lyapunov coefficient @xmath197 , where the negative sign of @xmath198 shows that the hopf bifurcation is supercritical .
+ are chosen as in fig .
[ fig4].,width=377 ] + without a further zoom into the bifurcation parameter , one would conclude from the observation of a large stable limit cycle for @xmath199 ( see below ) that the small stable limit cycle for @xmath200 has continuously grown to this large size .
what happens instead is described next .
* @xmath201 : two coexisting stable limit cycles and one unstable spiral .
+ figure [ fig6 ] shows a snapshot in bifurcation parameter space , just after the second limit cycle is born .
the two black stripes in the inset consists of two limit cycles , a smaller and a larger one , and the white space in between is filled by the trajectories , independently of the initial conditions , as long as this space is sufficiently small .
+ are chosen as in fig .
[ fig4].,width=377 ] + the second limit cycle pops up just after the small stable limit cycle from fig .
[ fig5 ] has grown until @xmath202 .
the phase space between the smaller and the larger limit cycle acts like an attractor as it is seen in fig .
[ fig7 ] : trajectories , starting from initial conditions in this intermediate space ( two initial conditions are indicated by the gray and black dot ) follow limit cycles in the area between the small and the large limit cycle , indicated as gray dashed lines in the lower right part of the figure , see fig .
the zoom presented in the lower and upper right part shows two time snapshots up to time @xmath203 and @xmath204 , respectively , where @xmath205 denotes the total sum over elementary time intervals .
+ = 257.12248 and other parameters chosen as in fig . [ fig4 ]
. trajectories starting between the lower and upper limit cycle remain in this area bounded by the red ( gray ) dashed lines in the lower right part of the figure . , width=491 ] + the area between the limit cycles rapidly grows ( fig .
[ fig8 ] left ) , until only one large stable limit cycle remains ( fig .
[ fig8 ] right ) . + on , only one large limit cycle survives.,title="fig:",width=245 ] on , only one large limit cycle survives.,title="fig:",width=245 ] * @xmath206 : one large stable limit cycle and one unstable spiral .
+ from @xmath207 on only one large stable limit cycle is left along with the unstable fixed point which stays a spiral until @xmath208 . *
@xmath209 : one large stable limit cycle and one unstable node .
+ at @xmath210 , the imaginary part of the complex conjugate eigenvalues become zero , leaving two real different positive eigenvalues , so that the fixed point inside the large stable limit cycle has turned into an unstable node .
this situation holds over a large interval in @xmath35 , until the unstable node turns back into an unstable spiral , not displayed .
* @xmath211 : one large stable limit cycle and one unstable spiral .
+ the return to an unstable spiral happens at @xmath212 , where the eigenvalues become again complex conjugate .
the spiral remains unstable until the second hopf bifurcation happens .
* @xmath213 : one stable spiral , one small unstable limit cycle and one large stable limit cycle . + the second hopf bifurcation happens at @xmath214 , where the real part of the complex conjugate eigenvalues crosses the abscissa , the formerly unstable spiral changes into a stable one .
along with that a small unstable limit cycle is born , while the large stable limit cycle is still alive " .
so we have a stable spiral , surrounded by a small unstable limit cycle , surrounded by the large stable limit cycle ( not visible in fig .
[ fig9 ] ) , see fig .
+ are chosen as in fig .
[ fig4].,width=377 ] + this coexistence holds only for a tiny interval , as long as @xmath215 .
the reason for the small size of this interval is that the fixed point and the small unstable limit cycle are located close to the trajectory of the large stable limit cycle , so that a tiny increase of @xmath35 leads to a growth of the radius of the small unstable cycle that is sufficient for the small cycle to collide with the large cycle and this way to induce its cancelation .
the collision of the two limit cycles amounts to a saddle - node bifurcation of two periodic orbits with the result that the periodic orbits disappear .
this suggests the vicinity of a bautin bifurcation @xcite ; the bautin bifurcation is characterized by two bifurcation conditions : ( i ) @xmath216 and ( ii ) the critical first lyapunov coefficient @xmath217 .
we have in our interval @xmath218 that @xmath216 and @xmath219 at the hopf bifurcation point . *
@xmath220 : one stable spiral , no limit cycle anymore . +
after the stable and unstable limit cycles disappear at @xmath221 , there remains a stable spiral as long as @xmath222 .
* @xmath223 : one stable node .
+ at @xmath224 the two complex conjugate eigenvalues become real again with two different negative values , so that finally the stable spiral returns to a stable node and remains like that for increasing @xmath35 .
while the linear stability analysis reveals the change from a stable spiral to an unstable spiral , and an unstable node back to an unstable spiral , to a stable spiral and a stable node , the numerical integration of the equations ( [ eqdetfastgenes1]),([eqdetfastgenes ] ) shows subtleties in how the large stable limit cycle is born and destroyed .
the creation happens discontinuously and in coexistence with a small stable limit cycle , the annihilation very likely via a bautin bifurcation .
j. martiel , a. goldbeter , biophys .
j. * 52 * , 807 ( 1987 ) .
b. novak , j. j. tyson , j. cell sci . * 106 * , 1153 ( 1993 ) .
j. r. pomerening , s. y. kim , j. e. jr .
ferrell , cell * 122 * , 565 ( 2005 ) .
j. j. tyson , proc .
88 * , 7328 ( 1991 ) .
l. qiao , r. b. nachbar , i. g. kevrekidis , s. y. shvartsman , plos computational biology * 3 * , e184 ( 2007 ) .
j. m. g. vilar , h. y. kueh , n. barkai , s. leibler , proc .
u.s.a.*99 * , 5988 ( 2002 ) .
n. t. ingolia , a. w. murray , current biology * 14 * , r771 ( 2004 ) .
a. garai , b. waclaw , h. nagel and h. meyer - ortmanns , j. stat .
p01009 * ( 2012 ) .
r. guantes , j. f. poyatos , plos comp . biol .
* 2 * , 0188 ( 2006 ) .
r. fitzhugh , biophys .
j. * 1 * , 445 ( 1961 ) .
j. s. nagumo , s. arimoto , s. yoshizawa , proc .
ire * 50 * , 2061 ( 1962 ) .
a. m. walczak , m. sasai , p. g. wolynes , biophys .
j. * 88 * , 828 ( 2005 ) .
d. t. gillespie , j. phys .
chem . * 81 * , 2340 ( 1977 ) .
n. g. van kampen , _
stochastic processes in physics and chemistry _ , ( north holland , amsterdam , 2005 ) .
p. kaluza , h. meyer - ortmanns , chaos * 20 * , 043111 ( 2010 ) .
e. m. izhikevich , int .
j. bifurcation chaos appl .
* 10 * , 1171 ( 2000 ) .
y. a. kuznetsov , _ elements of applied bifurcation theory _ , ( springer , heidelberg , 1998 ) .
|
from a coarse - grained perspective the motif of a self - activating species , activating a second species which acts as its own repressor , is widely found in biological systems , in particular in genetic systems with inherent oscillatory behavior .
here we consider a specific realization of this motif as a genetic circuit , in which genes are described as directly producing proteins , leaving out the intermediate step of mrna production .
we focus on the effect that inherent time scales on the underlying fine - grained scale can have on the bifurcation patterns on a coarser scale in time .
time scales are set by the binding and unbinding rates of the transcription factors to the promoter regions of the genes .
depending on the ratio of these rates to the decay times of the proteins , the appropriate averaging procedure for obtaining a coarse - grained description changes and leads to sets of deterministic equations , which differ in their bifurcation structure .
in particular the desired intermediate range of regular limit cycles fades away when the binding rates of genes are of the same order or less than the decay time of at least one of the proteins .
our analysis illustrates that the common topology of the widely found motif alone does not necessarily imply universal features in the dynamics .
|
the investigation of nonlinear properties of josephson junctions ( jjs ) is very important due to their broad applications in logic devices .
recently , a lot of attention was payed to josephson logic devices with high damping . in papers [ rylyakov @xmath0 likharev , 1999 ; bunyk _ et al . _ ,
2001 ] a description and analysis of the entire system of single flux quantum logic elements are presented .
the noise properties of systems consisting of many elements may be well understood from the noise properties of a single jj .
the processes occurring in such devices are based on a reproduction of quantum pulses due to @xmath1 spasmodic change of the phase difference of the overdamped jjs .
the use of high-@xmath2 jjs creates many new problems , such as the thermally induced digital errors and the effect of spreading of the switching speed due to thermal fluctuations .
the analysis of influence of thermal fluctuations on different characteristics of jjs , such as the current - voltage characteristic or the life time of superconductive state , has been made before on the basis of langevin approach [ barone @xmath0 patern@xmath3 , 1982 ; likharev , 1986 ; ortlepp @xmath0 uhlmann , 2004 ; ortlepp @xmath0 uhlmann , 2005 ] . recently experimental and theoretical work has been done on the mean switching time or mean escape time from the metastable state in periodically driven josephson junctions [ yu @xmath0 han , 2003 ; pankratov @xmath0 spagnolo , 2004 ; peltonen_et al . _ , 2006 ; sun _ et al . _ , 2007 ] .
the process of switching of a single jj with high damping under periodic driving was considered in [ pankratov @xmath0 spagnolo , 2004 ] . in that paper
the influence of the resonant activation ( ra ) and the noise enhanced stability ( nes ) phenomena in the accumulation or suppression of timing errors in rapid single flux quantum ( rsfq ) devices was analyzed .
specifically an interval of frequencies was found where the switching time increases with increasing noise intensity .
this effect is called noise enhanced stability [ mantegna @xmath0 spagnolo , 1996 ; agudov @xmath0 spagnolo , 2001 ; dubkov , agudov @xmath0 spagnolo , 2004 ; fiasconaro _ et al . _ , 2005 ;
spagnolo _ et al .
_ , 2007 ] .
this paper presents the analysis of the effects of thermal fluctuations on the switching time for a single josephson element with high damping , within the well - known resistive model of a josephson junction [ likharev , 1986 ] .
the goal of this work is to investigate in detail how the nes phenomenon depends on the parameters of the periodical driving signal such as the amplitude and the frequency .
the dynamics of a short overdamped jj , under a current @xmath4 is given by the following langevin equation @xmath5 where @xmath6 is the phase difference of the order parameter [ barone @xmath0 patern@xmath3 , 1982 ] , @xmath7 is the dimensionless potential profile ( see fig . 1 . )
, @xmath8 is the characteristic frequency of the jj , @xmath9 is the critical current , @xmath10 is the normal conductivity of the jj , @xmath11 is the electron charge and @xmath12 , with @xmath13 the planck constant . here
@xmath14 is the total current across the junction , @xmath15 is the constant bias current , @xmath16 is the driving signal , @xmath17 , @xmath18 , and @xmath19 is the random component of the current . because of thermal fluctuations
, the random current may be represented by white gaussian noise @xmath20 where @xmath21 is the dimensionless intensity of fluctuations , @xmath22 is the temperature and @xmath23 is the boltzmann constant .
initially , the jj is biased with a current across the junction smaller than the critical one , that is @xmath24 , so as the initial condition we take the location of the phase in a potential minimum , @xmath25 . a current signal @xmath16 , such that @xmath26 , switches therefore the junction into the resistive state . in fig . 1 we show the periodical potential profile of the jj and its extreme positions
within which it varies in time .
the switching occurs not immediately , but at the later time , which is called the switching time .
due to the noise the switching time is a random quantity .
we investigate therefore the mean switching time ( mst ) @xmath27 and its standard deviation ( sd ) @xmath28 . as a driving signal
we choose a sinusoidal signal @xmath29 , where @xmath30 is the oscillation frequency and @xmath31 is the signal amplitude . according to the definition [ malakhov @xmath0 pankratov , 2002 ] the mean switching time @xmath27 and its standard deviation @xmath28 are @xmath32 where @xmath33
, @xmath34 is the probability to find @xmath6 within the interval @xmath35 .
( 220,150)(10,10 ) in fig .
2 the mst is shown versus the signal frequency @xmath30 for @xmath36 , @xmath37 and for different values of the noise intensity , namely @xmath38 . a frequency range , from @xmath39 to @xmath40 , where the mst increases by increasing the noise intensity is clearly visible . in all the figures
the mst @xmath27 and the frequency @xmath30 are normalized to @xmath41 and @xmath42 , respectively .
( 220,150)(10,10 ) in fig . 3 we report the same quantity of fig . 2 , that is the mst versus the signal frequency , for a smaller signal amplitude @xmath43 and different noise intensities , namely @xmath44 , and @xmath45 .
we see that , the frequency range where the nes effect is observed is reduced , and it is shifted towards smaller values of the driving frequency @xmath30 , namely from @xmath46 to @xmath39 .
( 220,150)(10,10 ) in both figs . 2 and 3 , the mst has a minimum as a function of the driving frequency .
this is the signature of the resonant activation phenomenon , investigated in a jj in refs .
[ yu @xmath0 han , 2003 ; pankratov @xmath0 spagnolo , 2004 ; sun _ et al .
_ , 2007 ] .
the enhancement of the switching time may be considered in detail by plotting the dimensionless time mst as a function of the noise intensity for different values of the signal frequencies .
different behaviors of mst , depending on the values of the signal frequency , occur for physical systems with metastable states [ agudov @xmath0 spagnolo , 2001 ; dubkov , agudov @xmath0 spagnolo , 2004 ] .
specifically a nonmonotonic behavior and a monotonic one may be observed .
let us start our consideration from small to large frequencies for the case @xmath36 and @xmath37 , see fig . 2 . for frequencies smaller than 0.2
the character of the curves is similar : the mst decreases monotonically with increasing noise intensity .
this case is presented in fig . 4 by two values of frequency : @xmath47 and 0.1 .
if the value of the noise intensity becomes greater or equal to the height of the barrier , which is approximately equal to @xmath48 in the middle configuration ( see fig . 1 ) , the particle does not see the fine structure of the potential .
this is the reason why we restrict to @xmath49 the values of the dimensionless noise intensity in the plots ( figs .
so , we can see that the mst decreases , with respect to the level of deterministic switching time ( i.e. for @xmath50 ) , by increasing the noise intensity .
the explanation of this behavior is the following . for small noise intensity and very low frequencies ,
the particle lies in the minimum for a long enough time due to the slow variation of the potential profile .
the switching from the superconductive state to the resistive one does not occur until the potential barriers disappear almost completely . by increasing the noise intensity ,
the probability of the thermal activated switching increases and as a result the mst decreases .
the next interval of frequencies which we consider is from 0.2 to , approximately , 0.52 , because for such frequencies the behavior of the curves is similar . in this frequency range
the nes effect appears , the mst(@xmath51 ) - curve behaviour becomes nonmonotonic in contrast with the previous case ( @xmath52 ) . in fig . 4 this behavior
is shown for three frequency values , namely @xmath53 , and @xmath40 .
we note that in these three curves , after the nonmonotonic behavior , the mst increases again with the noise intensity @xmath51 ( for @xmath54 ) . due to the increasing values of the signal frequency
the escape process from the metastable state becomes more rapid for low noise intensities and the mst for @xmath55 is lower with respect to the previous frequency range ( @xmath52 ) . due to the periodic variation of the potential profile and for
noise intensities smaller than the barrier height , the particle starting from the minimum reaches a position near the top of the barrier .
then there is an optimum range of the system parameters , including the driving parameters , for which the particle turns back at the metastable state because of the noise [ agudov @xmath0 spagnolo , 2001 ; dubkov , agudov @xmath0 spagnolo , 2004 ] .
an enhancement of the lifetime of the metastable state produces an enhancement of the mean switching time . by increasing the noise intensity the thermally activated escape increases too and the mst decreases . for a further increase of the noise , for intensity values greater than the barrier height ( @xmath54 ) , the potential profile `` _ _ seen _ _ '' by the particle is a linear one , which fluctuates between the extreme positions of fig . 1
. the particle can reach more easily positions near the state at @xmath56 and the mst , therefore , restarts to increase again with the noise intensity .
( 220,150)(10,10 ) the nonmonotonic behavior shown in fig . 4 may be explained also by considering the competition of two factors in the above mentioned frequency range : the influence of the returning force from the lefthand side of the potential profile and the increase of the variance of the phase position with increasing temperature .
this explanation is related to the static case when the switching event starts because the value of the bias current is larger than the critical one [ malakhov @xmath0 pankratov , 1996 ] . by increasing the noise intensity
the phase variance increases too . when the phase variance is not so large , the main part of the phase probability distribution is located on the flat part near the point @xmath57 , where the influence of the returning force is low .
thus , the influence of variance dominates , and the mean switching time increases with increasing of the noise intensity , because the particle stays more near the metastable state . further increase of the phase variance , due to the noise , produces a larger probability distribution of the particle , which reaches high slope positions of the potential profile .
then the returning force begins to affect from left to right because of the asymmetry induced by the potential profile and does not permit , in this frequency range , the probability distribution to move into the direction where the potential profile goes up quickly . therefore , the probability distribution expands only into the right direction and the mst decreases because the particle escapes .
finally , the noise intensity becomes so large that it is possible for particle to move upstairs along the potential .
the returning force is not able to hold the increasing of the variance anymore and the switching time increases again . by further increasing of the signal frequency ( @xmath58 , for the parameter values used to obtain the curves in fig .
4 ) a trapping phenomenon occurs . a threshold frequency @xmath59 exists which does not allow the particle to move to the next valley during one period of the signal .
this means that for driving frequency @xmath60 , the particle is trapped within one period of the potential profile and , as a consequence , the mst diverges ( tends to an infinite value ) without noise .
the value of the threshold frequency increases with increasing the bias current and/or the maximal current across the junction [ agudov @xmath0 spagnolo , 2001 ; dubkov , agudov @xmath0 spagnolo , 2004 ] .
the frequency dependence of the switching time , for zero noise intensity , is presented in the inset of fig .
5 . to be trapped the particle , starting at @xmath61 from the position @xmath62 ( see fig .
1 ) of the potential profile in the middle configuration ( @xmath63 ) , should not reach the position at @xmath64 , in the lowest configuration of the potential profile ( @xmath65 ) , within a quarter of the signal period . as a consequence
the switching event from the superconductive state to the resistive one does nt occur . to estimate the threshold frequency we calculate , in absence of noise , the time @xmath66 spent by the particle to go from the point @xmath62 to the point @xmath64 , in the most favorable case , when the potential profile is fixed and is in the lowest configuration , with the current @xmath65 ( see fig . 1 ) .
this time @xmath66 , which is well approximated by one quarter of the threshold period @xmath67 , can be found from the solution of eq .
( [ le ] ) without random current @xmath68 and for a constant value of the total current @xmath4 exceeding the critical one [ malakhov @xmath0 pankratov , 1996 ] @xmath69 from here the threshold frequency @xmath70 .
the values of @xmath59 obtained from eq .
( [ ae ] ) are marked by the vertical lines in the inset of fig .
( 220,150)(10,10 ) specifically we have : @xmath71 , for @xmath45 and @xmath37 ; @xmath72 , for @xmath45 and @xmath43 ; @xmath73 , for @xmath74 and @xmath43 . as it is seen from the inset of fig . 5 , these values are very close to the exact values calculated by numerical simulations of eq . ( [ le ] ) in the absence of noise ( @xmath75 ) .
the most prominent nes effect is observed when the frequency of the periodic signal is near the threshold frequency ( see fig .
5 ) . for @xmath74 and @xmath43
the maximum mst is approximately three times greater than its value at @xmath76 . in the next fig .
6 the curves of mst are shown vs the noise intensity for five frequency values near and larger than the threshold @xmath59 , that is for @xmath77 .
( 220,150)(10,10 ) two distinct transient dynamical regimes are visible .
for @xmath78 we see a nonmonotonic behavior , with a finite value of @xmath27 for @xmath79 . for frequencies larger than the threshold value @xmath59 ,
all the curves are characterized by a monotonic divergent behavior in the limit of small noise intensity ( @xmath79 ) . in both dynamical regimes
a minimum of mst is present for a noise intensity value of the order of the barrier height . by increasing the signal frequency
the curves approach the behavior of the mst obtained with a fixed potential profile , corresponding to a constant total current @xmath63 . for high frequency values , in fact
, the fluctuations of the potential profile are so rapid that the potential `` _ _ seen _ _ '' by the particle is the average potential , that is the middle configuration of fig . 1 , with @xmath63 . in fig .
6 the bold curve corresponds to the case of fixed potential profile with constant total current @xmath80 , and it was calculated by using the exact analytical expression of mst , obtained in [ malakhov @xmath0 pankratov , 1996 ] .
( 220,150)(10,10 ) in the regime of frequencies above the threshold one the conditions of stochastic resonance ( sr ) observation are fulfilled ( see e.g. [ borromeo @xmath0 marchesoni , 2000 ] ) .
sr is manifested when the mst does not exceed the period of signal and thus , the system response becomes periodic owing to noise ( see fig . 5 , curve @xmath81 for which minimal mst is equal to @xmath82 and fig . 6 , curve @xmath83 , minimal mst @xmath84 ) . in fig .
7 the mst and its sd for four values of the signal frequency are shown .
we note that , rising from low to high frequencies near the threshold @xmath59 , monotonic and nonmonotonic behavior of the mst correspond to the same behavior of the sd , respectively .
moreover , in the limit of small noise intensity , all the sd behaviors show lower values with respect to the corresponding mst values , but with different behavior depending on how much the frequency is close to the threshold @xmath59 . for frequency @xmath85 ,
in fact , the sd increases slowly in such a way that it takes the same value of the mst for large noise intensity ( @xmath86 ) .
but for greater frequency the noise intensity value , for which the two curves of mst and sd cross , decreases .
specifically we have the following cross point values : @xmath87 for @xmath88 and @xmath89 for @xmath90 .
this means that when the driving frequency is close to the threshold one the system becomes more unstable and even a small variation of the noise intensity has a strong influence on the system .
the curve , shown in fig . 7 by the bold line , is calculated for the case of the total current @xmath91 and demonstrates a behavior of square root of the noise intensity ( @xmath92 ) .
this curve was obtained from the asymptotic analytical expansion of the sd in the small noise limit @xmath93 and for a fixed potential ( @xmath94 ) [ pankratov @xmath0 spagnolo , 2004 ; gordeeva @xmath0 pankratov , 2006 ] .
( 220,150)(10,10 ) in the limit of @xmath95 the curves of the sd , obtained by our simulations , approach the asymptotic curve obtained for fixed potential . the behavior of the sd curve for @xmath96 , obtained by numerical simulations of eq .
( [ le ] ) , is very close to the theoretical asymptotic curve ( @xmath91 ) not only in the low noise limit but also at larger values of @xmath51 ( @xmath97 ) .
this feature of the sd(@xmath51 ) - curve behavior is shown in more detail in the next logarithmic plot ( fig . 8) , which demonstrates the large rising of the sd for small noise intensities with increasing frequency .
8 the curves of the sd are shown for the same frequency values and parameters used in fig . 5 for the mst behaviors .
again we note that the cross point between the mst and sd curves , for frequencies close to the threshold one ( @xmath98 in this case ) , is at very low noise intensities .
for example for @xmath99 , the cross point is at @xmath100 .
after the threshold frequency , we recover for the sd the same divergent behavior of the mst , but with larger values .
this fact confirms that the switching event in this region of parameters is due to the noise .
it should be noted also that for large noise intensities ( @xmath101 ) all graphs of the mst and sd collapse to the same curve , not depending on the signal frequency ( see fig . 5 and fig .
this means that the new regime of switching is unaffected by the signal and it is only due to the noise .
( 220,150)(10,10 ) in fig .
9 the curves of mst vs @xmath51 for different values of the bias current and signal amplitude are shown .
we consider two cases : ( i ) the bias current is fixed , namely @xmath36 and @xmath102 , and ( ii ) the signal amplitude is fixed , namely @xmath103 and @xmath104 .
we choose @xmath90 as a signal frequency for which the nes effect is more pronounced .
we see that changes of the bias current modify the curves of mst more than changes of the signal amplitude .
we analyzed the transient dynamics of a single overdamped josephson junction driven by a periodic signal .
the conditions of existence of the nes effect are investigated .
we find that the enhancement of the switching time is larger for frequencies of the periodic signal close to the threshold frequency . in the region of the noise intensity values
in which the mean switching time begins to increase , the standard deviation increases too but with different behavior depending on the proximity of the signal frequency to the threshold @xmath59 .
the interval of frequencies , in which the nes effect is observed , depends on the maximal value of the total current across the junction and on the value of the bias current .
there are two ways to change the value of the total current : by changing the amplitude of the driving signal or the value of the bias current .
we find that the switching time changes more significantly due to the bias current variation , than due to the signal amplitude .
we note that for small but nonzero values of the mccumber - stewart parameter the mst increases by several percent but qualitative character of the curves remains the same , as in the considered overdamped case .
peltonen , j. t. , timofeev , a. v. , meschke , m. @xmath0 pekola , j. p. [ 2006
] `` detecting current noise with a josephson junction in the macroscopic quantum tunneling regime '' , _ journal of low temp .
* 146 * , 135 - 159 .
spagnolo , b. , dubkov , a. a. , pankratov , a. l. , pankratova , e. v. , fiasconaro , a. @xmath0 ochab - marcinek , a. [ 2007 ] `` lifetime of metastable states and suppression of noise in interdisciplinary physical models '' , _ acta physica polonica b _ * 38 * ( 5 ) , 1925 - 1950 .
|
we present the analysis of the mean switching time and its standard deviation of short overdamped josephson junctions , driven by a direct current and a periodic signal .
the effect of noise enhanced stability is investigated .
it is shown that fluctuations may both decrease and increase the switching time .
|
the negatively charged pions and kaons can be trapped in the coulomb potential of atomic nucleus forming so called pionic ( kaonic ) atoms .
observations of such atoms allows for studies of strong interaction of pions and kaons with atomic nuclei on the basis of shifts and widths of the energy levels @xcite .
it is also conceivable that a neutral meson could be bound to a nucleus . in this case
the binding is exclusively due to the strong interaction and hence such object can be called a _
mesic nucleus_. here the most promising candidate is the @xmath0-mesic nucleus since the @xmath0n interaction is strongly attractive . as discussed in the contribution by wycech @xcite the nuclear states of @xmath0
may be bound by a similar mechanism as states of @xmath1 with the difference that in the case of @xmath1 nucleons are excited to the @xmath2 and in the case of @xmath0 meson to the @xmath3 resonant state .
we may thus picture @xcite the formation and decay of the @xmath0-mesic nucleus as the @xmath0 meson absorption by one of the nucleons leading to the creation of the @xmath3 and then its propagation through the nucleus until it decays into the pion - nucleon pair which escapes from the nucleus . predicted values of the width of such states range from @xmath47 to @xmath440 mev @xcite .
however , it is important to stress that the width and binding energy of the eta - nucleus bound states depends strongly on the not well known subthreshold @xmath0-nucleon interaction @xcite making impossible an unambiguous prediction as regards the bound states , and , therefore , direct measurements are necessary @xcite .
the search of the @xmath0 - mesic nucleus was conducted in many inclusive experiments @xcite via reactions induced by pions @xcite , protons @xcite , and photons @xcite .
many promising indications of the existence of such an object were reported @xcite , but so far none was independently confirmed .
experimental investigations with high statistical sensitivity and the detection of the n@xmath5 decay products are being continued at the cosy @xcite , jinr @xcite , j - parc @xcite , and mami @xcite laboratories . in this contribution
we report on searches of the @xmath0 - mesic helium in exclusive measurements carried out at the cooler synchrotron cosy by means of the wasa - at - cosy @xcite and cosy-11 @xcite detector setups .
we present also suggestion of studies of @xmath0-mesic tritium with quasi - free reactions at the cosy - tof facility .
we consider the study of the @xmath0-mesic nuclei as interesting on its own account , but additionally it is useful for investigations of ( i ) the @xmath0n interaction , ( ii ) the n@xmath5 properties in nuclear matter @xcite , ( iii ) the properties of the @xmath0 meson in the nuclear medium @xcite , and ( iv ) the flavour singlet component of the @xmath0 meson @xcite .
in 1985 bhalerao and liu @xcite performed a coupled - channel analysis of the @xmath6 , @xmath7 and @xmath8 reactions and came to conclusion that the interaction between the nucleon and the @xmath0 meson is attractive . based on this finding
haider and liu postulated the existence of the @xmath0mesic nuclei @xcite , in which the electrically neutral @xmath0 meson might be bound with the nucleons by the strong interaction .
the formation of such a bound state can only take place if the real part of the @xmath0-nucleus scattering length is negative ( attraction ) , and the magnitude of the real part is greater than the magnitude of the imaginary part @xcite . in the 1980 s the @xmath0mesic nuclei were considered to exists for a @xmath9 12 only @xcite due to the relatively small value of the @xmath0n scattering length ( @xmath10 fm @xcite ) .
however , recent theoretical investigations of hadronic- and photo - production of the @xmath0 meson result in values of the real part of @xmath11 which depending on the analysis method range from 0.25 fm up to 1.05 fm @xcite , and which do not exclude the formation of bound @xmath0-nucleus states for such light nuclei as helium @xcite or even for deuteron @xcite . according to the calculations including multiple scattering theory @xcite or skyrme model
@xcite an especially good candidate for binding is the @xmath12 system .
recent calculations by haider @xcite or tryasuchev and isaev @xcite also indicate the binding in the @xmath12 system , while they rather exclude the existence of the @xmath13 state . on the other hand
there are promising experimental signals which may be interpreted as indications of the the @xmath14 bound state , as for example the shape of the excitation function for the @xmath15 reaction @xcite determined by the spes-4 @xcite , spes-2 @xcite , cosy-11 @xcite , and cosy - anke @xcite collaborations ( fig .
[ fig1](left ) ) . reaction as determined by the cosy - anke @xcite ( open circles ) and the cosy-11 @xcite ( full dots ) and @xcite ( triangles ) .
the solid and dashed lines represent the scattering length fit to the cosy-11 and cosy - anke data , respectively .
( right ) angular asymmetry parameter @xmath16 .
closed and open circles represent results of cosy - anke @xcite and cosy-11 @xcite , respectively .
the dashed and solid lines denote results @xcite of the fit ( allowing for the phase variation ) to the cosy-11 and cosy - anke data , respectively .
the dotted line denotes result of the fit without the phase variation .
the figure is adapted from ref .
@xcite.,title="fig : " ] it has been indicated by wilkin @xcite that a steep rise of the total cross section in the very close - to - threshold region followed by a plateau may be due to the existence of a pole of the @xmath17 scattering amplitude in the complex excess energy plane @xmath18 with @xmath19 @xcite .
this reference shows that the occurrence of the pole changes the phase and the magnitude of the s - wave production amplitude . and indeed ,
the momentum dependence of the asymmetry in the angular distributions of @xmath20 expressed in terms of the asymmetry parameter @xmath16 @xcite , can only be satisfactorily described ( solid and dashed lines in fig .
[ fig1 ] ( right ) ) if a very strong phase variation associated with the pole is included in the fits @xcite .
otherwise there is a significant discrepancy between the experimental data and the theoretical description ( dotted line in fig .
[ fig1 ] ( right ) ) .
details of the experimental technique @xcite and a comprehensive description of results concerning the search of the @xmath0-mesic @xmath23 nucleus conducted by the cosy-11 group were described elsewhere @xcite .
therefore , here we only summarize the main results .
the measurements were carried out using the deuteron beam of cosy which was circulating through the stream of the internal hydrogen target of the cluster - jet type @xcite .
data were taken during a slow acceleration of the beam from 3.095 gev / c to 3.180 gev / c , crossing the kinematical threshold for the @xmath0 production in the @xmath24 reaction at 3.141 gev / c .
the corresponding variation of the excess energy in the @xmath25 system ranged from -10 mev to + 9 mev .
for the scanned beam momentum interval we determined excitation functions for the pion production in the @xmath26 and @xmath27 reaction which were chosen out of possible decay channels of the @xmath0-mesic @xmath28 .
a signature of existence of the @xmath21 bound state was based on observation of a resonance like structure with the center lying below the @xmath0 production threshold in the measured excitation curves . for the @xmath26 channel we concentrated on differential cross sections for the forward pion angles ( @xmath29 ) .
this choice was dictated by the fact that the @xmath30 cross section is up to two orders of magnitude smaller at the forward angles than at the most backward angles .
assuming that the searched structure is produced isotropically , one can expect that it can be best seen just at the forward angles since it appears on the level of small `` non - resonant '' cross section .
the determined excitation function for the @xmath31 process does not show any structure which could originate from a decay of @xmath32 bound state @xcite .
the estimated upper limit for the cross section of the @xmath33 reaction chain is equal to 70 nb .
this limit appears not very restrictive at least under assumption that the cross sections for the @xmath34 bound state formation are of the same order as the @xmath35 cross sections near threshold ( 0.4 @xmath36b ) , and that other possible decay channels like @xmath37 or @xmath38are more favorable .
we expect that @xmath39 is one of the favorable decay channels of the @xmath32 bound state since it corresponds to one step process of absorption of the @xmath0 meson on the neutron inside the @xmath40 nucleus in the reaction @xmath41 . in the @xmath42 rest frame the pion and the proton
are emitted back - to - back with momenta of about 430 mev / c . in the center - of - mass ( c.m . ) system
these momenta are smeared due to the fermi motion of the neutron inside the @xmath40 nucleus .
however , they are significantly larger than the momenta of the two remaining protons which are in the order of 100 mev / c , and which play the role of `` spectators '' .
the counting rate of all identified @xmath27 events including the quasi - free @xmath43 production stays constant in the scanned range of the beam momentum ( see fig . [
ppppi_counts]a ) .
events as a function of the beam momentum without any cuts ( a ) and after rejection of events corresponding to the quasi - free @xmath43 production ( b ) .
the dashed line represents calculations of the @xmath0 absorption described in the text.,title="fig : " ] however , after rejection of the quasi - free events , the number of @xmath27 counts in the beam momentum interval above the @xmath0 threshold is higher than the number of counts in the interval of equal width below the threshold ( see fig .
[ ppppi_counts]b ) .
this difference is equal to @xmath44 and its statistical significance is of @xmath45 .
the observation of this effect was reported in ref .
@xcite . as a possible reaction mechanism explaining the observed enhancement we suggested the production of `` on - shell '' @xmath0 mesons in the reaction @xmath46 which subsequently are absorbed in the @xmath40 nucleus and converted to negatively charged pions in the interaction with the neutron . for testing the @xmath0 absorption hypothesis we estimated the counting rate of the @xmath27 events assuming that the absorption cross section @xmath47 is equal to the cross section for the @xmath46 reaction ( for more details see ref .
this assumption is justified by the observation that the real and imaginary part of the @xmath34 scattering length @xcite have comparable values .
the result of our estimation multiplied by a factor of 10 is shown by the dashed line in fig .
[ ppppi_counts]b .
it underestimates the experimental counts roughly by an order of magnitude and thus it does not corroborate the @xmath0 absorption hypothesis .
for testing this hypothesis we analysed also momentum and angular distributions of the final state particles in the @xmath48 reaction .
the experimental distribution of the pion momentum in the c.m .
system is centred around 430 mev / c as expected for pions originating from the decay of the @xmath3 .
it agrees with results of simulations of the @xmath0 absorption .
however , the momentum distribution of the proton with the highest momentum , which we call the leading proton , is centred at around 300 mev / c and is lower than the corresponding distribution for the @xmath0 absorption grouped around 430 mev / c .
a disagreement takes place also in the case of the distribution of the center - of - mass angles between the pion momentum vector and the momentum vector of the leading proton which in the case of simulations of the @xmath0 absorption are close to @xmath49 and for the experiment lie around @xmath50 .
one can expect , that due to very similar kinematical conditions , absorption of the @xmath0 mesons bound in the @xmath40 nucleus is characterized by differential distributions which are very close to ones predicted in our simulations for the absorption of `` on - shell '' @xmath0 mesons . in particular one can expect that the @xmath51 pairs originating from the decay of the @xmath0-mesic @xmath40 are emitted at c.m .
angles concentrated predominately in the range @xmath52 as it is the case for the @xmath0 meson absorption . in this angular range , there are only two experimental counts . assuming , that these two counts originate from the decay of the @xmath53 bound state
, we estimated the cross section for the production of such a state in @xmath22 collisions close to the @xmath0 production threshold .
the resulting cross section of @xmath54 @xmath36b should be considered as an upper limit for the production cross section of the @xmath0-mesic @xmath55 since the observed two events might originate from other processes than the bound state decay .
data collected in the cosy-11 measurements with the ramped deuteron beam were also used to investigate the cusp effect observed at saturne in the threshold excitation curve for the process @xmath56 @xcite . as suggested by wilkin @xcite , the cusp visible in the saturne data below the @xmath0 threshold
could be caused by an interference between an intermediate state including the @xmath0 meson and the non - resonant background corresponding to the multi - pion production . in the saturne experiment
the @xmath55 ejectiles were measured with the spes iv spectrometer which was set in such a way that it registered the @xmath55 which was approximately at rest in the c.m . frame .
since the cosy-11 momentum and angular acceptance was much larger than one of the spes iv spectrometer , the limitation on the c.m .
momenta of @xmath55 was realized by means of corresponding cuts during the data analysis @xcite .
contrary to the saturne result no cusp below the @xmath0 threshold was observed @xcite .
the installation of the wasa detector at cosy opened a unique possibility to search for the @xmath58 bound state with high statistics and high acceptance .
we conduct a search via an exclusive measurement of the excitation function for the @xmath59 reaction varying continuously the beam momentum around the threshold for the @xmath60 reaction .
ramping of the beam momentum and taking advantage of the large acceptance of the wasa detector reaction the acceptance of the wasa - at - cosy detector equals to almost 70% . ] allows to minimize systematical uncertainties making the wasa - at - cosy a unique facility @xcite for such kind of exclusive experiments .
the @xmath61 bound state should manifest itself as a resonant like structure below the threshold for the @xmath62 reaction .
if a peak below the @xmath63 threshold is found , then the profile of the excitation curve will allow to determine the binding energy and the width of the @xmath64 bound state .
if , however , only an enhancement around the threshold is found , then it will enable to establish the relation between width and binding energy @xcite .
finally , if no structure is seen the upper limit for the cross section of the production of the @xmath0-helium nucleus will be set .
in addition , when searching for the signal of the @xmath0-mesic state we may take advantage of the fact that the distribution of the relative angle between the @xmath65 pair for the background ( due to the prompt @xmath66 reaction ) is much broader than the one expected from the decay of the bound state .
this is because the relative angle between the outgoing @xmath65 pair originating from the decay of the n@xmath5 resonance is equal to 180@xmath67 in the n@xmath68 reference frame and it is smeared only by about 30@xmath67 in the reaction center - of - mass frame due to the fermi motion of the nucleons inside the @xmath69 nucleus .
[ angle ] shows the distribution of the relative @xmath70 angle as expected for the signal and for the background due to the prompt @xmath66 reaction .
distribution of the relative @xmath71 angle seen in the reaction c.m .
system as simulated for the processes leading to the creation of the eta - helium bound state : @xmath72 ( left ) , and for the prompt production of the @xmath73 system assuming a homogeneous population of the phase space for the @xmath74 reaction ( right ) .
the figure is adapted from reference @xcite.,title="fig : " ] distribution of the relative @xmath71 angle seen in the reaction c.m .
system as simulated for the processes leading to the creation of the eta - helium bound state : @xmath72 ( left ) , and for the prompt production of the @xmath73 system assuming a homogeneous population of the phase space for the @xmath74 reaction ( right ) .
the figure is adapted from reference @xcite.,title="fig : " ] in the first experiment conducted in june 2008 , we used a deuteron pellet target and the cosy deuteron beam with a ramped momentum corresponding to a variation of the excess energy for the @xmath64 system from -51.4 mev to 22 mev . at present
the data are evaluated and preliminary excitation curves determined for few intervals of the @xmath75 angle are shown in the left panel of fig .
[ fig4 ] .
reaction measured in the 20 degrees intervals of the @xmath75 angle .
courtesy to @xcite .
( right ) ratio of excitation functions angular ranges as indicated in the figure .
[ fig4],title="fig : " ] the figure indicates no structure in the angular range close to the 180 degree where the signal is expected .
the ratio of excitation functions from various angular ranges is also constant as indicated in the right panel of fig .
[ fig4 ] . therefore , taking into account the luminosity and the detector acceptance we preliminary estimated that an upper limit for the eta - mesic helium production via the @xmath76 reaction is equal to about 20 nb on a one sigma level
. the experiment will be continued in november 2010 .
however , we can not estimate the statistics required for the observation of the signal since it is a priori impossible to predict in a model independent way a cross section for the creation of the eta - mesic nucleus , which strongly depends on the strength of the relatively poorly known @xmath0-nucleon interaction , and which involves computations of the behaviour of the many - body system .
therefore , based on the fact that the real and imaginary part of the helium-@xmath0 scattering length are almost of the same value we estimate the cross section for the creation of the eta - mesic helium applying the assumption that the probability of the production of the @xmath0 meson in continuum and its absorption on the helium nucleus is of the same order .
this leads to the hypothesis that the cross section for the creation of the bound state below the threshold ( which is than connected with the absorption of the @xmath0 meson on the helium nucleus ) is in the first order the same as the close to threshold cross section for the @xmath0 meson production .
therefore , for the estimation of the counting rate in the future measurements we assume that the cross section in the maximum of the breit - wigner distribution for the @xmath77 production is about 15 nb as measured for the @xmath78 reaction @xcite .
further on , we assume , that the probability of decay of the @xmath61 bound state into the @xmath79 channel is equal to @xmath80 . the factor @xmath81 takes into account the fact that there are four possible @xmath0 absorption channels : @xmath82 , @xmath83 , @xmath84 , @xmath85 . in turns ,
the factor @xmath86 represents our guess of the probability that the three observer nucleons ( @xmath87 ) , in the process of the @xmath0 absorption on the neutron in @xmath88 , form @xmath55 in the final state .
estimation of this effect needs , in our understanding of the problem , projection of the @xmath88 wave function on the wave function of the @xmath89 pair and inclusion of the pion and proton rescattering on the observer nucleons .
since , this kind of estimations is nt at present in our reach , we justify our guess using an analogy between the decay of the @xmath61 nuclei and the @xmath90-hypernuclei . for the later case
it was observed namely that in the @xmath43 decay channel the decay mode @xmath91 is dominant @xcite .
the above considerations implies that the cross section for the @xmath92 reaction should be in the order of 2 nb .
this level of sensitivity can be reached in the experiment which will be continued in november 2010 @xcite .
the systematic uncertainties in establishing the shape of the excitation functions discussed in the previous sections are significantly reduced by using the momentum ramping technique since the energy range of interest for the search of the @xmath0-mesic nucleus is scanned in each cosy cycle .
-@xmath93 reaction .
the fermi momentum of the nucleons inside the deuteron is presented by the dotted arrows and the beam momentum by the dashed one .
the figure is adapted from reference @xcite .
[ fig2],width=359,height=124 ] a scan of the energy can also be achieved in the case of the fixed beam energy available for the external beam experiments .
this may be realised by means of the quasi - free reactions as already successfully used at cosy for the study of the meson production in the quasi free proton - neutron collisions @xcite .
for instance , a search of the @xmath0-mesic tritium may be realized by studying the excitation function of the @xmath94 reaction using a deuteron beam and tagging the @xmath95 reactions by the measurement of the spectator protons ( @xmath96 ) from the @xmath97 reaction .
this reaction , shown schematically in fig .
[ fig2 ] , may be measured using the cosy - tof detector .
the fermi motion of nucleons inside the deuteron beam will allow to scan a large range ( in the order of 100 mev ) of excess energies in the @xmath95 reaction .
the work was partially supported by the european community - research infrastructure activity under the fp6 and fp7 programmes ( hadron physics , rii3-ct-2004 - 506078 , primenet no .
, by the polish ministry of science and higher education under grants no .
3240/h03/2006/31 and 1202/dfg/2007/03 , by the german research foundation ( dfg ) , by the ffe grants from the research center jlich , and by the virtual institute `` spin and strong qcd '' ( vh - vi-231 ) .
_ , _ arxiv:1002.4277 _ ( 2010 )
. j. zmeskal _
_ , _ nucl .
phys . _ * a835 * , 410 ( 2010 ) . s. wycech , _ acta
. phys . pol .
b _ * 41 * , ( 2010),this issue
. g. a. sokol _ et al .
, arxiv : nucl - ex/0106005 _ ( 2001 ) . c. garcia - recio _ et al .
_ , _ phys .
lett . _ * b550 * , 47 ( 2002 ) . q. haider , l.c .
liu , _ phys .
rev . _ * c66 * , 045208 ( 2002 ) .
q. haider , l.c .
liu , _ acta phys
* 2 * , 121 ( 2009 ) .
r. e. chrien _ et al .
_ , _ phys .
_ * 60 * , 2595 ( 1988 ) .
b. j. lieb _ et al .
_ , _ proc .
conf . , sao paulo , brazil _ ( 1989 ) .
g. a. sokol _ et al .
_ , _ arxiv : nucl - ex/9905006 _ ( 1999 )
. m. kh .
anikina _ et al .
_ , _ arxiv : nucl - ex/0412036 _ ( 2004 ) .
a. gillitzer , _ acta phys . slovaca _ * 56 * , 269 ( 2006 ) .
a. budzanowski _ et al .
_ , _ phys .
* c79 * , 061001(r ) ( 2009 ) . m. pfeiffer , _ et al .
_ , _ phys .
lett . _ * 92 * , 252001 ( 2004 ) .
p. moskal _
et al . _ ,
cosy proposal no .
186.1 ( 2008 ) .
w. krzemien , p. moskal , j. smyrski , _ acta phys .
* 2 * , 141 ( 2009 ) .
h. fujioka , k. itahashi , _ acta .
phys . pol .
b _ * 41 * , ( 2010 ) , this issue , _
arxiv:1002.0201_. b. krusche , i. jaegle , _ acta phys . pol .
* 2 * , 51 ( 2009 ) .
j. smyrski , p. moskal , w. krzemien , _ acta phys .
* 2 * , 133 ( 2009 ) .
w. krzemie _ et al .
phys . _ * a24 * , 576 ( 2009 ) .
j. smyrski _ et al .
_ , _ nucl .
phys . _ * a790 * , 438 ( 2007 ) . j. smyrski _ et al .
_ , _ acta phys
. slovaca _ * 56 * , 213 ( 2006 ) .
d. jido , h. nagahiro , s. hirenzaki , _ phys .
rev . _ * c66 * , 045202 ( 2002 ) .
t. inoue , e. oset , _ nucl .
* a710 * , 354 ( 2002 ) .
s. d. bass , a. w. thomas , _ phys .
_ * b634 * , 368 ( 2006 ) .
s. d. bass , a. w. thomas , _ acta .
phys . pol .
b _ * 41 * , ( 2010 ) , this issue , _
arxiv:1007.0629_. r. s. bhalerao , l.c .
liu , _ phys .
lett . _ * 54 * , 865 ( 1985 ) .
q. haider , l.c .
liu , _ phys .
_ * b172 * , 257 ( 1986 ) .
a. sibirtsev _ et al .
_ , _ phys .
rev . _ * c70 * , 047001 ( 2004 ) .
a. m. green , s. wycech , _ phys .
* c71 * , 014001 ( 2005 ) . c. wilkin , _ phys .
_ * c47 * , r938 ( 1993 ) . c. wilkin , _ acta
phys . pol .
b _ * 41 * , ( 2010 ) , this issue , _ arxiv:1008.0482_. s. wycech , a. m. green and j. a. niskanen , _ phys .
_ * c52 * , 544 ( 1995 ) .
a. m. green _ et al .
_ , _ phys .
rev . _ * c54 * , 1970 ( 1996 ) .
n. n. scoccola and d. o. riska , _ phys .
lett . _ * b444 * , 21 ( 1998 ) .
v. a. tryasuchev , a. v. isaev , _ arxiv:0901.3242 _ ( 2009 ) .
j. berger , _ et al .
_ , _ phys .
lett . _ * 61 * , 919 ( 1988 ) .
b. mayer , _ et al .
_ , _ phys .
rev . _ * c53 * , 2068 ( 1996 ) .
j. smyrski _ et al .
_ , _ phys .
* b 649 * , 258 ( 2007 ) .
t. mersmann _ et al .
_ , _ phys .
lett . _ * 98 * , 242301 ( 2007 ) .
adam _ et al .
_ , _ phys .
rev . _ * c75 * , 014004 ( 2007 ) . c. wilkin _ phys .
lett . _ * b654 * , 92 ( 2007 ) . c. wilkin _ aip conf .
proc . _ * 950 * , 23 ( 2007 ) .
s. brauksiepe _ et al .
_ , _ nucl .
instr . & meth .
_ * a376 * , 397 ( 1996 ) .
p. klaja _ et al .
_ , _ aip conf
. proc . _ * 796 * , 160 ( 2005 ) .
p. moskal , hep - ph/0408162 ( 2004 ) .
f. plouin , _ production and decay of light mesons _ , ed . by p. fleury ; world scientific , singapore 1988 , p.114 . c. wilkin , _ production and decay of light mesons _ , ed . by p. fleury ; world scientific , singapore 1988 , p.187 .
adam _ et al .
_ , arxiv : nucl - ex/0411038 ( 2004 ) . w. krzemien , private communication ( 2010 ) .
r. frascaria _ et al .
_ , phys .
* c 50 * , 537 ( 1994 ) .
n. willis _ et al .
_ , phys .
lett.*b 406 * , 14 ( 1997 ) .
a. wronska _
_ , acta phys . pol . * 56 * , 279 ( 2006 ) .
j. g. fetkovich _
et al . _ ,
phys . rev .
* d6 * , 3069 ( 1972 ) .
p. moskal
_ , _ phys .
_ * c79 * , 015208 ( 2009 ) . j. klaja _ et al .
_ , _ phys .
rev . _ * c81 * , 035209 ( 2010 ) .
m. abdel - bary _ et al .
_ , _ eur .
j _ * a29 * , 353 ( 2006 ) .
m. skurzok , diploma thesis , jagiellonian university , cracow ( 2010 ) .
|
we review status and perspectives of the search for the light eta - mesic nuclei using cosy-11 , wasa - at - cosy and cosy - tof detector systems .
|
at first glance , the motion of the gas - liquid interface along the solid surface is a purely hydrodynamic problem .
however , it attracted significant attention from the physicists since the work @xcite , which showed an unphysical divergence that appears in the hydrodynamic treatment if a motion of a wedge - shaped liquid slides along the solid surface .
the reason for this divergence lies in the no - slip condition ( i.e. zero liquid velocity ) at the solid surface .
being so common in hydrodynamics , this boundary condition is questionable in the vicinity of the contact line along which the gas - liquid interface joins the solid . in the absence of mass transfer between the gas and the liquid ,
the no - slip condition requires zero velocity for the contact line that is supposed to be formed of the liquid molecules in the contact with the solid .
it means that , for example , an oil drop can not move along the glass because of the no - slip condition ! of course , this contradicts the observations .
the experiment @xcite demonstrated that the velocity on the liquid - gas interface is directed towards the contact line during the contact line advance .
the authors interpreted this result by the rolling ( `` caterpillar '' ) motion of the drop @xcite .
however , later theoretical study @xcite shows that such a motion is compatible with the no - slip condition on the non - deformable solid surface only for the contact angles close to @xmath0 .
the justification of the no - slip condition is well known @xcite : it is the excess of the attractive force between the solid and the liquid molecules over the force between two liquid molecules .
this attraction has a tendency to prevent the motion of the liquid molecules adjacent to the solid .
obviously , the same forces resist when these molecules are forced to move . in other words , some relatively large ( with respect to viscous dissipation ) energy should be spent for this forcing .
numerous microscopic theories ( see e.g. @xcite ) propose different phenomena as to be responsible for the contact line motion .
however , no general theory has been agreed upon .
this situation is partly due to the scarceness of the information that can be extracted from the experiments .
most of them deal with either drops with cylindrical symmetry ( circular contact lines ) or the climbing of the contact line over a solid immersed into a liquid ( straight contact line ) . in these experiments , the contact line velocity @xmath1 measured in the normal direction does not vary along the contact line on a macroscopic scale larger than the size of the surface defects .
the experiments with non - spherical drops where such a variation exists can give additional information .
this information can be used to test microscopic models of the contact line motion . to our knowledge
, there are only two kinds of investigated situations that feature the non - spherical drops .
the first one is the sliding of the drop along an inclined surface @xcite .
the second concerns the relaxation of the sessile drops of complicated shape towards the equilibrium shape of spherical cap .
this latter case was studied experimentally in @xcite for water drops on the silanized silicon wafers at room temperature .
the present article deals with this second case .
the principal results of @xcite can be summarized as follows : \(i ) the relaxation of the drop from the elongated shape towards the spherical shape is exponential .
the characteristic relaxation time @xmath2 is proportional to the drop size .
the drop size can be characterized by the contact line radius @xmath3 at equilibrium when the drop eventually relaxes towards a spherical cap .
\(ii ) the dependence of @xmath2 on the equilibrium contact angle @xmath4 is not monotonous : @xmath5 and @xmath6 .
\(iii ) the relaxation is extremely slow .
the capillary number @xmath7 is of the order of @xmath8 , where @xmath9 is the surface tension and @xmath10 is the shear viscosity .
since the motion is not externally forced , a small @xmath11 shows that the energy dissipated in the vicinity of the contact line is much larger than in the bulk of the drop .
the contact line motion is characterized by the normal component @xmath1 of its velocity .
many existing theories result in the following relationship between @xmath1 and the dynamic contact angle @xmath4 : @xmath12 where @xmath13 is the static contact angle , @xmath14 is a constant characteristic velocity and @xmath15 is a function of two arguments , the form of which depends on the model used . for all existing models ,
the following relation is satisfied @xmath16 which implies the trivial condition @xmath17 .
it means simply that the line is immobile when @xmath18 .
the theories of voinov @xcite and cox @xcite correspond to @xmath19 there are many theories ( see e.g. @xcite , @xcite ) which result in @xmath20 in a recent model by pomeau @xcite , it is proposed that @xmath21 with the coefficient @xmath14 that depends on the direction of motion ( advancing or receding ) but not on the amplitude of @xmath1 .
since the drop evolution is extremely slow , the drop shape can be calculated using the quasi - static argument according to which at each moment the drop surface can be calculated from the constant curvature condition and the known position of the contact line .
the major problem is how to find this position .
independently of the particular contact line motion mechanism , at least two approaches are possible .
the first of them is the
local " approach @xcite , which consists in the determination of the position of a given point of the contact line from eq .
( [ vnl ] ) where @xmath4 is assumed to be the _ local _ value of the dynamic contact angle at this point .
another , non - local approach is suggested in sec .
[ nonloc ] .
certainly , both of this approaches should give the same result when @xmath1 does not vary along the contact line .
however , we show that the result is different in the opposite case .
the influence of surface defects on the contact line dynamics is considered in sec .
in this section we generalize another approach , suggested in @xcite , for an arbitrary drop shape .
this approach postulates neither eq . ( [ vnl ] ) nor a particular line motion mechanism .
it simply assumes that the energy dissipated during the contact line motion is proportional to its length and does not depend on the direction of motion ( advancing or receding ) .
then , at low contact line velocity , the leading contribution to the energy dissipated per unit time ( i.e. the dissipation function ) can be written in the form @xmath22 where the integration is performed over the contact line and @xmath23 is the constant dissipation coefficient . according to the earlier discussed experimental results @xcite , the dissipation in the bulk is assumed to be much smaller than that in the vicinity of the contact line . since we assume that most dissipation takes place in the region of the drop adjacent to the contact line , our discussion is limited to the case where the prewetting film ( that is observed for zero or very low contact angles ) is absent
this situation corresponds to the conditions of the experiment @xcite where the dropwise ( as opposed to filmwise ) condensation shows the absence of the prewetting liquid film .
the above assumption also limits the description to the partial wetting case .
this assumption is also justified by the experimental conditions under which it is extremely difficult to obtain macroscopic convex drops for the contact angles @xmath24 because of the contact line pinning @xcite .
the main reason is that the potential energy @xmath25 of the drop from eq .
( [ u2 ] ) goes to zero as the contact angle goes to zero . at small contact angles @xmath25 is not large enough to overcome the pinning forces that originate from the surface defects , see sec .
therefore , the macroscopic convex drops under consideration can not be observed at small contact angles .
generally speaking , the behavior of the drop obeys the lagrange equation @xcite @xmath26 where the lagrangian @xmath27 is the function of the generalized coordinates @xmath28 and of their time derivatives which are denoted by a dot .
the current time is denoted by @xmath29 , @xmath30 is the kinetic energy , and @xmath31 is the potential energy . since there is no externally forced liquid motion in this problem and the drop shape change is extremely slow , we can neglect the kinetic energy by putting @xmath32
. then eq .
( [ lagr ] ) reduces to @xmath33 the expression applied first to the contact line motion in @xcite .
the potential energy of a sessile drop is @xcite @xmath34 where @xmath9 is the liquid surface tension , @xmath35 and @xmath36 are the areas of the vapor - liquid and liquid - solid interfaces respectively and @xmath37 is the equilibrium value of the contact angle .
we neglect the contribution due to the van der waals forces because we consider macroscopic drops and large contact angles @xmath38 . for such drops the van der waals
forces influence the interface shape only in the very close vicinity of the contact line and this influence can be neglected . in general
the static contact angle @xmath13 is not equal to @xmath37 because of the presence of the defects , a problem which will be treated in the section [ def ] .
meanwhile , we assume that @xmath39 .
the gravitational contribution is neglected in eq .
( [ uu ] ) because the drops under consideration are supposed to be small , with the radius much smaller than the capillary length
. the volume of a sessile drop is fixed .
its calculation provides us with another equation , which closes the problem provided that the shape of the drop surface is known .
the drop shape is determined from the condition of the quasi - equilibrium which results in the constant curvature of the drop surface .
usually , the wetting dynamics is observed either for the spreading of droplets with the shape of the spherical cap , or for the motion of the liquid meniscus in a cylindrical capillary , or for the extraction of a solid plate from the liquid @xcite . in all these cases ,
the contact line velocity @xmath1 does not vary along the contact line and the dissipation function in the form eq .
( [ diss ] ) results in the expression @xcite @xmath40 which is equivalent to eq .
( [ vnl ] ) , with the function @xmath15 taking the usual form ( [ fcos ] ) .
one might think that this equivalence confirms the universal nature of this expression . in the next section
we show that it is not exactly so because the non - local approach results in a different expression when @xmath1 varies along the contact line .
let us now apply the algorithm described above to the problem of drop relaxation .
a shape for a non - spherical drop surface of constant curvature can be found only numerically . in order to treat the problem analytically , we approximate the drop shape by a spheroidal cap that is described in cartesian @xmath41 coordinate system by the equation @xmath42 at @xmath43 , the plane xoy corresponding to the solid surface .
the symmetry of the problem allows only quarter of the drop ( see fig . [ quarter ] ) to be considered . .
the local contact angles at the points m and n are shown too.,height=226 ] since one of the parameters @xmath44 is fixed by the condition of the conservation of the drop volume that can be calculated as @xmath45 there are only two free parameters left .
the time - dependent parameters @xmath46 and @xmath47 can be taken as generalized coordinates .
however , it is more convenient to use another set of parameters , @xmath48 and @xmath49 , which are the half - axes of the ellipse that form the base of the drop ( see fig . [ quarter ] ) , @xmath50 .
they are related to @xmath46 and @xmath47 by the equations @xmath51 that follow from eq .
( [ ell ] ) . at the end of the relaxation @xmath52 where @xmath53 is the final radius of curvature of the drop .
therefore , during the late stage @xmath54 with @xmath55 .
some points of the contact line advance , some points recede .
the dynamic contact angle changes its value along the contact line .
in particular , the point n@xmath56 in fig .
[ quarter ] advance and m@xmath57 recede .
these points are extreme and their velocities have the maximum absolute values , positive for n and negative for m. the dynamic contact angles ( @xmath58 : dynamic advancing contact angle in n and @xmath59 : dynamic receding contact angle in m ) also have the extreme values there .
they can be found from the equations @xmath60 that reduce for @xmath61 to @xmath62/3 .
\end{array}\label{coslin}\ ] ] eq .
[ ut ] , written for the generalized coordinates @xmath63 and @xmath64 together with the expression for the dissipation function ( see appendix [ a2 ] ) , implies the set of equations @xmath65 where @xmath66 and the coefficients @xmath67 and @xmath68 are given by eq .
( [ ab ] ) in appendix [ a2 ] .
the solutions of eqs .
[ kin ] read @xmath69/2 , \label{rx}\\ r_{y}(t ) = [ ( r_{y}^{(i)}-r_{x}^{(i)})\,\exp(-t/\tau_s)+ \nonumber\\(r_{x}^{(i)}+r_{y}^{(i)})\,\exp(-t/\tau_n)]/2 , \label{ry}\end{aligned}\ ] ] where @xmath70 and @xmath71 are the initial ( @xmath72 ) values for @xmath73 and @xmath74 respectively , and the relaxation times @xmath75,\\ \tau_n= 45\,\tau_0\,(1+\cos\theta_s)/[(108 + 41\,\cos\theta_s+\nonumber\\14\,\cos^2\theta_s + 17\,\cos^3\theta_s)(1-\cos\theta_s)].\label{taun}\end{aligned}\ ] ] the variables @xmath73 and @xmath74 are defined in eq .
( [ rr ] ) in such a way that when @xmath76 the drop surface remains spherical during its relaxation .
one can see from eqs .
( [ rx],[ry ] ) that the evolution is defined entirely by the characteristic time @xmath77 ( spherical " ) in this case . when @xmath78 , only @xmath79 ( non - spherical " ) defines the drop evolution . in the real experimental situation
where @xmath80 , the relaxation time @xmath81 alone defines the relaxation of the drop as it follows from eqs .
( [ rx ] , [ ry ] ) .
therefore @xmath81 should be associated with the experimentally observed relaxation time .
the functions @xmath82 are plotted in fig .
[ times ] assuming that @xmath23 is independent of @xmath13 . versus the static contact angle @xmath13.,height=226 ] clearly , both @xmath77 and @xmath79 increase monotonically with @xmath13 in agreement with the observed tendency for large contact angles .
it is interesting to check whether or not by applying the local approach of eq .
( [ cos ] ) we recover the non - local result for @xmath1 .
this is easy to do for the case @xmath78 , i.e. when @xmath83 . in this case
, the non - local model ( [ kin ] ) implies @xmath84 , and the contact line velocities at the points m and n are @xmath85 since eq .
( [ coslin ] ) results in @xmath86 the local approach ( [ cos ] ) implies that @xmath87 the comparison of eqs .
( [ taun]-[vn ] ) with eqs .
( [ vml],[vnl ] ) show that the results of the local and the non - local approaches are different .
however , one can verify that the results are the same in the limit of very small @xmath13 .
for finite contact angles , the non - local approach is not equivalent to the local approach .
the main difference can be summarized as follows .
the @xmath1 value that is obtained with our non - local approach can be presented in the form ( [ vnl ] ) common for the local approach . however , while the characteristic velocity @xmath14 from eq .
( [ vnl ] ) is constant in the local approach , it is _ a function of the position on the contact line _ in the non - local approach .
indeed , the comparison of eqs .
( [ vm]-[cn ] ) with eqs .
( [ vnl ] , [ fcos ] ) shows that @xmath88 \;\mbox{at the point m,}\label{vmc}\\ v_c=&3\,{\sigma\over\xi}{\tau_0\over\tau_n}/[\sin^2\theta_s\,(5\,\cos\theta_s-2 ) ] \;\mbox{at the point n.}\label{vnc}\end{aligned}\ ] ] we do not expect our model to be a good description for the contact angles close to @xmath89 .
the reason is the limitation of the spheroid model for the drop shape .
the spheroidal shape necessarily fixes @xmath90 when @xmath91 independently of the contact line velocity , which is incorrect .
in addition , the spheroid model does not work at all for @xmath92 .
one needs to find the real shape of the drop ( which is defined by constant curvature condition ) to overcome these difficulties . in order to estimate the limiting value for @xmath13
for which the spheroid model works well we mention that the dynamic advancing and receding contact angles defined by eq .
( [ coslin ] ) must satisfy the inequality @xmath93 . by putting @xmath78 in eqs .
( [ rx ] , [ ry ] ) one finds that this inequality is satisfied when @xmath94 .
the last inequality provides us with the limit of the validity for the spheroidal model . to conclude this section we note that our non - local approach to the dynamics of wetting is not equivalent to the traditional local approach .
both approaches allow the relaxation time to be calculated for a given contact angle provided that the contact angle dependence of the dissipation coefficient @xmath23 is known .
additional experiments are needed to reveal which approach is the most suitable . under the assumption that the @xmath95 dependence ( if any ) is weak , we find that the relaxation time decreases with the contact angle .
this result explains the decrease of the relaxation time at large contact angles observed in @xcite .
we think that the opposite tendency observed for the small contact angles is related to the influence of the surface addressed in the next section .
the motion of the contact line in the presence of defects has been frequently studied ( see @xcite for a review ) . however ,
little is understood at the moment because the problem is very sophisticated .
most of its studies deal with the influence of the defects on the static contact line ( see , e. g. @xcite ) when they are responsible for the contact angle hysteresis .
the latter was studied in @xcite and @xcite for the wedge geometry which assumes the external forcing of the contact line .
when the contact line moves under the action of a force @xmath96 , it encounters pinning on the random potential created by surface defects .
thus the motion shows the stick - slip " behavior .
it is characteristic for a wide range of physical systems where pinning takes place and is the basis of the theory of dynamical critical phenomena , in which the average contact line velocity is @xmath97 where the exponent @xmath98 is universal and @xmath99 is the pinning threshold .
this expression is often applied ( see @xcite and refs . therein ) to the contact line motion in the systems , where the geometry of the meniscus does not depend on the dragging force .
however , the values of @xmath98 vary widely depending on the experimental conditions and do not correspond to the theoretical predictions .
the motion of the contact line during the coalescence of drops is even more complicated because the geometry of the meniscus is constantly changing .
therefore , application of the expression ( [ v ] ) is even more questionable in this situation . in this section
we employ the formalism developed in the previous section in order to understand the influence of the surface defects on the relaxation time of the drop where the contact line is _ not forced externally_. the surface defects are modeled by the spatial variation of the local density of the surface energy , which can be related to the local value of the equilibrium contact angle @xmath100 by the young formula as was suggested in @xcite to describe the static contact angles .
the expression ( [ uu ] ) can be rewritten for this case in the form @xmath101 the contribution of the defects and thus the deformation @xmath102 of the contact line due to the defects is assumed to be small .
then , in the first approximation that corresponds to the horizontal averaging " approximation from @xcite @xmath103 the superscript @xmath104 means that the corresponding quantity is calculated for @xmath105 and for the constant value of the contact angle @xmath106 defined by the expression @xmath107 .
\label{bt}\ ] ] then @xmath108 it can be shown that the first - order correction to this value , which appears due to the defects is @xmath109 we accept the following model for defects because it is , on one hand , simple and , on the other hand , proven @xcite to be a good description for the advancing and the receding contact angles in the approximation considered .
the defects are supposed to be similar circular spots of radius @xmath110 arranged in a _
regular _ spatially periodic pattern , @xmath111 being the spatial period , the same in both directions , see fig . [ defects ] . ) and outside them ( @xmath112 ) are also shown.,height=226 ] the spots and the clean surface have the values of the equilibrium contact angle @xmath113 and @xmath114 respectively .
for this pattern , eq .
[ bt ] yields @xmath115 , \label{cpp}\ ] ] the parameter @xmath116 being the fraction of the surface covered by the defect spots .
we consider the case @xmath117 in the following .
then it is obvious from fig .
[ defects ] , that @xmath118 in the following , we will for simplicity treat a 2d sessile drop , i. e. a liquid stripe of infinite length , the cross - section of which is the segment of a circle as shown in fig .
[ 2ddrop ] . is shown too.,height=151 ] the volume @xmath119 of the stripe per its length @xmath120 does not change with time : @xmath121 where @xmath48 is the half - width of the stripe , see fig .
[ 2ddrop ] .
the dynamic contact angle @xmath4 can be calculated from eq .
( [ vs ] ) , provided that @xmath48 is known .
it can be shown by the direct calculation of @xmath122 ( [ g0 ] ) and the dissipation function @xmath123 ( [ diss ] ) that eq .
( [ ut ] ) with the substitution @xmath124 reduces to the equation @xmath125 the first - order correction to the drop energy @xmath126 can be calculated from eq .
( [ dg ] ) by following the guidelines of @xcite .
its explicit expression for the chosen geometry is given in appendix [ a3 ] .
the kinetics of the relaxation is shown in fig .
[ kinet ] .
the relaxation kinetics for the drop on the ideal substrate with the equilibrium value of the contact angle equal to the value of @xmath106 is also shown for comparison .
the half - width of the drop on the ideal substrate relaxes to its equilibrium value @xmath3 that is related to the volume of the drop through eq .
( [ vs ] ) written for @xmath127 and @xmath128 .
we chose @xmath129 for fig .
[ kinet ] .
the stick - slip motion is illustrated in the insert in fig .
[ kinet ] . of the drop with and with no defects with the same initial ( @xmath72 ) value of @xmath130 and for @xmath131 and @xmath132 .
the latter value corresponds to the defect radius @xmath133 , @xmath134 , and @xmath135.,height=226 ] note that the contact line in its final position for the non - ideal case is pinned in a metastable state so that the final 2d radius of the drop is larger than @xmath3 .
the final contact angle ( the equilibrium receding contact angle ) thus differs from that for an ideal surface .
because the contact line is being stuck on the defects , its motion is slowed down .
however , the presence of defects does not change strongly the relaxation time .
it remains of the order of @xmath136 because this deceleration is compensated by the acceleration during the slip motion .
[ kinet ] shows the impact of the defects on the relaxation .
the relaxation time appears to be _ smaller _ in the presence of defects than in the ideal case ( no defects ) because the contact line is pinned by defects whereas it would have continued to move on the ideal surface .
it should be noted that this model is just a first step towards the description of contact line kinetics on a non - ideal substrate . in reality , the different portions of the contact line slip at different moments in time ( cascades of slips are observed e. g. in @xcite ) .
this means that the liquid flows in the direction parallel to the contact line to the distances much larger than the defect size , i. e. the first - order approximation is not adequate .
the direction of this flow reverses frequently .
this effect can lead to the expression like eq .
( [ v ] ) and to a large relaxation time .
this article deals with two important issues concerning contact line dynamics .
first , it discusses the local versus non - local approaches to contact line motion . while the local approach consists in postulating a direct relationship between the normal contact line velocity and the dynamic contact angle _ at a given point of the contact line _ , the nonlocal approach starts from a more general hypothesis about the form of the dissipation function of the droplet .
these approaches give the same results for very small contact angles or for the normal contact line velocity that does not vary along the contact line , which is the case of a drop that has the shape of a spherical cap .
in other cases ( large contact angle , non - spherical drops ) the results of these two approaches differ .
we carried out calculations assuming that the drop surface is a spheroid . in reality , its surface is not a spheroid and has a constant curvature .
more theoretical work is needed to overcome this approximation .
the second issue treated in this article is the influence of surface defects on contact line dynamics . in the approximation of a 2d drop
, it is assumed that the contact line remains straight during its motion . in this approximation , the stick - slip microscopic motion does not influence the average dynamics strongly .
the defects manifest themselves by changing the final position of the contact line by pinning it in a metastable state .
therefore , the relaxation is more rapid than that on an ideally clean surface simply because it is terminated earlier .
we used the @xmath137 system for the analytical computations .
we find first the dissipation function @xmath123 .
the contact line can be described by the equation @xmath138 where @xmath48 and @xmath49 are time - dependent . by using the well - known formula of differential geometry
@xmath139 , the integral ( [ diss ] ) can be written in an explicit form . in order to obtain the first order approximation for @xmath123
, one can use the expansion ( [ rr ] ) .
we need to keep only the second - order terms . since the integrand is a quadratic form with respect to @xmath140
, one can put @xmath141 in it .
the resulting expression can be integrated to obtain the explicit expression for the dissipation function @xmath142 it is easy to find out that @xmath143 always holds as it should be .
it is more difficult to obtain the drop interface area @xmath144 where the function @xmath145 is defined by eq .
( [ ell ] ) . after the integration over @xmath146 ,
( [ avl ] ) reduces to @xmath147 where @xmath148 .
the subsequent development of the integrand into the series over @xmath149 and its integration term - by - term results in @xmath150\label{avl2}.\end{aligned}\ ] ] this expression can be developed into a series with respect to @xmath151 by using eqs .
( [ vv ] , [ rel]-[rr ] ) .
its substitution into eq .
( [ uu ] ) leads to the explicit expression for @xmath25 : @xmath152\bigr\ } \label{u2},\end{aligned}\ ] ] where @xmath153/[45(1+\cos\theta_s)],\\ b=\left[(72 + 409\,\cos\theta_s+346\,\cos^2\theta_s+73\,\cos^3\theta_s)\cdot\right.\\\left.(1-\cos\theta_s)\right]/[45(1+\cos\theta_s ) ] .
\end{array } \label{ab}\ ] ] it is easy to show that the expression in the square brackets in eq .
( [ u2 ] ) is positive for an arbitrary @xmath13 .
it means that the function @xmath154 has its minimum at the point @xmath155 , i.e. for the drop that has the shape of the spherical cap .
this result was expected . the substitution of eqs .
( [ u2 ] , [ tres ] ) into eq .
( [ ut ] ) written for @xmath156 results in the set of eqs .
( [ kin ] ) and thus concludes their derivation .
the accepted assumptions facilitate calculation of the @xmath157 . the resulting function is periodical with the period @xmath158 , so that for @xmath117 it can be presented in the form @xmath159/\lambda , & 0\leq \rho < r \\
\varepsilon^2\left(\rho-\lambda/4\right ) , & r\leq \rho
< { \frac{\lambda}{2}}-r \\ \varepsilon^2\left(\rho-\lambda/2\right)+\\\bigl\{r^2\arcsin\left[\left ( \lambda/2- \right.\right.\\\left.\left .
\rho\right)/r\right]+ \left(\lambda/2-\rho\right)\bigl[r^2-\\\left(\lambda/2 -\rho\right)^2 \bigr]^{1/2}\bigr\}/\lambda , & { \frac{\lambda}{2}}-r\leq \rho<{\frac{\lambda}{2 } } \end{array } \right .
\label{dg1}\ ] ] where @xmath160 is the fractional part of @xmath161 , multiplied by @xmath158 , and @xmath162 .
since @xmath163 , the presence of defects generates many local minima of the function @xmath164 near its global minimum .
these minima represent the metastable states . according to this model
, the contact line is pinned in the minimum closest to its initial position .
|
we present a theoretical study related to a recent experiment on the coalescence of sessile drops .
the study deals with the kinetics of relaxation towards equilibrium , under the action of surface tension , of a spheroidal drop on a flat surface .
for such a non - spherical drop under partial wetting conditions , the dynamic contact angle varies along the contact line .
we propose a new non - local approach to the wetting dynamics , where the contact line velocity depends on the geometry of the whole drop .
we compare our results to those of the conventional approach in which the contact line velocity depends only on the local value of the dynamic contact angle . the influence on drop dynamics of the pinning of the contact line by surface defects
is also discussed .
|
black phosphorus ( bp ) is a thermodynamically stable allotrope of phosphorus with a layered structure .
the layers of covalently bonded atoms are held together by the van der waals interaction .
similar to obtaining graphene from graphite by mechanical exfoliation , bp can be isolated to a few layers @xcite .
the resulting structure is a recent addition to the family of two - dimensional ( 2d ) materials called multilayer phosphorene .
monolayer phosphorene has a direct band gap of 1.45 ev @xcite . in multilayer phosphorene ,
the gap decreases with increasing number of layers owing to relatively strong van der waals interactions between the layers and remains direct @xcite .
this makes phosphorene a promising candidate for electronic and optical applications @xcite .
moreover , the phosphorene crystal structure is highly anisotropic , which gives rise to phenomena such as anisotropic electronic and thermal transport @xcite , linear dichroism @xcite , and anisotropic plasmons @xcite . additionally , owing to its heavily puckered structure , phosphorene is highly tunable by strain @xcite and electric field @xcite . in particular
, applying an electric field normal to the layers reduces the band gap , leading to a transition from a moderate - gap semiconductor to a semimetal @xcite . applying an electric field
also leads to the emergence of more exotic features of phosphorene , including nontrivial topological phases @xcite .
however , the transition of the landau levels ( lls ) under the influence of electric field as a trademark of the topological phase transition has not been fully understood .
furthermore , little is known about the possible practical use of electric field modulation in tuning the electronic characteristics of nanostructures based on phosphorene . in this paper
, we investigate the effects of a vertical electric field on the electronic properties ( band structure and electronic transport ) of multilayer phosphorene and its nanoribbons . in multilayer phosphorene at low fields , electrons with momenta in the zigzag direction [ fig .
[ fig1](a ) ] have parabolic bands , but in the armchair direction they behave as massive dirac fermions with a gap - dependent effective mass . at a critical electric field
@xmath0 , the gap closes , and electrons exhibit a massive - to - massless dirac fermion transition . above @xmath0
, there are two dirac points , and the band structure is that of anisotropic massless dirac fermions .
this continuous massive - to - massless dirac fermion transition could be observed in hall measurements , as the ll energy dependence on the magnetic field would change from linear below @xmath0 , to the novel 2/3-power at @xmath0 , to the square - root dependence above @xmath0 . in phosphorene nanoribbons ( pnrs ) , we show that there are twofold - degenerate bands within the bulk gap , which govern electronic transport .
the associated wave functions are localized near the ribbon edges .
we propose a dual - edge - gate structure that affects these midgap states and drives the conducting - to - insulating transition in pnrs , thus enabling field - effect transistor action compatible with modern nanoelectronics , and potentially leading to new pnr - based devices .
schematic of bilayer phosphorene .
( a ) top view .
( b ) side view .
the left and the right edges are zigzag , while the top and bottom edges are armchair .
yellow ( blue ) circles are the phosphorene atoms in the upper ( lower ) layer while the atoms in the upper ( lower ) sublayer are portrayed a bright ( dark ) color .
the gray rectangle denotes the unit cell of bilayer phosphorene .
higher multilayer systems will have the unit cell with the same top view and with additional 4 atoms per layer . ]
the band structure of multilayer phosphorene is described by a tight - binding ( tb ) hamiltonian @xmath3 which was parametrized on the basis of first - principles calculation within the @xmath4 approximation @xcite .
as the tb parametrization was benchmarked for mono- , bi- and trilayers , we restrict our study here to these three systems , but note that tb may be suitable for larger multilayers , as well .
a phosphorene monolayer contains two sublayers in a puckered structure and has four atoms per unit cell .
each additional layer brings four more atoms to the unit cell ( see the structure of bilayer phosphorene in fig . [ fig1 ] ) , with odd - numbered layers aligned with other odd - numbered ones , and analogously for even - numbered layers . a unit cell for bilayer phosphorene
is denoted by the shaded box in fig .
[ fig1](a ) . in the nearest - neighbor approximation
, there are fifteen relevant tight - binding hopping parameters ( table [ t1 ] ) , ten intralayer and four interlayer .
the multilayer phosphorene structure is highly anisotropic .
cutting phosphorene in the ( horizontal ) @xmath5 direction ( see fig . [ fig1 ] ) would result in an armchair edge , while cutting along the ( vertical ) @xmath6 direction would result in a zigzag edge .
we refer to the @xmath5-direction as armchair and the @xmath6-direction as zigzag .
ccccccccc @xmath7 & @xmath8 & @xmath9 & @xmath7 & @xmath10 & @xmath9 & @xmath7 & @xmath10 & @xmath9 + 1 & @xmath11 & @xmath12 & 6 & @xmath13 & @xmath14 & @xmath15 & @xmath16 & @xmath17 + 2 & @xmath18 & @xmath19 & 7 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 + 3 & @xmath25 & @xmath26 & 8 & @xmath27 & @xmath28 & @xmath29 & @xmath30 & @xmath31 + 4 & @xmath32 & @xmath33 & 9 & @xmath34 & @xmath35 & @xmath36 & @xmath37 & @xmath38 + 5 & @xmath39 & @xmath40 & 10 & @xmath41 & @xmath42 & @xmath43 & @xmath44 & @xmath45 +
the band structure of trilayer phosphorene is depicted in fig . [ fig2](a ) ; note the rectangular brillouin zone ( bz ) and its high - symmetry @xmath46 and @xmath47 points along the @xmath6 and @xmath5 directions .
the band gap , @xmath48 , is at the @xmath49 point and has a value of @xmath50 ev . along the zigzag direction , both the conduction band ( cb ) and the valence band ( vb ) show quadratic dependencies on the wave vector @xmath51 [ fig . [ fig2](a ) ] .
we calculate the effective masses of @xmath52 ( cb ) and @xmath53 ( vb ) , where @xmath54 is the free - electron rest mass . in the armchair direction , however , the dispersion has an asymptotic linear trend . to describe this highly anisotropic band structure near the band gap
, we propose a low - energy two - band hamiltonian as @xmath55.\ ] ] here , @xmath56 are the pauli matrices , @xmath57 and @xmath58 are the effective masses in the zigzag direction , and @xmath59 is the fermi velocity in the armchair direction . in fig .
[ fig2](b ) , we see that the fit from this effective two - band hamiltonian ( red dots ) agrees very well with the cb and vb dispersions obtained by tb ( solid curve ) within @xmath60 ev of midgap .
finding the dispersion for @xmath61 from the two - band hamiltonian ( [ foo ] ) leads to @xmath62 this is a characteristic dispersion of massive relativistic particles @xcite , where @xmath63 plays the role of rest energy . upon employing the relativistic definition of the effective mass as @xmath64 , we obtain an effective mass for both cb and vb in the armchair direction to be @xmath65 , a value close to the experimentally obtained @xmath66 @xcite , which could not be explained previously based on the parabolic approximation .
thus , it is important to consider electrons in the armchair direction as massive dirac fermions .
( a ) band structure of unbiased trilayer phosphorene over the first brillouin zone .
cuts normal to ( 10 ) and ( 01 ) through the dirac point are projected onto the side walls , with the high - symmetry @xmath49 , x , and y points denoted .
the solid lines are from tb , the red dotted line from the low - energy two - band hamiltonian ( [ foo ] ) .
( b ) magnetic field dependence of cb and vb for trilayer phosphorene in absence of electric field bias .
the linear blue lines are fitted to the yellow dots obtained from the harper s equation . ] since unbiased phosphorene is a conventional 2d electron gas , cb and vb are expected to have a linear dependence on an applied vertical magnetic field @xcite . in order to check this hypothesis , magnetic field is incorporated in the tight - binding model through peierls substitution , which adds a phase to the hopping term between any two sites , @xmath67 where @xmath68 is the vector potential and @xmath69 is the magnitude of the magnetic induction in @xmath70-direction . by this substitution , one will arrive at harper s equation @xcite , and lls are obtained by numerical calculation of its eigenvalues . for unbiased trilayer phosphorene ,
the lls were calculated at various magnetic fields .
the bands corresponding to the cb and vb are shown in fig .
[ fig2](b ) . predictably , cb and vb change linearly as a function of magnetic field strength .
next , we extend our model to the case of a vertical electric field @xmath71 [ applied normal to the layers , i.e. , in the @xmath70-direction in fig .
[ fig1](a ) ] . in the tb calculations
, @xmath71 is accounted for by assuming a linear potential drop across the structure extending from @xmath72 to @xmath73 ( @xmath74 is the thickness ) : a potential energy @xmath75 is added to the diagonal terms of the tb hamiltonian in eq .
( [ eq1 ] ) according to _ i_-th atom s @xmath70 coordinate .
figure [ fig3](c ) shows that , as @xmath71 increases , the cb and vb shift toward each other due to the stark effect .
the band gap closes at a critical electric field , @xmath0 , @xcite which decreases with increasing number of layers : @xmath76 v / for bilayer ( also reported in ) and @xmath77 v / for trilayer phosphorene .
the full band structure of trilayer phosporene at @xmath0 is depicted in fig . [ fig3](a ) .
electric field affects the curvature of the parabolic cb and vb bands along the zigzag direction [ fig .
[ fig3](b ) ] : @xmath57 ( @xmath58 ) decreases ( increases ) with increasing @xmath71 , but @xmath78 for all values of @xmath71 , in keeping with the high hole mobility reported in experiment @xcite .
application of the vertical electric field and the corresponding gap reduction have a strong effect on dispersion in the armchair direction : as @xmath79 , @xmath80 drops with increasing @xmath71 and reaches zero at @xmath81 [ fig.[fig3](c ) ] . therefore , there is a smooth transition from massive to massless dirac fermions in the armchair direction under a vertical electric field . at @xmath0 , the dispersion of cb and vb near
the gap becomes @xmath82 owing to the anisotropic dispersion at the critical electric field
massless dirac [ eq .
( [ eq4 ] ) ] along @xmath5 , parabolic along @xmath6 the electron density of states ( dos ) has a peculiar energy dependence : @xmath83 .
@xcite if such an electron system were placed in a uniform vertical magnetic field @xmath69 , it would exhibit a unique dependence of the ll energy on the @xmath69 field ( stemming from the square - root energy dependence of the dos @xcite ) that has never before been observed in experiment : @xmath84^{2/3},\ ] ] we obtained this dependence [ see fig . [ fig3](d ) ] by numerically solving harper s equation , as described earlier .
the @xmath85-power dependence is distinct from the linear dispersion in two - dimensional electron gases , @xmath86 [ shown in fig .
[ fig2](b ) ] , and from the square - root dependence characteristic of dirac fermions , @xmath87 .
biased multilayer phosphorene would be the first realization of this ll dispersion in a real material , though the phenomenon was thoroughly investigated on a parametric honeycomb lattice @xcite . while yaun _
@xcite , recently investigated lls in perpendicular electric fields , they did not solve the harper s equation @xcite in an extended system , as we did , and thus did not note this peculiar behavior at @xmath0 .
we propose an extension to experiment @xcite , where the gap is closed via doping and mimics applying @xmath0 : a measurement of the hall conductance at this condition would experimentally confirm the peculiar form ( [ eqllec ] ) of the ll @xmath69-field dependence . ( a ) band structure of biased trilayer phosphorene at the critical electric field ( @xmath88 ) .
( b ) effective mass in the conduction and valence bands as a function of electric field along the zigzag direction .
( c ) band gap and relativistic effective mass ( both cb and vb ) in the armchair direction versus electric field .
( d ) energy of the lowest ll in the cb and vb versus magnetic field @xmath69 , obtained by solving harper s equation ( dots ) .
solid lines are fits based on eq .
( [ eqllec ] ) . ] increasing the electric field beyond the critical value ( @xmath89 ) splits the dirac point into two .
the dirac points ( @xmath90 ) move away from the @xmath49 point along the x direction as a function of electric field [ see fig . [
fig4](a ) ] . in the low - energy , two - band hamiltonian ( [ foo ] )
, @xmath91 takes the form @xmath92,\ ] ] where @xmath93 is the value of the inverted gap shown in fig .
[ fig4](a ) .
at each dirac point , the hamiltonian can be expanded in terms of @xmath94 , which gives us the low - energy form @xmath95 , a generic dirac hamiltonian with anisotropic fermi velocity @xmath96 and ultimately the dispersion relation @xmath97 based on this dispersion , the berry phase , @xmath98 , can be calculated by integrating the berry potential , @xmath99 $ ] , over a closed path around each dirac point ; the integration path chosen here is a circle around each dirac point , depicted in fig . [ fig4](a ) .
@xmath100 is obtained from both the two - band low - energy and tb hamiltonians .
@xcite this means that electrons in biased multilayer phosphorene carry an extra degree of freedom called the pseudospin .
dirac fermions exhibit unconventional quantum hall effect , where lls vary with @xmath69 as @xmath87 .
@xcite . magnetic field dependency of cb and vb of multilayer phsosphorene biased at @xmath89 are shown in figs . [
fig4](b ) .
as can be seen , both cb and vb ( yellow circles ) follow the square root power law of dirac fermions ( blue lines ) .
( a ) band structure of biased trilayer phosphorene at @xmath101 .
a pair of dirac points appear in the band structure ( @xmath102 ) .
( b ) cb and vb ( yellow dots ) as a function of magnetic field obtained from diagonalization of harper s equation at ( @xmath89 ) .
the data is fitted with the square root power law of dirac fermions ( blue lines ) . ]
in order to obtain metallic pnrs , we can choose to confine multilayer phosphorene along two different directions , which results in skewed - armchair and zigzag pnrs shown in figs .
[ fig5](a ) and [ fig5](b ) , respectively .
other types of pnrs are disregarded due to insulating behavior @xcite .
the band structures of mono- , bi- , and trilayer zigzag pnrs are shown in figs . [ fig6](a)-(c ) and those corresponding to skewed - armchair pnrs are shown in figs .
[ fig6](d)-[fig6](f ) , respectively .
one can see the presence of midgap bands ( red curves ) completely detached from the bulk bands ( shown in blue ) . for zigzag pnrs , each layer of phosphorene contributes one band of twofold - degenerate midgap states , making a total of two , four , six midgap states for mono- , bi- and trilayer pnrs , respectively .
however in case of skewed - armchair pnrs , each layer provides two , twofold - degenerate midgap states . as a result , a total of four , eight , twelve midgap bands are present for mono- , bi- and trilayer skewed - armchair pnrs , respectively .
crystal structure of bilayer skewed - armchair ( a ) and zigzag ( b ) pnrs .
corresponding unit cells are outlined by the solid line boxes . ] in the absence of a vertical electric field , zigzag pnrs are metallic since the fermi level passes through all midgap bands and is far from the bulk states . in a similar manner ,
metallicity of skewed - armchair pnrs is caused by the two midgap states closest to the fermi level without any contribution from the bulk states [ see figs . [ fig6](d)-[fig6](f ) ] .
therefore , near - equilibrium electronic transport in metallic pnrs is governed by midgap states . in figs . [ fig6](g ) and [ fig6](h ) , we plot the probability density for two @xmath103 wave functions marked in figs . [ fig6](a ) and [ fig6](d ) , respectively , one from a midgap band ( shown in red ) and another from the bulk cb ( shown in blue ) .
in contrast to the states from the bulk bands , whose probability density peaks near the pnr middle , the midgap state s probability density is highest by the edges ( near one edge in the top sublayer and near the other end in the bottom sublayer of a given monolayer ) .
therefore , electronic transport in pnrs should be tunable by the manipulation of the midgap states , confined to the edges .
band structure of ( a ) monolayer , ( b ) bilayer , and ( c ) trilayer zigzag pnrs .
band structures of their skewed - armchair counterparts are shown in ( d)(f ) , respectively .
all ribbons are @xmath104 nm wide .
( g ) and ( h ) probability densities of the states whose energies are marked by `` x '' in panels ( a ) and ( d ) , respectively .
the red ( blue ) circles denote the probability densities of the midgap ( bulk ) states shown by the same color in panels ( a ) and ( d ) . ] based on this finding , we propose a field - effect - transistor ( fet ) structure shown in fig .
[ fig7](a ) , where the conductance in pnrs via the midgap states ( which are confined near the edges ) is modulated locally by two edge gates , with voltages @xmath105 and @xmath106 .
the width of the ribbons ( @xmath107 ) is chosen to be @xmath104 nm , wide enough for experimental realization yet narrow enough to be computationally feasible via atomistic tb and nonequilibrium green s functions ( negf ) @xcite .
the width of each edge gate ( @xmath108 ) is @xmath109 nm and the drain voltage ( @xmath110 ) is fixed at @xmath111 v. figure [ fig7](b ) and ( d ) shows the band dispersion for zigzag and skewed armchair pnrs for @xmath112 v respectively .
the applied bias strongly affects the midgap states , which have a high probability density right under the gate .
the field induced by the bias shifts the bands whose wave function is localized in the upper sublayer upwards and shift the bands whose wave function is in the lower sublayer downward .
( a ) schematic of the proposed dual - edge - gate field - effect structure to control the conductance in pnrs .
( b ) & ( c ) [ ( d ) & ( f ) ] band structure of zigzag ( skewed - armchair ) bilayer pnrs under applied bias on the edge gates .
plots in ( b ) and ( d ) are for bias @xmath112 v , while in ( c ) and ( d ) @xmath113 v. red curves represent the dispersions of the twofold degenerate midgap bands , whose wave functions are confined near the pnr edges .
the dashed horizontal line represents the fermi level .
( f ) current in single and bilayer pnrs as a function of applied bias on the edge gates , indicating a transition from a conductor to an insulator .
the results are limited to single and bilayer pnrs since in trilayer pnrs the van der waals interaction between layers reduces the energy difference between the midgap states and the bulk states [ see fig . [ fig6](c ) and ( f ) ] . as a result
, bulk states are also affected by the bias of the gates and a transition to the `` off '' state is not accessible at room temperature in trilayer pnrs . ]
these edge gates have essentially no effect on the bulk bands .
although the midgap states have moved toward the bulk bands compared to their initial position in the gap [ see figs . [ fig6](b ) and [ fig6](e ) ] , the ribbon remains metallic since the fermi level [ which is fixed by the potentials on the source and drain on the left and right , fig . [ fig7](a ) ] still passes through the midgap bands . increasing the bias voltages to @xmath113 v in figs .
[ fig7](c ) and [ fig7](e ) leads to a further energy separation of the midgap bands , so there are no states at the fermi level , which denotes a transition from conducting [ fig .
[ fig6](b ) and ( d ) ] to insulating behavior [ fig .
[ fig6](c ) and [ fig6](e ) ] .
this transition is also demonstrated by calculating the current for single and bilayer pnrs at room temperature ( @xmath114k ) [ fig .
[ fig7](f ) ] using the ballistic negf formalism . @xcite . at low bias on the edge gates , the current @xmath115 [ fig .
[ fig7](f ) ] is relativily high , indicating that midgap states are crossing the fermi level or located at the vicinity of fermi level . as the voltage magnitude increases , the midgap states are pushed apart . above a threshold bias on the edge gates ,
the current becoms vanishingly small , demonstrating a transition to an insulator [ fig .
[ fig7](f ) ] .
this dual - edge - gate structure is compatible with current fabrication technology , as similar multigate structures were realized previously @xcite . also , the topologically protected midgap states are robust against bulk vacancies @xcite .
a recent calculation showed that edge functionalization can lead to a dramatic flattening of the midgap - band dispersions @xcite , which would significantly reduce the threshold voltage required for the conductor - to - insulator transition .
in summary , we showed that the electron band dispersion for multilayer phosphorene transitions from parabolic in the zigzag and massive dirac fermion in the armchair direction at vertical electric fields below @xmath0 to a dispersion of an anisotropic massless dirac fermion at electric fields above @xmath0 .
we posited that this transition would be observed in quantum hall experiments as an electric - field - dependent change in the landau - level energy vs @xmath69 : @xmath1 below @xmath0 , @xmath2^{2/3}$ ] at @xmath0 , and @xmath2^{1/2}$ ] above @xmath0 .
moreover , we showed that pnrs feature degenerate edge states with energies belonging to midgap bands , which govern low - field conduction .
we proposed a structure with dual edge gates that strongly affect the edge states but leave the bulk ones inert .
the dual - edge - gate structure can induce a transition from a conducting `` on '' state to an insulating `` off '' state by moving the midgap bands away from the fermi level , thereby realizing field - effect transistor action on pnrs .
electric - field modulation of phosphorene and pnrs is a versatile concept that can enable access to new physics and establish a framework for further investigation of phosphorene - based devices .
this work was primarily supported by nsf through the university of wisconsin materials research science and engineering center ( dmr-1121288 ) .
the work was performed using the resources of the uw - madison center for high throughput computing ( chtc ) .
50ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , vol .
( , ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( )
|
we study the effects of a vertical electric field on the electronic band structure and transport in multilayer phosphorene and its nanoribbons .
we find a massive - to - massless dirac fermion transition along the armchair direction at a critical field @xmath0 at which the gap closes .
this transition is observable in quantum hall measurements , as the power - law dependence of the landau - level energy on the magnetic field goes from @xmath1 below @xmath0 , to @xmath2^{2/3}$ ] at @xmath0 , to @xmath2^{1/2}$ ] above @xmath0 .
multilayer phosphorene nanoribbons ( pnrs ) have edge states that govern electrical conduction .
we propose a dual - edge - gate pnr structure that works as a quantum switch .
|
* optical set - up*. ultrashort pulses ( 130 fs , 800 nm , 76 mhz repetition rate , mira , coherent ) are steered onto a @xmath19 pixels reflective slm ( boulder nonlinear systems ) .
the slm is imaged on the back focal plane of the focusing lens ( achromatic lens , f=19 mm , ac127 - 019-b - ml , thorlabs ) with the scattering medium placed after it .
the nanoktp crystals are imaged by an objective ( 40x , 0.75 na , nikon ) on a 12-bit cmos camera ( flea3 , point grey ) and on a large area photon counting pmt ( mp 953 , perkinelmer ) .
the shg signal is spectrally separated with suitable dichroic mirror ( 560 nm longpass , ahf analysentechnik ) , shortpass ( 700 nm , fesh0700 , thorlabs ) and bandpass ( 400 @xmath20 10 nm , chroma technology ) filters .
additionally , a longpass filter ( 650 nm , felh0650 , thorlabs ) and a zero - order half - wave plate ( 46 - 555 , edmund optics ) placed in a motorized rotation stage are used for the excitation beam before the focusing lens .
an analyser ( wp25m - ub , thorlabs ) is used to select the detected polarization state whenever stated .
* broadband transmission matrix measurement , wavefront shaping and nonlinear imaging methodology*. a thorough description of the methods used for acquisition of the transmission matrix can be found elsewhere @xcite .
it consists of a two - step process : first the broadband transmission matrix of the system is acquired using ultrashort pulses . once acquired , the transmission matrix elements are used for selective refocusing in a roi .
briefly , for each input - output channel , phase and amplitude of the corresponding transmission matrix element are obtained by phase - shifting interferometry .
the wavefront phase of a hadamard base ( input channel ) is shifted in respect to a reference field in the range [ @xmath21 with its intensity recorded by a cmos pixel ( output channel ) .
a fourier transform is applied on the output channel interferogram of the @xmath22th hadamard basis thus retrieving its phase and amplitude .
once all the hadamard basis are measured , an unitary transformation is applied to obtain the transmission matrix in the canonical basis @xcite . for acquiring the nonlinear images two methodologies were used . in fig .
[ fig4 ] , we raster scan the refocus at each spatial position in the roi containing the nanoktp and , in parallel , collecting the shg signal integrated within the imaged plane with the large area pmt . in fig .
[ fig5 ] , we raster scan the refocus using the memory effect : a linear phase ramp is added to the wavefront , thus displacing the refocus in the image plane ( with the shg signal acquired in parallel ) .
number of controlled slm segments for the experiments : @xmath23 ( fig .
[ fig4 ] ) and @xmath24 ( figs .
[ fig2 ] , [ fig3 ] , [ fig5 ] ) * sample preparation*. various types of scattering media were used : dispersions of @xmath25m - diameter polystyrene beads ( fig .
[ fig2].a and [ fig3 ] ) , 1-mm thick brain slice ( fig .
[ fig2].b ) both fixed in agarose solution 1 mm - thick parafilm film @xcite ( fig .
[ fig4 ] ) and a commercial diffuser ( fig .
[ fig5 ] , @xmath26 light shaping diffuser , newport ) .
the rat tendon ( fig . [ fig5 ] ) was placed between two coverlips separated by a 120 @xmath27m - thick spacer and filled with agarose solution .
the 1 mm - thick mouse brain coronal cross - section ( fig . [ fig2].b ) followed the same protocol , except the spacer was 1 mm - thick .
the nanoktp crystals ( 150 nm diameter ) were previously characterized by various methods @xcite and were drop cast on a coverslip ( 170 @xmath27m - thickness ) .
the scattering mean free path , @xmath28 , of the scattering media in fig .
[ fig2].a and [ fig3 ] ( @xmath29 m ) were determined by measuring the extinction of the laser through a thin dispersion slab of known thickness , with scattered light and ballistic light spatially isolated in the fourier space @xcite .
anisotropy values @xmath30 for calculating @xmath31 were obtained using the scattering pattern from mie theory .
the authors thank esben andressen and herve rigneault for fruitful discussions and support , arnaud malvache and rosa cossart for providing the brain slices , thierry gacoin and ludovic meyer for providing the nanoktp particles .
this work has been supported by the femto network and contracts anr-10-inbs-04 - 01 ( france - bioimaging infrastructure network ) , anr-11-insb-0006 ( france life imaging infrastructure network ) , and the a*midex project anr-11-idex-0001 - 02 .
is funded by the european research council ( grant 278025 - comedia ) .
h.b.a . and s.b .
conceived and developed the ideas with fundamental input of s.g .
developed and performed all experiments , processed and analyzed the data .
all authors contributed to the writing and editing of the manuscript .
s. m. popoff , g. lerosey , r. carminati , m. fink , a. c. boccara , and s. gigan .
measuring the transmission matrix in optics : an approach to the study and control of light propagation in disordered media .
lett . _ , * 104 * , 100601 , 2010 .
d. j. mccabe , a. tajalli , d. r. austin , p. bondareff , i. a walmsley , s. gigan , and b. chatel .
spatio - temporal focusing of an ultrafast pulse through a multiply scattering medium .
_ nature commun .
_ , * 2 * , 447 , 2011 .
b. judkewitz , y. min wang , r. horstmeyer , a. mathy , and c. yang .
speckle - scale focusing in the diffusive regime with time reversal of variance - encoded light ( trove ) .
_ nature photon .
_ , * 7 * , 300305 , 2013 .
t. chaigne , o. katz , a. c boccara , m. fink , e. bossy , and s. gigan . controlling light in scattering media non - invasively using the photoacoustic transmission matrix
. _ nature photon .
_ , * 8 * , 5864 , 2013 .
s. tripathi , r. paxman , t. bifano , and k. c. toussaint .
vector transmission matrix for the polarization behavior of light propagation in highly scattering media .
express _ , * 20 * , 1606716076 , 2012 .
j. duboisset , d. at - belkacem , m. roche , h. rigneault , and s. brasselet . generic model of the molecular orientational distribution probed by polarization - resolved second - harmonic generation .
a _ , * 85 * , 043829 , 2012 . f .- z .
bioud , p. gasecka , p. ferrand , h. rigneault , j. duboisset , and s. brasselet .
structure of molecular packing probed by polarization - resolved nonlinear four - wave mixing and coherent anti - stokes raman - scattering microscopy . _
a _ , * 89 * , 013836 , 2014 .
hsieh , y. pu , r. grange , g. laporte , and d. psaltis . imaging through turbid layers by scanning the phase conjugated second harmonic radiation from a nanoparticle .
express _ , * 18 * , 2072320731 , 2010 .
l. mayer , a. slablab , g. dantelle , v. jacques , a .- m .
lepagnol - bestel , s. perruchas , p. spinicelli , a. thomas , d. chauvat , m. simonneau , t. gacoin , and j .- f .
single ktp nanocrystals as second - harmonic generation biolabels in cortical neurons .
_ nanoscale _ , * 5 * , 84668471 , 2013 .
|
wavefront shaping has revolutionized imaging deep in scattering media @xcite , being able to spatially @xcite and temporally @xcite refocus light through or inside the medium . however , wavefront shaping is not compatible yet with polarization - resolved microscopy given the need of polarizing optics to refocus light with a controlled polarization state . here , we show that wavefront shaping is not only able to restore a focus , but it can also recover the injected polarization state without using any polarizing optics at the detection . this counter - intuitive effect occurs up to several transport mean free path thick samples , which exhibit a speckle with a completely scrambled state @xcite .
remarkably , an arbitrary rotation of the input polarization does not degrade the quality of the focus . this unsupervised re - polarization out of the originally scrambled polarization state paves the way for polarization - resolved structural microscopy @xcite at unprecedented depths .
we exploit this phenomenon and demonstrate second harmonic generation ( shg ) structural imaging of collagen fibers in tendon tissues behind a scattering medium .
+ accessing the structural organization of molecular assemblies is an important aspect of biological imaging .
for instance , this organization determines important functions such as the formation of bio - filaments , but also contributes to disorders such as in the formation of amyloids in neurodegenerative diseases .
polarization - resolved microscopies are able to provide molecular structural insights beyond the diffraction - limited scale @xcite , and , being non - invasive and non - destructive , could in principle be used for deep _ in - vivo _ structural imaging .
unfortunately , light propagation through multiply scattering media , like thick biological tissues , inherently scrambles polarization states .
after injection of a well - defined polarization state and wavevector direction , each successive scattering event rotates the polarization @xcite .
thus , after multiple scattering events the polarization is ultimately scrambled and the wavefront concomitantly distorted generating a speckle ( schematically sketched in fig . [ fig1].a ) .
this unavoidable scrambling makes molecular structural imaging impossible deep in biological media .
most importantly , it is highly detrimental for polarization - resolved studies because no predetermined relation with the incoming polarization state exists anymore : this is indeed a prerequisite for the analysis of the polarized response and its relation to structural modeling @xcite .
wavefront shaping has been introduced to overcome penetration depths limits in scattering media @xcite . by coherently controlling the speckle pattern generated from multiple scattering events ,
one can increase the energy density at targeted positions building a constructive interference locally .
wavefront shaping allows refocusing light through @xcite , or inside @xcite , scattering media thanks to the manipulation of the wavefront phase or amplitude @xcite .
nevertheless , the analysis of polarization effects on wavefront shaping experiments has been often disregarded because , in principle , there are no correlations among the @xmath0 elements of the vectorial transmission matrix @xmath1 @xcite , relating input field @xmath2 with polarization state @xmath3 to output field @xmath4 with polarization state @xmath5 .
thus , in order to control the output polarization state , one needs to perform polarization sensitive measurements of the transmission matrix @xcite . here
, we demonstrate an effect which enables polarization - resolved imaging through scattering media , and forms the basis for prospective deep structural molecular imaging : polarization revival out of a polarization - state - scrambled speckle ( fig . [
fig1].b ) .
we first discuss aspects related to the speckle without wavefront shaping .
we performed experiments at controlled optical thickness ( @xmath6 ) using phantoms with scattering properties similar to biological media ( see methods ) .
to aid the discussion , the inset of fig .
[ fig2].a shows simplified schematics of the polarization state combinations used . fig .
[ fig2].a ( red dashed ) shows the spatially output - averaged degree of linear polarization ( dolp=@xmath7 , where i is intensity ) of the speckle observed outside the medium . as @xmath6 increases , the speckle polarization state decreases its correlation with the initial linearly polarized one as evidenced by the decrease in @xmath8 .
conversely , when performing wavefront shaping the outcome is very different . to find the optimal wavefronts , a _
broadband _ transmission matrix is first acquired @xcite ( see methods ) _ without any selection of output polarization state _ using the experimental layout shown in fig .
[ fig4].a ( fig .
[ fig2].a inset , continuous line ) . in practice , this is equivalent to acquire experimental transmission matrix elements @xmath9 .
once the transmission matrix is known , we can selectively refocus light at targeted positions ( fig .
[ fig2].a , upper right panel ) .
surprisingly , upon rotation of the input polarization state the refocus does not degrade .
we further quantified this resilience to a change in input state as intensity ratio between the two injected polarization states ( @xmath10 ) shown in fig .
[ fig2].a ( continuous line ) .
clearly , only minute degradation of the refocus intensity ratio is seen reaching depths in the diffusive regime ( @xmath11 ) , in striking contrast with the dolp measurements .
a similar effect was observed through an opaque 1 mm - thick brain slice , where the dependence of the focus intensity with respect to the incident polarization is almost constant ( fig [ fig2].b , top left panel ) .
remarkably , the optimally shaped wavefront is able to refocus light with a well - defined polarization state , even though no output polarization state is privileged .
the refocus is indeed highly polarized and correlated with the input one , as evidenced by the complete extinction of its intensity upon placement of an analyzer , as shown in fig .
[ fig2].b ( bottom left panel , see supplementary section 1 for polarization state evaluation of the refocus ) . the results above indirectly reveal correlations of the polarization - related part of the vectorial transmission matrix elements because they imply @xmath12 .
these correlations are formally shown in fig .
[ fig3 ] as cross - correlation of another set of measurements where @xmath13 and @xmath14 are independently measured for the same realization of disorder ( with the polarization state combination used shown in fig .
[ fig3 ] left panels ) . in this analysis ,
which procedure is explained in supplementary section 2 , the peak above the noise confirms that the diagonal elements of the vectorial transmission matrix are highly correlated .
we conclude from these observations that @xmath15 , since the speckle from @xmath16 has a smaller contrast and is moreover weakly correlated with @xmath13 ( see supplementary section 2 for correlation among elements ) .
we further explored the reason for these effects and found out that the transmission matrix correlations and polarization state revival are related to the broadband nature of the acquired transmission matrix .
indeed , these effects are considerably decreased when the transmission matrix is taken with a monochromatic source , as expected @xcite ( see supplementary section 3 for similar experiments performed with monochromatic light ) .
we then exploited these correlations for demonstrating nonlinear structural imaging .
fig .
[ fig4 ] shows the experimental layout and representative results of a wavefront shaping experiment . in a first step ,
a polarization combination is used where analyzer and incoming polarization state are aligned together , that is , a configuration measuring the @xmath13 elements . note that the presence of the analyzer is not necessary for a wavefront shaping experiment , but it supports an unambiguous proof of structural imaging ( see below ) . after acquiring the broadband transmission matrix ( see methods ) , we spatially refocus light in a raster scanning fashion and acquire in parallel the enhanced shg signal from nanocrystals of potassium titanyl phosphate ( nanoktp ) .
the dipolar character of nanoktp ensures that no ambiguity exists in the polarization state evaluation .
the refocused light generates efficient shg from the nanoscopic sources as seen by the two nanoparticles in fig .
[ fig4].b ( left panel ) . in a second step , a second imaging scan using the very same optimal wavefronts is taken ( fig .
[ fig4].b , right panel ) , however with the incoming polarization state rotated 90@xmath17 .
the refocus after the scattering medium still persists and highlights the second particle revealing the orientation - dependent nonlinear efficiency .
this highlights the presence of nanoktp crystals with different orientations , which are identified using the same wavefront .
any oriented nonlinear source could be probed in the same way , highlighting its molecular order organization in a scattering medium .
a similar situation is found , for example , in coherent raman imaging of biological membranes @xcite or myelin sheath in neurons @xcite .
the implication of these observations for structural nonlinear imaging in biologically relevant specimens are shown in fig .
[ fig5 ] . in order to mimic a nonlinear microscopy experiment , we showcase shg imaging of collagen fibers placed after the scattering medium with a characterized transmission matrix .
different from the previous experiment , here we exploit the memory effect to raster scan the refocus @xcite .
this methodology is able to provide highly contrasted shg images of rat tendon specimens where clear fibers can be seen running along the diagonal of the images ( fig .
[ fig5 ] , left panel ) . in this demonstration
, we do not use an analyzer to detect the shg : in an eventual refocusing inside a scattering medium the nonlinearly generated photons polarization state would be scrambled @xcite . therefore , structural imaging is only insightful if performed in a non - analyzed configuration . by fixing the refocus position in a fiber and rotating the excitation polarization state , we can still perform a polarization - resolved study @xcite .
fig .
[ fig5 ] ( right panel ) shows the measured shg intensity ( green circles ) and a fit ( black continuous line ) .
the retrieved nonlinear optical constants of collagen are in excellent agreement with previously reported values @xcite ( see supplementary section 4 for further information on the shg analysis ) . finally , we expect these correlations to have profound implications for deep structural imaging in various approaches . for instance , acoustic - based methods @xcite are able to refocus in the diffuse regime ( @xmath18 ) with cellular resolution .
this means that the acoustic signal from the obtained refocus could be able to perform structural analysis : if the absorbers are organized in an ordered way , a rotation of the refocus optical polarization , using the very same wavefront , should lead to a change in the acoustic signal .
|
in past few decades , network science has discovered a plethora of novel phenomena while trying to mimic real world systems in a better manner .
one of such discovery is an observation of the chimera state .
it was first reported by kuramato et .
al . in 2002 while investigating non locally coupled identical oscillators in a ring network @xcite .
later , it was analyzed and christened by abrams and strogatz in 2006 as chimera state @xcite .
like , its counterpart in greek mythology , a chimera state has come to be referred as a mathematical hybrid state in which coherent and incoherent dynamics coexist in non - locally coupled identical oscillators in a structurally symmetric network .
chimera has been extensively investigated both theoretically @xcite and experimentally @xcite .
it has been observed in plenty of networks including phase oscillators @xcite , chemical @xcite , mechanical oscillators @xcite , neuron models @xcite , planar oscillators @xcite , boolean networks @xcite , 1d superconducting meta material @xcite , etc .
chimera was originally reported for non - locally coupled oscillators , recently it has been reported in feed back delayed networks @xcite , globally coupled networks @xcite , time varying networks @xcite and networks with purely local coupling @xcite . moreover , different types of chimera has been reported including multi cluster chimera @xcite , virtual @xcite , breathing @xcite and two dimensional chimera @xcite .
a recent work suggests emergence of chimera , dependent on nonhyperbolicity of dynamical systems for both the time - discrete and time continuous cases @xcite .
the chimera state has also been reported for various real world networks models such as rosenzweig - macarthur oscillators for ecological networks @xcite .
chimera has also been characterized by the state of the dynamical evolution of the network .
i chimera is characteristic of the hyper chaotic behavior with many positive lyapunov exponents @xcite .
this type of chimera has primarily been observed for time - continuous systems like complex ginzburg - landau oscillators or kuramoto oscillators . in type - ii chimera ,
only spatial chaos has been observed with a rather simple temporal behavior ( mostly periodic ) .
though , this type of chimera has been reported mainly for the time - discrete systems ( maps ) @xcite , it has recently been observed for the time - continuous stuart - landau oscillators as well @xcite .
further , modeling real world complex systems under the multiplex network framework is one of the recent advancements in the network theory @xcite .
we consider a multiplex network consisting of the same nodes across the layers ( fig . [ fig.mul_pic ] ) and investigate the occurrence of chimera state in the multiplex ring networks .
in this structure , each node has exactly the same connection architecture .
we observe type - ii chimera with spatial chaos and periodic temporal behavior .
though , the chimera state , upon multiplexing , remains of the same type as observed for the isolated network , the multiplexing changes the region of incoherence .
dependence of the chimera state on initial conditions is observed for the multiplex networks as already observed for the isolated networks .
additionally we present a measure in terms of distance variable to distinguish between the coherent and the chimera state and to identify the transition point for the coherent - incoherent bifurcation .
we also investigate role of the size of the network in determining the critical coupling strength for the symmetry breaking and thus emergence of the chimera state .
we consider a multiplex @xmath6 ring network with n nodes in each layer , where @xmath6 represents one dimensional symmetric cyclic group with elements being invariant to the permutation operation @xcite . considering @xmath7 as a real dynamical variable at time @xmath8 for the @xmath9 node
, the dynamics of the network can be described as , @xmath10 \label{eq.evol}\ ] ] where @xmath11 represents the coupling strength , @xmath12 represent number of layers , @xmath13 represents the coupling radius defined by @xmath14 , with @xmath15 signifying the number of neighbours in each direction in a layer . the elements of the adjacency matrix a of a network is defined as @xmath16 or @xmath17 depending upon whether @xmath9 and @xmath18 nodes are connected or not .
the diagonal entries @xmath19 represents no self connection in the network .
the adjacency matrix @xmath20 for the multiplex network can be written as , @xmath21 where @xmath22,@xmath23, ... ,@xmath24 represent the adjacency matrix of the first , second,
,@xmath25 layer , respectively and @xmath26 is an unit nxn matrix .
note that the number of nodes in each layer of the multiplex network is same .
a mismatch in the network size of the layers will yield nodes in different layers having different interaction patterns and hence we can not define the chimera state .
we use logistic map @xmath27 with the bifurcation parameter @xmath28 at which individual logistic map exhibits chaotic behavior .
we consider coupling radius @xmath29 for each layer indicating degree of each node being 64 .
a state of the network is defined as spatially coherent if for any node @xmath30 , the spatial distance between the neighboring nodes approaches to zero for @xmath31 @xmath32 geometrically , this signifies a smooth profile of the spatial curve .
smoothness of the curve , signifying the correlated spatial values of the neighboring nodes , is defined with the absence of any discontinuity in the spatial curve . whereas , temporal coherence ( synchronization ) is defined as , @xmath33 for @xmath34 .
therefore , temporal coherence can be written as @xmath35 which leads to a straight line for the spatial curve associated with the temporal coherence .
ring network .
( a ) and ( f ) are for @xmath36 , ( b ) and ( g ) @xmath37 , ( c ) and ( h ) @xmath38 , ( d ) and ( i ) @xmath39 , ( e ) and ( j ) @xmath40 .
number of nodes in each layer remains @xmath41 and coupling radius @xmath42 .
initial transient is taken as 5000 .
results are presented for time range 5000 to 5015.,width=326,height=336 ] appearance of discontinuity in the smooth spatial curve implies coexistence of the coherence and incoherence regions . to demonstrate the absence of smoothness ,
we define a measure based on the spatial distance as follows , @xmath43 which captures the difference of the spatial distances between the neighboring nodes . for smooth spatial profile @xmath44 ,
signifying a symmetric distribution of the distances between neighboring nodes , while discontinuity in the spatial profile , signifying the transition point , is indicated as kinks in the distribution .
we use this measure to find the critical coupling strength for the symmetry breaking and thus resulting the chimera state .
ring network ( a),(c ) for @xmath40 and ( b),(d ) for @xmath45 .
network size is @xmath41 in each layer and coupling range is @xmath42.,width=336,height=192 ] we evolve eq .
[ eq.evol ] starting with a set of special initial conditions and after an initial transient , study the spatio temporal patterns of the multiplex network .
note that uniform or the gaussian distributed random initial conditions lead to either a completely coherent , spanning all the nodes , or a completely incoherent state depending upon the coupling strength .
for @xmath46 leads to the incoherent evolution of all the nodes . motivated from @xcite
, we use a hump back function to generate initial conditions as follows .
we choose an uniform random number @xmath47 for initial state for @xmath9 node within some interval which varies like a gaussian function as : @xmath48 \label{ini_eq}\ ] ] the variance @xmath49 is chosen depending upon the size of the network such that the random variable lies between @xmath17 and @xmath50 .
the same initial condition is used for both the layers in the multiplex network .
a very narrow width of the function leads to almost very close value of the initial conditions for a fraction of nodes leading to the coherent state . and @xmath42
. number of nodes @xmath41 for each layer.,width=336,height=192 ] in the absence of any coupling between the nodes ( @xmath51 ) or for weak couplings , all the nodes evolve independently and no spatial coherence is observed .
for instance , as demonstrated in fig .
[ fig.main](a ) , for @xmath36 , the evolution of the nodes in the multiplex network yields an incoherent state with no correlations in the neighboring nodes . as the coupling strength is increased ,
a partially coherent state emerges at @xmath37 with correlated spatial values of the neighboring nodes in the end and in the middle regions of each layer , however , the spatial range of the incoherent region is more than the coherent region ( fig . [ fig.main](b ) ) .
this coexistence of the coherent and incoherent dynamics corresponds to the chimera state in the multiplex network . .
figure are plotted for @xmath41 and @xmath42 , width=336,height=192 ] the dynamical behaviour of two layers of a multiplex network is a replica of each other manifesting exactly the same spatio - temporal patterns ( fig .
[ fig.main ] ) .
exactly same behaviour is observed for multiplex networks having more than two layers ( fig.[fig.mul_lay ] ) . at the same coupling value , the spatio - temporal dynamics ( fig . [ fig.main ] ( g ) ) reflects non - regular skeletal type pattern in the incoherent regions .
this irregularity of the pattern suggests that , in the multiplex network framework , a node may get attracted to either of the upper or lower region depending on its initial value as reported for the isolated network @xcite .
as we increase the coupling strength further , range of the incoherent region decreases as depicted by fig .
[ fig.main](c ) for @xmath52 . at @xmath45 ,
we observe a sharp discontinuity in the otherwise smooth profile of @xmath53 and the incoherency appears at two distinct points in each layer .
this is a bifurcation point for the coherent - incoherent transition . above this coupling value
, all the nodes in the multiplex network acquire the complete coherent state as indicated by the appearance of a smooth geometric profile at @xmath40 ( fig .
[ fig.main](e ) ) .
fig.[fig.main ] depicts spatial regions of incoherent nodes and thus indicates a non - zero spatial entropy with the periodic temporal dynamics , representing a type ii chimera state .
further , the regions of incoherence in the spatial profile continues to exist for narrower intervals with an increase in the coupling strength ( fig .
[ fig.main ] ) . furthermore ,
in the chimera state , the time evolution of all the nodes in the network depict periodic behaviour with the periodicity two depicting temporal regularity .
the coupled dynamics displays the spatial chaos which is defined by the non - zero spatial entropy given by @xmath54 @xmath55 , where @xmath56 represents fraction of the incoherent nodes in network @xcite .
we show that the distance measure ( eq.[eq.dmsr ] ) is able to easily distinguish between coherent and chimera state . the discontinuous spatial profile ( fig .
[ fig.dismsr1](b ) ) at @xmath45 gives rise to the kinks ( fig .
[ fig.dismsr1](d ) ) in the distance measure distribution signifying transition to the chimera state .
we calculate the spatial entropy as a function of the coupling strength ( @xmath11 ) in order to demonstrate the transition between the chimera to the coherent state .
a transition from the chimera to coherent state is indicated by the discontinuous change in spatial entropy of the network ( fig.[fig.trans ] ) .
ring of multiplexed network are shown .
parameters are @xmath57 and @xmath42 . number of nodes @xmath41 for each layer .
initial transient is taken as 5000 .
results are presented for time range 5000 to 5015.,width=345,height=192 ] in order to see the impact of multiplexcity on the dynamical behavior of nodes , we compare the dynamical state of an isolated 1d lattice with that of the multiplexed with another 1d lattice .
we find that while the chimera state is retained after multiplexing , the dynamical evolution of the network differs .
the multiplexing may enhance or suppress the incoherency .
for instance , at @xmath38 , the isolated network displays the chimera state with incoherence in the middle regions of the spatio temporal pattern ( fig .
[ fig : mul_effect ] ( a ) and ( c ) ) . after multiplexing
, the region of incoherence shrinks to a point discontinuity in the middle and intermediate regions ( fig .
[ fig : mul_effect ] ) .
thus , multiplexing here retains the chimera state as well as the type of the chimera state , but leads to a change in the region of incoherence as well as in the dynamical evolution .
rings for different realizations of the initial conditions . parameters are @xmath58 and @xmath42 .
number of nodes @xmath41 for each layer .
initial transient is taken as 5000 .
results are presented for time range 5000 to 5015.,width=307,height=192 ] furthermore , similar to the isolated networks , in the multiplex networks as well , the chimera exhibits dependency on the initial conditions .
as already discussed in the section iii , the spatial profile displayed by the chimera state can be very different for different profile of initial conditions for the same set of control parameters , namely @xmath59 and @xmath13 .
but interestingly , even for the initial condition given by the same profile as eq .
[ ini_eq ] with a constant value @xmath49 for a given network size @xmath60 , different realizations of the initial conditions can lead to a different incoherent regions .
for example , at @xmath38 , for three different realizations of the initial conditions , all given by eq .
[ ini_eq ] , three different spatio temporal patterns are observed ( fig .
[ fig : initial ] ) .
though , the multi - chimera state is evident for all the realizations , the region of incoherence differs without any consistent behavior .
rings with coupling strength ( a ) @xmath61 ( b ) @xmath62 ( c ) @xmath37 and ( d ) @xmath63 .
coupling range @xmath42 and network size for ( a),(c ) @xmath64 and ( b),(d ) @xmath65 .
initial transient is taken as 5000 .
results are presented for time range 5000 to 5015.,width=316,height=192 ] ring network .
( a),(c ) for @xmath40 and ( b),(d ) for @xmath66 .
network size @xmath64 and coupling range @xmath42 in each layer.,width=307,height=192 ] moreover , we investigated the impact of network size on emergence of chimera state as well as on the critical coupling strength below which one observes incoherent - coherent regime .
[ fig : n_size ] presents chimera state for a multiplex ring network with two different network size .
for both the network sizes , the coupled evolution exhibits the coexistence of coherence - incoherence dynamics .
however , the critical coupling strength for the coherent - incoherent bifurcation increases to @xmath67 ( fig.[fig : n_sizemsr ] ) as compared to @xmath68 for network size @xmath69 as indicated by fig .
[ fig.main ] .
to summarize , we report an emergence of the chimera in the multiplex networks with the layers being represented by 1-d lattice architecture having non - local couplings .
we find that an emergence of the chimera is identical in the mirror layers arising due to the underlying symmetry of the network .
furthermore , while the temporal behavior of the network remains periodic even after multiplexing , the range of the coupling strength for which chimera is observed changes . the chimera in the multiplex network is found to be sensitive to the changes in the initial conditions , which is revealed through the changes in the incoherent region of the dynamical evolution for different sets of initial conditions .
we also show that the critical coupling strength increases with the size of the network .
the results presented here may provide a better understanding to the peculiar nature of the chimera state observed in many natural systems like unihemispheric sleep , ventricular fibrillation , brain networks which incorporates multilayer network architecture @xcite .
sj is grateful to department of science and technology ( dst ) , government of india and council of scientific and industrial research , government of india project grants emr/2014/000368 and 25(0205)/12/emr - ii respectively for financial support .
sg acknowledges dst , government of india , for the inspire fellowship ( if150149 ) as well as the complex systems lab members for timely help and useful discussions .
authors acknowledge conference on nonlinear systems and dynamics 2015 held at iiser mohali where this work was initiated .
omelchenko , o. e. , wolfrum , m. , yanchuk , s. , maistrenko , y. l. & sudakov , o. [ 2012 ] `` stationary patterns of coherence and incoherence in two - dimensional arrays of non - locally - coupled phase oscillators , '' _ phys .
e. _ * 85 * 036210 .
rosin , d. p. , rontani , d. , haynes , n. d. , sch@xmath70ll , e. & gauthier , d. j. [ 2014 ] `` transient scaling and resurgence of chimera states in networks of boolean phase oscillators , '' _ phys .
e. _ * 90 * 030902 .
schmidt , l. , sch@xmath70nleber , k. , krischer , k. & garc@xmath72a - morales , v. [ 2014 ] `` coexistence of synchrony and incoherence in oscillatory media under nonlinear global coupling , '' _ chaos _ * 24 * 013102 .
omelchenko , i. , omelchenko , o. e. , h@xmath70vel , p. & sch@xmath70ll e. [ 2013 ] `` when nonlocal coupling between oscillators becomes stronger : patched synchrony or multichimera states , '' _ phys .
_ , * 110 * 224101 .
semenova , n .
, zakharova , a. , sch@xmath70ll , e . & anishchenko , v . [ 2015 ] `` does hyperbolicity impede emergence of chimera states in networks of nonlocally coupled chaotic oscillators ?
, '' _ epl _ * 112 * 40002 .
omelchenko , i. , riemenschneider , b. , h@xmath70vel , p. , maistrenko , y. l. & sch@xmath70ll , e. [ 2012 ] `` transition from spatial coherence to incoherence in coupled chaotic systems , '' _ phys .
e. _ * 85 * 026212 .
|
chimera is a relatively new emerging phenomenon where coexistence of synchronous and asynchronous state is observed in symmetrically coupled dynamical units .
we report observation of the chimera state in multiplex networks where individual layer is represented by 1-d lattice with non - local interactions . while , multiplexing does not change the type of the chimera state and retains the multi - chimera state displayed by the isolated networks , it changes the regions of the incoherence .
we investigate emergence of coherent - incoherent bifurcation upon varying the control parameters , namely , the coupling strength and the network size .
additionally , we investigate the effect of initial condition on the dynamics of the chimera state . using a measure based on the differences between the neighboring nodes which distinguishes smooth and non - smooth spatial profile , we find the critical coupling strength for the transition to the chimera state . observing chimera in a multiplex network with one to one inter layer coupling is important to gain insight to many real world complex systems which inherently posses multilayer architecture . * emergence of chimera in multiplex network * + * saptarshi ghosh@xmath0 , sarika jalan@xmath1 , + _ @xmath2 complex systems lab , discipline of physics , indian institute of technology indore , khandwa road , simrol , indore 452020 , india + _ @xmath3 centre for biosciences and biomedical engineering , indian institute of technology indore , khandwa road , simrol , indore 452020 ,
india + @xmath4 email : [email protected] + @xmath5 email : [email protected] ( corresponding author ) _ _ *
|
the current census of nearby white dwarfs is complete to within 20 parsecs from the sun , and contains 126 white dwarfs @xcite . since it represents a sample too small for detailed statistical analysis ,
there is a need to extend the complete sample of white dwarfs to a larger volume .
nearby white dwarfs have been traditionally found in catalogs of stars with large proper motions . in order to improve the statistics of the local white dwarf population
, we have been hunting for white dwarfs in the superblink catalog .
+ the superblink proper motion database @xcite is based on a re - analysis of the poss - i and poss - ii plates of the digitized sky survey ( 20 - 45 yr baseline ) .
its 1.3 million stars in the northern sky were identified with a detection level > @xmath2 down to @xmath3 , and a proper motion limit as low as 0.04@xmath4 . because of its low proper motion limit
, the superblink sample effectively eliminates the kinematic bias of the luyten catalogs , while detecting all white dwarfs down to the luminosity function turn - off .
the first step of our analysis was to gather information about suberblink stars , as a complement to photographic @xmath5 , @xmath6 and @xmath7 measurements , j2000 coordinates , and proper motions available for each superblink star ( see @xcite ) .
catalogs were queried in order to obtain galex far - uv ( @xmath8 ) and near - uv ( @xmath9 ) @xcite , sdss @xmath10 @xcite or 2mass @xmath11 @xcite photometry .
apparent @xmath12 magnitudes are estimated using the empirical relation @xmath13 .
stars are then placed in @xmath14 vs @xmath15 , @xmath16 vs @xmath17 or @xmath16 vs @xmath18 reduced proper motion diagrams , depending on the available photometry for each object . when possible , cross - correlations between the diagrams are performed .
selection criteria are applied based on each star s position in the diagram ( more details will be given in limoges et al .
2011 , in preparation ) . finally , confirmed white dwarfs from @xcite are used to obtain color - magnitude calibrations , from which we determine an absolute magnitude and photometric distance for each candidate .
candidates with distance estimates less than 50 parsecs were selected for spectroscopic follow - up .
+ .spectral types of our new white dwarfs sample [ cols="<,^,^,^,^,^,^,^,>",options="header " , ] all spectra have been obtained at kitt peak with the steward observatory 2.3-m telescope , and the noao mayall 4-m and 2.1-m telescopes .
the adopted configurations allow a spectral coverage of @xmath1932005300 and @xmath1938006700 , at a mean resolution of @xmath20 fwhm .
a subset of da and dc stars from our sample of new white dwarfs is shown in figure 1 . in figure 2
, a color - color diagram displays the confirmed white dwarfs and candidates for which @xmath10 data were available .
a summary of the spectral types for the 137 new white dwarfs is shown in table 1 .
finally , a sample of the detailed list of our new white dwarfs , including astrometry and spectral type is given in table 2.limoges/ ]
from our list of candidates , 137 white dwarfs were found .
most of these objets were already part of the literature , without spectral type or white dwarf status confirmation .
the atmospheric parameters ( @xmath21 and @xmath22 ) of 114 stars were measured .
da stars were analyzed using the spectroscopic fitting technique described at length in @xcite and references therein .
detailed photometric analyses of their energy distributions @xcite were performed for cooler objects showing only the @xmath23 hydrogen line and featureless dc stars .
our results are summarized in figure [ correltm ] in a mass vs @xmath24 diagram .
distances were then derived from the atmospheric parameters .
it was found that 82 stars are closer than 50 parsecs , and 57 are within 40 parsecs .
the remaining 23 cool stars will be analyzed as soon as we acquire the photometric measurements .
figures 2 and [ correltm ] demonstrate the effectiveness of our method in detecting cool white dwarfs .
we are now reaching the bottom of the luminosity distribution , and on our way to obtaining a statistically complete sample of nearby white dwarfs .
we would like to thank the director and staff of steward observatory for the use of their facilities .
this work was supported in part by the nserc canada and by the fund fqrnt ( qubec ) .
p.b . is a cottrell scholar of the research corporation for science advancement .
holberg , j. b. , sion , e. m. , oswalt , t. , mccook , g. p. , foran , s. , & subasavage , j. p. 2008 , _ apj _ , 135 , 1225 lpine , s. , & shara , m. m. 2005 , _ aj _ , 129 , 1483 adelman - mccarthy , j.k . ; et al .
2009 , vizier on - line data catalog , 2294 gil de paz et al .
2009 , vizier on - line data catalog , 21730185 mccook , g.p . , @xmath25 sion , e.m .
2006 vizier online data catalog , 3235 skrutskie , m.f .
2006 , _ aj _ , 131 , 1163 liebert , j. , bergeron , p. & holberg , j. b. 2005 _ apj _ , 156 47 bergeron , p. , saffer , r. & liebert , j. 1992 , _ apj _ , 394 247 bergeron , p. , leggett , s. k. , & ruiz , m. t. 2001 , _ apjs _ , 133 , 413
|
our aim is to compile a catalog of white dwarfs within 40 parsecs of the sun , in which newly discovered objects would significantly increase the completeness of the current census .
white dwarf candidates are identified from the superblink proper motion database ( @xcite ) , which allows us to investigate stars down to a proper motion limit as low as 40 @xmath0 .
the selection criteria and distance estimates are based on a combination of color - magnitude and reduced proper motion diagrams .
candidates with distances less than 50 parsecs are selected for spectroscopic follow - up .
we present our preliminary sample of spectroscopically confirmed white dwarfs , as well as their atmospheric parameters .
these parameters are obtained using the spectroscopic technique developed in @xcite for da stars .
db , dq , and dz stars are also analyzed spectroscopically . for featureless spectra as well as those showing only @xmath1 , we perform a detailed photometric analysis of their energy distribution . address = dpartement de physique , universit de montral , c.p .
6128 , succ .
centre - ville , montral , qubec h3c 3j7 , canada address = dpartement de physique , universit de montral , c.p . 6128 , succ .
centre - ville , montral , qubec h3c 3j7 , canada , address = department of astrophysics , division of physical sciences , american museum of natural history , central park west at 79th street , new york , ny 10024 ,
|
despite the great success of einstein s general theory of relativity , it is generally believed that it has to be modified at small distances , incorporating quantum effects of spacetime . to achieve this goal and arrive at a consistent theory of quantum gravity ,
a number of different approaches have been proposed , including string theory and loop quantum gravity as best known examples . for another prominent approach ,
see the contributions on causal dynamical triangulations to this proceedings .
the aim of these models is to provide a microscopic description of quantum spacetime subsequently to make contact to more macroscopic phenomena , like e.g. our universe . in doing
so it turns out that it is quite hard to connect the very small length scales on which these models are defined with the large scales on which observable physics takes place , e.g. cosmic inflation or particle physics .
a complementary approach towards quantum gravity is to construct effective theories as an intermediate step between general relativity and a full theory of quantum gravity and study physical applications within it .
these results can then possibly be used to connect full quantum gravities to physical phenomena .
there have been many approaches in this direction , from which we choose the approach of noncommutative ( nc ) gravity based on a deformation of symmetries @xcite ( see also the review @xcite ) . for a generalization to nc supergravity ,
see @xcite .
the main idea behind this formalism is to replace the classical symmetries of general relativity ( i.e. the diffeomorphisms ) by a twist deformed hopf algebra of diffeomorphisms , which can be interpreted as quantum symmetries . as a result of this postulate one obtains that these theories naturally give rise to noncommutative spacetimes .
thus we do not deal with a quantized metric field , but rather with a metric field on a quantum spacetime .
we will give an introductory overview of the noncommutative gravity theory developed in @xcite .
for a deeper discussion we refer to the original works and the review @xcite . after this
, we will shortly review noncommutative symmetry reduction , discussed previously in @xcite .
we will keep the discussion very brief and refer to the original paper for more details . as a next step
, we will show under which conditions deformed symmetric models solve noncommutative einstein equations and discuss some phenomenology of deformed cosmological and black hole models .
more details on exact solutions in nc gravity can be found in @xcite , or in the related works of schupp and solodukhin @xcite and aschieri and castellani @xcite .
the basic idea of noncommutative geometry is to replace the usual coordinate functions @xmath1 on a manifold by noncommuting operators @xmath2 , satisfying nontrivial commutation relations [ x^,x^ ] = i^(x ) .
analogously to quantum mechanics , the nontrivial commutators imply uncertainty relations x^x^0 , leading to an upper bound on the resolution in spacetime . for our purposes , a more suitable approach to noncommutative geometry
is to use @xmath0-products instead of operators .
the basic idea is to modify , i.e. to deform , the point - wise product in the algebra of functions on the manifold , leading to an associative , but noncommutative product .
the @xmath0-products we will use in this work , the so - called rjs products @xcite , are given by [ eqn : rjsprod ] ( f g)(x ) : = f(x ) ( ^ ) g(x ) , where @xmath3 are mutually commuting vector fields on the manifold satisfying @xmath4}=0 $ ] and @xmath5 is a constant and antisymmetric matrix , which can be chosen in the canonical ( i.e. darboux ) form .
the most widely studied @xmath0-product , the moyal - weyl product on @xmath6 , is part of this class .
this can be seen by choosing the partial derivatives as twist generating vector fields , i.e. @xmath7 .
note that due to the noncommutativity of the @xmath0-product , we can have nontrivial commutators as well [ fg ] : = fg - gf0 . as a next step ,
we want to discuss the symmetries of theories including @xmath0-products . for this
, we have to note that the rjs @xmath0-product ( [ eqn : rjsprod ] ) can be obtained by using a so - called ( abelian or rjs type ) drinfeld twist given by [ eqn : twist ] = ( - ^ x_x _ ) . the @xmath0-product ( [ eqn : rjsprod ] ) is then given by [ eqn : rjsprodtw ] fg = ( ^-1 fg ) , where @xmath8 is the usual pointwise multiplication map and @xmath9 is the inverse twist .
furthermore , @xmath10 denotes the action of vector fields on functions and is defined via the lie derivative @xmath11 .
having identified the twist , we can construct the twisted symmetries of rjs deformed spaces . in order to do this , we start with the classical infinitesimal diffeomorphisms given by the lie algebra of vector fields on the manifold @xmath12})$ ] . from
that we can construct in a canonical way a hopf algebra @xmath13 , where @xmath14 is the universal envelopping algebra of vector fields , @xmath15 is the related multiplication and @xmath16 is the coproduct ( leibniz rule ) .
we will not use the antipode @xmath17 and the counit @xmath18 in the following and refer to @xcite for a definition .
the coproduct is an algebra homomorphism and acts on the generators as follows ( u):= u1 + 1u , ( 1):= 11 . in order to deform the diffeomorphisms , such that they are compatible with the rjs product , we just have to deform the coproduct as follows _ ( ) : = ( ) ^-1 , where @xmath19 .
it can be shown that @xmath20 is still a hopf algebra , but now in general a noncocommutative one .
we obtain the compatibility condition of the coproduct @xmath21 ( fg ) = _ ( fg ) = _ ( _ ( ) fg ) , where @xmath22 is the @xmath0-multiplication map .
the @xmath0-product ( [ eqn : rjsprodtw ] ) can be generalized to @xmath0-tensor products , @xmath0-wedge products , etc . , by using the corresponding action of vector fields on the involved objects , i.e. using the appropriate lie derivatives @xcite .
as it was shown in @xcite , one can construct covariant derivatives and riemannian curvature on twist deformed spaces .
these objects are deformed covariant under the hopf algebra symmetries by using suitable combinations of the classical objects and the twist . in this article
, we will not review the nc riemannian geometry in a basis free form as presented in @xcite , but we will give all formulae in a particular simple basis of vector fields and one - forms .
this will lead to a better readability of this review article .
it can be shown that for all rjs twists ( [ eqn : twist ] ) , which are analytical almost everywhere , we can find a densely defined basis of vector fields @xmath23 , such that [ eqn : basis ] [ e_a , x_]=0,[e_a , e_b]=0 , for all @xmath24 .
here @xmath25 denotes the dimension of the manifold .
this means in particular that the twist acts trivially on the basis vector fields .
the associated dual basis @xmath26 defined by the pairing @xmath27 has the same property . making use of the natural basis ( [ eqn : basis ] )
, we obtain the following formulae for the geometrical quantities : given a metric @xmath28 , the inverse metric @xmath29 fulfills the condition g_abg^ca= g^acg_ba= ^c_b , thus it is simply the @xmath0-inverse matrix of @xmath30 .
the associated torsion - free and metric compatible connection , the @xmath0-levi - civita connection , is given by [ eqn : connection ] _
ab^ c= ( e_a(g_bd ) + e_b(g_ad ) -e_d(g_ab))g^cd , where @xmath31 denotes the action of the vector field @xmath32 on _ functions _
( i.e. the lie derivative on functions ) .
the @xmath0-covariant derivative of a tensor field is given by ( ^_e_c)_b_1 b_l^a_1
a_n = e_c(^a_1 a_n_b_1 b_l ) -_cb_1^ b^a_1
b_l - + ^a_1 a_b_1
b_l_c a^ a_n .
furthermore , we obtain the @xmath0-riemann tensor [ eqn : riemann ] _ abc^ d = e_a(_bc^ d ) - e_b(_ac^ d ) + _
bc^ e_ae^ d - _
ac^ e_be^ d , the @xmath0-ricci tensor @xmath33 and the @xmath0-curvature scalar @xmath34 .
one way of writing the nc dynamics of the metric field are the nc einstein equations [ eqn : einstein ] g_ab:= _ ab- g_ab = 8 g_n t_ab , where we have introduced the nc einstein tensor @xmath35 , a stress - energy tensor @xmath36 and the newton constant @xmath37
. other possibilities of defining the dynamics of the nc metric field , like e.g. constructing a deformed version of the einstein - hilbert action , will be discussed elsewhere . in the following
, we will construct solutions of the nc einstein equations ( [ eqn : einstein ] ) .
this is in general a very hard task , since these equations are nonlinear and also nonlocal due to the @xmath0-products . as in the classical case
, symmetry reduction will turn out to be a helpful tool for constructing symmetric solutions .
in classical gravity , it turned out that splitting the problem of describing physical situations into finding background solutions and afterwards perturbing them by field fluctuations on those fixed backgrounds is very fruitful for practical applications .
in particular , if the problem obeys some approximate symmetries , like for example isotropy and homogeneity in cosmology , this approach has found many successful applications . using the example of cosmology , the typical steps are the restriction of the metric and inflaton field to spherical symmetric and translation invariant configurations , leading to the friedmann equations , which are in general easier to solve than the full einstein equations . after solving for the background fields
, one allows small fluctuations of the metric and scalar field in order to determine , e.g. , the fluctuations in the cosmic microwave background ( cmb ) . in this article
, we will focus on the first step in the approach described above in a noncommutative setting .
the second step , i.e. the ( quantum ) field theory of fluctuations , will be discussed elsewhere . to reach our goal
, we have to generalize the usual lie algebra symmetries to quantum lie algebra symmetries , which will be the basic ingredients of nc symmetry reduction .
furthermore , we will discuss the application of nc symmetry reduction to physical problems , in particular to cosmology and black holes .
while in classical gravity the symmetries of physical problems can often be described by lie algebras @xmath38})$ ] , in noncommutative gravity this mathematical structure has to be replaced by a so - called quantum or @xmath0-lie algebra @xcite . for the twist deformed case ,
@xmath0-lie algebras can be easily constructed by using the twist @xcite .
for example , the quantum analogon of the lie algebra of vector fields @xmath12})$ ] is given by @xmath12}_\star)$ ] , where the @xmath0-lie bracket @xmath39}_\star = { [ \bar f^\alpha ( u),\bar f_\alpha(v)]}$ ] is constructed using the commutative lie bracket and the inverse twist @xmath40 .
the same construction can be performed for every classical lie algebra @xmath38})$ ] , if the twist is constructed from elements of @xmath41 . in nc gravity ,
the twist is i.g .
constructed using general vector fields , not necessarily related to the symmetries one wants to implement .
this means that we have to look for compatibility conditions among a given classical symmetry lie algebra @xmath41 and a twist @xmath42 . for the case of rjs twists ( [ eqn :
twist ] ) , this will reduce to compatibility conditions among the vector fields @xmath43 and the lie algebra @xmath41 .
this was done in @xcite and we found out that the condition [ x _ , ] , _ is a necessary and sufficient condition to turn @xmath38}_\star)$ ] into a quantum lie subalgebra of @xmath12}_\star)$ ] . based on this condition we were able to give a classification of compatible @xmath44-pairs for rjs twists and cosmological and black hole symmetries @xmath41 .
we refer to @xcite for the results and only state some models of particular physical interest in section [ sec : examples ] .
this method can also be applied to study deformations of other symmetric spaces , such as for example ( anti ) de sitter spaces or black brane scenarios . having a quantum lie algebra symmetry
, we have to study the solution of the quantum invariance condition on tensor fields @xmath45 given by [ eqn : invariance ] _
^ ( ) : = _ |f^ ( ) |f_()=0 , which is a deformed version of the statement that the symmetries should annihilate symmetric tensors ( this is the very definition ) . in this expression
we used the deformed lie derivative @xmath46 constructed by using the twist .
we were able to prove that the quantum invariance condition and the classical one @xmath47 are equivalent , such that the @xmath0-invariant tensor fields are just the ones invariant under the classical counterpart of the symmetries .
in table [ tab : comparison ] we compare the classical and deformed version of symmetry reduction .
.[tab : comparison ] comparison between classical ( cl . ) and nc symmetry reduction .
[ cols="<,^,^",options="header " , ]
in this section , we will study deformed @xmath41-symmetric solutions of the nc einstein equations ( [ eqn : einstein ] ) .
for this we use the framework explained in section [ sec : symmetryreduction ] , see in particular table [ tab : comparison ] , and use as ansatz @xmath0-symmetric tensor fields satisfying @xmath48 .
this results in consistently deformed equations for the symmetry reduced tensor fields , like e.g. deformed friedmann equations in cosmology .
as it was recognized and explained in @xcite , as well as in the related works @xcite , the deformed symmetric equations of motion reduce to the undeformed ones , if the twist ( [ eqn : twist ] ) satisfies [ eqn : semikilling ] ^ x_x _ + .
twists of this form are called semi - killing , since using the canonical form of @xmath5 , half of the vector fields @xmath43 have to be killing vector fields , i.e. @xmath49 .
the fact that all @xmath0-products between @xmath0-symmetric tensor fields drop out can be easily understood , since acting with ( [ eqn : semikilling ] ) on @xmath0-symmetric tensor fields is trivial due to their remaining classical symmetries @xmath50 , see table [ tab : comparison ] .
since , at least in principle , we know how to solve the classical equations of einstein gravity coupled to matter , this leads to solutions of the nc einstein equations ( [ eqn : einstein ] ) .
exact solutions beyond the semi - killing case are more complicated , but we have found one particular example in @xcite , see also section [ sec : examples ] .
furthermore , in @xcite affine killing twists have been considered , leading to another class of exactly solvable models . as a last note ,
we want to add that the reduction of the deformed symmetry reduced dynamics to the undeformed one in the semi - killing case does not mean that these models are trivial .
in contrast , these models will indeed receive distinct nc effects from field fluctuations on the background solutions , as well as from uncertainties of the coordinate operators .
in this section we will give some examples of deformed cosmological ( friedmann - robertson - walker ) models and deformed schwarzschild black holes .
for a complete classification of these models see @xcite .
furthermore , see @xcite for a deeper discussion of what follows .
consider the following families of deformed frw models @xmath51 and black hole models @xmath52 : [ eqn : models ] [ eqn : c22 ] & _
22:x_1=x_1 ^ 0(t)_t , x_2=d_+ f_2 r_r , + [ eqn : c32 ] & _
32:x_1=x_1 ^ 0(t)_t+d _ , x_2=x_2 ^ 0(t)_t + f_2 r_r , + [ eqn : b12 ] & _ 12:x_1 = c_1 ^ 0 _ t + _ 1 _ , x_2 = c_2 ^ 0(r ) _ t + _ 2 _ + f(r ) _ r .
here we used ( comoving ) spherical coordinates @xmath53 , as well as some freely chooseable parameter(function)s .
these models are examples within the classification of admissible @xmath44-pairs given in @xcite .
the associated @xmath0-commutation relations among appropriate coordinate functions are given by [ eqn : commutators ] @xmath54}&=-2\exp i\phi~\sinh\bigl(\frac{\lambda d}{2 } x_1 ^ 0(t)\partial_t\bigr ) t\\ { [ t\stackrel{\star}{,}r ] } & = 2 i r ~\sin\bigl(\frac{\lambda f_2}{2 } x_1 ^ 0(t)\partial_t\bigr)t \end{aligned } \right .
\\ \label{eqn : a32 } \mathfrak{c}_{32}:\ ; & \left\ { \begin{aligned } { [
t\stackrel{\star}{,}\exp i\phi]}&= 2 \exp i\phi~\sinh\bigl(\frac{\lambda d}{2 } x_2 ^ 0(t)\partial_t\bigr ) t\\ { [ t\stackrel{\star}{,}r ] } & = 2 i r ~\sin\bigl(\frac{\lambda f_2}{2 } x_1 ^ 0(t)\partial_t\bigr)t\\
\exp i\phi\star r & = e^{-\lambda d f_2}~r\star \exp i\phi \end{aligned } \right . \\
\label{eqn : a12 } \mathfrak{b}_{12}:\ ; & \left\ { \begin{aligned } { [ t\stackrel{\star}{,}\exp i\phi ] } & = \exp i\phi~ \bigl(2 \sinh\bigl(\frac{\lambda\kappa_1}{2}\bigl(c_2 ^ 0(r)\partial_t + f(r)\partial_r\bigr)\bigr ) t -\lambda \kappa_2 c_1 ^ 0\bigr)\\ { [ t\stackrel{\star}{,}r ] } & = i\lambda c_1 ^ 0 f(r)~,\\ { [ \exp i\phi\stackrel{\star}{,}r ] } & = -2 \exp i\phi ~\sinh\bigl(\frac{\lambda\kappa_1}{2}f(r)\partial_r\bigr)r \end{aligned } \right .. \end{aligned}\ ] ] it turns out that these models are semi - killing for the choice @xmath55 in model @xmath56 , @xmath57 in model @xmath58 , as well as for all choices of parameters in model @xmath59 .
this leads to exact solutions describing nc cosmologies and black holes .
next , we will discuss physical implications of the nontrivial coordinate algebras for some of our models .
consider for example the model @xmath56 ( [ eqn : c22 ] ) with @xmath55 and for simplicity @xmath60 .
then the coordinate algebra ( [ eqn : a22 ] ) reduces to the algebra of a quantum mechanical particle on the circle , i.e. [ e , t]=e , where we introduced the abstract operators @xmath61 and @xmath62 and set @xmath63 .
this algebra previously appeared e.g. in the context of noncommutative field theory @xcite and the noncommutative btz black hole @xcite .
it is well known that the operator @xmath61 can be represented as a differential operator acting on the hilbert space @xmath64 of square integrable functions on the circle and the spectrum can be shown to be given by @xmath65 , where @xmath66 labels unitary inequivalent representations .
the spectrum should be interpreted as possible time eigenvalues .
thus our model with discrete time can be used to realize singularity avoidance in cosmology .
consider for example an inflationary background with scale factor @xmath67 , where @xmath68 is a parameter .
this so - called power - law inflation can be realized by coupling a scalar field with exponential potential to the geometry even in our nc model , since the symmetry reduced riemannian geometry reduces to the undeformed one as explained above .
note that the scale factor goes to zero at the time @xmath69 and leads to a singularity in the curvature scalar .
but as we discussed above , the possible time eigenvalues are @xmath70 , which for @xmath71 does not include the time @xmath69 . taking a look at the coordinate algebra of the black hole ( [ eqn : a12 ] ) , we observe that it includes in particular the algebra of a quantum mechanical particle on the circle for time and angle variable , if we choose @xmath72 and @xmath73 .
this leads to discrete times .
another simple choice is @xmath74 , @xmath72 , @xmath75 and @xmath76 .
the radius spectrum in this case is @xmath77 , describing a fine grained geometry around the black hole .
the phenomenological problem with this model is that the spacings between the radius eigenvalues grow exponentially in @xmath78 .
this can be fixed by considering a modified twist like e.g. @xmath74 , @xmath72 , @xmath75 and @xmath79 , where @xmath80 is some length scale .
the essential modification is to choose a bounded @xmath81 .
consider the coordinate change @xmath82 , then the algebra ( [ eqn : a12 ] ) in terms of @xmath83 becomes [ e,]=- e , leading to the spectrum @xmath84 .
the spectrum of @xmath85 is then given by @xmath86 .
this spectrum approaches constant spacings between the eigenvalues for large @xmath78 .
we will close this section by giving an example of a non - semi - killing solution of the nc einstein equations ( [ eqn : einstein ] ) .
consider the model @xmath56 with @xmath87 ( [ eqn : c22 ] ) .
it turns out that both @xmath88 , therefore the riemannian geometry does not to reduce to the undeformed one .
thus we expect corrections in @xmath89 to the nc einstein equations ( [ eqn : einstein ] ) and its solutions .
consider now the ( undeformed ) de sitter space given by the scale factor @xmath90 , where @xmath91 is the hubble parameter .
it turns out that all @xmath0-products entering the deformed geometrical quantities ( see section [ sec : ncriemann ] ) reduce to the undeformed ones , if @xmath60 .
thus the undeformed de sitter space solves nc einstein equations ( [ eqn : einstein ] ) for this particular choice of twist and an undeformed cosmological constant . note that in contrast to the solutions above , we required the explicit form of the scale factor @xmath92
in this paper we gave an introductory overview of the noncommutative gravity theory defined in @xcite .
this theory is constructed in such a way that it is covariant under a deformed diffeomorphism symmetry , mathematically described by a twist deformed hopf algebra of diffeomorphisms . in section [ sec : symmetryreduction ] we reviewed our approach to noncommutative symmetry reduction @xcite .
we emphasized the role of quantum lie algebras and constructed compatibility conditions among the twist @xmath42 and the desired symmetry @xmath41 .
based on this approach , we discussed the issue of solvability of the noncommutative einstein equations in section [ sec : dynamics ] .
we found out under which conditions the nc einstein equations reduce to the undeformed ones , thus simplifying the search for exact solutions .
we gave explicit examples of quantum lie algebra symmetric cosmological and black hole models solving the noncommutative einstein equations in section [ sec : examples ] .
in a future work @xcite , we will investigate classical and quantum field fluctuations on background solutions .
this is a step required in order to make contact to physics , such as for example the cosmic microwave background ( cmb ) . in order to formulate and quantize field theories in a deformed covariant way
, we will use deformed poisson algebras as defined in @xcite .
i want to thank the organizers and participants of the 49 .
cracow school of theoretical physics 2009 in zakopane for this very interesting conference .
furthermore , i want to thank the faculty of physics and astronomy of the university of wrzburg for the financial support through the `` wilhelm - conrad - rntgen - studienpreis '' . this research is supported by deutsche forschungsgemeinschaft through the research training group 1147 _ theoretical astrophysics and particle physics_. 10 p. aschieri , c. blohmann , m. dimitrijevic , f. meyer , p. schupp and j. wess , `` a gravity theory on noncommutative spaces , '' class .
* 22 * , 3511 ( 2005 ) [ arxiv : hep - th/0504183 ] .
p. aschieri , m. dimitrijevic , f. meyer and j. wess , `` noncommutative geometry and gravity , '' class .
* 23 * , 1883 ( 2006 ) [ arxiv : hep - th/0510059 ] .
t. ohl and a. schenkel , `` symmetry reduction in twisted noncommutative gravity with applications to cosmology and black holes , '' jhep * 0901 * , 084 ( 2009 ) arxiv:0810.4885 [ hep - th ] . t. ohl and a. schenkel , `` cosmological and black hole spacetimes in twisted noncommutative gravity , '' arxiv:0906.2730 [ hep - th ] .
p. aschieri and l. castellani , `` noncommutative d=4 gravity coupled to fermions , '' arxiv:0902.3817 [ hep - th ] ; p. aschieri and l. castellani , `` noncommutative supergravity in d=3 and d=4 , '' arxiv:0902.3823 [ hep - th ] .
n. reshetikhin , `` multiparameter quantum groups and twisted quasitriangular hopf algebras , '' lett . math .
phys . * 20 * , 331 ( 1990 ) . c. jambor and a. sykora ,
`` realization of algebras with the help of * -products , '' [ arxiv : hep - th/0405268 ] .
m. chaichian , a. demichev , p. presnajder and a. tureanu , _ space - time noncommutativity , discreteness of time and unitarity , _ eur .
j. c * 20 * , 767 ( 2001 ) [ arxiv : hep - th/0007156 ] ; a. p. balachandran , t. r. govindarajan , a. g. martins and p. teotonio - sobrinho , _ time - space noncommutativity : quantised evolutions , _ jhep * 0411 * , 068 ( 2004 ) [ arxiv : hep - th/0410067 ] .
p. aschieri , f. lizzi and p. vitale , `` twisting all the way : from classical mechanics to quantum fields , '' phys .
d * 77 * , 025037 ( 2008 ) [ arxiv:0708.3002 [ hep - th ] ] ; p. aschieri , `` star product geometries , '' arxiv:0903.2457 [ math.qa ] .
|
we review the noncommutative gravity of wess et al .
@xcite and discuss its physical applications .
we define noncommutative symmetry reduction and construct deformed symmetric solutions of the noncommutative einstein equations .
we apply our framework to find explicit deformed cosmological and black hole solutions and discuss their phenomenology .
this article is based on a joint work with thorsten ohl @xcite .
|
one of the major goals of cosmic ray physics has always been the discovery and understanding of the _ end of the cosmic ray spectrum_. until the end of the 60s , this search was mainly aimed to understand the limits to the acceleration processes and nature of the sources responsible for the production of the particles with the highest energies .
however , after the discovery of the cosmic microwave background ( cmb ) , it became soon clear that the observed spectrum of the cosmic radiation had to be cut off at a `` natural '' energy , even if an ideal class of sources existed , able to accelerate particles to infinite energy .
in fact , if the sources are distributed homogenously in the universe , the photopion production in the scattering of particles off the cmb photons imply a cutoff in the observed spectrum of the cosmic rays at an energy @xmath1 ev , close to the kinematic threshold for that process @xcite . this cutoff has become known as the _ gzk cutoff _ and particles with @xmath2
are usually named ultra - high energy cosmic rays ( uhecrs ) .
several experiments have been operating to detect the flux of uhecrs , starting with volcano ranch @xcite and continuing with haverah park @xcite and yakutsk @xcite to the more recent experiments like agasa @xcite , fly s eye @xcite and hires @xcite .
the search for the gzk cutoff , instead of confirming the simple picture illustrated above , has provided stronger and stronger evidence for the existence of events corresponding to energies well in excess of @xmath3 : the gzk cutoff has not been found .
this problem hides many issues on plasma physics , particle physics and astrophysics , that in their whole represent the puzzle of uhecrs . in section
[ sec : observations ] the present status of observations of uhecrs is summarized ; in section [ sec : serious ] some caveats in the arguments that are often used to address the problem of uhecrs are considered , for the purpose of stating the problem in a clear way . in sections
[ sec : acceleration ] and [ sec : td ] the bottom - up and top - down models of uhecr origin are summarized . in section [ sec : liv ] some speculations are discussed of new physics scenarios that might play a role not only for the explanation of uhecrs , but also for the understanding of other current puzzles in high energy astrophysics .
conclusions are reported in section [ sec : conclusion ] .
the cosmic ray spectrum is measured from fractions of gev to a ( current ) maximum energy of @xmath4 ev .
the spectrum above a few gev and up to @xmath5 ev ( the knee ) is measured to be a power law with slope @xmath6 , while at higher energies and up to @xmath7 ev ( the ankle ) the spectrum has a steeper slope , of @xmath8 . at energy
larger than @xmath9 ev a flattening seems to be present .
the statistics of events is changing continuously : the latest analysis of the `` all experiments '' statistics was carried out in @xcite where 92 events were found above @xmath10 ev .
47 events were detected by the agasa experiment .
a more recent analysis @xcite of the agasa data , carried out expanding the acceptance angle to @xmath11 , has increased the number of events in this energy region to 59 .
in @xcite the directions of arrival of the agasa events ( with zenith angle smaller than @xmath12 ) above @xmath10 ev were studied in detail : no appreciable departure from isotropy was found , with the exception of a few small scale anisotropies in the form of doublets and triplets of events within an angular scale comparable with the angular resolution of the experiments ( @xmath13 for agasa ) .
this analysis was repeated in @xcite for the whole sample of events above @xmath10 ev , and a total of 12 doublets and 3 triplets were found within @xmath14 angular scales .
the attempt to associate these multiplets with different types of local astrophysical sources possibly clustered in the local supercluster did not provide evidence in that direction @xcite .
recently , the agasa collaboration reported on the study of the small scale anisotropies in the extended sample of events with zenith angle @xmath15 : 5 doublets ( chance probability @xmath16 ) and 1 triplet ( chance probability @xmath17 ) were found .
the information available on the composition of cosmic rays at the highest energies is quite poor .
a study of the shower development was possible only for the fly s eye event @xcite and disfavors a photon primary @xcite .
a reliable analysis of the composition is however possible only on statistical basis , because of the large fluctuations in the shower development at fixed type of primary particle .
the fly s eye collaboration reports of a predominantly heavy composition at @xmath18 ev , with a smooth transition to light composition at @xmath7 ev .
this trend was later not confirmed by agasa @xcite .
recently in ref .
@xcite the data of the haverah park experiment on highly inclined events were re - analyzed : this new analysis results in no more than @xmath19 of the events with energy above @xmath9 ev being consistent with photons or iron ( at @xmath20 confidence level ) and no more than @xmath21 of events being photons above @xmath10 ev .
recently a new mass of data has been presented by the hires experiment @xcite .
only two events with energy above @xmath22 ev have been detected by this experiment insofar , compatible with the presence of a gzk cutoff .
this discrepancy with the results of several years of agasa operation needs further investigation .
several sistematics have been identified that might considerably affect the determination of the energies and fluxes of fluorescence experiments versus the ground array techniques @xcite .
these issues will not be discussed further in the present paper .
the puzzle of uhecrs can be summarized in the following points : * _ : _ the generation of particles of energy @xmath23 ev requires an excellent accelerator , or some new piece of physics that allows the production of these particles in a non - acceleration scenario . * _ : _ observations show a remarkable large scale isotropy of the arrival directions of uhecrs , with no correlation with local structures ( e.g. galactic disk , local supercluster , local group ) . *
_ : _ the small ( degree ) scale anisotropies , if confirmed by further upcoming experiments , would represent an extremely strong constraint on the type of sources of uhecrs and on magnetic fields in the propagation volume . *
_ : _ the gzk cutoff is mainly a geometrical effect : the number of sources within a distance that equals the pathlength for photopion production is far less than the sources that contribute lower energy particles , having much larger pathlength ( comparable with the size of the universe ) .
the crucial point is that the cutoff is present even if plausible nearby uhecr engines are identified . *
_ : _ it is crucial to determine the type of particles that generate the events at ultra - high energies .
the composition can be really considered a smoking gun either in favor or against whole classes of models .
the five points listed above are most likely an oversimplification of the problem : some other issues could be added to the list , such as the spectrum , but at least at present this can not be considered as a severe constraint . on the other hand ,
specific models make specific predictions on the spectral shape , so that when the results of future observations will be available , this information will allow a strong discrimination among different explanations for the origin of uhecrs .
any model that aims to the explanation of the problem of uhecrs must address all of the issues listed above ( and possibly others ) . in this section we consider in some more detail the issue of the gzk cutoff and the seriousness of its absence in the observed data .
it is often believed that the identification of one or a class of nearby uhecr sources would explain the observations and in particular the absence of the gzk cutoff .
this is not necessarily true .
the ( inverse of the ) lifetime of a proton with energy @xmath24 is plotted in fig . 1 ( left panel ) together with the derivative with respect to energy of the rate of energy losses @xmath25 ( right panel ) [ the figure has been taken from ref .
the flux per unit solid angle at energy @xmath24 in some direction is proportional to @xmath26 , where @xmath27 is the density of sources ( assumed constant ) , @xmath28 and @xmath29 is the source spectrum .
this rough estimate suggests that the ratio of detected fluxes ( multiplied as usual by @xmath30 ) , at energies @xmath31 and @xmath32 is @xmath33 where in the last term we assumed that the source spectrum is a power law @xmath34 .
if for instance one takes @xmath35 ev ( below @xmath3 ) and @xmath36 ev ( above @xmath3 ) , from fig . 1
one obtains that @xmath37 for @xmath38 and @xmath39 for @xmath40 .
the ratio @xmath41 gives a rough estimate of the suppression factor at the gzk cutoff and its dependence on the spectrum of the source .
for flat spectra ( @xmath42 ) the cutoff is less significant , but it is more difficult to fit the low energy data @xcite ( at @xmath43 ev ) .
steeper spectra make the gzk cutoff more evident , although they allow an easier fit of the low energy data . the simple argument illustrated above
can also be interpreted in an alternative way : if there is a local overdensity of sources by a factor @xmath44 , the gzk cutoff is attenuated with respect to the case of homogeneous distribution of the sources .
the question of whether we are located in such a large overdensity of sources was recently addressed , together with the propagation of uhecrs , in @xcite .
assuming that the density of the ( unknown ) sources follows the density of galaxies in large scale structure surveys like pscz @xcite and cfa2 @xcite , the authors estimate the local overdensity on scales of several mpc to be of order @xmath45 , too small to compensate for the energy losses of particles with energy above the threshold for photopion production .
there is however another issue that the calculations in @xcite address , which is related to the statistical fluctuations induced by the process of photopion production .
the large inelasticity of this process can be taken into account properly only through the use of montecarlo calculations . when the montecarlo is applied to simulate the small statistics of events typical of current experiments , the fluctuations in the simulated fluxes above @xmath46 ev are very large , so that for flat spectra ( @xmath47 ) the discrepancy between observations and simulations on the total number of events above @xmath46 ev is at the level of @xmath48 ( in agreement with the conclusions of ref .
the situation is represented in fig .
2 @xcite where the hatched regions show the uncertainties in the simulated fluxes .
the data points are from agasa @xcite .
the bottom line of this section can be summarized in the following few points : \1 ) the gzk cutoff is not avoided by finding sources of uhecrs that lie within the pathlength of photopion production , unless these sources are located only or predominantly nearby and are less abundant at large distances .
\2 ) the significance of the gzk feature depends on the fluctuations in the photopion production , and can be addressed properly only with a enhanced statistics of events with energy @xmath23 ev .
models for the origin of uhecrs can be strongly constrained on the basis on the criteria illustrated in the previous section .
the challenge to conventional acceleration models , that are supposed to work at lower energy scales , induced an increasing interest for more exotic generation mechanisms , eventually requiring new particle physics . in this section the main ideas on production scenarios and their signatures will be summarized .
[ [ sec : acceleration ] ] the accelerators able to reach maximum energies of order 1 zev have been named zevatrons @xcite . the challenge for zevatrons was recently discussed in detail by several authors @xcite : the main concept in this class of models is that the energy flux embedded in a macroscopic motion or in magnetic fields is partly converted into energy of a few very high energy particles .
this is what happens for instance in shock acceleration .
a discussion of all the models in the literature is not the purpose of the present paper , and in a sense we think it may not be very interesting .
nevertheless , it is instructive to understand at least which classes of models may have a chance to explain the acceleration to zev energies . in this respect ,
a pictorial way of proceeding is based on what is known as the hillas plot @xcite .
our version of it is reported in fig .
4 . in the hillas plot
the maximum energy is taken to be in its simplest form , as determined by the local magnetic field @xmath49 and the size @xmath50 of the accelerator : @xmath51 . here
@xmath52 is the electric charge of the accelerated particles . from fig .
4 it is evident that only 4 classes of sources have the potential to accelerate protons to ultra - high energies : 1 ) neutron stars ; 2 ) radio lobes ; 3 ) active galactic nuclei ; 4 ) clusters of galaxies . in the case of iron ,
the situation becomes more promising for other sources , like the galactic halo or extreme white dwarfs .
these sources would however have other problems that make them unlikely sources of uhecrs .
the hillas plot does not include the effect of energy losses in the acceleration sites .
photopion production limits the maximum energy achievable in clusters of galaxies to @xmath53 ev .
these sources will therefore not be considered any longer as sources of cosmic rays above the gzk energy .
we also do not discuss here the so - called bursting sources .
the prototypical example of these sources are gamma ray bursts , that have been proposed as sources of uhecrs @xcite .
we refer the reader to recent literature treating this topic @xcite . in the following we
briefly summarize the situation with the other three classes of objects listed above .
the possibility that neutron stars may be accelerators of uhecrs was discussed in detail in ref .
@xcite ( and references therein ) .
the main problem encountered in reaching the highest energies is related to the severe energy losses experienced by the particles in the acceleration site @xcite .
most of the mechanisms discussed in the literature refer to acceleration processes in the magnetosphere of the neutron star , where curvature radiation limits the maximum energy to a value much smaller that @xmath22 ev .
an alternative approach is to think of acceleration processes that occur outside the light cylinder of young neutron stars @xcite . rapidly rotating
, newly formed neutron stars can induce the acceleration of iron nuclei through mhd winds outside the light cylinder @xcite .
although the mechanism through which the rotation energy of the star is converted into kinetic energy of the wind is not yet completely understood , it seems from the observations of the crab nebula that a relativistic wind does indeed exist , with a lorentz factor of @xmath54 @xcite .
possible nuclei with charge @xmath55 can be accelerated in young neutron stars to a maximum energy @xmath56 , as estimated in ref .
@xcite . here
@xmath57 is the surface magnetic field in units of @xmath58 g and @xmath59 is the rotation frequency of the star .
energies gradually smaller are produced while the star is spinning down , so that a spectrum @xmath60 is produced by a neutron star .
the process of escape of the accelerated particles becomes efficient about a year after the neutron star birth .
particles that are generated earlier can not escape , but can produce high energy neutrinos in collisions with the ambient particles and photons @xcite .
the issue of the anisotropy due to the galactic disk is currently under investigation @xcite .
active galaxies are thought to be powered by the accretion of gas onto supermassive black holes .
acceleration of particles can occur in standing shocks in the infalling gas or by unipolar induction in the rotating magnetized accretion disk @xcite . in the former scenario energy losses and size of the acceleration region are likely to limit the maximum energy of the accelerated particles to @xmath61 ev .
in the latter case , the main limiting factor in reaching the highest energies is represented by curvature energy losses , that are particularly severe @xcite unless moderately high magnetic fields can be kept with a small accretion rate .
this may be the case of dormant supermassive black holes , possibly related to the so - called dark massive objects ( dmo ) .
some 32 of these objects have been identified in a recent survey @xcite and 14 of them have been estimated to have the right features for acceleration of uhecrs @xcite . had this model to be right , it would not be suprising that bright counterparts to the uhecr events were not found , since dmos are in a quiescent stage of their evolution .
very little is known of dmos as cosmic ray accelerators : the spectrum is not known , and neither is known their spatial distribution .
it is therefore hard to say at present whether dmos can satisfy the criteria listed in section [ sec : serious ] . in @xcite an interesting prediction
was proposed : if uhecrs are accelerated by unipolar induction , they have to radiate part of their energy by synchrotron emission , resulting in the sources to become observable at tev energies .
one of the most powerful sites for the acceleration of uhecrs is the termination shock of gigantic lobes in radio galaxies .
of particular interest are a subclass of these objects known as fanaroff - riley class ii objects ( fr - ii ) , that can in principle accelerate protons to @xmath62 ev and explain the spectrum of uhecrs up to the gzk cutoff @xcite .
these objects are on average on cosmological distances .
the accidental presence of a nearby source of fr - ii type might explain the spectral shape above the gzk energy , but it would not be compatible with the observed anisotropy @xcite .
nevertheless , it has been recently proposed that a nearby source in the virgo cluster ( for instance m87 ) and a suitable configuration of a magnetized wind around our own galaxy might explain the spectrum and anisotropy at energies above @xmath46 ev @xcite as measured by agasa .
this conclusion depends quite sensibly on the choice of the geometry of the magnetic field in the wind .
several additional tests to confirm or disprove this model need to be carried out . [ [ sec : td ] ] an alternative to acceleration scenarios is to generate uhecrs by the decay of very massive particles . in these _ particle physics inspired models _
the problem of reaching the maximum energies is solved _ by construction_. the spectra of the particles generated in the decay
are typically flatter than the astrophysical ones and their composition at the production point is dominated by gamma rays , although propagation effects can change the ratio of gamma rays to protons .
the gamma rays generated at distances larger than the absorption length produce a cascade at low energies ( mev - gev ) which represents a powerful tool to contrain td models @xcite .
there are basically two ways of generating the very massive particles and make them decay at the present time : 1 ) trapping them inside topological defects ; 2 ) making them quasi - stable ( lifetime larger than the present age of the universe ) in the early universe .
we discuss these two possibilities separately in the next two sections .
symmetry breakings at particle physics level are responsible for the formation of cosmic topological defects ( for a review see ref .
topological defects as sources of uhecrs were first proposed in the pioneering work of hill , schramm and walker @xcite .
the general idea is that the stability of the defect can be locally broken by different types of processes ( see below ) : this results in the false vacuum , trapped within the defect , to fall into the true vacuum , so that the gauge bosons of the field trapped in the defect acquire a mass @xmath63 and decay .
several topological defects have been studied in the literature : ordinary strings @xcite , superconducting strings @xcite , bound states of magnetic monopoles @xcite , networks of monopoles and strings @xcite , necklaces @xcite and vortons @xcite . only strings and necklaces will be considered here , while a more extended discussion can be found in more detailed reviews @xcite .
strings can generate uhecrs with energy less than @xmath64 ( the scale of symmetry breaking ) if there are configurations in which microscopic or macroscopic portions of strings annihilate .
it was shown @xcite that self - intersection events provide a flux of uhecrs which is much smaller than required .
the same conclusion holds for intercommutation between strings .
the efficiency of the process can be enhanced by multiple loop fragmentation : as a nonintersecting closed loop oscillates and radiates its energy away , the loop configuration gradually changes . after the loop has lost a substantial part of its energy , it becomes likely to self - intersect and fragment into smaller and smaller loops , until the typical size of a loop becomes comparable with the string width @xmath65 and the energy is radiated in the form of x - particles .
although the process of loop fragmentation is not well understood , some analytical approximations @xcite show that appreciable uhecr fluxes imply utterly large gamma ray cascade fluxes ( see however @xcite ) .
another way of liberating x - particles is through cusp annihilation @xcite , but the corresponding uhecr flux is far too low @xcite compared with observations .
the idea that long strings lose energy mainly through formation of closed loops was recently challenged in the simulations of ref .
@xcite , which show that the string can produce x - particles directly and that this process dominates over the generation of closed loops . this new picture
was recently questioned in refs .
@xcite . even if the results of ref .
@xcite are correct however , they can not solve the problem of uhecrs @xcite : in fact the typical separation between two segments of a long string is comparable with the hubble scale , so that uhecrs would be completely absorbed . if by accident a string is close to us ( within a few tens mpc )
then the uhecr events would appear to come from a filamentary region of space , implying a large anisotropy which is not observed . even if the uhecr particles do not reach us , the gamma ray cascade due to absorption of uhe gamma rays produced at large distances imposes limits on the efficiency of direct production of x - particles by strings .
necklaces are formed when the following symmetry breaking pattern is realized : @xmath66 . in this case
each monopole gets attached to two strings ( necklace)@xcite .
the critical parameter that defines the dynamics of this network is the ratio @xmath67 where @xmath68 is the monopole mass and @xmath69 is the typical separation between monopoles ( e.g. the length of a string segment )
. if the system evolves toward a state where @xmath70 , the distance between the monopoles decreases and in the end the monopoles annihilate , with the production of x - particles and their decay to uhecrs .
the rate of generation of x - particles is easily found to be @xmath71 .
the quantity @xmath72 is upper limited by the cascade radiation , given by @xmath73 ( @xmath74 ) .
the typical distance from the earth at which the monopole - antimonopole annihilations occur is comparable with the typical separation between necklaces , @xmath75 kpc .
clearly , necklaces provide an example in which the typical separation between defects is smaller than the pathlength of gamma rays and protons at ultra - high energies .
hence necklaces behave like a homogeneous distribution of sources , so that the proton component has the usual gzk cutoff .
this component dominates the uhecr flux up to @xmath46 ev , while at higher energies gamma rays take over .
the fluxes obtained in ref .
@xcite are reported in fig .
4 , where the susy - qcd fragmentation functions @xcite were used .
the dashed lines are for @xmath76 gev , the dotted lines for @xmath77 gev and the solid lines for @xmath78 gev .
the two curves for gamma rays refer to two different assumptions about the radio background at low frequencies @xcite .
super heavy particles with very long lifetime can be produced in the early universe and generate uhecrs at present @xcite . in order to keep the same symbolism used in previous sections , we will call these particles x - particles .
the simplest mechanism of production of x - particles in the early universe is the _ gravitational production _
@xcite : particles are produced naturally in a time variable gravitational field or indeed in a generic time variable classical field . in the gravitational case
no additional coupling is required ( all particles interact gravitationally ) .
if the time variable field is the inflaton field @xmath79 , a direct coupling of the x - particles to @xmath79 is needed . in the gravitational production inflation
is not required a priori , and indeed it reduces the effect .
it can be shown that at time t , gravitational production can only generate x - particles with mass @xmath80 , where @xmath81 is the hubble constant and @xmath82 is the inflaton mass .
the authors in ref .
@xcite and @xcite demostrated that the fraction of the critical mass contributed by x - particles with @xmath83 gev produced gravitationally may be @xmath84 .
if the x - particles are directly coupled to the inflaton field , they can be effectively generated during preheating @xcite .
alternative mechanisms for the production of x - particles are based on non - equilibrium thermal generation during the preheating stage @xcite . as mentioned in the beginning of this section , in order for x - particles to be useful dark matter candidates and generate uhecrs they need to be long lived ( for a possible annihilation scenario see ref .
@xcite ) . the gravitational coupling by itself induces a lifetime much shorter than the age of the universe for the range of masses which we are interested in . therefore , in order to have long lifetimes , additional symmetries must be postulated : for instance discrete gauge symmetries can protect x - particles from decay , while being very weakly broken , perhaps by instanton effects @xcite .
these effects can allow decay times larger than the age of the universe , as shown in @xcite .
the slow decay of x - particles produces uhecrs .
the interesting feature of this model is that x - particles cluster in the galactic halo , as cold dark matter @xcite .
hence uhecrs are expected to be produced locally , with no absorption , and as a consequence the observed spectra are nearly identical to the emission spectra , and therefore dominated by gamma rays .
the very flat spectra and the gamma ray composition are two of the signatures .
the calculations of the expected fluxes have been performed in @xcite .
the strongest signature of the model is a slight anisotropy due to the asymmetric position of the sun in the galaxy @xcite .
more recently a detailed evaluation of the amplitude and phase of the first harmonic has been carried out in @xcite and @xcite .
the two papers agree that the present data is still consistent with the anisotropy expected in the model of x - particles in the halo .
small scale anisotropies do not find an easy explanation in td models , with possibly the exception of the sh relics @xcite .
the possibility that at sufficiently high energies some deviations from known physics may occur is of particular interest for uhecrs .
some attempts have been made to explain the events above @xmath22 ev as a manifestation of some kind of physics beyond the standard model of particle interactions .
these suggestions became even more interesting after the recent claims for correlations of the arrival directions of uhecrs with objects at large redshift .
some of these correlations are still subject of debate @xcite .
more recent studies result in a quite intriguing correlation with bl lacs @xcite .
two of the bl lacs in the sample used in @xcite are in the error box of the two triplets of events detected by agasa , and correspond to a distance of @xmath85 mpc , much larger than the pathlength for photopion production of protons ( however more than half of the objects used in ref .
@xcite have unknown redshift [ tkachev , private communication ] ) .
the absence of bright nearby counterparts to the uhecr events has first inspired theoretical proposals of neutrinos as primary particles , since these particles are not affected by the presence of the cmb and can therefore propagate on cosmological distances without energy losses other than those due to the cosmic expansion .
however , the small cross section of neutrinos makes them unlikely primaries : they simply fail to generate the observed showers in the atmosphere .
there is one caveat in this argument : the neutrino nucleon cross section at center of mass energies above the electroweak ( ew ) scale has not been measured , so that the argument above is based on the extrapolation of the known cross sections , and simply limited by the weak unitarity bound @xcite .
it has been proposed @xcite that an increase in the number of degrees of freedom above the ew scale would imply the increase of the neutrino nucleon cross section above the standard model prediction . in particular , in the theories that predict unification of forces at @xmath86 tev scale with large extra dimensions , introduced to solve the hierarchy problem @xcite , these additional degrees of freedom arise naturally , and imply a neutrino - nucleon cross section that increases linearly with energy , reaching hadronic levels for neutrino energies of @xmath46 ev ( for string scale at 1 tev ) .
the calculations in ref .
@xcite show that the cross section remains too small to explain vertical showers in the atmosphere ( for possible upper limits on this cross section see ref .
@xcite ) .
an alternative , even more radical proposal to avoid the prediction of a gzk cutoff in the flux of uhecrs consists in postulating a tiny violation of lorentz invariance ( li ) .
the main effect of this violation is that some processes may become kinematically forbidden @xcite .
in particular , photon - photon pair production and photopion production may be affected by li violation .
the absence of the gzk cutoff would then result from the fact that the threshold for photopion production disappears and the process becomes kinematically not allowed .
it is suggestive that the possibility of a small violation of li has also been proposed to explain the apparent absence of an absorption cutoff in the tev gamma ray emission from markarian - like objects @xcite ( see @xcite for a critical view of this possibility ) .
the increasing evidence for a flux of uhecrs exceeding the theoretical expectations for extragalactic sources has fueled interest in several models .
these models aim to satisfy all the requirements imposed by observations on fluxes , spatial anisotropies and composition . at present
, however there is no obvious successful model .
the situation might change with the availability , soon to come , of quite larger statistics of events , that will be achievable by experiments like the pierre auger project @xcite and euso / airwatch / owl @xcite
. these future experimental efforts will be crucial mainly in three respects : 1 ) the increase of statistics by a factor 100 for auger and even more for the space based experiments will allow to strongly constrain theoretical models , and check whether the present excess is a @xmath87 fluctuation or a physical effect .
moreover the small scale anisotropies , if real , will become stronger and a correlation function approach will definitely become appropriate to the analysis of the events ; 2 ) the full sky coverage will finally allow a test of models based on local extragalactic sources and on galactic sources of uhecrs , through the measurement of the large scale anisotropy , that is currently spoiled by the limited spatial exposure ; 3 ) a better determination of the composition of the uhecr events will represent a smoking gun either in favour of or against models : td models would be ruled out if no gamma rays are found or if heavy nuclei represent the main component . on the other hand , iron - dominated composition would point toward a possible galactic origin , possibly related to neutron stars .
the confirmation , on statistical grounds , of the association of uhecrs to distant cosmological objects like bl lacs would represent a very strong indication of physics beyond the standard model .
either new interactions or some modification of fundamental physics would be needed to explain such a result .
j. alvarez - muniz , r. engel and t. stanev , in the proceedings of icrc2001 , og1.03 , hamburg , germany , 07 - 15 august 2001 ; s. oneill , a.v .
olinto and p. blasi , in the proceedings of icrc2001 , og1.03 , hamburg , germany , 07 - 15 august 2001 ( preprint astro - ph/0108401 ) .
farrar and p.l .
biermann , phys .
81 * , 3579 ( 1998 ) ; c.m .
hoffman , phys .
. lett . * 83 * , 2471 ( 1999 ) ; g.r .
farrar and p.l .
biermann , phys .
lett . * 83 * , 2472 ( 1999 ) ; g. sigl _
et al _ , preprint astro - ph/0008363 ; a. virmani _ et al _ , preprint astro - ph/0010235 .
s. coleman and s.l .
glashow , nucl .
b574 * , 130 ( 2000 ) ; l. gonzales - mestres , nucl .
b ( proc . suppl . )
* 48 * , 131 ( 1996 ) ; r. aloisio , p. blasi , p. ghia and a.f .
grillo , phys .
rev . * d62 * , 053010 ( 2000 ) .
streitmatter , proc . of _ workshop on observing giant cosmic air showers from @xmath88 ev particles from space _ ,
krizmanic , j.f . ,
ormes , j.f . , and streitmatter , r.e .
( aip conference proceedings 433 , 1997 ) .
|
the universe should be dark at energies exceeding @xmath0 ev .
this simple but solid prediction of our best known particle physics is not confirmed by observations , that seem to suggest a quite different picture .
numerous events have in fact been detected in this energy region , with spectra and anisotropy features that defy many conventional and unconventional explanations .
is there a problem with known physics or is this a result of astrophysical uncertainties ?
here we try to answer these questions , in the light of present observations , while discussing which information future observations may provide on this puzzling issue .
= 11.6pt
|
quantum monte carlo methods , adopted for nuclei in the last decade , have proven to be very useful in obtaining exact solutions of the nuclear schrdinger equation for the ground and low - energy excited states of up to six nucleons @xcite , and it is likely that many states of seven and eight nucleons will be exactly calculated with realistic nuclear forces in the near future . a variational approximation , @xmath3 , to the lowest - energy state of the desired spin , parity , and isospin , @xmath4 , is first obtained by a variational monte carlo ( vmc ) calculation @xcite . the exact eigenstate , @xmath5 , belonging to the lowest - energy eigenvalue , @xmath6 , for that @xmath4 is then projected out using @xmath7 |\psi_v\rangle\ ] ] with the green s function monte carlo ( gfmc ) method @xcite .
the computation yields a population of configurations @xmath8 , where @xmath9 specifies the positions of all the nucleons and @xmath10 labels the configurations , distributed with the probability @xmath11 \psi_v({\bf r})| \ .\ ] ] mixed estimates , defined as : @xmath12_m = \frac { \langle\psi_v| o exp[-(h - e_0)\tau ] |\psi_v\rangle } { \langle\psi_v| exp[-(h - e_0)\tau ] |\psi_v\rangle } \ , \ ] ] are then calculated from this population .
obviously @xmath13_m$ ] converges to the eigenvalue @xmath6 in the limit @xmath14 , however , the variance of @xmath13_m$ ] depends primarily on that of the local variational energy : @xmath15 expectation values of operators other than h are generally calculated only up to first order in the difference @xmath16 from @xmath17_m - \frac { \langle\psi_v| o |\psi_v\rangle } { \langle\psi_v|\psi_v\rangle } \ .\ ] ] it is clear that accurate variational wave functions are needed for the success of this approach .
the variance of @xmath18 decreases and the terms neglected in eq.(5 ) become smaller as @xmath19 .
state - of - the - art variational wave functions @xcite of few - body nuclei have the form : @xmath20 |\psi_p\rangle \\ ] ] with the pair wave function @xmath21 given by : @xmath22 |\psi_{j}\rangle \ .\ ] ] the @xmath23 represents a symmetrized product , and the jastrow wave function , @xmath24 , is given by : @xmath25 |\phi\rangle \ .\ ] ] here @xmath26 is an antisymmetric product of single - particle wave functions with the desired @xmath4 , the @xmath27 is a two - body central correlation , and the operator @xmath28 is defined as : @xmath29 u_p(r_{ij } ) o^p_{ij } \ , \ ] ] @xmath30\otimes [ 1 , \tau_i\cdot\tau_j ] \ .\ ] ] the @xmath27 and @xmath28 correlations are generated by the static parts of the two - nucleon interaction .
the spin - orbit correlations are defined as : @xmath31 ( { \bf l}\cdot{\bf s } ) _ { ij } \ , \ ] ] and the eight radial functions @xmath27 , @xmath32 , @xmath33 , and @xmath34 are obtained from approximate two - body euler - lagrange equations with variational parameters @xcite . the factor : @xmath35 \ , \label{eq : fpijk}\ ] ] with @xmath36 suppresses spin - isospin correlations between nucleons @xmath37 and @xmath38 when a third nucleon @xmath39 comes close to either @xmath37 or @xmath38 , and @xmath40 are variational parameters .
finally , @xmath41 with @xmath42 , represent correlations induced by the three - nucleon interaction @xmath43 . its form is suggested by perturbation theory , and @xmath44 and @xmath45 are variational parameters .
the exact ground - state energy of @xmath0h for the argonne v@xmath46 two - nucleon @xcite and urbana model viii three - nucleon @xcite interactions , calculated using a faddeev wave function @xcite and monte carlo integration , is @xmath47 , while the energy for @xmath2he , calculated by the gfmc @xcite method , is @xmath48 mev , where the numbers in parentheses denote the statistical error .
vmc calculations @xcite with the wave function of eq.(6 ) give instead the energies @xmath49 and @xmath50 mev that are too high by @xmath51 .
this difference , reflecting the deficiencies in @xmath3 , could be because the two - body functions @xmath52 , @xmath53 , @xmath54 , and @xmath55 are not optimally determined by the approximate equations used to construct them , or the true @xmath5 contains many - body correlations in addition to @xmath56 and @xmath57 .
note that a @xmath3 containing two- and three - body correlations can in principle reproduce the exact @xmath5 of @xmath0h .
we carried out extensive comparisons between the 34-channel faddeev @xmath58 of the los alamos - iowa group @xcite , one of the best available approximations to the exact @xmath5 , and the @xmath59 for @xmath0h , to test the accuracy of the two - body functions in @xmath59 and study the three - body correlations in @xmath58 , and therefore in @xmath5 .
these comparisons are discussed in sec .
slight improvements in the two - body functions did not significantly reduce the difference @xmath60 .
however , a few three - body correlation operators can reduce this difference substantially .
vmc calculations , for the binding energy of the three and four - body systems , with wave functions including the new correlations are presented in sec .
the error in these is approximately half of that in the best previous variational calculations @xcite .
the new variational wave function we adopt is given by : @xmath61 [ \prod_{i < j < k}f^c_{ijk } ] |\psi_p\rangle \ , \ ] ] and our purpose is to study the new three - body correlations , @xmath62 and @xmath63 , and also to improve the @xmath56 correlations in @xmath21 .
we consider the triton wave function expressed in the spin - isospin basis @xmath64 where @xmath65 and @xmath66 are respectively the total spin and isospin of the pair of particles ( 12 ) , @xmath67 is the total spin of the three nucleons , and @xmath68 is its projection . allowed values for @xmath67 are @xmath69 ( for @xmath70 ) and @xmath71 ( for @xmath72 ) .
the total isospin of the system is @xmath69 , its projection is @xmath73 , and @xmath66 can have the values @xmath74 or @xmath75 .
consequently there are 16 components in the wave function , corresponding to the 8 spin states @xmath76 labeled by @xmath77 and the 2 isospin states @xmath78 labeled by @xmath79 .
these are defined as : @xmath80 and @xmath81 the triton wave function at any configuration @xmath82 can then be written as @xmath83 the main components of the wave function are @xmath84 and @xmath85 , which are the only nonzero components in the jastrow wave function .
other components are generated mostly by tensor and three - body correlations . in order to determine the @xmath62 and @xmath63 we compared @xmath59 with @xmath58 , and studied the deviations of the former relative to the latter . using the decomposition of eq.([eq : psive ] ) for both wave functions
, we defined a number of `` deviation '' quantities . for each spatial configuration , @xmath86 , the deviation @xmath87 is given by @xmath88 separately , we define a deviation @xmath89 for each spin - isospin component @xmath90 where @xmath91 is the total number of configurations considered .
finally , we define an average deviation @xmath92 @xmath93 the exact @xmath94 satisfies @xmath95 at all @xmath82 , however the @xmath96 does not satisfy this equation in all regions of configuration space , presumably due to the truncation in sums over partial waves and to interpolation between mesh points . in some regions
@xmath97 can be different from @xmath98 by as much as 25% .
for the study of the deviations listed above , a set of 1,000 configurations @xmath86 for which @xmath97 is within 5% of @xmath99 was chosen from a larger set of random configurations distributed with the probability @xmath100 .
we studied a wide variety of three - body correlations of the form @xmath101 . for each @xmath102
, values of @xmath103 were obtained by minimizing @xmath87 .
simple functions of @xmath82 were then chosen to approximate the extracted values of @xmath103 . in principle , we can choose a set of sixteen operators @xmath102 and define the complete three - body correlation as : @xmath104 instead of using eq .
the @xmath105 can then be calculated by solving the matrix equation : @xmath106 obtained from @xmath107 , at each value of @xmath82 .
however , all our attempts led to unacceptable , rapidly fluctuating , functions @xmath105 presumably dominated by the small differences between @xmath108 and the exact @xmath109 and inappropriate choices of the sixteen operators @xmath102 . in contrast , the @xmath105 obtained by minimizing the deviation @xmath87 are smooth and useful for some of the operators @xmath102 . from the many operators
@xmath102 considered only a few produced a significant reduction of the deviations ; we list below the successful ones . a spatial three - body correlation , @xmath110 reduces the probability for particles @xmath37 , @xmath38 , and @xmath39 to be in a line for @xmath111 .
the parameters of the new three - body correlations of type @xmath112 are denoted by @xmath113 ; thus @xmath114 and @xmath115 are the parameters of @xmath116 . a spin - orbit three - body correlation @xmath117 \ ,
\label{eq : uls}\ ] ] where @xmath118 may be generated by the two - body spin - orbit interaction @xmath119 operating on the jastrow correlations @xmath120 and @xmath121 .
it is expressed as a function of @xmath122 and @xmath123 to underline its motivation , however @xmath122 differs significantly from the bare @xmath124 potential and both @xmath122 and @xmath123 are determined by minimizing the @xmath87 . an isospin three - body correlation @xmath125 -2q^{\tau}_{1}exp(-2q^{\tau}_{2}x_{ijk}^{2 } ) \right\ } \tau_{i}\cdot\tau_{j } \ , \label{eq : utau}\ ] ] with @xmath126 r_{ijk } \ .\ ] ] enhances the probability for @xmath127 when nucleons @xmath37 and @xmath38 are far from @xmath39 .
an example of a three - body correlation that we tried and found to be of marginal utility is given by : @xmath128 i\tau_{i}\cdot(\tau_{j}\times\tau_{k } ) \
, \label{eq : uxt}\ ] ] where @xmath129 such a correlation may be generated by tensor interactions between pairs @xmath130 and @xmath131 .
this correlation was omitted in our final energy calculations .
finally , an improved parametrization of @xmath56 was found , and eq.([eq : fpijk ] ) is replaced by : @xmath132 initial values for the various parameters were obtained by minimizing @xmath87 for the test set of 1,000 faddeev configurations .
the final values of the parameters , which are somewhat different , are given in table i ; they were obtained by minimizing the binding energy of @xmath0h with the argonne @xmath133 and urbana model viii interactions for samples of 10,000 variational configurations .
to illustrate the procedure used to determine the form of the new correlation functions , we show in fig .
1 the data points for the three - body central correlation @xmath134 obtained by minimizing @xmath135 at each value of @xmath82 . these are for isosceles configurations @xmath136 with @xmath137 fm , plotted against @xmath138 , where @xmath139 is the angle formed by @xmath140 and @xmath141 .
the functional form of @xmath62 , eq.([eq : fcijk ] ) , was chosen because it can fit these numerical values with appropriate parameters @xmath142 and @xmath143 .
the curve in fig .
1 shows @xmath134 with the slightly different values of @xmath142 and @xmath143 that minimize the energy ; the data points and curve are quite similar .
the improvement achieved in the main components of @xmath144 can also be seen in fig . 2 where we plot the difference @xmath145 for equilateral configurations as a function of @xmath146 .
the @xmath147 is also shown in fig . 2 for comparison . in this calculation
the particles are placed in the xz - plane , where @xmath148 are real , and for equilateral configurations @xmath149 holds exactly for antisymmetric @xmath150 .
the difference @xmath151 is @xmath152 1% of @xmath153 for these configurations .
the average values of @xmath154 , @xmath89 [ eq.([eq : eab ] ) ] , @xmath155 and the total deviation @xmath156 [ eq.([eq : e ] ) ] for the test set of 1,000 faddeev configurations are listed in table ii for the new and old @xmath144 .
they are normalized so that @xmath155 has unit average value for these configurations ; note that this does not correspond to the standard normalization @xmath157 , however , it clearly exhibits the relative contributions and errors .
the main components , @xmath158 = 31 and 12 , have a rather small relative error , defined as @xmath159 , of @xmath160 in the new @xmath144 .
next in importance is the group of @xmath161 states ( @xmath162 = 5 to 8) which have total orbital angular momentum @xmath163 and account for @xmath164 of @xmath165 .
the largest of these have @xmath158 = 61 and 72 , and a relative error of @xmath166 .
the states @xmath158 = 71 and 51 have smaller contributions , and a larger relative error of @xmath167 .
the next group has @xmath168 , @xmath169 where the pair of nucleons 1 and 2 is in an odd parity state .
these states account for @xmath170 of @xmath165 and have the largest relative error of @xmath171 .
the last group has @xmath168 and @xmath172 , for an @xmath173 .
it also accounts for @xmath170 of @xmath165 , and these components of the new @xmath144 have an error of @xmath174 .
the @xmath89 of the new @xmath144 are smaller than those of the old @xmath144 for all values of @xmath158 other than 62 .
the total @xmath92 of 0.00062 is reduced to 0.00047 by the new @xmath62 .
it is further reduced to 0.00035 by the inclusion of @xmath175 . the subsequent reduction to @xmath92 = 0.00028
is due to @xmath176 . the @xmath177 correlations and the changes in the @xmath56 correlations have very small ( @xmath178 ) effect on the total @xmath92 .
as mentioned in the previous section , the final values of the parameters of the new three - body correlations were determined by minimizing the energy . in table iii , we present the results for the binding energy of @xmath0h , for the faddeev wave function and various variational wave functions moving from the old to the new .
a set of 50,000 configurations sampled from @xmath179 was used for these calculations .
the first column contains the values of the binding energy and the second the difference @xmath180 .
we can clearly see that this difference is reduced approximately by a factor of two , when one goes from the old to the new @xmath144 , and that the @xmath116 and @xmath175 give almost 80% of the improvement .
the last column of table iii gives @xmath181 , where @xmath182 and @xmath183 .
the @xmath184 gives the admixture of excited states in @xmath59 .
the mean energy of these excited states is given by : @xmath185 for both old and new @xmath144 .
for this reason , the gfmc energy : @xmath186 = \frac { \langle\psi_v| h exp[-(h - e_0)\tau ] |\psi_v\rangle } { \langle\psi_v| exp[-(h - e_0)\tau ] |\psi_v\rangle } \ , \ ] ] of @xmath0h ( and also @xmath2he ) converges to the true @xmath6 by a rather small @xmath187 mev@xmath188 . in contrast the @xmath144 of six - body nuclei seem to have admixtures of states with much smaller @xmath189 , and their @xmath190 does not converge so rapidly to @xmath6 as @xmath191 increases @xcite .
these new correlations , obtained by comparing the @xmath108 and @xmath144 for the argonne @xmath133 and urbana viii hamiltonian , seem to have more general applicability . results of variational calculations for @xmath0h and @xmath2he , with different interaction models , using either the old @xmath144 ( ref . [ 2 ] ) or the new @xmath144 with the same three - body correlation parameters given in table i are listed in tables iv and v. ( the difference between the old argonne @xmath133 and urbana viii value given in table iv and the old @xmath56 given in table iii is of marginal statistical significance ; it can be attributed to slight changes to the the @xmath21 in the present work compared to ref . [ 2 ] , and the different random walks followed in the two calculations . ) a significant improvement is seen in each case , with @xmath192 3050% of the energy difference with exact calculations , where available , recovered . in principle , the parameters of the new correlations should be optimized separately for each case , but this has not been done in the present work .
the faddeev energies shown in table iv have been calculated from the los alamos - iowa @xmath108 using monte carlo integration , and thus have a quoted statistical error .
these energies are consistent with those published by the los alamos - iowa group @xcite . for @xmath2he ,
the current vmc results are fairly competitive with other recently developed methods , faddeev - yakubovsky ( fy ) @xcite and correlated hyperspherical harmonics ( chh ) @xcite , as shown in table v. however , all these methods fall short of the exact gfmc calculations @xcite in @xmath2he . if the wave function @xmath193 were an exact eigenstate of h , then it should satisfy the eigenvalue equation : @xmath194 and this equation should be satisfied for any space configuration , meaning that any local energy @xmath195 should be equal to the eigenvalue @xmath196 . for the three - body system we considered sets of different space configurations , each set being composed of configurations having the same @xmath146 ( the perimeter of the triangle formed by the three particles ) , and computed the average of the energies returned by eq.([eq : eigenrr ] ) .
these average values , @xmath197 , are then a function of @xmath146 and can be plotted against it . in figs . 3 , 4 , and 5 we show the results respectively for the faddeev wave function , and the old and new variational wave function .
these calculations were done with the argonne @xmath133 two - body and urbana model viii three - body interactions . from fig .
3 we can clearly see that , even for the faddeev wave function , there are regions , in particular for small and large perimeters , where the average values differ from the energy expectation value .
moreover , the variance is finite at all @xmath146 , whereas it would be zero for an exact eigenfunction .
the deviations of @xmath198 and the variance of the faddeev local energy are perhaps due to the errors introduced by numerical interpolation and the application of the permutation operators , required to go from the faddeev amplitude to the full wave function , as explained in the previous section .
the application of these permutation operators in the los alamos - iowa faddeev solution is implemented by means of partial wave series truncated at 34 channels .
this problem led us to consider , in the procedure to search for the three - body correlations , only those regions in configuration space where the faddeev wave function gives an @xmath199 close to the exact energy . a comparison of the old and new variational wave function results ( fig . 4 to fig .
5 ) shows that the new correlations clearly tend to push the average values in the right direction and to reduce the variance . in variational calculations of six - body nuclei
@xcite the new three - body correlations continue to lower the variational upper bound .
one simple generalization that can be made is to let the @xmath62 correlation have different strengths depending on whether none , one , or two of the nucleons @xmath37 , @xmath38 , and @xmath39 are in the p - shell . at present
it appears best to let this correlation be zero if one or two are in the p - shell , although an alternate structure might be even better in these cases .
it is also beneficial to reduce the overall strength of the @xmath175 and @xmath176 correlations . with these steps , the new three - body correlations lower the energy by @xmath192 0.25 mev in @xmath200li , and also reduce the variance . in conclusion
we have reported significant progress in the construction of accurate variational wave functions for the nuclear few - body problem .
we believe that the new three - body correlations will lead to clear improvement in gfmc wave functions for light nuclei , as well as in variational wave functions of heavier systems .
we wish to thank j.l .
friar , b.f .
gibson , and g.l .
payne for the use of their faddeev wave functions .
one of the authors ( aa ) would like to thank the kind hospitality of the physics department of the university of illinois at urbana - champaign , where part of this work has been performed .
the work of aa is supported by the university of lisbon , junta nacional de investigao cientfica e tecnolgica under contract no .
pbic / c / cen/1108/92 , that of vrp by the national science foundation via grant phy 94 - 21309 , and that of rbw by the u.s .
department of energy , nuclear physics division , under contract no .
w-31 - 109-eng-38 .
cdl @xmath114 & 0.19 & @xmath201 + @xmath115 & 0.83 & fm@xmath188 + @xmath203 & 0.20 & + @xmath204 & 0.06 & fm@xmath188 + @xmath205 & 0.12 & @xmath206 + @xmath207 & 0.12 & @xmath206 + @xmath208 & 0.85 & @xmath206 + @xmath209 & 0.014 & fm@xmath188 + @xmath210 & 0.016 & @xmath206 + @xmath211 & 1.2 & fm + @xmath212 & 0.25 &
|
a detailed comparison of faddeev and variational wave functions for @xmath0h , calculated with realistic nuclear forces , has been made to study the form of three - body correlations in few - body nuclei .
three new three - body correlations for use in variational wave functions have been identified , which substantially reduce the difference with the faddeev wave function .
the difference between the variational upper bound and the faddeev binding energy is reduced by half , to typically @xmath1 .
these three - body correlations also produce a significant lowering of the variational binding energy for @xmath2he and larger nuclei .
= 10000
|
the recent growing evidences for a vacuum energy component in the universe @xcite have lead to the development of a number of models , most of which tend to alleviate the theoretical difficulties of the well known `` cosmological constant problem '' ( see @xcite for a review ) .
all these models are replacing the cosmological constant with a `` dynamical vacuum energy '' provided by an evolving scalar field , often named `` quintessence '' .
this field should provide the present cosmic accelerating phase through its negative equation of state .
such `` dark energy '' component has also been modeled in the general framework of scalar - tensor theories of gravity : in these `` extended quintessence '' ( eq ) scenarios @xcite-@xcite , the quintessence scalar field is directly coupled to the ricci scalar @xmath0 in the lagrangian of the theory : the canonical ricci scalar term @xmath1 , with @xmath2 , is replaced by @xmath3 , where @xmath4 is a function of the quintessence field @xmath5 . here
, @xmath6 is a `` bare '' gravitational constant , generally different from the newton s constant g as it is measured by cavendish - type experiments @xcite . in eq models ,
a scalar field has thus a double role : at any epoch , its value determines the effective gravitational constant , and the contribution in dark energy density .
of course , such a coupling is not arbitrary : it is instead subject to a number of constraints , mainly arising from the bounds on the time variations of the constants of nature , which must be taken into account .
eq scenarios have been analyzed from different points of view , and there are robust predictions about their possible imprints on the power spectrum of cmb anisotropies @xcite . however , the important role of dark energy in modern cosmology opens the question of its relation with the other main dark component currently under investigation , the dark matter .
one major point with quintessence scenarios is whether a dark energy component could affect , in some way , the formation of dark matter clumps . in particular , one fundamental question is : can the dark energy itself form clumps on certain scales ?
if we restrict ourselves to minimally - coupled scalar fields , we have to deal with a very relativistic component : indeed , the sound speed of such component turns out to be identical to that of radiation @xcite . as a consequence ,
an `` ordinary '' ( i.e. , minimally coupled ) quintessence component becomes homogeneous on scales smaller than the horizon , as its energy density perturbations are rapidly damped out .
for this reason , clumps of `` ordinary '' quintessence can not form on such scales . in eq scenarios , however , the situation can dramaticaly change , since the physical properties of an eq are deeply modified with respect to those of an ordinary quintessence field @xcite , as is the case for its effective sound speed ; one natural question is whether dark energy structures could form in these scenarios , starting from initial small overdensities in the linear regime . in order to give an answer to this question
, we start from a definition of the stress - energy tensor @xmath7 of the quintessence field , which includes a term of purely gravitational origin , in addition to the standard contributions from minimal coupling and non - minimal coupling @xcite : @xmath8 = t_{\mu \nu}^{mc}[\phi]+ t_{\mu \nu}^{nmc}[\phi ] + t_{\mu \nu}^{grav}[\phi ] \ \ , \ ] ] with @xmath9 = \left ( { 1\over \kappa}-f\right ) g_{\mu \nu}\ .\ ] ] the so - defined stress - energy tensor of the quintessence field is a conserved quantity , as a consequence of the contracted bianchi identities ; drawing the scalar field energy density @xmath10 from this tensor , it will satisfy the continuity equation .
most importantly , also in @xmath10 we can distinguish the gravitational contribution from the others , analogously to the separation performed into @xmath11 : @xmath12 in the last expression , @xmath13 is the conformal expansion rate , and @xmath14 is the scalar field potential .
let us focus on the last term in the rhs of this equation : noticeably , when @xmath15 , the gravitational contribution to @xmath10 switches on , feeding @xmath10 with a term proportional to the total energy density of the universe via @xmath16 .
in practice , the quintessence energy density is dragged by the energy density of the dominant component at any epoch : this effect , which we call `` gravitational dragging '' , is also active on the time evolution of the perturbations .
indeed , we could perform an analogous division between the various contributions to the scalar field density contrast @xmath17 , finding @xmath18 being the gravitational contribution @xcite @xmath19 looking at this expression , it is evident that the quintessence field density contrast can be non - vanishing even when the quintessence field is homogeneous by itself , provided @xmath20 ; in this case , indeed , @xmath21 is sourced by the @xmath22 term , containing the total cosmic fluid energy density perturbations , which are generally non vanishing .
while the scalar field perturbations @xmath23 contribute to @xmath24 and @xmath25 only , the ricci scalar perturbations are responsible for the gravitational contribution to @xmath21 ; the latter can dominate the dynamics of @xmath21 whenever @xmath26 are negligible and @xmath20 .
in other words , the gravitational dragging mechanism forces the scalar field perturbations to behave as the density perturbations of the dominant cosmic fluid component : noticeably , in the matter dominated era , an eq field can even behave as pressureless dark matter , thanks to the explicit coupling netween @xmath5 and @xmath0 .
the quintessence field overdensities receive an injection of power which may potentially enhance them up to the exit from the linear regime , following the matter behaviour . in fig .
[ powerspectrum ] we plot the logarithmic power of density fluctuations , defined in the fourier space as @xmath27 , both for cdm perturbations and for eq perturbations , in a specific example with @xmath28 ( where @xmath29 is an observationally constrained coupling constant ) .
note that , at the present time , the quantity @xmath30 reduces to @xmath31 ( where @xmath32 is the field value at @xmath33 ) , which is generally a non - vanishing term .
the plots in figure [ powerspectrum ] refers to redshifts @xmath34 and @xmath35 ; because of the gravitational dragging , linear perturbations in the quintessence energy density are close to those in the cdm component up to redshifts relevant for structure formation , while they appear to decrease at @xmath36 , when the quintessence field comes to dominate the expansion and the term @xmath37 is no more the dominant term in the denominator of @xmath38 .
however , scales on which perturbations reached the non - linear regime at some redshift are no more adequately described by the linear approach . at that point
, it is quite obvious that a non - linear approach is required to further investigate the quintessence perturbation evolution .
+ the same result can be obtained with a different approach , based on the properties of the `` effective sound speed '' @xmath39 of a fluid @xcite : such a quantity , defined as the sound speed in the fluid rest frame , encloses both the adiabatic and entropic contributions to the sound speed , and it well characterizes the total pressure response to the gravitational collapse . while it is equal to the speed of light for minimally - coupled scalar fields , it turns out to be a scalar field dependend quantity in eq models ; in the gravitational dragging regime , the effective sound speed is inversely proportional to @xmath21 , which can be at the level of perturbations in the cdm component , resulting in low values of @xmath39 . also the sign of @xmath39 may be reversed , depending on the sign of the coupling constant , strongly modifying the response of pressure to the variation in the energy density , and radically changing the gravitational collapse scenario for the quintessence component .
the fact that the quintessence sound speed gets strongly modified by the coupling to the ricci scalar , a result which is analogous to the gravitational dragging mechanism described above , allows for the possibility of quintessence clumps ; the characteristic scales for this process have not been investigated yet , and the further evolution of quintessence energy density perturbations in the non - linear regime deserve a future investigation .
the intriguing possibility of the gravitational dragging of matter perturbations on quintessence perturbations is one of the possible grounds where to probe the relation between dark matter ( dm ) and dark energy ( de ) . in recent years
, the possibility of de clumps formation , the stability of spherically symmetric scalar field structures and their relation to the dm component in galaxies , have received increasing attention @xcite . in the scenario here proposed , such relation can only be of gravitational type , and may possibly be detected on the galactic scales .
the question of how the dm is distributed inside dark halos and cluster is still matter of debate .
for example , it is still unknown whether the dark halos density profile is universal , or it depends on properties of the galaxy itself .
kinematic properties , such as rotation curves of disk galaxies , are an excellent probe for dm on galactic scales ; it has been shown that independent high - resolution optical rotation curves observations , can probe the halo mass distribution and resolve their structure .
an important contribution has been obtained @xcite by the analysis of low - luminosity late - type spirals ; the surprising result of this analysis has been that the halo velocity profiles increase linearly with the radius , at least up to the stellar disk edge ( the visible galaxy " ) .
this implies that , over all the stellar disk , the dark halo density is about a constant ; similar properties were previously found also for dwarf and low surface brightness galaxies , which are both dm objects @xcite .
the presence of a constant density central core is in contrast with the cdm scenario , which predicts a steeply cusped density distribution in the same region .
the presence of a central singularity in the density profile of dm has to be excluded also as a result of the analysis of gravitational lensing on the cluster scales .
this disparity between the model and the data on the galaxy rotation curves requires the introduction of different fields and/or particles , including the de as a candidate .
indeed , this disparity follows directly from the nature of collisionless particles , which have no physically associated length scale , with the result that dm halos should have zero effective core radii @xcite . in order to generate cores in the dm distribution , additional physical properties
are required , such as could be provided by hot dark matter ( in this case , the central density may be naturally limited by phase space constraints ) or an universe dominated by dissipative baryonic matter .
even though these models have to be excluded because they fail in accounting for the observed large scale structure , they are both able to provide a mechanism that limits the central density of the cdm particles , introducing a characteristic scale . in this context
, it is extremely interesting the question of how , and whether , the de has a key role into the structure of the dm halos .
if this is the case , we expect that this role is based on the purely gravitational interaction between dm and de , which may imply a redistribution of dm in the halos , particularly in the central core .
the gravitational dragging mechanism described above opens new perspectives on the relation between dm and de in collapsed structures , from galactic scales to the scales of galaxy clusters .
in particular , predictions for the galactic rotation curves in eq models should be drawn and compared to observations .
a non - linear approach to the perturbation growth in gravitational dragging regime will allow to check for the possibility of constant density central cores in galaxies , and for determining the characteristic scales of eventual quintessence clumps . c. wetterich , phys.lett .
b 522 , 5 ( 2001 ) ; f. s. guzman , t. matos , class.quant.grav .
17 , l9 ( 2000 ) ; t. matos , f. s. guzman , d. nunez , phys.rev .
d62 061301 ( 2000 ) ; a. arbey , j. lesgourgues , p. salati , phys.rev .
d64 123528 ( 2001 ) ; e.l .
lokas , y. hoffman , astro - ph/0108283 , submitted to mnras ( 2001 ) ; t. padmanabhan , t. roy choudhury , hep - th/0205055 ( 2002 )
|
in the general context of scalar - tensor theories , we consider a model in which a scalar field coupled to the ricci scalar in the gravitational sector of the lagrangian , is also playing the role of an `` extended quintessence '' field , dominating the energy content of the universe at the present time . in this framework , we study the linear evolution of the perturbations in the quintessence energy density , showing that a new phenomenon , named here `` gravitational dragging '' , can enhance the scalar field density perturbations as much as they reach the non - linear regime . the possibility of dark energy clumps formation is thus discussed .
|
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.